TRNSYS 16

a TRaNsient SYstem Simulation program

Volume 8
Programmer's Guide

Solar Energy Laboratory, Univ. of Wisconsin-Madison

http://sel.me.wisc.edu/trnsys

TRANSSOLAR Energietechnik GmbH

http://www.transsolar.com

CSTB Centre Scientifique et Technique du Btiment

http://software.cstb.fr

TESS Thermal Energy System Specialists, LLC

http://www.tess-inc.com

TRNSYS 16 Programmer's Guide

About This Manual

The information presented in this manual is intended to provide a reference to users who want to
create their own TRNSYS 16 components or want to understand the TRNSYS 16 simulation
engine. This manual is not intended to provide information about other utility programs in
TRNSYS 16. More information concerning these programs can be found in other parts of the
TRNSYS documentation set. The latest version of this manual is always available for registered
users on the TRNSYS website (see here below).

Revision history
2004-09

For TRNSYS 16.00.0000

2005-02

For TRNSYS 16.00.0037

Where to find more information

Further information about the program and its availability can be obtained from the TRNSYS
website or from the TRNSYS coordinator at the Solar Energy Lab:
TRNSYS Coordinator
Solar Energy Laboratory, University of Wisconsin-Madison
1500 Engineering Drive, 1303 Engineering Research Building
Madison, WI 53706 U.S.A.

Email: trnsys@engr.wisc.edu
Phone: +1 (608) 263 1586
Fax:
+1 (608) 262 8464

TRNSYS website: http://sel.me.wisc.edu/trnsys

Notice
This report was prepared as an account of work partially sponsored by the United States
Government. Neither the United States or the United States Department of Energy, nor any of
their employees, nor any of their contractors, subcontractors, or employees, including but not
limited to the University of Wisconsin Solar Energy Laboratory, makes any warranty, expressed
or implied, or assumes any liability or responsibility for the accuracy, completeness or usefulness
of any information, apparatus, product or process disclosed, or represents that its use would not
infringe privately owned rights.
2004 by the Solar Energy Laboratory, University of Wisconsin-Madison
The software described in this document is furnished under a license agreement. This manual
and the software may be used or copied only under the terms of the license agreement. Except
as permitted by any such license, no part of this manual may be copied or reproduced in any form
or by any means without prior written consent from the Solar Energy Laboratory, University of
Wisconsin-Madison.

82

TRNSYS 16 Programmer's Guide

TRNSYS Contributors

S.A. Klein

W.A. Beckman

J.W. Mitchell

J.A. Duffie

N.A. Duffie

T.L. Freeman

J.C. Mitchell

J.E. Braun

B.L. Evans

J.P. Kummer

R.E. Urban

A. Fiksel

J.W. Thornton

N.J. Blair

P.M. Williams

D.E. Bradley

T.P. McDowell

M. Kummert

Additional contributors who developed components that have been included in the Standard
Library are listed in Volume 5.

Contributors to the building model (Type 56) and its interface (TRNBuild) are listed in Volume 6.

Contributors to the TRNSYS Simulation Studio are listed in Volume 2.

83

TRNSYS 16 Programmer's Guide

84

TRNSYS 16 Programmer's Guide

Table of contents
8. PROGRAMMER'S GUIDE

89

8.1. Introduction

89

8.2. How to update TRNSYS 15 components

89

8.2.1. Steps to use TRNSYS 15 components in "legacy mode"

810

8.2.1.1. Export the name of the routine (Typennn)

810

8.2.1.2. Version sign the Type

810

8.2.1.3. Additional considerations

811

8.2.2. Instructions to convert TRNSYS 15 Types to the TRNSYS 16 standard

812

8.2.2.1. Export the name of the routine (Typennn)

812

8.2.2.2. Special calls identified by INFO(7:8) and INFO (13)

812

8.2.2.3. Double precision variables

814

8.2.2.4. Access to global constants: the "TrnsysConstants" module

815

8.2.2.5. Access to Kernel variables: Access Functions and the "TrnsysFunctions" module

816

8.2.2.6. Data storage (the "S array")

817

8.2.2.7. The "Messages" subroutine

817

8.2.2.8. Utility subroutine calls

818

8.3. How to create new components

8.3.1. Fundamentals
8.3.1.1. The INFO Array and component control

819
819
820

8.3.2. The first line

821

8.3.3. DLL Export

822

8.3.4. USE Statements

822

8.3.5. Variable Declarations

823

8.3.6. Data Statements

824

8.3.7. Global Constants and Variables

825

8.3.8. Version Signing

825

8.3.9. Last Call Manipulations

826

8.3.10. Post Convergence Manipulations

827

8.3.10.1. Resetting counters

827

8.3.10.2. Updating storage variables

827

8.3.11. Initialization Call Manipulations

828

8.3.12. TIME = TIME0 Manipulations

829

8.3.13. Multiple Unit Manipulations

831

8.3.14. Every Time Step Manipulations

832

8.3.14.1. Retrieve Stored Values

832

85

TRNSYS 16 Programmer's Guide

8.3.14.2. Retrieve Input Values

832

8.3.14.3. Perform Calculations

833

8.3.14.4. Set Storage Values

834

8.3.14.5. Set Outputs

834

8.3.14.6. Return Control

834

8.3.15. Complete Code

835

8.3.16. Including your new component in an external DLL

837

8.4. Reference
8.4.1. Global constants

839

8.4.1.1. Constants settings limits on TRNSYS simulations

839

8.4.1.2. Constants used for string length

840

8.4.2. Access functions

86

839

841

8.4.2.1. function getMaxDescripLength( )

841

8.4.2.2. function getmaxFileWidth( )

841

8.4.2.3. function getMaxLabelLength( )

841

8.4.2.4. function getMaxPathLength( )

842

8.4.2.5. function getnMaxStorageSpots( )

842

8.4.2.6. function CheckStability( )

842

8.4.2.7. function ErrorFound( )

843

8.4.2.8. function getConvergenceTolerance( )

843

8.4.2.9. function getDeckFileName( )

843

8.4.2.10. function getFormat(i,j )

843

8.4.2.11. function getLabel(i,j)

844

8.4.2.12. function getListingFileLogicalUnit( )

844

8.4.2.13. function getMinimumTimestep( )

844

8.4.2.14. function getNextAvailableLogicalUnit( )

844

8.4.2.15. function getnMaxIterations( )

845

8.4.2.16. function getnMaxWarnings( )

845

8.4.2.17. function getnTimeSteps( )

845

8.4.2.18. function getNumberOfErrors( )

845

8.4.2.19. function getNumericalSolver( )

845

8.4.2.20. function getSimulationStartTime( )

846

8.4.2.21. function getSimulationStartTimeV15( )

846

8.4.2.22. function getSimulationStopTime( )

846

8.4.2.23. function getSimulationTimeStep( )

846

8.4.2.24. function getTrnsysExeDir( )

846

8.4.2.25. function getTrnsysInputFileDir( )

847

TRNSYS 16 Programmer's Guide

8.4.2.26. function getTrnsysRootDir( )

847

8.4.2.27. function getTrnsysUserLibDir( )

847

8.4.2.28. function getVariableDescription(i,j)

847

8.4.2.29. function getVariableUnit(i,j)

848

8.4.2.30. function getVersionNumber( )

848

8.4.2.31. function LogicalUnitIsOpen(LU)

848

8.4.3. The INFO array Typical calling sequence

849

8.4.3.1. INFO(1:5)

849

8.4.3.2. INFO(6)

849

8.4.3.3. INFO(7)

850

8.4.3.4. INFO(8)

850

8.4.3.5. INFO(9) and Types calling sequence

851

8.4.3.6. INFO(10)

852

8.4.3.7. INFO(11)

852

8.4.3.8. INFO(12)

852

8.4.3.9. INFO(13)

852

8.4.3.10. Typical calling sequence for an iterative Type

852

8.4.4. Utility subroutines

855

8.4.4.1. Calling External Programs

855

8.4.4.2. DFIT (FIT)

856

8.4.4.3. DIFFERENTIAL_EQN (DIFFEQ)

859

8.4.4.4. DynamicData (DYNDATA and DATA)

861

8.4.4.5. ENCLOSURE (ENCL)

864

8.4.4.6. FLUID_PROPS (FLUIDS)

866

8.4.4.7. LINKCK

869

8.4.4.8. MATRIX_INVERT (INVERT and DINVRT)

871

8.4.4.9. MESSAGES

872

8.4.4.10. ParRead

874

8.4.4.11. PSYCHROMETRICS (PSYCH)

875

8.4.4.12. RCHECK

878

8.4.4.13. REWIND

884

8.4.4.14. SOLCEL

885

8.4.4.15. STEAM_PROPS (STEAM)

887

8.4.4.16. Storage of Data Between Time Steps

889

8.4.4.17. TABLE_COEFS (TABLE)

892

8.4.4.18. TAU_ALPHA (TALF)

894

8.4.4.19. TYPECK

896

87

TRNSYS 16 Programmer's Guide

8.4.4.20. VIEW_FACTORS (VIEW)

8.4.5. Solving Differential Equations

897
8103

8.4.6. Use of the Alternate Return (RETURN 1)

8104

8.4.7. Use of the ICNTRL Array

8105

8.4.8. Converting Types from earlier TRNSYS Versions

8107

8.4.8.1. Converting components from version 13.x to version 14.1

8107

8.4.8.2. Converting components from version 14.1 to version 14.2

8107

8.4.8.3. Converting components from version 14.2 to version 15.x

8108

8.4.8.4. Converting components from version 15.x to version 16.x

8108

8.4.9. Instructions to rebuild TRNDll.dll with Compaq Visual Fortran 6.6

8.4.9.1. Cvf66 Workspace

8109

8.4.9.2. TRNDll project

8109

8.4.9.3. Project settings

8109

8.4.9.4. Project files

8111

8.4.9.5. Building the TRNSYS DLL (TRNDll.dll)

8111

8.4.9.6. Debugging TRNSYS

8112

8.4.9.7. Known issues with CVF 6.6

8112

8.4.10. Instructions to add a new project to the CVF66 workspace

8114

8.4.10.1. Cvf66 workspace

8114

8.4.10.2. New Project

8114

8.4.10.3. Project settings

8114

8.4.10.4. Project files

8115

8.4.10.5. Building the new DLL

8116

8.4.10.6. Troubleshooting multiple DLL's

8116

8.4.11. Instructions to rebuild TRNDll.dll with Intel Visual Fortran 8.1

8117

8.4.11.1. Ivf81-Mvs2003 Solution

8117

8.4.11.2. TRNDll project

8117

8.4.11.3. Project settings

8118

8.4.11.4. Project files

8119

8.4.11.5. Building TRNDll.dll

8120

8.4.11.6. Known issues with IVF 8.1

8120

8.5. References

88

8109

8121

TRNSYS 16 Programmer's Guide

8. PROGRAMMER'S GUIDE
8.1. Introduction
There are 3 main sections in this guide:
Section 8.2 explains how to update TRNSYS 15 components in order to use them in
TRNSYS 16. All TRNSYS 15 Types can be used in legacy mode after a few steps that do
not require any programming knowledge.
Section 8.3 describes the process of creating a new component from scratch
Section 8.4 Provides detailed reference on the TRNSYS kernel:
Calling sequence of Types and structure of the INFO array (8.4.3)
Available global constants (8.4.1). Some constants are especially interesting for
programmers since they set standard string lengths in TRNSYS (Fortran 90 does not
handle variable-length strings), see section 8.4.1.2.
Access functions that allow Types to access kernel variables (8.4.2)
Utility subroutines (0)
Instructions to recompile the TRNSYS DLL and add new DLL's to UserLib using different
compilers (Sections 8.4.9, 8.4.10 and above)

8.2. How to update TRNSYS 15 components

Existing TRNSYS 15 components can be updated to TRNSYS 16 in two ways:
After 2 easy steps that do not require any programming knowledge, you can run
existing Types in a "legacy TRNSYS 15" mode, where all the information exchange with
the Kernel is identical to what was happening in TRNSYS 15. This implies that you will not
benefit from some of the new features:
General use of double precision
Added security and user-friendliness thanks to access functions, etc.
Drop-in DLL's: Types running in "legacy TRNSYS 15 mode" must be linked to the
main TRNSYS DLL (TRNDll.dll)
You will also have to be careful if you are interested in the system response just after
initialization: Because of the change in the definition of the initial time step, TRNSYS 15
Types might give an incorrect answer just after the simulation start.
You can also fully convert an existing component to the TRNSYS 16 standard by
following a systematic procedure. The difficulty of that procedure and the time required to
perform the conversion will depend on the way the existing Type is coded: heavy access to
Kernel variables through common blocks, use of Fortran functions specific to single precision
variables, etc.

89

TRNSYS 16 Programmer's Guide

8.2.1.

Steps to use TRNSYS 15 components in "legacy

mode"

The following 2 steps are required in order to use TRNSYS 15 Types in TRNSYS 16. If you want
to use your Type in a "TRNSYS 15 legacy mode", this is all you need to do. If you want to fully
convert your Types to the TRNSYS 16 standard, you will find additional instructions in section
8.2.1.3. Note that the steps listed in this section are also required to fully convert existing
TRNSYS 15 Types.

8.2.1.1.

Export the name of the routine (Typennn)

In order for your component to be found by the TRNSYS Kernel, it needs to broadcast its name.
This is done by adding the following syntax to the component just underneath the subroutine line
at the top of the file.
!DEC$ATTRIBUTES DLLEXPORT :: TYPEnnn

Where nnn is the Type number. Thus Type205 would have the following line added:
!DEC$ATTRIBUTES DLLEXPORT :: TYPE205

It is important that the syntax begin in the leftmost column. In other words, do not put any
tabs or spaces to the left of the first exclamation point. As far as the Fortran compiler itself is
concerned, the line that has just been added is a comment. However, the pre-processor
recognizes the syntax and includes the Type in a file called an export library. Essentially this
means that the subroutine will be accessible by other subroutines not contained within the same
DLL.
Note: The line here above is required due to the new calling mechanism in TRNSYS 16, even
though TRNSYS 15 Types have to be linked to the main DLL. The "DLLEXPORT" directive is
required but it does not mean your Type will be usable in an external DLL.

8.2.1.2.

Version sign the Type

A number of structural changes have been made to TRNSYS components. Types in TRNSYS 16
are called in a different manner and with new INFO array codes, allowing them much more
flexibility in when and how they perform their calculations. Many of these changes are not
compatible with Types written for TRNSYS 15 or earlier so to avoid forcing users to completely
rewrite all their components, TRNSYS 16 contains two parallel calling structures, one for new
TRNSYS 16 components, and one for legacy components. In order for TRNSYS to determine the
manner in which it should call a component, each component must be "signed" with a version
number for which it was written. To sign your component, you need to add the following syntax to
it, at the top of the executable section (just below the variable declarations)
! Set the version information for TRNSYS
if (INFO(7) == -2) then
INFO(12) = 15
return 1
endif

By adding the above, you have signaled to TRNSYS that your component was written using the
TRNSYS 15 calling structure. TRNSYS will treat it accordingly. There should be no further
changes that you need to make to your Type in order for it to run under version 16.

810

TRNSYS 16 Programmer's Guide

8.2.1.3.

Additional considerations

If your Type refers to "param.inc", the include file that declares global TRNSYS 15 constants, you
need
to
update
the
path
to
that
file.
Param.inc
is
now
located
in
%TRNSYS16%\SourceCode\Include (%TRNSYS16% is your installation directory). Note that
Param.inc is obsolete and that TRNSYS 16 Types should use the global constants declared in
the TrnsysConstants module (see section 8.2.2.4 for details).

811

TRNSYS 16 Programmer's Guide

8.2.2.

Instructions to convert TRNSYS 15 Types to the

TRNSYS 16 standard

The steps required in order to convert an existing Type to the TRNSYS 16 standard are given in
the following sections. The first two steps are the same as the steps required to run TRNSYS 15
Types in "legacy mode" (except that the Type is signed as a "Version 16" Type). Please note that
if you have a Type that was written for a TRNSYS version prior to TRNSYS 15, steps for
converting those Types are listed in the Reference section of this manual.

8.2.2.1.

Export the name of the routine (Typennn)

See section 8.2.1.1 here above for details. You need to add the following syntax to the
component just underneath the subroutine line at the top of the file.
!DEC$ATTRIBUTES DLLEXPORT :: TYPEnnn

Where nnn is the Type number. Thus Type205 would have the following line added:
!DEC$ATTRIBUTES DLLEXPORT :: TYPE205

It is important that the syntax begin in the leftmost column. In other words, do not put any
tabs or spaces to the left of the first exclamation point.

8.2.2.2.

Special calls identified by INFO(7:8) and INFO (13)

Note that detailed explanations of the INFO array and a full example of a typical calling sequence
for an iterative call are provided in section 8.4.3. The following sections just outline the changes
to existing TRNSYS 15 Types.

VERSION SIGN CALL (INFO(7) = -2)

See section 8.2.1.2, here above, for details. To sign your component, you need to add the
following syntax to it, at the top of the executable section (just below the variable declarations):
! Set the version information for TRNSYS
if (INFO(7) == -2) then
INFO(12) = 16
return 1
endif

By adding the above, you have signaled to TRNSYS that your component was written using the
TRNSYS 16 standard.

INITIALIZATION (PRE-SIMULATION) CALL (INFO(7) = -1)

The INFO(7) = -1 call is still intended to allow components to perform initializations, as in
TRNSYS 15. However, due to the new specification of TIME in TRNSYS, the manipulations
required at INFO(7) = -1 have changed. The manipulations that your component should do at
INFO(7) = -1 are the following:
Set INFO(6) to the number of output spots needed by the component.
Set INFO(9) so that your Type will be called in the proper manner (see section 8.4.3 for
details).

812

TRNSYS 16 Programmer's Guide

Handle data storage initialization:

Set INFO(10) to the number of storage spots required in the single precision storage
structure (in other words, uses the S array). (Deprecated)
Call SetStorageSize to reserve space in the double precision storage structure. See
section 8.2.2.6 for more information.
Call TYPECK to have TRNSYS check whether the correct number of Inputs, Outputs and
Parameters were specified in the deck.
Call RCHECK to have TRNSYS check the units of the Input Output connections in the deck.
return 1: no other manipulations should be made during this call.
Note that reading parameters and process them can be realized either during this call or during
the initial time step (see "TIME = TIME0 call" here below). You should be careful to handle all
operations that must be done only once (e.g. file opening) in only one of those initial calls.

NEW SIGNIFICANCE OF THE FIRST TIME STEP (TIME = TIME0)

In TRNSYS 16, the time specified in the simulation cards ("Simulation Start Time" in the
Simulation Studio) is the exact moment at which the simulation starts. This is different from
TRNSYS 15, where the so-called simulation start was the time at the end of the first time step. In
TRNSYS 16, components are called at TIME = TIME0 and they should just output their initial
values at that call. There are no iterations for that time step.
The manipulations that a component should make at TIME = TIME0 are:
read parameter values and set them to local variable names
check parameters for validity. For example, you might want to check to see if the specified
specific heat is negative and generate an error if it is. Note that
set outputs to initial conditions or 0
return 1: no other manipulations should be made during this call.

"POST CONVERGENCE" CALL (INFO(13) = 1)

All components that are signed as having been written for version 16 are called once more at the
end of every time step after convergence has been reached. This modification was made in order
to simplify the use of storage (the TRNSYS 15 S array) and other operations that must be done
after convergence has been reached, such as printing or integration. If your component does not
use storage and is not a printer or an integrator, then you may simply add the lines:
!

Perform post-convergence operations

if(INFO(13) > 0) then
return 1
endif

If your component does use storage, then you should add lines that set up your local storage
array, then send the values to global storage using a call to the setStorageVars subroutine as
explained in section 8.2.2.6.

FINAL CALL AFTER THE SIMULATION IS DONE (INFO(8) = -1)

This call has not changed between TRNSYS 15 and 16. At the very end of a TRNSYS simulation,
each Type is called one last time with INFO(8) = -1. If your Type does not contain any handling of
INFO(8) = -1, it will simply run through its calculations one last time and probably no harm will

813

TRNSYS 16 Programmer's Guide

come of it. However, it is a good idea to specifically handle INFO(8). You can add the following
lines if your component does not have to do anything:
!

Perform last call manipulations

if(INFO(8) == -1) then
return 1
endif

You may want your components to actually do something at the end of the simulation. They
could, for instances print a message to the list file or close logical units that were used during the
simulation. Standard component Type 22 (Iterative Feedback controller) performs end of the
simulation manipulations that you might use as an example.
The INFO(8)=-1 call also occurs if the simulation terminates with a fatal error. In that case, the
component that generates the error returns control to TRNSYS, which calls all components one
last time in order to perform their "end of simulation" operations. You can check if the very last
call occurs because of an error or as part of the normal simulation process by calling
getNumberOfErrors(). Some end of simulation operations are unnecessary or might crash
TRNSYS if the simulation ends with a fatal error. E.g. Type 22 handles INFO(8)=1 as follows:
if (info(8) == -1) then
! Exit immediately if this call is the result of a fatal error
if (getNumberOfErrors() > 0) then
return 1
! Otherwise, print the nb of "stuck" timesteps
else
...
printing manipulations
...
endif
return 1
! Exit
endif

Notes:
If the simulation ends without errors, the INFO(8) = -1 call happens after the user has
allowed the simulation to terminate by clicking on the "yes" or "continue" button at the end
of the simulation
The lines of code here above should be placed before the normal instructions so the
return occurs before those instructions are executed.

8.2.2.3.

Double precision variables

One of the other focuses of version 16 development was to move away from single precision
variables toward double precision variables. Four of the arguments sent to each Type are now
double precision (TIME, PAR, T, and DTDT). Previously only XIN and OUT were double
precision. You will need to go through and declare TIME, PAR, T and DTDT as double precision
instead of real. It is also recommended that you do this for all your variables. If you prefer to keep
your variables single precision, it is important that you help Fortran make the conversion
correctly. The following incorrect code will not be caught by the Fortran compiler and may result
in incorrect answers:
RealVariable = DoublePrecisionVariable

The correct way of performing the above is the following:

RealVariable = sngl(DoublePrecisionVariable)

To go the other way, you need the following:

DoublePrecisionVariable = dble(RealVariable)

814

TRNSYS 16 Programmer's Guide

8.2.2.4.

Access to global constants: the "TrnsysConstants"

module

TRNSYS 16 Types should access global constants through the "TrnsysConstants" data module.
The available constants are described in section 8.4.1. Many new constants have been added in
order to promote consistency in TRNSYS (e.g. by making sure strings are sized in a consistent
way), and the existing constants previously found in the "param.inc" file have been transferred
into the new data module. Note that "param.inc" is still included in the TRNSYS distribution for
backwards compatibility, but it should not be used by TRNSYS 16 Types. To use the
"TrnsysConstants" module in a Type, add the following line at the top of the declaration section
(i.e. just below the !DEC$ATTRIBUTES instruction):
use TrnsysConstants

This will declare all the constants in the module in your Type. Alternatively, if you only want to use
the constant defining the maximum length of a Label string (maxLabelLength), you can use the
following syntax:
use TrnsysConstants, only: maxLabelLength

This will only declare the constant(s) listed after the "only" keyword. You can list several
constants in one "only:" instruction by using a comma-separated list.

A NOTE ON STRINGS
Fortran 77/90/95 do not provide variable-length strings, so each individual string variable must be
sized according to its expected maximum length. This can cause problems when different parts of
the code expect different maximum lengths for the same string. If your Type manipulates strings,
we strongly recommend that you check section 8.4.1 to see if a suitable constant is already
available for the size of your string, rather than using a new constant or a hard-coded value.
It is also recommended to use the generic "a" format when printing strings, rather than "ann"
where nn is the length of the string. Where an old print instructions for a variable descriptor
(column title in a printer) might say:

1000

character*10 myString
myString = 'Hello world'
write(*,1000) myString
format(a20)

The recommended way is now:

use TrnsysConstants
! defines integer, parameter ::maxDescripLength
character (len=maxDescripLength) :: myString
myString = 'Hello world'
write(*,'(a)') trim(myString)

Note that the generic "a" format does not work for character arrays (which are not exactly the
same as strings). In that case it is necessary to adjust the format instruction with a variable, e.g.:
use TrnsysConstants
! defines integer, parameter ::maxDescripLength
character (len=100) :: myFormatString
character :: myString(maxDescripLength)
! myString is a character array,
not a string
write(myFormatString,'(i)') maxDescripLength
myFormatString = '(' // trim(adjustl(myFormatString)) // 'a1)'
myString = 'Hello world'
write(*,myFormatString) trim(myString)

We recommend that you use string variables rather than character arrays when possible.

815

TRNSYS 16 Programmer's Guide

8.2.2.5.

Access to Kernel variables: Access Functions and the

"TrnsysFunctions" module

WHY ACCESS FUNCTIONS?

TRNSYS 16 Types should access global (Kernel) variables through access functions only.
Access functions allow Types to use Kernel variables in a much safer way than the Common
Blocks that were used in TRNSYS 15. Common Blocks are inherently dangerous because they
did not ensure name nor even type consistency for Kernel variables among the Types. There is
also no way of preventing a Type from changing the simulation stop time, or even the current
simulation time, with Common Blocks.
The principle of Access Functions is to provide functions that allow access to Kernel variables in
the way they are intended to be used. Foe example, a function is provided to "Get" the simulation
time step, but no function is provided to change it.
All access functions are listed in section 8.4.2.

HOW TO USE ACCESS FUNCTIONS

First, the module with all the function declarations must be Used in the Type:
use TrnsysFunctions

That line should be inserted with the other "Use" instructions at the top of the declaration section.
Here again, it is possible to add "only:" instructions if only some of the functions are needed and if
name conflicts may occur (it is strongly discouraged to re-use TRNSYS function names in
Types)
Then, when a Kernel variable is needed, it can be accessed with the corresponding function. For
example, the simulation time step (DELT variable in the Kernel) is accessed through the
"getSimulationTimeStep()" function. Let's say a Type is integrating a value x over time, the result
being put in "integral".

TRNSYS 15

subroutine Type205(TIME,XIN,OUT,T,DTDT,PAR,INFO,ICNTRL,*)
common/SIM/TIME0,TFINAL,DELT,IWARN
real TIME0,TFINAL,DELT
integer IWARN

real x,s

s = s+x*DELT

subroutine Type205(TIME,XIN,OUT,T,DTDT,PAR,INFO,ICNTRL,*)
use TrnsysFunctions

TRNSYS 16

double precision :: delt, x, s

delt = getSimulationTimeStep()

s = s+x*delt

NOTE ON EXTERNAL DLL'S

816

TRNSYS 16 Programmer's Guide

The TRNSYS 15 Common Blocks are still present in TRNSYS 16 in order to allow existing Types
to run in "legacy mode". However, they are not exported from the main DLL (TRNDll.dll). If you
do not replace Common Blocks with the use of Access Functions, you will not be able to use your
Type in an external DLL.

8.2.2.6.

Data storage (the "S array")

STORAGE IN TRNSYS 15
One of the most unintuitive utilities in TRNSYS 15 was storage using the S array. The underlying
idea of storage is that sometimes you want your calculations not to be based upon the value of a
variable at the last iteration but the value of that variable at the end of the last time step. To use
the S array, one would bring it in as part of a COMMON block and declare how many places were
needed by setting INFO(10) to that number. Then every time a stored variable was needed, the
user would get a pointer into the S array by setting INFO(10) to a local integer variable (such as
ISTORE), then would start pulling out the stored values by referencing them as S(ISTORE),
S(ISTORE+1) etc. At the end of a time step, the user would do the reverse, again getting a
pointer into the S array, then setting its values to the correct local variables.
The problem was that Types did not know when it was the end of a time step, only that a new
time step had begun. Consequently, users had to notice that a new time step had started, then
update their S array with the most recent values (from the end of the last time step).In addition,
there was no check that a given Unit was not writing values in parts of the S array that had been
allocated to a different Unit.

STORAGE IN TRNSYS 16
In TRNSYS 16, the S array still exists in the COMMON block /STORE/ for legacy components.
However, three new utility subroutines were created, allowing for a hopefully more intuitive
approach to variable storage. The functions are described in details in section 8.4.2.1.
At the initialization call (INFO(7) = -1), the Type should call SetStorageSize to allocate
storage for the current Unit.
When the user wants to set storage at the end of a time step, a call is made to the
setStorageVars subroutine when INFO(13) = 1
When the user wants to obtain a previously stored value, a call is made to the
getStorageVars subroutine. This usually occurs during the first call of a new time step,
identified by INFO(7) = 0.

8.2.2.7.

The "Messages" subroutine

One of the enhancements in TRNSYS 16 is the introduction of a Messages subroutine that

handles all notices and error messages that component print to the listing and log files. This
makes it easier to parse a listing or log file for errors and it ensures a consistent handling of
messages by the Simulation Studio's "Error Manager".
Where Type previously wrote error messages to the listing file (accessed throught the logical unit
LUW in the LUNITS Common Block), they should now call the Messages subroutine:
call Messages(errorCode,message,severity,unitNo,typeNo)

817

TRNSYS 16 Programmer's Guide

errorCode is a standard TRNSYS error number (if available), message is the string that has to be
printed, severity indicates the severity of the message (e.g. warning or error) and UnitNo and
TypeNo indicate the calling unit and Type numbers.
Please refer to section 8.4.4.9 for more information on the Messages subroutine.

8.2.2.8.

Utility subroutine calls

One of the focuses of version 16 development was to move away from single precision variables
toward double precision variables. As a consequence, it is important to correctly handle the
arguments to utility subroutines (such as PSYCH) that your components may be calling. In
TRNSYS 15, the call to PSYCH was:
call PSYCH(TIME,INFO,IUNITS,MODE,WBMODE,PSYDAT,0,STATUS,N)

The arguments TIME and PSYDAT were both declared as real (single precision). In TRNSYS 16,
both TIME and PSYDAT should be declared in components as double precision and the call
should be made to the double precision version of PSYCH, which is called Psychrometrics. Thus
the syntax would be:
call Psychrometrics(TIME,INFO,IUNITS,MODE,WBMODE,PSYDAT,0,STATUS,N)

The single precision PSYCH subroutine still exists in TRNSYS 16 but it is merely a shell that
correctly converts single precision arguments to double precision, calls Psychrometrics, and
reconverts the results from double precision back to single precision. Table 8.2.2.8-1 shows the
TRNSYS 15 utility subroutine names in column 1, the corresponding TRNSYS 16 utility
subroutine name in column 2 and the arguments that must be converted from single to double
precision in column 3.
Table 8.2.2.8-1: TRNSYS 15 (single precision) and TRNSYS 16 (double precision) utility routines

Single precision
routine

Double precision
routine

Arguments that need to be changed

from single to double precision

PSYCH

Psychrometrics

TIME, PSYDAT

DATA

DynamicData

X, Y

TALF

TauAlpha

THETA,XKL, REFRIN, ALPHA, RHOD

DIFFEQ

Differential_Eqn

TIME, AA, BB, TI, TF, TBAR

ENCL

Enclosure

SPAR, WPAR, FV

VIEW

ViewFactor

A, B, C, D, E, F, G, H, X3, Y3, Z3

TABLE

Table_Coefs

B, C, D

INVERT

Matrix_Invert

DINVRT

Matrix_Invert

FLUIDS

Fluid_Props

PROP

STEAM

Steam_Props

PROP

Additional information on Utility Subroutines can be found in section 0.

818

TRNSYS 16 Programmer's Guide

8.3. How to create new components

One of TRNSYSs major strengths is the ease with which users may write new components to
expand upon the capabilities of the program. Three features provide the foundation of this
expandability. First is TRNSYSs open architecture. With one exception, all standard components
are provided along with their source code to act as a reference and to act as the basis for adding
new components. Second, all components whether standard components or user written
components are formulated in the same manner and follow the same steps and progression
throughout the code; writing new components becomes a matter of using a template and adding
the appropriate functions, utility calls and equations to write a model of your own. Thirdly, and
perhaps most importantly, the TRNSYS kernel does not impose any hierarchy whatsoever on the
components that are used in a simulation. Nor does it make any assumptions regarding the order
in which components should be solved to simulate the system. Many other simulation packages
do impose such a hierarchy. To take the example of simulating a building, they often solve the
building loads using one part of their kernel, then proceed to solve the system with another part of
their kernel and finally solve the plant aspect of the system last. Many do not revise a previous
step based on calculations in a subsequent step. Thus the system could well not meet the load at
a given time step. However, the building, at the next time step, has no idea that the load was not
met. TRNSYS, by contrast, continues to iterate through all components (whether by successive
substitution or Powells method) until convergence is reached. While this can make controls in a
TRNSYS simulation somewhat more cumbersome, it has a major advantage from the point of
view of someone writing a new component; that is that there is no need for the component writer
to modify the TRNSYS kernel in any way in order to accommodate and include their new
component. There is no need, for example to decide where in the management hierarchy the new
component that you are writing fits best. Components in TRNSYS are automatically called by the
kernel as soon as they are found in a given simulation. By assigning a number to each
component, the TRNSYS kernel anticipates and automatically calls all component numbers that it
finds in a simulation input file. If a particular Type number is found in the input file but is not found
in the compiled code by the TRNSYS kernel, a dummy place holding component is called in its
stead and the simulation ends with an error.
The following sections of this manual will step the new TRNSYS Type programmer through the
process of filling in a Type template in order to create a new model for use in TRNSYS
simulations.

8.3.1.

Fundamentals

At the most fundamental level, a component (referred to as a Type) in TRNSYS is merely a black
box. The TRNSYS kernel feeds inputs to the black box and in turn, the black box produces
outputs. To delve a little deeper, however, TRNSYS makes a distinction between inputs that
change with time and inputs that do not change with time. Examples of inputs that might change
with time are temperature, flow rate, or voltage. Example of inputs that do not change with time
are area or a rated capacity. In the early days of TRNSYS, this distinction was critical because
computing time was very costly and inputs that do not change with time can be set once at the
beginning of a simulation thereby saving a good bit of calculation time. With modern computing
power, this distinction is less critical but TRNSYS retains the designation of time dependent
inputs and time independent inputs. Time dependent inputs are referred to as INPUTS while time
independent inputs are referred to as PARAMETERS.

819

TRNSYS 16 Programmer's Guide

At each iteration and at each time step, a component turns the current values of the INPUTS and
PARAMETERS into OUTPUTS. No distinction is made among OUTPUTS; all OUTPUTS are
assumed to be time dependent and are recomputed by a component whenever appropriate.
TRNSYS has one more input / output distinction; that is DERIVATIVES. Components that solve
differential equations numerically will often have DERIVATIVES as well as INPUTS, OUTPUTS
and PARAMETERS. From the simulation users point of view, the DERIVATIVES for a given
Type specify initial values, such as the initial temperatures of various nodes in a thermal storage
tank or the initial zone temperatures in a multi zone building. At any given point in time, the
DERIVATIVES of a Type hold the results of the solved differential equation. If these
DERIVATIVES are set in a component as OUTPUTS, the simulation user can see, plot and
output these values.

8.3.1.1.

The INFO Array and component control

Much of the control as to what a component is supposed to do at a given point in the simulation
(particularly during initialization) is controlled by the contents of a data structure called the INFO
array. This array carries in its 15 spots, a great deal of information about what is going on
globally in the simulation at a given time. For example, the first INFO spot (INFO(1)) carries the
current UNIT number that the kernel is calling. The second INFO spot (INFO(2)) carries the
current TYPE number that the kernel is calling. The seventh INFO spot has a value of 1 if the
current call from the kernel intends for the called Type to initialize itself. The same spot has a
value of 0 at the very first call to a Type in a given time step. Thereafter it carries the total number
of iterations that have gone by in the current time step. The eighth INFO spot carries the total
number of calls to each Type that have gone by in the simulation at any given time. The thirteenth
INFO spot is normally 0 but is set to 1 after convergence has been reached in a given time step.
A complete reference to the contents of the INFO array is available in section 8.4.3 of this
manual. The INFO array will be used heavily in this section as a method for controlling the
functionality of a Type at various points in a simulation.
The remainder of this section of the Programmers Guide will use the example of a simple liquid
heating device to illustrate the process of writing a new component for TRNSYS.
Consider a simple heater that raises the variable inlet temperature of a flowing liquid to a user
defined temperature Tset. Writing an energy balance on the fluid of inlet temperature, Tin, at a
mass flow rate m:

Q& = m& C p (Tset Tin )

Eq. 8.3.1-1

The first decision that must be made by the Type programmer is: what are going to be the
PARAMETERS, INPUTS and OUTPUTS of the model. Two possibilities become apparent. If the
goal of the component is for the end user to be able to specify an inlet temperature and a desired
outlet temperature, and in return, find out how much energy was required to bring the liquid from

& should obviously be an output, while m& , Tset, and

its inlet condition to its outlet condition, then Q
Tin should be INPUTS. Cp could be designated as either a PARAMETER or as an INPUT
depending upon whether or not the liquid specific heat can be assumed to be constant or whether
it varies significantly with liquid temperature. If, on the other hand, the goal of the component is
for the end user to provide inlet conditions, a control signal and a heater output that is full ON
whenever the control signal is set to ON, then perhaps the output of the model would be Tout,
& , Tin, and Q& .
and the inputs would be m

820

TRNSYS 16 Programmer's Guide

For the purposes of this example, the INPUTS will be Tin, Tset, and m, the OUTPUTS of interest

& , the instantaneous heating rate and the outlet temperature, Tout. The PARAMETERS
will be Q
characterizing the heater will be Tset, Cpf, and Cap, allowing the heater to be capacity limited in
& will also be an OUTPUT
the amount of energy that it can deliver to the liquid stream. Note that m
so that this information is available to other components. We begin at the top of the new
component subroutine.

8.3.2.

The first line

Every Type begins with the same line:

SUBROUTINE TYPEn(TIME,XIN,OUT,T,DTDT,PAR,INFO,ICNTRL,*)

Where n in TYPEn is a number between 1 and 999. To avoid problems with more than one
component being assigned the same number, the following conventions have been adopted:
Type1 150: reserved for library of Standard TRNSYS Components
Type151 200 : reserved for user written components
Type201 300: reserved for add on components commercially available from Transsolar
Type301 400: reserved for add on components commercially available from CSTB
Type401 500: reserved for add on components commercially available from Aiguasol
Type501 Type800: reserved for add on components commercially available from Thermal
Energy System Specialists
Type801 Type999: not reserved
For this example, Type number 151 will arbitrarily be chosen. Thus the first line of the subroutine
should be:
SUBROUTINE TYPE151(TIME,XIN,OUT,T,DTDT,PAR,INFO,ICNTRL,*)

In the subroutine parameter sequence shown above, the following variables are used:
TIME

(double precision) the simulation time

XIN

the double precision array containing the component INPUTS

OUT

the double precision array which the component fills with its
appropriate OUTPUTS

the double precision array containing the dependent variables for

which the derivatives are evaluated in the particular model

DTDT

the double precision array containing the derivatives of T which

are evaluated by the model

PAR

the double precision array containing the PARAMETERS that

characterize the component

INFO

the integer array described in Section 8.4.3

ICNTRL

the integer array described in Section 8.4.7

the alternate return described in Section 8.4.6

821

TRNSYS 16 Programmer's Guide

8.3.3.

DLL Export

With the release of TRNSYS v. 16.0, a great deal of effort was put into allowing Types to exist
either in the standard TRNDll.dll along with the other standard Types or to exist in an External
DLL. The TRNSYS 16 kernel examines the contents of a particular user library directory at the
beginning of each simulation where all external DLLs must be located if they are to be used. It
then loads all files with the extension *.dll, and examines them to find out if there are any
exported Type subroutines contained therein. If it finds exported Type routines it loads them into
memory for the duration of the simulation. The main advantage to end users is that they no longer
have to recompile the TRNDll.dll file when they receive a new component that they wish to use in
their simulation. Instead, they merely need to drop a DLL containing that Type into the
appropriate directory. As a Type programmer, there are certain steps that you must follow in
order to compile your Type and link it as an external DLL. For more information on linking your
component into a DLL, please refer to sections 8.3.16 and 8.4.10 of this manual. From a code
point of view, however, you need only add one line to your component, right at the top of the file
after the SUBROUTINE declaration and before any USE statements. The syntax of the line is:
!DEC$ATTRIBUTES DLLEXPORT :: TYPEn

Where n is the Type number. For our example (Type151) we would have the following line added:
...
!DEC$ATTRIBUTES DLLEXPORT :: TYPE151
...

It is important that the syntax begin in the leftmost column. In other words, do not put any
tabs or spaces to the left of the first exclamation point. The exclamation mark is recognized as a
comment by the Fortran code. It is, however, noticed by the Fortran compiler as an indication that
this subroutine should be exported for access by other DLLs.

8.3.4.

USE Statements

Whereas previous versions of TRNSYS relied heavily on Fortran COMMON blocks, TRNSYS 16
(and later version) Types should access global constants through various Data Modules that are
included with the source code. A Data Module is really nothing more than a file that contains a
series of variable declarations; they are a convenient way of declaring all in one place, the
variables that are accessible by various subroutines (Types, and kernel routines) within a
program. If a variable contained within a data module is needed by a given component, it is
accessed by adding a USE statement directly after the DLL Export instruction (see section 8.3.3
above). The syntax of a USE statement is as follows:
USE NameOfDataModule

The above has the effect of declaring ALL of the variables contained within the module in your
Type. Alternatively, if you only want to use a particular variable that is contained within the
module but do not want access to all of the variables therein, you can use the syntax:
USE NameOfDataModule, ONLY: nameOfVariable1,nameOfVariable2

This will only declare the variable(s) listed after the "ONLY" keyword. The advantage of listing
only those variables to which we want access is that we can reuse the names of global variables
locally as long as we do not request access to them through a USE statement. A complete list of
Data Modules accessible from a given Type or subroutine of a Type are listed in sections 8.4.1,
8.4.2. The most common Data Modules that you will need to access from your Type are:
TrnsysFunctions, TrnsysConstants, and TrnsysData. TrnsysFunctions contains a series of
functions that return commonly useful data such as the simulation start time or time step, or which
perform common functions such as reserving a new logical unit number in a global list for use by

822

TRNSYS 16 Programmer's Guide

an external file accessed by your Type. The TrnsysConstants module contains values that do not
change throughout a simulation but which are again, of common interest. The TrnsysData module
carries a series of data structure sizing constants such as the number of allowable UNITS in a
given simulation or the number of allowable EQUATIONS in a given simulation.
For the purposes of our example, the liquid heater will only need some of the TRNSYSs Access
Functions. Thus, in the USE statement section of the code, we need only add the following line:
...
! USE STATEMENTS
USE TrnsysFunctions
...

As previously mentioned, it is permissible to list the names of those functions that we wish to
access from the TrnsysFunctions module. However, since there is little chance that we will be
writing Functions of our own in this component, the chances that the name of a local function and
the name of a global function will conflict, is minimal.

8.3.5.

Variable Declarations

Having defined the subroutine arguments, exported the component, and given the component
access to global TRNSYS variables, we next need to declare all of the local variables that we
intend to use in the Type. Local variables can be split into two categories; local variables that are
required by every Type, and local variables that are only required by a given Type. Those local
variables that are required by every Type include the array that contains the Types parameters,
the array that contains the Types inputs, the array that contains the Types outputs, the simulation
time, etc. This block of variable declarations can simply be copied from an existing Type and can
be pasted into your new one. You only need make one modification to the following; that is to
change the values of NP, NI, NO, ND, and NS to match the number of parameters, inputs,
outputs, derivatives, and storage spots that will be required for your Type. The block of local
variables required by every Type is as follows.
...
IMPLICIT NONE !FORCE DECLARATION OF ALL VARIABLES
!

TRNSYS DECLARATIONS
DOUBLE PRECISION XIN,OUT,TIME,PAR,T,DTDT,TIME0, DELT,STORED
INTEGER*4 INFO(15),NP,NI,NOUT,ND,IUNIT,ITYPE,ICNTRL
CHARACTER*3 OCHECK,YCHECK

!
!

SET THE MAXIMUM NUMBER OF PARAMETERS(NP),INPUTS(NI),OUTPUTS(NO),

DERIVATIVES(ND), AND STORAGE SPOTS(NS) THAT MAY BE SUPPLIED FOR THIS TYPE
PARAMETER (NP=a,NI=b,NO=c,ND=d,NS=e)

REQUIRED TRNSYS DIMENSIONS

DIMENSION XIN(NI),OUT(NO),PAR(NP),YCHECK(NI),OCHECK(NO),STORED(NS)
...

For the purposes of our liquid heater component, NP will be set to 4 (four parameters are
required), NI will be set to 5 (five inputs are required), NO will be set to 5 (five outputs will be
calculated), ND will be set to 0 (no derivatives are required), and NS will be set to 0 (no storage
spots are required).
With the local variables that apply to all components declared, we next declare the variables that
pertain only to our liquid heater. From a practical point of view, there is no way to anticipate and
simply enter all of the variables that will be needed during the course of developing a new
component. The following list is almost always created as the Type is being written and as local
variables are needed.

823

TRNSYS 16 Programmer's Guide

...
LOCAL VARIABLE DECLARATIONS
INTEGER IGAM
DOUBLE PRECISION TIN, & !temperature of fluid at heater inlet [C]
TOUT, &
!temperature of fluid at heater outlet [C]
TBAR, &
!average temperature of fluid in heater [C]
TAMB, &
!ambient temperature of heater surroundings [C]
TSET, &
!heater setpoint temperature [C]
TON, &
!set temporarily to TOUT before check on TOUT>TSET [C]
QMAX, &
!heater capacity [kJ/hr]
QAUX, &
!required heating rate [kJ/hr]
QLOSS, & !rate of thermal losses to surroundings [kJ/hr]
FLOW, &
!fluid flow rate through heater [kg/hr]
CP, &
!fluid specific heat [kJ/kg.K]
HTREFF, & !heater efficiency [-]
UA, &
!overall loss coeff. for heater during operation [kJ/hr.K]
QFLUID
!rate of energy delivered to fluid [kJ/hr]
...

8.3.6.

Data Statements

It is advantageous in some Types to set some constants that will be used throughout the Type
outside of the executable portion of the code using a DATA statement. The advantage of doing so
is that by defining a constant in a DATA statement, that definition is loaded at the simulation start
and never is there a time when simulation time has to be spent setting or resetting those values.
Often times Type programmers will set a variable called PI to the value 3.1415926 in a DATA
statement, or they will set a variable for the conversion factor between degrees and radians in a
DATA statement. Any variable whose value is set in a DATA statement must be declared among
the Local Variable declarations discussed in section 8.3.5 above.
One other common set of constants that are declared in DATA statements before the beginning
of the executable portion of a Type are the three letter codes that set the units of the Types
INPUT and OUTPUT variables. Refer to section 8.4.4.12 for additional information on these three
letter codes.
We have already declared and sized the two CHARACTER*3 arrays that will hold the input and
output unit codes for the liquid heater among the definition of local variables that are required by
all Types (see section 8.3.5). Our list of inputs, their units, and our list of outputs and their units is
as follows:
Table 8.3.1.1-1: Inputs, Outputs, and their Units

Input
Number

Input Description

Units

Output
Number

Output Description

Units

Liquid inlet temperature

Liquid outlet temperature

Liquid inlet mass flow rate

kg/hr

Liquid outlet flow rate

kg/hr

Control signal

0..1

Required heating rate

kJ/hr

Temperature set point

Energy lost to ambient

kJ/hr

Ambient temperature

Energy to liquid stream

kJ/hr

Referring to Table 8.3.1.1-1 and to Table 8.4.4.12-1 through Table 8.4.4.12-22, the following
DATA statements can be added to the Type, thus setting the units of the first input as a
temperature in degrees C, the units of the second input as mass flow rate in kilograms per hour,
the third input as a control signal having a value between 0 and 1, etc.

824

TRNSYS 16 Programmer's Guide

...
DATA STATEMENTS
DATA YCHECK/'TE1','MF1','CF1','TE1','TE1'/
DATA OCHECK/'TE1','MF1','PW1','PW1','PW1'/
...

8.3.7.

Global Constants and Variables

We have now arrived at the executable section of the Type; from here on, each line of the code
will be executed sequentially, starting at the top. It is therefore necessary to put those lines that
should be executed every single time step and every single iteration right at the beginning of the
executable section. In the case of almost all Types, the only thing that needs to be done this
regularly is to retrieve the values of some global variables that are used in controlling what the
Type does at what point in the simulation. Much of this information has already been passed to
the component. Variables such as TIME, and arrays such as INFO, PAR, and XIN are already
passed to us. Others, such as the time step, the simulation start time, and the simulation stop
time are not passed but must be retrieved on an as needed basis. This is the point in the Type
where AccessFunctions will be used most heavily. Almost all Types, this one included, need to
know the simulation start time and the simulation time step. These two variables are used later to
control Type functionality (see section 8.3.12 here below). Again, the local variable names for
these two variables have already been declared in the block of local variables that are required by
every Type. The local variable name for the simulation time step is DELT and the local variable
name for the simulation start time is TIME0 The code that should be added to Type151 is as
follows:
!----------------------------------------------------------------------------!
GET GLOBAL TRNSYS SIMULATION VARIABLES
TIME0
= getSimulationStartTime()
DELT
= getSimulationTimeStep()

8.3.8.

Version Signing

Version signing is one of a series of special calls that is made to a Type when the Type is not
intended to perform any manipulations of inputs or outputs. Once, at the very beginning of a
simulation, a Type needs to log itself as having been written for a given version of TRNSYS.
Throughout TRNSYSs history, new requirements have periodically been added to the
recommended practices in coding components. These requirements reflect new enhancements
that have been added or reflect changes is recommended coding practices that come from the
developers of Fortran compilers. Prior to the release of TRNSYS version 16, it was left to the
Type programmer to simply make the required modifications before being able to use an existing
Type with the new version of TRNSYS. With the release of TRNSYS 16.0, an effort was made to
allow for backwards compatibility in Types. That is that the TRNSYS 16 kernel was designed in
such a way that it was able to deal with Types that were written to conform with the TRNSYS 16
coding requirements, and to deal with Types that had been written for TRNSYS 15 (which have
different coding requirements). In order for the kernel to know how to treat a Type, however, each
Type must log itself as having been written with TRNSYS 16 conventions or as having been
written with TRNSYS 15 (and earlier) conventions. For additional information about the
differences in conventions, please refer to section 8.2 of this manual.
For the purposes of the liquid heater example, we are writing a TRNSYS 16 style component so
we must add lines of code that version sign our Type as complying with the TRNSYS 16 coding
requirements. The code lines that must be added are:
...
!------------------------------------------------------------------------------

825

TRNSYS 16 Programmer's Guide

SET THE VERSION INFORMATION FOR TRNSYS

IF(INFO(7).EQ.-2) THEN
INFO(12)=16
RETURN 1
ENDIF

...

At the very beginning of a TRNSYS simulation, the TRNSYS kernel calls each Type present in
the input file with the seventh spot of the INFO array set to a value of 2. Every Type should
therefore have a line in it that watches specifically for this situation. The only thing that the Type
should do when it realizes that INFO(7) = -2 is to set the twelfth input spot to a value of 16 if the
Type was written using TRNSYS 16 coding conventions or to a value of 15 if the Type was
written using previous coding conventions. If a Type does not handle the INFO(7) = -2 call
correctly (does not set INFO(12) to either 15 or 16) the TRNSYS kernel will not allow the
simulation to proceed and will generate an error, writing it to the simulation list and log files.

8.3.9.

Last Call Manipulations

The second special case call to Types occurs at the very END of each simulation. Whether the
simulation ends correctly or ends in error, each Type is recalled by the TRNSYS kernel before the
simulation shuts down. This call allows Types to perform any last call manipulations that may be
necessary. These may include closing external data files that were opened by the Type during
the course of a simulation, or calculating summary information that may have been collected
during a simulation. The Type that we are writing, does not require any such last call
manipulations, so it is advisable to simply return control directly to the TRNSYS kernel. The code
that is added to Type151 is as follows:
...
!-----------------------------------------------------------------------------!
PERFORM LAST CALL MANIPULATIONS
IF (INFO(8).EQ.-1) THEN
RETURN 1
ENDIF
...

You may want your component to actually do something at the end of the simulation. It could, for
instances print a message to the list file or close logical units that were used during the
simulation. Standard component Type 22 (Iterative Feedback controller) performs end of the
simulation manipulations that you might use as an example.
The INFO(8)=-1 call also occurs if the simulation terminates with a fatal error. In that case, the
component that generates the error returns control to TRNSYS, which calls all components one
last time in order to perform their "end of simulation" operations. You can check if the very last
call occurs because of an error or as part of the normal simulation process by calling
getNumberOfErrors(). Some end of simulation operations are unnecessary or might crash
TRNSYS if the simulation ends with a fatal error. E.g. Type 22 handles INFO(8)=1 as follows:
if (info(8) == -1) then
! Exit immediately if this call is the result of a fatal error
if (getNumberOfErrors() > 0) then
return 1
! Otherwise, print the nb of "stuck" timesteps
else
... printing manipulations
...
endif
return 1
! Exit
endif

826

TRNSYS 16 Programmer's Guide

Notes:
If the simulation ends without errors, the INFO(8) = -1 call happens after the user has
allowed the simulation to terminate by clicking on the "yes" or "continue" button at the end
of the simulation
The lines of code here above should be placed before the normal instructions so the
return occurs before those instructions are executed.

8.3.10. Post Convergence Manipulations

At the end of each time step, each Type in a simulation is recalled with the thirteenth spot in the
INFO( ) array set to a value of 1. During all other calls (during the iterations of a given Type
during a given time step) this value is set to zero. In this case of the liquid heater example, no
post convergence manipulations are needed. Therefore, we can enter the following code at this
point in the component:
...
!-----------------------------------------------------------------------------!
PERFORM POST CONVERGENCE MANIPULATIONS
IF(INFO(13).GT.0) THEN
RETURN 1
ENDIF
...

In more complex components than the one detailed by this example, the post convergence call
allows a number of different manipulations to be performed. They are resetting counters and
updating storage variables.

8.3.10.1. Resetting counters

It is sometimes advantageous to count the number of times that a particular event has occurred
during a given time step. One example might be a controller that counts the number of times that
it has calculated a different value of its control signal during a given time step. After a certain
number of different decisions, the controller sticks and the calculated output state no longer
changes. At the end of the time step after all components have converged, it is necessary to reset
the iteration counter to zero. This could be done as one of the Post Convergence Manipulations
when INFO(13) is set to 1.

8.3.10.2. Updating storage variables

One of the most common Post Convergence Manipulations performed by Types is that of
updating storage variables. For background information on updating storage variables, please
refer to section 8.4.4.16. One common application of updating storage variables at the end of a
time step is for a Type to store two values of a variable; the initial value of the variable (at the
beginning of the time step) on which all calculations at the subsequent time step are based and
the final value of the variable, which is reevaluated at every iteration. At the end of the time step,
when all components have converged, it is necessary to replace the initial value with the
converged final value. The code employed might look like the following:
...
!-----------------------------------------------------------------------------!
PERFORM POST CONVERGENCE MANIPULATIONS
IF(INFO(13).GT.0) THEN
CALL getStorageVars(Stored,2,INFO)

827

TRNSYS 16 Programmer's Guide

OldInitialValue = Stored(1)
NewFinalValue = Stored(2) !NewFinalValue was calculated by the
!component at each iteration in the time
!step and was stored in the Stored(2) spot.
Stored(1) = NewFinalValue
CALL setStorageVars(Stored,2,INFO)
RETURN 1
ENDIF
...

8.3.11. Initialization Call Manipulations

At the very start of each simulation, after all Types have version signed themselves (see section
8.3.8 or section 8.4.3.3 for more information), each Type is called with the seventh spot of the
INFO( ) array set to 1, indicating to all Types that they should initialize themselves. There is a
very specific set of steps that each Type should take at this point in the simulation. With the
release of TRNSYS version 16, the steps taken by a Type during the initialization step changed
from what had been done previously. The most important thing to keep in mind concerning the
initialization call is that a Type should NOT read its parameter list, nor should it perform ANY
calculation of output values. It should perform the following operations
Reserve the required amount of space in the output array using INFO(6). Since we have already
set the expected number of outputs as a Fortran parameter in section 8.3.5, there is no need to
do anything except add a line that says:
INFO(6) = NO

Set how the Type should be called using INFO(9). Most Types will set INFO(9) to a value of 1,
indicating that the Type should be called every iteration whether or not its input values have
changed. Types that are integrators or printers use different INFO(9) settings. Users
programming this kind of Type should refer to section 8.4.3.5 for more information.
Set the number of single precision storage spots required using INFO(10). It is highly
recommended that all components make use of DOUBLE, not SINGLE precision storage
structures. Therefore, in all likelihood, you should simply set INFO(10) to a value of 0.
Set the number of double precision storage spots required by calling the setStorageSize
subroutine. Information on calling setStorageSize can be found in section 8.4.4.16. It is again
important to bear in mind that we are not setting initial values of the storage variables at this time,
we are merely reserving space for later usage.
Call the TYPECK routine to make sure that the TRNSYS user has supplied the correct number of
parameters, inputs, and derivatives in the input file. Again, since we have already set the
expected number of parameters, inputs and derivatives using a Fortran parameter statement in
section 8.3.5, we need only paste in a line of the proper syntax at this point as shown below. The
1 in the fist argument of the call to TYPECK is an indicator of the action that TYPECK should take
in this particular case. For additional information on the arguments to TYPECK and their
meanings, please refer to section 8.4.4.19.
CALL TYPECK(1,INFO,NI,NP,ND)

Call the RCHECK routine to make sure that the units of the outputs that are connected to the
inputs of this Type are correct. In section 8.3.6, we listed three letter codes for the units of each of
the inputs and outputs of the Type151 that we are writing. By calling RCHECK, we are allowing
TRNSYS to check the users connections between other Types and this one to make sure that
they did not inadvertently connect an output with units of power (PW1 for example) to an input
with units of temperature (TE1 perhaps)

828

TRNSYS 16 Programmer's Guide

Return control to the TRNSYS kernel. Again, there are NO iterations performed during this time
step. Therefore there should be NO calculations performed.
The code required for the above steps in the context of Type151 is as follows. Please note that
the following code is quite generic and can be copied for any Type that does not have a variable
number of parameters, inputs, or derivatives allowed in the input file.
...
!-----------------------------------------------------------------------------!
PERFORM INITIALIZATION MANIPULATIONS
IF (INFO(7).EQ.-1) THEN
!retrieve unit and type number for this component from the INFO array
IUNIT=INFO(1)
ITYPE=INFO(2)
!reserve space in the global OUT array
INFO(6) = NO
! this TYPE should be called at each iteration whether or not its
! inputs change
INFO(9) = 1
! reserve space in the single precision storage structure
INFO(10) = 0
! call the Type check subroutine to compare what this Type requires to
! what has been supplied in the input file.
CALL TYPECK(1,INFO,NI,NP,ND)
! call the input-output check subroutine to set the correct input and
!output variable units.
CALL RCHECK(INFO,YCHECK,OCHECK)
! return to the calling program.
RETURN 1
ENDIF
...

8.3.12. TIME = TIME0 Manipulations

Once the initialization steps have been completed, control is returned to the TRNSYS kernel. The
next time that each Type is called is when the simulation time is equal to the simulation start time.
Throughout much of TRNSYSs history (through version 15) there was no call to components at
the simulation start time. After being called with INFO(7) = -1 for initialization (as described in
section 8.3.11) Types were next called at the end of the first time step (TIME = InitialTime +
TimeStep). With the release of TRNSYS 16.0, an important distinction was made by now calling
all Types once after they have been initialized but before time has advanced in the simulation.
The point of this initial time call is to perform the following actions:
First, read the local parameter list and set each parameter to a local variable. As described in
section 8.3.2, one of the data structures passed by the TRNSYS kernel to each Type is a one
dimensional array containing that Types parameter values. Parameter values can (and should)
be read only at two places in a Type. The first is during this time = time0 call before time has
progressed. The other is during the multiple unit manipulations section (discussed below in
section 8.3.13) Only reading parameters at these two stages of the simulation saves a great deal
of computational time because in this manner parameters (which do not change with time) are
read only when they must be. In each of these two sections, each parameter value is set equal to
a local variable name that has been declared as a double precision variable among the Types
local variables as discussed in section 8.3.5. The generalized code for setting a parameter value
to a local variable name is
LocalVariable = PAR(n)

829

TRNSYS 16 Programmer's Guide

Type151 has four parameters, one each for the maximum heating rate, the specific heat of the
working liquid, the overall loss coefficient of the heater, and the heater efficiency. Referring to the
code at the end of this section, each of these four parameters (PAR(1) through PAR(4) is set to a
local variable name: QMAX, CP, UA, and HTREFF respectively.
The second step performed during the time = time0 manipulations section is to check each of
the parameter values for validity. Parameters whose values cannot be negative, or whose values
must be between 0 and 1, can be flagged as an error using a call to the TYPECK subroutine.
Take, for example, the specific heat of the working liquid for the heater. It does not make any
physical sense to allow the user to enter a negative value of specific heat. We could use the
following code to detect and report the problem, assuming that the second parameter is the
specific heat and has been set to a local variable called SpecHeat.
...
SpecHeat = PAR(2)
IF (SpecHeat.LT.0.0) CALL TYPECK(-4,INFO,0,2,0)
...

The -4 argument in the call to TYPECK indicates to TYPECK that we are flagging an incorrect
parameter value and that we want TYPECK to report the error for us using its generic error
message. The 2 argument indicates to TYPECK that it is the second parameter that is a
problem.
Once TYPECK has reported the flagged parameter and printed its error message to the
simulation log and list files, it returns control to the calling Type but unfortunately does not return
an argument that tells the Type that an error message was written. It is good practice for a Type
that has called TYPECK with a bad parameter to immediately return control to the TRNSYS
kernel without completing its TIME = TIME0 manipulations in case the bad parameter could
cause other problems before the end of the call. It is therefore recommended that once all the
parameters have been checked, programmers use the ErrorFound( ) Access Function to
determine whether TYPECK wrote any error messages to the simulation list and log files. For
more information on the ErrorFound( ) Access Function, please refer to section 8.4.2.6. The code
for returning control in this manner is:
...
IF (ErrorFound()) RETURN 1
...

Next, it is necessary to set the initial values of any storage variables that are used by the Type.
This is accomplished simply by assigning local variable values to each of the spots in the array
that is used to transfer information to and from the double precision storage structure, then calling
the setStorageVars subroutine. The code used to set two initial storage values might look like the
following:
...
storageArray(1) = LocalVar1
storageArray(2) = LocalVar2
CALL setStorageVars(storageArray,2,INFO)
...

Since our Type151 does not require any storage variable spots, there is no need to include the
above step in this section of the Type151 code.
The final step performed during the TIME = TIME0 Manipulations section is to set the initial
values of the outputs. It is essential that you set initial values here, your Type should not perform
any of its calculations at this point in time. If it does, then it will likely get one time step ahead of
where it is supposed to be at this point in the simulation. Because of this restriction on performing
calculations, it is often difficult to know what value to choose for an output initial value. A good
default value is zero since that usually indicates to the user that the output has not yet been
calculated. In the case of devices that have fluid flow through them, the outlet temperature and

830

TRNSYS 16 Programmer's Guide

outlet flow rate of the fluid can often be set to the inlet temperature and inlet flow rate
respectively. In the case of Type151, the first two outputs can be set to the input initial values in
this manner. The other outputs cannot be calculated easily and so are simply set to zero.
Below is the entire code for the TIME = TIME0 Manipulations section of Type151.
...
!-----------------------------------------------------------------------------!
PERFORM INITIAL TIMESTEP MANIPULATIONS
IF(TIME.LT.(TIME0+DELT/2.)) THEN
!set the UNIT and TYPE numbers
IUNIT = INFO(1)
ITYPE = INFO(2)
!read parameter values
QMAX
= PAR(1)
CP
= PAR(2)
UA
= PAR(3)
HTREFF = PAR(4)
!check the parameters for problems and RETURN if any are found
IF(QMAX.LT.0.) CALL TYPECK(-4,INFO,0,1,0)
IF(CP.LT.0.)
CALL TYPECK(-4,INFO,0,2,0)
IF(UA.LT.0.)
CALL TYPECK(-4,INFO,0,3,0)
IF((HTREFF.GT.1.).OR.(HTREFF.LT.0.)) CALL TYPECK(-4,INFO,0,4,0)
IF (ErrorFound()) RETURN 1
!perform any required calculations, set the outputs to appropriate
! initial values.
OUT(1) = XIN(1) !outlet temperature = inlet temperature [C]
OUT(2) = XIN(2) !mass flowrate out = mass flowrate in [kg/hr]
OUT(3) = 0.
!required heating rate [kJ/hr]
OUT(4) = 0.
!rate of losses to environment [kJ/hr]
OUT(5) = 0.
!rate of energy delivered to stream [kJ/hr]
RETURN 1
!the first timestep is for initialization - exit.
ENDIF
...

8.3.13. Multiple Unit Manipulations

Only one more special manipulations section is required before we begin actually simulating the
performance of the liquid heater. In this Multiple Unit Manipulations section, we must reread our
components parameter list if and only if there is more than one instance of the component in the
users simulation.
You may have noticed in some of the previous sections that often, one of the first few lines of
code in each section says:
!set the UNIT and TYPE numbers
IUNIT = INFO(1)
ITYPE = INFO(2)

What this does is take the first two spots of the INFO( ) array each and every time the Type is
called and save the current Type number and the current Unit number to local variables ITYPE
and IUNIT respectively. If there is only one instance of a given Type in a simulation, then the
value of IUNIT will always be equal to the value of INFO(1). Conversely, if there is more than one
instance of a given Type in a simulation, then whenever we begin simulating a new instance of
that Type, the value of IUNIT will not match the value of INFO(1). We make use of this very
feature at this stage of the manipulations. If the unit number has changed (IUNIT does not equal
INFO(1) then we know that we need to reread that parameter list so that we are sure that all local
variables are set to the current instance of the Types parameter lists. The code is as follows:

831

TRNSYS 16 Programmer's Guide

...
!-----------------------------------------------------------------------------!
RE-READ THE PARAMETERS IF ANOTHER UNIT OF THIS TYPE HAS BEEN CALLED SINCE
!
THE LAST TIME THEY WERE READ IN
IF(INFO(1).NE.IUNIT) THEN
!reset the unit number
IUNIT = INFO(1)
ITYPE = INFO(2)
!read the parameter values
QMAX
= PAR(1)
CP
= PAR(2)
UA
= PAR(3)
HTREFF = PAR(4)
ENDIF
...

There is no need to recheck the parameters for validity since this was done already during the
TIME = TIME0 manipulations section.

8.3.14. Every Time Step Manipulations

At last we have come to the meat of writing new components. It is in this final section that the
performance of the component is simulated. Because every component is different, the rules for
what steps go in this section become much more flexible. That said, the Every Time Step
Manipulations section can be broken up into four basic categories which proceed in order as
follows.

8.3.14.1. Retrieve Stored Values

At each iteration, a call should be made to the getStorageVars subroutine to retrieve stored
values that are needed in calculations. The call might look like the following:
...
CALL getStorageVars(storageArray,NS,INFO)
LocalVar1 = storageArray(1)
LocalVar2 = storageArray(2)
...

8.3.14.2. Retrieve Input Values

Retrieving input values follows much the same pattern as that of retrieving parameter values. The
inputs are passed to each Type in a double precision, one dimensional array called XIN( ). The
values of this array are typically read to local variable names for convenience and then can be
checked for validity in much the same way as parameters.
Our Type151 has five inputs. However, only two of them have restrictions on their values.
Namely, the flow rate should not be negtive, and the heater control signal cannot be less than
zero or greater than one. The section that reads them from the XIN( ) array, sets the inputs to
local variables and checks them might look like the following.
...
!-----------------------------------------------------------------------------!
RETRIEVE THE CURRENT INPUT VALUES FROM THE XIN ARRAY
TIN = XIN(1)
FLOW = XIN(2)
IGAM = JFIX(XIN(3)+0.1)

832

TRNSYS 16 Programmer's Guide

TSET = XIN(4)
TAMB = XIN(5)
CHECK THE INPUTS FOR VALIDITY AND RETURN IF PROBLEMS ARE FOUND
IF (FLOW.LT.0.)
CALL TYPECK(-3,INFO,2,0,0)
IF ((IGAM.GT.1.).OR.( IGAM.LT.0.))
CALL TYPECK(-3,INFO,3,0,0)
IF (ErrorFound() ) RETURN 1

...

Here the 3 in the TYPECK arguments indicates that we are flagging a bad input value
(previously we used 4 to indicate that we were flagging a bad parameter value.). The 2 and the
3 indicate the input number that is the problem in each case.

8.3.14.3. Perform Calculations

Once all of the input values have been retrieved, the heater performance calculations and control
logic can be performed. Each line in the following code is commented to show its purpose.
!

PERFORM CALCULATIONS
IF (FLOW .GT. 0.) GO TO 10 !if the inlet flow is greater than 0, skip to
!line 10

NO FLOW CONDITION
OUT(1) = TIN
OUT(2) = 0.
OUT(3) = 0.
OUT(4) = 0.
OUT(5) = 0.
RETURN 1

!set the outlet temperature to the inlet temperature

!set the outlet flow rate to 0.
!set auxiliary energy use to 0.
!set heater losses to 0.
!set the energy imparted to the liquid to 0.
!return control to the kernel

!
!
10

FLOW CONDITION
CHECK INLET TEMPERATURE AND CONTROL FUNCTION
IF ((TIN.LT.TSET).AND.(IGAM.EQ.1)) GO TO 20 !if the inlet temperature is
!below the set point temperature
!otherwise, the heater and the
!control signal is set to ON,
!skip to line 20. is OFF.

HEATER OFF CONDITION

TOUT
= TIN
!set the
QAUX
= 0.
!set the
QLOSS = 0.
!set the
QFLUID = 0.
!set the
GO TO 50
!skip to

!
!
!
20

outlet temperature to the inlet temperature

auxiliary energy use to 0.
heater losses to 0.
energy imparted to the liquid to 0.
the output setting section.

HEATER ON CONDITION
calculate the outlet temperature assuming that the heter is on and
running at capacity
TON =(QMAX*HTREFF+FLOW*CP*TIN+UA*TAMB-UA*TIN/2.d0)/(FLOW*CP+UA/2.d0)
!the outlet temperature is the lesser of TON and TSET the heater is
!able to modulate its energy output to reach the setpoint temperature.
TOUT = MIN(TSET,TON)
!calculate the average heater temperature losses are based on the
!average
TBAR = (TIN+TOUT)/2.d0
!calculate the auxiliary energy required, accounting for heater
!efficiency.
QAUX = (FLOW*CP*(TOUT-TIN)+UA*(TBAR-TAMB))/HTREFF
!calcualte the energy lost from the heater, including heater inefficiency
QLOSS = UA*(TBAR-TAMB) + (1.d0-HTREFF)*QAUX
!calculate the energy imparted to the liquid.
QFLUID = FLOW*CP*(TOUT-TIN)

833

TRNSYS 16 Programmer's Guide

8.3.14.4. Set Storage Values

Before exiting the Type at each iteration, it is good practice to update all appropriate storage
variables. Obviously, if one or more of the storage variables is an initial value to be used for all
iterative calculations at a given time step until convergence has been reached then it is not
appropriate to update this value until the INFO(13) = 1 call (see sections 8.3.10 and 8.4.3.9 for
additional information.

8.3.14.5. Set Outputs

In much the same way that we took the TRNSYS kernel provided parameter and input arrays and
set them to local variable names, we now have to set the calculated local variables to the
appropriate TRNSYS output array. The code to do so for Type151 is as follows:
...
!
50

SET OUTPUTS
OUT(1) = TOUT
OUT(2) = FLOW
OUT(3) = QAUX
OUT(4) = QLOSS
OUT(5) = QFLUID

...

8.3.14.6. Return Control

The very last requirement is that a Type return control to the TRNSYS kernel after its
computations are complete. The 1 following the RETURN statement is referred to as an
Alternate Return. In essence, a call to a FORTRAN subroutine can be set up so that under one
set of circumstances, control is returned to the line following the call line. Under other
circumstances, an alternate return to return control to some other point in the calling routine. For
additional information on the alternate return statement, please refer to section 8.4.6.
...
RETURN 1
END SUBROUTINE Type151
...

834

TRNSYS 16 Programmer's Guide

8.3.15. Complete Code

The entire code of Type151 is reproduced here below as a complete reference.
SUBROUTINE TYPE151(TIME,)
!-----------------------------------------------------------------------------!DEC$ATTRIBUTES DLLEXPORT :: TYPE151
! USE STATEMENTS
USE TrnsysFunctions
IMPLICIT NONE !FORCE DECLARATION OF ALL VARIABLES
!

TRNSYS DECLARATIONS
DOUBLE PRECISION XIN,OUT,TIME,PAR,T,DTDT,TIME0,TFINAL,DELT,STORED
INTEGER*4 INFO(15),NP,NI,NOUT,ND,IUNIT,ITYPE,ICNTRL
CHARACTER*3 OCHECK,YCHECK

!
!

SET THE MAXIMUM NUMBER OF PARAMETERS(NP),INPUTS(NI),OUTPUTS(NO),

DERIVATIVES(ND), AND STORAGE SPOTS(NS) THAT MAY BE SUPPLIED FOR THIS TYPE
PARAMETER (NP=a,NI=b,NO=c,ND=d,NS=e)

REQUIRED TRNSYS DIMENSIONS

DIMENSION XIN(NI),OUT(NO),PAR(NP),YCHECK(NI),OCHECK(NO),STORED(NS)

LOCAL VARIABLE DECLARATIONS

INTEGER IGAM
DOUBLE PRECISION TIN, & !temperature of fluid at heater inlet [C]
TOUT, &
!temperature of fluid at heater outlet [C]
TBAR, &
!average temperature of fluid in heater [C]
TAMB, &
!ambient temperature of heater surroundings [C]
TSET, &
!heater setpoint temperature [C]
TON, &
!set temporarily to TOUT before check on TOUT>TSET [C]
QMAX, &
!heater capacity [kJ/hr]
QAUX, &
!required heating rate [kJ/hr]
QLOSS, & !rate of thermal losses to surroundings [kJ/hr]
FLOW, &
!fluid flow rate through heater [kg/hr]
CP, &
!fluid specific heat [kJ/kg.K]
HTREFF, & !heater efficiency [-]
UA, &
!overall loss coeff. for heater during operation [kJ/hr.K]
QFLUID
!rate of energy delivered to fluid [kJ/hr]

DATA STATEMENTS
DATA YCHECK/'TE1','MF1','CF1','TE1','TE1'/
DATA OCHECK/'TE1','MF1','PW1','PW1','PW1'/

!----------------------------------------------------------------------------!
GET GLOBAL TRNSYS SIMULATION VARIABLES
TIME0
= getSimulationStartTime()
DELT
= getSimulationTimeStep()
!-----------------------------------------------------------------------------!
SET THE VERSION INFORMATION FOR TRNSYS
IF(INFO(7).EQ.-2) THEN
INFO(12)=16
RETURN 1
ENDIF
!------------------------------------------------------------------------------

835

TRNSYS 16 Programmer's Guide

PERFORM LAST CALL MANIPULATIONS

IF (INFO(8).EQ.-1) THEN
RETURN 1
ENDIF

!-----------------------------------------------------------------------------!
PERFORM POST CONVERGENCE MANIPULATIONS
IF(INFO(13).GT.0) THEN
RETURN 1
ENDIF
!-----------------------------------------------------------------------------!
PERFORM INITIALIZATION MANIPULATIONS
IF (INFO(7).EQ.-1) THEN
!retrieve unit and type number for this component from the INFO array
IUNIT=INFO(1)
ITYPE=INFO(2)
!reserve space in the global OUT array
INFO(6) = NO
! this TYPE should be called at each iteration whether or not its
! inputs change
INFO(9) = 1
! reserve space in the single precision storage structure
INFO(10) = 0
! call the Type check subroutine to compare what this Type requires to
! what has been supplied in the input file.
CALL TYPECK(1,INFO,NI,NP,ND)
! call the input-output check subroutine to set the correct input and
!output variable units.
CALL RCHECK(INFO,YCHECK,OCHECK)
! return to the calling program.
RETURN 1
ENDIF
!-----------------------------------------------------------------------------!
RE-READ THE PARAMETERS IF ANOTHER UNIT OF THIS TYPE HAS BEEN CALLED SINCE
!
THE LAST TIME THEY WERE READ IN
IF(INFO(1).NE.IUNIT) THEN
!reset the unit number
IUNIT = INFO(1)
ITYPE = INFO(2)
!read the parameter values
QMAX
= PAR(1)
CP
= PAR(2)
UA
= PAR(3)
HTREFF = PAR(4)
ENDIF
!-----------------------------------------------------------------------------!
RETRIEVE THE CURRENT INPUT VALUES FROM THE XIN ARRAY
TIN = XIN(1)
FLOW = XIN(2)
IGAM = JFIX(XIN(3)+0.1)
TSET = XIN(4)
TAMB = XIN(5)
!
CHECK THE INPUTS FOR VALIDITY AND RETURN IF PROBLEMS ARE FOUND
IF (FLOW.LT.0.)
CALL TYPECK(-3,INFO,2,0,0)
IF ((IGAM.GT.1.).OR.( IGAM.LT.0.))
CALL TYPECK(-3,INFO,3,0,0)
IF (ErrorFound() ) RETURN 1
!

836

PERFORM CALCULATIONS
IF (FLOW .GT. 0.) GO TO 10

TRNSYS 16 Programmer's Guide

NO FLOW CONDITION
OUT(1) = TIN
OUT(2) = 0.
OUT(3) = 0.
OUT(4) = 0.
OUT(5) = 0.
RETURN 1

!set the outlet temperature to the inlet temperature

!set the outlet flow rate to 0.
!set auxiliary energy use to 0.
!set heater losses to 0.
!set the energy imparted to the liquid to 0.
!return control to the kernel

!
!
10

FLOW CONDITION
CHECK INLET TEMPERATURE AND CONTROL FUNCTION
IF ((TIN.LT.TSET).AND.(IGAM.EQ.1)) GO TO 20

HEATER OFF CONDITION

TOUT
= TIN
!set the
QAUX
= 0.
!set the
QLOSS = 0.
!set the
QFLUID = 0.
!set the
GO TO 50
!skip to

outlet temperature to the inlet temperature

auxiliary energy use to 0.
heater losses to 0.
energy imparted to the liquid to 0.
the output setting section.

!
20

HEATER ON CONDITION
TON =(QMAX*HTREFF+FLOW*CP*TIN+UA*TAMB-UA*TIN/2.d0)/(FLOW*CP+UA/2.d0)
TOUT = MIN(TSET,TON)
TBAR = (TIN+TOUT)/2.d0
QAUX = (FLOW*CP*(TOUT-TIN)+UA*(TBAR-TAMB))/HTREFF
QLOSS = UA*(TBAR-TAMB) + (1.d0-HTREFF)*QAUX
QFLUID = FLOW*CP*(TOUT-TIN)

!
50

SET OUTPUTS
OUT(1) = TOUT
OUT(2) = FLOW
OUT(3) = QAUX
OUT(4) = QLOSS
OUT(5) = QFLUID
RETURN 1
END SUBROUTINE Type151

8.3.16. Including your new component in an external

DLL
This step is covered by section 8.4.10, which explains how to create and configure a new project
that will generate a DLL in the UserLib folder.
Note that if you start from the TRNSYS Studio and generate a skeleton for your Type using the
"Export as Fortran" command, the Studio will automatically create a Project with the correct
settings for you.

837

TRNSYS 16 Programmer's Guide

838

TRNSYS 16 Programmer's Guide

8.4. Reference
8.4.1.

Global constants

8.4.1.1.

Constants settings limits on TRNSYS simulations

Constant

Default

Comment

nMaxUnits

1000

Maximum number of units

nMaxEquations

500

Maximum number of equations

nMaxDerivatives

100

Maximum number of derivatives

nMaxOutputs

3000

Maximum number of Outputs

nMaxParameters

2000

Maximum number of Parameters

nMaxStorageSpots

10000

Number of Storage places in the S array

nMaxFiles

300

Maximum number of files that can be opened by TRNSYS

nMaxDescriptions

750

Maximum number of descriptions in an input file

nMaxVariableUnits

750

Maximum number of input variable types in an input file

nMaxLabels

100

Maximum number of labels in an input file

nMaxFormats

100

Maximum number of format statements in an input file

nMaxChecks

20

Maximum number of chek statements in an input file

nMaxCheckCodes

30

Maximum number of outputs per CHECK statement

nMaxUnitConversions 250

Max. number of unit conversions allowed in units.lab file

nMaxUnitTypes

250

Maximum number of unit types allowed in units.lab file

nMaxCardValues

250

Max. nb. of values allowed in one TRNSYS input file card

E.g. a component with 500 parameters would require
nMaxCardValues to be greater or equal to 500

minTimeStep

0. 1
3600

Minimum Simulation Time Step [in hours] (0.1 second)

nMaxTimeSteps

109

Maximum number of time steps in a simulation (1 billion)

nMaxPlottedTimeSteps 525601

Max. nb. of time steps plotted by the online plotter

Note: This constant is set in TRNExe. Do not edit

nMaxErrorMessages

Maximum number of standard error messages in TrnsysErrors

Note: If you increase this constant, additional messages
will not be initialized. Add lines in TrnsysErrors.f90

1000

839

TRNSYS 16 Programmer's Guide

8.4.1.2.

Constants used for string length

Constant

Default

Comment

maxPathLength

300

Maximum length of variables containing path- and filenames

Note: DFWIN defines MAX_PATH as 260 but Windows XP
can
actually
use
much
longer
pathnames.
ATTENTION: TRNExe must be adapted if this constant is
modified. Users should not change maxPathLength

maxFileWidth

1000

Maximum file width, i.e. maximum length of any line in a text

file that must be read from / written to by TRNSYS
maxFileWidth should be >= maxPathLength. This constant is
also used for strings, e.g. error messages

maxDescripLength

25

Maximum length of a variable description, e.g. descriptions for

printers and plotters

maxVarUnitLength

20

Maximum length of units associated with variables

maxEqnNameLength

20

Maximum length of variable names in equations

maxLabelLength

300

Maximum length of labels. Some labels are file- and

pathnames so a suggested value is maxLabelLength =
maxPathLength

maxMessageLength

800

Maximum length of notices, warnings and error messages

Error messages and any text printed on the same line must
fit within the maximum file width (maxFileWidth).
It is recommended to set maxMessageLength to
maxFileWidth-200

840

TRNSYS 16 Programmer's Guide

8.4.2.

Access functions

This section lists all functions declared in the "TrnsysFunctions" module. Note that the
subroutines used to handle data storage are mentioned here for the record, but detailed
explanations on their use can be found in section 8.4.4.16.
Please note that the first group of access functions (getMaxDescripLength(), etc.) provide access
to global constants. Those constants are declared in the TrnsysConstants module, so Frotranwritten Types can access them more easily through an "use" statement: use TrnsysConstants.
Those access functions are provided for non-Fortran Types.

8.4.2.1.

function getMaxDescripLength( )

[ maxDescripLength is also available through "use TrnsysConstants" ]

An integer function that returns the maximum allowable length of a variable description, e.g. input
descriptions for printers and plotters. The value of the maximum allowable length of a variable
description can be set in the TrnsysConstants file. If the value is modified, the trnlib.dll must then
be recompiled in order for the change to take effect.
Example Usage
INTEGER DescripLen
...
DescripLen = getMaxDescripLength ()

8.4.2.2.

function getmaxFileWidth( )

[ maxFileWidth is also available through "use TrnsysConstants" ]

An integer function that returns the maximum allowable width (in characters or columns) of line in
any text file that must be read by TRNSYS. This includes both the input file and external data files
read by the DynamicData utility routine. The value of the maximum allowable line length can be
set in the TrnsysConstants file. If the value is modified, the trnlib.dll must then be recompiled in
order for the change to take effect.
Example Usage
INTEGER FileWidth
...
FileWidth = getmaxFileWidth()

8.4.2.3.

function getMaxLabelLength( )

[ maxLabelLength is also available through "use TrnsysConstants" ]

An integer function that returns the maximum allowable length (in characters) of variables that
contain pathnames and filenames. The value of the maximum allowable label length can be set in
the TrnsysConstants file. If the value is modified, the trnlib.dll must then be recompiled in order
for the change to take effect.
Example Usage

841

TRNSYS 16 Programmer's Guide

INTEGER LabLen
...
LabLen = getMaxLabelLength()

8.4.2.4.

function getMaxPathLength( )

[ maxPathLength is also available through "use TrnsysConstants" ]

An integer function that returns the maximum allowable length (in characters) of pathnames. The
value of the maximum allowable path length can be set in the TrnsysConstants file. If the value is
modified, the trnlib.dll must then be recompiled in order for the change to take effect.
Example Usage
INTEGER PathLen
...
PathLen = getMaxPathLength()

8.4.2.5.

function getnMaxStorageSpots( )

[nMaxStorageSpots is also available through "use TrnsysConstants" ]

An integer function that returns the maximum total number of storage spots that may be
requested in a given simulation. Note that this function is used by the setStorageSize subroutine
and that therefore it is not necessary for the Type programmer to check whether the required
number of storage spots for his/her Type will exceed the maximum allowable number of storage
spots. The value of the maximum allowable number of storage spots can be set in the
TrnsysConstants file. If the value is modified, the trnlib.dll must then be recompiled in order for
the change to take effect.
Example Usage
INTEGER StorSpots
...
StorSpots = getnMaxStorageSpots()

8.4.2.6.

function CheckStability( )

This function is only used with Solver 1 (Powell's method). You can find more
information on TRNSYS Solvers in Volume 07, TRNEdit (check the section on
the Solver Statement).
An integer function that returns a 1 if the last time step converged and a 0 if the last time step did
not converge. The CheckStability function is used in SOLVER 1, which tries more than one
control strategy and then backs up a time step in order to try something else if it did not find a
stable solution. TRNSYS data reading components need to know that they should not continue
reading the data file but should back up as well.
Example Usage
if (CheckStability() < 1) then
backspace(LU_DATA)
endif

842

TRNSYS 16 Programmer's Guide

8.4.2.7.

function ErrorFound( )

Logical function that returns a value of TRUE if TRNSYS errors have been found. This function
is most useful in Types that call one or more of the TRNSYS Utility subroutines (such as
TYPECK, PSYCHROMETRICS, or STEAM) which may call find and flag errors. For example, say
a user written Type calls TYPECK to indicate that one of the INPUTS had an inappropriate value.
TYPECK would call the TRNSYS utility subroutine MESSAGES, which would print out the error,
would log that an error occurred and would return control to TYPECK. TYPECK would in turn
return control to the Type, which can then avoid the remainder of its calculations by accessing the
ErrorFound function. An example follows in which calculations cease if an illegal value of the
input variables FLOW_CHW, FLOW_CW, or FLOW_HW is found.
Example Usage
IF(FLOW_CHW.LT.0.) CALL TYPECK(-3,INFO,2,0,0)
IF(FLOW_CW.LT.0.) CALL TYPECK(-3,INFO,4,0,0)
IF(FLOW_HW.LT.0.) CALL TYPECK(-3,INFO,6,0,0)
IF( ErrorFound() ) RETURN 1

8.4.2.8.

function getConvergenceTolerance( )

Double precision function that returns the user specified value of a simulations convergence
tolerance (the second argument in the TOLERANCES keyword statement).
Example Usage
DOUBLE PRECISION ErrTol
...
ErrTol = getConvergenceTolerance()

8.4.2.9.

function getDeckFileName( )

A character function that returns the name (NOT including the path) of the input file being run.
This function is useful in printer and other output devices in which results that pertain to a
particular input file are to be stored. The length of the returned string is maxPathLength.
Example Usage
USE TrnsysConstants, ONLY: maxPathLength
CHARACTER (len=maxPathLength) DeckName
...
DeckName = getDeckFileName()

8.4.2.10. function getFormat(i,j )

A character function that returns the jth format statement associated with unit i. The length of the
returned string is maxFileWidth. For example, in certain modes, Type9 takes a single Format
statement.. To access the format of the data file line that is to be read, Type9 contains the
following code:
Example Usage
use TrnsysConstants, only: maxFileWidth
...
character (len=maxFileWidth) :: myFormat
...

843

TRNSYS 16 Programmer's Guide

myFormat = getFormat(INFO(1),1)
...
! Print or read using myFormat
write(luPrint,myFormat) ... variables to be printed

8.4.2.11. function getLabel(i,j)

A character function that returns the jth label of unit i. The length of the returned string is
maxLabelLength.. For example, Type66 takes two LABELs; the first is the location of the EES
executable, the second is the name and location of the EES file that is to be solved. To access
these labels, Type66 contains the following code:
Example Usage:
USE TrnsysConstants, ONLY: maxLabelLength
INTEGER ThisInstance
CHARACTER (len=maxLabelLength) EESLocation,EESFile(10)
...
EESLocation = getLabel(INFO(1),1)
EESFile(ThisInstance) = getLabel(INFO(1),2)

8.4.2.12. function getListingFileLogicalUnit( )

An integer function that returns the logical unit of the TRNSYS list file.
Example Usage
INTEGER LUW
...
LUW = getListingFileLogicalUnit()

8.4.2.13. function getMinimumTimestep( )

Double precision function that returns the minimum allowable TRNSYS time step. The minimum
allowable TRNSYS time step can be set in the TrnsysConstants file. If the value is modified, the
trnlib.dll must then be recompiled in order for the change to take effect.
Example Usage
DOUBLE PRECISION minStep
...
minStep = getMinimumTimestep()

8.4.2.14. function getNextAvailableLogicalUnit( )

Integer function that returns an available logical unit number and logs the logical unit number in a
global list of assigned logical unit numbers. The getNextAvailableLogicalUnit function can be
used to find an available logical unit number for temporary files, scratch files, or for data files
accessed by components that do not require a logical unit number among their list of parameters.
Example Usage
LU = getNextAvailableLogicalUnit()
! Write some data to a temporary file
open(UNIT=LU,FILE='TmpFileForMyType.tmp',status='new')
write(LU,*) ... Data to be written ...

844

TRNSYS 16 Programmer's Guide

close(LU,status='keep')

8.4.2.15. function getnMaxIterations( )

An integer function that returns the maximum allowable number of iterations in one time step
before the simulation continues with the current values and a warning is issued. The maximum
allowable number of iterations can be set in a Limit statement (or in the Studio's Control Cards).
Example Usage
INTEGER maxIter
...
maxIter = getnMaxIterations()

8.4.2.16. function getnMaxWarnings( )

An integer function that returns the maximum allowable number of warnings written to the
simulation list and log files before the simulation stops in error. The maximum allowable number
of warnings can be set in a Limit statement (or in the Studio's Control Cards).
Example Usage
INTEGER maxWarns
...
maxWarns = getnMaxWarnings()

8.4.2.17. function getnTimeSteps( )

An integer function that returns the total number of time steps in the simulation being run.
Example Usage
INTEGER nSteps
...
nSteps = getnTimeSteps()

8.4.2.18. function getNumberOfErrors( )

An integer function that returns the total number of errors that have occurred thus far in the
simulation.
Example Usage
INTEGER NumErrs
...
NumErrs = getNumberOfErrors()

8.4.2.19. function getNumericalSolver( )

An integer function that returns the Solver used by TRNSYS. More information on TRNSYS
solvers can be found in Volume 07, TRNEdit (check the section on the Solver statement).
Handling of discrete control signals is different with Solver 0 and Solver 1. See Type 2 for a n
example.
Example Usage

845

TRNSYS 16 Programmer's Guide

integer TRNSolver
...
if (TRNSolver == 0) then
... Do things required for successive substitution
else
... Do things required for Powell's solver
endif

8.4.2.20. function getSimulationStartTime( )

Double precision function that returns the hour of the year at which the simulation began.
Example Usage
DOUBLE PRECISION TIME0
...
TIME0 = getSimulationStartTime()

8.4.2.21. function getSimulationStartTimeV15( )

Double precision function that returns the hour of the year at which the simulation began. The
returned value is the original start time in a Version 15 deck file (Start time is changed by
TRNSYS 16 to adapt it to the new start time definition)
Example Usage
DOUBLE PRECISION TIME0V15
...
TIME0V15 = getSimulationStartTimeV15()

8.4.2.22. function getSimulationStopTime( )

Double precision function that returns the hour of the year at which the simulation is set to end.
Example Usage
DOUBLE PRECISION TFINAL
...
TFINAL = getSimulationStopTime()

8.4.2.23. function getSimulationTimeStep( )

Double precision function that returns the simulation time step (in hours).
Example Usage
DOUBLE PRECISION DELT
...
DELT = getSimulationTimeStep()

8.4.2.24. function getTrnsysExeDir( )

A character function that returns the path in which the TRNSYS executable (TrnExe.exe) is
located (default = "C:\Program Files\Trnsys16\Exe"). The length of the returned string is
maxPathLength.

846

TRNSYS 16 Programmer's Guide

Example Usage
USE TrnsysConstants, ONLY: maxPathLength
CHARACTER (len= maxPathLength) ExeLoc
...
ExeLoc = getTrnsysExeDir()

8.4.2.25. function getTrnsysInputFileDir( )

A character function that returns the path in which the input file being run is contained. This
function is useful in printer and other output devices in which results that pertain to a particular
input file are to be stored. The length of the returned string is maxPathLength.
Example Usage
USE TrnsysConstants, ONLY: maxPathLength
CHARACTER (len=maxPathLength) DeckLoc
...
DeckLoc = getTrnsysInputFileDir()

8.4.2.26. function getTrnsysRootDir( )

A character function that returns the path to the root TRNSYS directory (default = "C:\Program
Files\Trnsys16 "). The length of the returned string is maxPathLength.
Example Usage
USE TrnsysConstants, ONLY: maxPathLength
CHARACTER (len=maxPathLength) RootDir

RootDir = getTrnsysRootDir()

8.4.2.27. function getTrnsysUserLibDir( )

A character function that returns the path and directory in which all of the loadable user dlls are
located (default = "C:\Program Files\Trnsys16\UserLib"). The length of the returned string is
maxPathLength.
Example Usage
USE TrnsysConstants, ONLY: maxPathLength
CHARACTER (len=maxPathLength) UsrLibDir
...
UsrLibDir = getTrnsysUserLibDir()

8.4.2.28. function getVariableDescription(i,j)

String function that returns the variable descriptor for the jth variable of Unit i. Variable descriptors
are strings that give the "name" of the variable in printers and other output devices. The length of
the returned string is maxDescripLength.
Example usage
character (len=maxDescripLength) :: columnHeader

847

TRNSYS 16 Programmer's Guide

...
do j = 1, nVariables
columnHeader = getVariableDescription(INFO(1),j)
... write column header to a file, etc...
enddo

8.4.2.29. function getVariableUnit(i,j)

String function that returns the variable unit string for the jth variable of Unit i. Variable descriptors
are strings that give the "name" of the variable in printers and other output devices. The length of
the returned string is maxVarUnitLength.
Example usage
character (len=maxVarUnitLength) :: varUnitString
...
do j = 1, nVariables
varUnitString = getVariableDescription(INFO(1),j)
... write units to a file, etc...
enddo

8.4.2.30. function getVersionNumber( )

Integer function that returns the TRNSYS version number (15, 16, etc.). This function is primarily
useful in allowing for backward compatability between versions of a component. Say, for example
that a Type required 14 parameters in an input file written for TRNSYS version 15 but requires 15
parameters in an input file written for TRNSYS version 16. The Type programmer could access
the getVersionNumber function and know whether the input file being run was written for
TRNSYS version 15 or 16, and could read the corresponding parameter list.
Example Usage
INTEGER VERSNUM
...
VERSNUM = getVersionNumber()

8.4.2.31. function LogicalUnitIsOpen(LU)

Logical function that returns true if the Logical unit has been assigned in the deck file and opened
by TRNSYS.
Example usage
logical :: luOk
integer :: lu
...
lu = nint(par(1))
luOk = LogicalUnitIsOpen(lu)
if (luOk) then
... it's OK to print to / read from the file
else
... Probably an error in parameters
endif

848

TRNSYS 16 Programmer's Guide

8.4.3.

The INFO array Typical calling sequence

The INFO array is an integer array of 15 values. It conveys to the component subroutines (Types)
and to the TRNSYS kernel information about the current UNIT. The contents of the INFO array
for a given unit are given in Table 8.4.2.31-1 (see text for detailed explanations).
Table 8.4.2.31-1: INFO array contents

INFO(n)

Role

Unit number

Type number

Number of Inputs in the input file (.dck) for this Unit

Number of Parameters in the input file (.dck) for this Unit

Number of Derivatives in the input file (.dck) for this Unit

Number of outputs (set by the Type)

Number of iterative calls to this Unit in the current time step. Special values: -1, -2

Total number of calls to this unit. Special values: -1

Indicate whether or not the Type depends on the passage of time and tells TRNSYS
where the Type should go in the calling sequence

10

Used in storage management for TRNSYS 15 Types (not used by TRNSYS 16

Types)

11

Indicates number of discrete control variables

12

TRNSYS version for which the Type was written (15 or 16)

13

Indicates when all Units have converged in the current time step

14

Reserved for future use

15

Reserved for future use

8.4.3.1.

INFO(1:5)

The fist 5 spots in INFO are set by TRNSYS from information in the input file (.dck). They are
used to distinguish the current Unit / Type from the other ones and to detect possible
configuration errors with the TYPECK routine. It is possible to check that the number of
parameters in the input file is acceptable, etc.

8.4.3.2.

INFO(6)

INFO(6) indicates the number of outputs used by the Type. A minimum of 20 outputs is reserved
for each unit in the simulation, but INFO(6) should always be set to provide an accurate Trace
and insure efficient error checking.

849

TRNSYS 16 Programmer's Guide

8.4.3.3.

INFO(7)

INFO(7) indicates how many times the Unit has been called in the current time step. This
information is very useful for building stability into user written controller routines and for
eliminating unneeded calculations at each time step. INFO(7) also carries out information through
special values (-1 and -2). Possible values of INFO(7) are listed in Table 8.4.3.3-1.
Table 8.4.3.3-1: INFO(7) values

INFO(7)

Role

-2

Special call at the very beginning of the simulation to allow for Type version signing.
Types should only set the value of INFO(13) during this call

-1

Initial call in simulation for this Unit. The Type should only perform initialization
operations here: array sizing, memory allocation, file opening, parameter checking,
etc. The Type should also set the value of INFO(6), INFO(9) and INFO(10) (if used)

First call in the current time step for this Unit.

First iterative call (second call) in the current time step for this Unit

Second iterative call (third call) in the current time step for this Unit

+n

8.4.3.4.

nth iterative call ((n+1)th call) in the current time step for this Unit

INFO(8)

INFO(8) indicates how many times the Unit has been called in the simulation. INFO(8) also
carries out information through special values (-1). Possible values of INFO(8) are listed in Table
8.4.3.4-1.
Table 8.4.3.4-1: INFO(8) values

INFO(8)

850

Role

Very first call in the simulation (version signing call)

Initialization call (INFO(7) = -1)

Initial conditions at Time = Start time (no iteration for this time step)

First call of the first time step (Time = Start time + Time step)

Second call of the first time step (Time = Start time + Time step)

Third call of the first time step, or first call of next time step

Etc

-1

Very last call of the simulation. This call occurs after the "OK" or "Continue? Yes"
button has been pressed in TRNExe. Types should close files and perform other
post-simulation tasks during this call.

TRNSYS 16 Programmer's Guide

8.4.3.5.

INFO(9) and Types calling sequence

INFO(9) indicates if the Type depends on the passage of time and tells TRNSYS where it should
go in the calling sequence. Different options are available: normal (iterative) component,
integrator, printer, etc. INFO(9) is initialized to 1 for all Types by the TRNSYS Kernel. Each Type
can change that value during the initialization call (when INFO(7) = -1).
Possible values of INFO(9) are listed in Table 8.4.3.5-1. Please note that the introduction of a
post-convergence call (flagged with INFO(13) = 1) makes some of the options offered by INFO(9)
redundant. They have been kept for backwards compatibility.
Table 8.4.3.5-1: INFO(9) values

INFO(9)

Role

The Type's outputs only depend upon its input values and not explicitly upon time.
Note that failure to properly identify such a component (keeping the default value of
INFO(9) = 1) will only results in some unnecessary calls to the component

1
(default)

The Type's outputs depend upon the passage of time and the Type must therefore
be called at least once every time step even if the values of inputs do not change

The Type should be called after all other components have converged and before
the integrators and printers. User-written statistic subroutines that send their output
to integrators is a situation where this may be appropriate.

Standard integrators

Standard printers

The Type should be called after all other components have converged and after the
integrators and printers. User-written output subroutines and non-iterative
controllers are situations for which this would be appropriate.

Note that the order in which components are called does not match the logical progression
in INFO(9) values for backwards compatibility reasons. The calling sequence is as follows:
Iterations during a time step: Units with INFO(9) = 1 (and Units with INFO(9) = 0 if their
inputs have changed) are called iteratively.
Convergence has been reached. All INFO(9) = 1 and 0 are called again with INFO(13) = 1
so they can perform post-convergence manipulations (e.g. store values)
INFO(9) = 3 components (user-written routines to be called after convergence but before
integrators) are called once
INFO(9) = 4 components (standard integrators) are called once
INFO(9) = 5 components (standard printers) are called once
INFO(9) = 2 components (user-written components to be called after integrators and
printers) are called once
This calling sequence only applies to normal time steps. All components are called once, in an
order according to their Unit numbers, during the following calls:
Version signing call (INFO(7) = -2)
Initialization call (INFO(7) = -1)

851

TRNSYS 16 Programmer's Guide

Initial time step (INFO(7) = 0, TIME = Start time). No iterations are performed, Types
should just output initial conditions. and the very
Very last call (INFO(8) = -1)

8.4.3.6.

INFO(10)

INFO(10) was used for storage management in TRNSYS 15. It was used to send the number of
required storage spots to TYPECK and then receive the index of the first reserved spot in the
global storage array. That way of handling storage is obsolete in TRNSYS 16 and should not be
used anymore. It is still present to allow TRNSYS 15 Types to run in legacy mode. Please note
that the storage array accessed by those Types is single-precision, unlike the storage array
manipulated through the new access functions (SetStorageSize, SetStorageVars,
GetStorageVars). The standard way of handling variable storage in TRNSYS 16 is described in
section 8.4.4.16.

8.4.3.7.

INFO(11)

INFO(11) sets the number of discrete control variables in the Type. It should only be set to a nonzero value if Solver 1 must be used with the Type. It is very important to make sure that the
Solver used in a simulation is handled properly by all Types. For more details about Solver 1,
please refer to the Solver Statement description in Volume 07, TRNEdit manual.

8.4.3.8.

INFO(12)

INFO(12) indicates the TRNSYS version for which the Type was written (15 or 16). It must be set
at the very first call during the simulation (INFO(7) = -2).

8.4.3.9.

INFO(13)

INFO(13) is a flag that indicates that all the Units have converged during the current time step. It
is 0 during iterative calls and 1 during the post-convergence call. Special units for which the Type
has set INFO(9) to any value greater than or equal to 2 will only be called once per time step, with
INFO(13) = 1.

8.4.3.10. Typical calling sequence for an iterative Type

Table 8.4.3.10-1 shows the time and INFO array values for an iterative Type in a typical
simulation. Please note the following:
INFO(1:5) are set by TRNSYS before calling the Type from information in the input (.dck) file
and should not be modified by the Type. For that reason they are omitted in the table.
It is assumed that convergence is reached after 2 iterations (after the 3rd call in a time steps)
for all time steps.
Start, Stop and Step are respectively the simulation start time, stop time and time step. nO is
the number of outputs set by the Type.
S is the value set by the Type for INFO(9).
Values that change are in bold and especially meaningful values have a dark background.

852

TRNSYS 16 Programmer's Guide

It is assumed that the total number of iterative calls to the Type in the simulation (which is
reported in the "Iterative call breakdown" in the listing and log files) is nCalls.
The Type is assumed to follow TRNSYS 16 Standard (i.e. it is not signed as a "15" Type and
it does not use the old storage array)
Table 8.4.3.10-1: Typical INFO calling sequence for an iterative Type

Call description

Time

Version signing call

INFO element
6

10

11

12

13

Start

-2

-10

Initialization call

Start

-1

-10

16

Start (initial conditions)

Start

nO

16

st

st

Start+Step

nO

16

st

nd

st

Start+Step

nO

16

st

rd

nd

Start+Step

nO

16

Start+Step

nO

16

1 step, 1 call
1 step, 2 call (1 iter.)
1 step, 3 call (2 iter.)
st

1 step, post-convergence
nd

st

Start+2*Step

nO

16

nd

nd

st

Start+2*Step

nO

16

nd

rd

nd

Start+2*Step

nO

10

16

2 step, post-convergence

Start+2*Step

nO

11

16

etc.

16

2 step, 1 call
2 step, 2 call (1 iter.)
2 step, 3 call (2 iter.)
nd

st

Stop

nO

nCalls-3

16

nd

st

Stop

nO

nCalls-2

16

rd

nd

Last step, 3 call (2 iter.)

Stop

nO

nCalls-1

16

Last step, post-converg.

Stop

nO

nCalls

16

Very last call (after "OK")

Stop

nO

-1

16

Last step, 1 call

Last step, 2 call (1 iter.)

853

TRNSYS 16 Programmer's Guide

854

TRNSYS 16 Programmer's Guide

8.4.4.

Utility subroutines

8.4.4.1.

Calling External Programs

In order to allow calls to external programs, a utility subroutine named CALLPROGRAM has
been added to TRNSYS with the release of TRNSYS 15. CALLPROGRAM uses several Win32
Application Programming Interface (API) commands to start the external program. API
commands directly control the Windows operating system. Depending on the mode specified in
the call statement, CALLPROGRAM will either wait for the second program to finish its task
before proceeding or will merely start the second program and then proceed with other TRNSYS
calculations. The call statement for CALLPROGRAM is:
...
CALL CALLPROGRAM(CMDLINE,bwait,prochand, thrdhand)
...

Where:
CMDLINE

is the command line text string containing the path and name for
the executable program

BWAIT

is a logical variable that determines whether TRNSYS will wait

for the second program to finish running or not.
-If BWAIT = .TRUE., then TRNSYS does not proceed until the
second program is finished and closed.
-If BWAIT = .FALSE., then TRNSYS proceeds once the second
program is done initializing but still running.

PROCHAND

is the Windows process handle for the new program. This is a

value returned by the CALLPROGRAM routine.

THRDHAND

is the Windows thread handle for the new program. This is a

value returned by the CALLPROGRAM routine.

When CALLPROGRAM is used to start another program and leave it running (BWAIT =
.FALSE.), then the program needs to be terminated at the end of the simulation before leaving
TRNSYS. This can be done by using the following commands:
...
res = TerminateProcess(PROCHAND,1)
IF (res.EQ.0) THEN
WRITE(CHCKWHY,*) GETLASTERROR()
ENDIF
...

See the source code of Type 66 (CALL EES MODELS) for an example of the use of
CALLPROGRAM.

855

TRNSYS 16 Programmer's Guide

8.4.4.2.

DFIT (FIT)

GENERAL DESCRIPTION
Subroutines FIT and DFIT use linear least-squares regression to determine coefficients of a user
defined correlation. FIT requires single precision arguments, while DFIT uses double precision
arguments. Since a majority of the kernel routine and standard components use double precision
variables throughout, it is recommended that user written routines employ the double precison
version DFIT. A call to DFIT is of the form:
INTEGER :: NROW,NCOL,NCOEF,NDATA,IFLAG
DOUBLE PRECISION :: X,Y,PHI
...
CALL FIT(NROW,NCOL,NCOEF,NDATA,X,Y,PHI,IFLAG,*N)

where,
NROW

(integer) Row dimension of X in routine calling DFIT.

NCOL

(integer) Column dimension of X in routine calling DFIT.

NCOEF

(integer) Number of coefficients in model.

NDATA

(integer) Number of data points for the regression.

Double precision two dimensional array containing the values of

the independent variables to the model. X(I,J) is the value of the
independent variable associated with the ith coefficient of the
model and the jth data point.

(double precision) vector containing values of the dependent

variable corresponding to independent variable data points.

PHI

(double precision) vector containing the coefficients determined

from the regression.

IFLAG

Integer flag returned indicating the error condition (0 - no error, 1

- dimension error, 2 - coefficients cannot be determined).

the corresponding line number to jump to if the subprogram is

present (see Section 3.3.5)

Least-squares regression is used to minimize a function J, where:

J=

N data

(Y
j =1

Ymod el , j =

Ymod el , j )

Eq. 8.4.4-1

N coef

X
i =1

i, j

Eq. 8.4.4-2

where,

856

Yj

jth data point for dependent variable

Ymodel,j

dependent variable of the user's model for the jth data point

ith coefficient to be determined with FIT

TRNSYS 16 Programmer's Guide

Xi,j

user defined functions of the independent variable or variables.

Example
Consider the equation used in the Type53 Chiller component. The dimensionless
power consumption is correlated with performance data using the following bi-quadratic equation:

G = a0 + a1 E + a2 E 2 + a3 F + a4 F 2 + a5 EF

Eq. 8.4.4-3

This is the same form as Eq. 8.4.4-2 with:

1-5 = a0-a5
X1,j = 1.0
X2,j = E
X3,j = E2
X4,j = F
X5,j = F2
X6,j = EF
If the following data existed for E, F, and G:
E

3.0

0.2

10.0

5.0

0.4

12.5

7.0

0.6

13.1

Then DFIT would be called with

NROW =

6 or more (minimum is the # of coefficients,i's)

NCOL =

3 or more (minimum is the # of data points,j's)

NCOEF =

6 (number of coefficients, i's)

NDATA =

3 (number of data points, j's)

857

TRNSYS 16 Programmer's Guide

X=

1.0

1.0

1.0

E1

E2

E3

E1 2 E2 2 E3 2
F1

F2

F3

F1 2 F2 2 F3 2

1.0 1.0 1.0

3.0 5.0 7.0
9.0 25.0 49.0
0.2 0.4 0.6
.04 .16 .36
0.6 2.0 4.2

Eq. 8.4.4-4

EF1 EF2 EF3

10.0
Y = 12.5
13.1

Eq. 8.4.4-5

The DFIT subroutine would be called with the above input data and return values for PHI and
IFLAG. PHI would be a vector containing the values for a0-a5.

858

TRNSYS 16 Programmer's Guide

8.4.4.3.

DIFFERENTIAL_EQN (DIFFEQ)

The double precision utility routine Differential_Eqn and its single precision companion DIFFEQ
provide analytical solutions to first-order linear differential equations that can be written as

dT
= aT + b
dt

Eq. 8.4.4-6

The designation double precision means that the arguments and results of the utility routine
Differential_Eqn must be declared in the calling Type as DOUBLE PRECISION variables as
opposed to REAL or INTEGER variables. In TRNSYS version 15.x and below, only a single
precision version of this utility was available. That single precision version (called DIFFEQ) was
maintained for completeness and backward compatibility but it is highly recommended that user
employ the double precision version (called Differential_Eqn) whenever possible.
Users intending to utilize the Powells method TRNSYS solver (SOLVER 1) should use the
Differential_Eqn subroutine discussed here instead of the optional DTDT array discussed in
section REVISIT. The form of the call to Differential_Eqn is:
CALL DIFFERENTIAL_EQN(TIME, AA, BB, TI, TF, TBAR)

where
TIME

(double precision) current value of time

AA

(double precision) is the a coefficient of the differential equation

BB

(double precision) is the b coefficient of the differential equation

TI

(double precision) is the value of the dependent variable (T) at

the beginning time step

TF

(double precision) is the value of the dependent variable (T) at

the end of the time step

TBAR

(double precision) is the average value of the dependent variable

over the time step.

The Differential_Eqn routine does not require link checking since it is considered to be a kernel
routine; those routines which must be present to successfully run TRNSYS.
Differential_Eqn solves differential equations at each call based on parameters and the current
set of inputs. If the inputs have not converged, the iteration proceeds by direct substitution
provided that the Successive Substitution solver (SOLVER 0) is being used. Additional notes on
the analytical solution of differential equations are as follows:
1)

Differential_Eqn requires the arguments AA, BB, and TI, while returning TF and TBAR.
For calls to Differential_Eqn at the start time of the simulation (TIME= TIME0), the initial
condition TI is returned for both TF and TBAR.

2)

It is necessary to save the final values of dependent variables each time step as initial
conditions for the following interval. This can be done using calls to the getStorageVars
and setStorageVars routine discussed in sections 0 and 0.

859

TRNSYS 16 Programmer's Guide

3)

Inputs will often be a function of the dependent variable of the differential equation. For
instance, collector outlet temperature, which might be an input to a storage component,
might also be a function of the storage temperature. In order to get a good estimate of
the average inputs over each interval, TBAR should be set as an output if required as an
input of other components. To be consistent with this formulation and the rectangular
integration provided by Type24, etc., any outputs that are rates and are expressed as a
function of the dependent variable T should be evaluated at TBAR.

As an example of a model formulation, suppose that one wishes to study the temperature
response of a one-node house to ambient conditions. The instantaneous energy balance for this
situation is given as

dTR
= (UA) L (TR Ta )
dt

Eq. 8.4.4-7

This can be written in the form of Eq. 8.4.4-6 if

a=

(UA) L
C

and

b=

(UA) L Ta
C

Eq. 8.4.4-8

The following code would establish initial conditions, solve the differential equation for the final
and average temperatures, and determine the average energy loss rate from the house for each
time interval.
DOUBLE PRECISION :: STORED(2),TI,AA,BB,UAL,TA,C,QLOSS,TF,TBAR
INTEGER :: NS,INFO(15)
...
!this section is executed at the end of each timestep. In this case it is
!used to update the storage structures for the following time step.
IF(INFO(7).EQ.13) THEN
CALL getStorageVars(NS,STORED,INFO)
STORED(1) = STORED(2)
CALL setStorageVars(NS,STORED,INFO)
ENDIF
...
CALL getStorageVars(NS,STORED,INFO)
TI = STORED(1)
AA = -UAL/C
BB = UAL*TA/C
CALL DIFFERENTIAL_EQN(TIME,AA,BB,TI,TF,TBAR)
STORED(2) = TF
QLOSS = UAL*(TBAR - TA)
CALL setStorageVars(NS,STORED,INFO)
...

For this particular example, the component would not be called by the kernel during any time step
after its inputs have converged within the algebraic error tolerance. There is no need for a
differential equation convergence check, since the solution will always be consistent with the
current set of inputs. A DERIVATIVES control statement should not be used for the component
in the simulation input file; the initial value of the dependent variable at the beginning of the
simulation should be supplied to the subroutines as a PARAMETER. In this respect the example
is similar to Types 12, 21 and 23.

860

TRNSYS 16 Programmer's Guide

8.4.4.4.

DynamicData (DYNDATA and DATA)

The DynamicData utility routine is available to read user supplied data from external text based
files that have been assigned a FORTRAN logical unit number in the TRNSYS input file. The
DynamicData routine is able to interpolate the data found in the file in up to four dimensions
during the course of the simulation. For each data set, up to five dependent (Y) functions may be
specified in terms of up to four independent (X) variables. At each call to the DynamicData
routine, the calling Type sends the values of each independent variable. DynamicData performs
the multi dimensional interpolation and returns the interpolated values of as many dependent
variables as have been provided in the data file and requested by the calling Type. DynamicData
is NOT able to extrapolate beyond the range of data given in the external file. If the calling Type
sends a value of an independent variable that falls above or below the range given in the data
file, DynamicData will return the values of the dependent variables that correspond to the
maximum or minimum of the range.
DynamicData is a double precision routine, meaning that the independent and dependent
variables found in the external file will be treated as FORTRAN double precision variables.
Similarly, the X and Y arrays in the calling Type must be declared as double precision. In
TRNSYS versions 15.x and below, DynamicData was known as either DATA or DYNDATA,
both of which were single precision routines. DATA and DYNDATA were retained for backwards
compatibility but every effort should be made to use the double precision DynamicData routine.
Examples of standard components that use DynamicData are the Type1 Solar Collector and the
Type44 Conditioning Equipment model. The form of a call to DynamicData is:
INTEGER :: LU,NIND,NX(:),NY(:),INFO(15)
DOUBLE PRECISION :: X(:),Y(:)
...
CALL DynamicData (LU, NIND, NX, NY, X, Y, INFO,*N)

where
LU

(integer) is the FORTRAN logical unit number through which the

performance data is accessed.

NIND

(integer) is the number of independent variables (1 NIND 4)

NX

is an integer array dimensioned to NIND containing the number

of values of each independent variable to be read from LU.
NX(1), NX(2), NX(3), and NX(4) are the number of values of X1,
X2, X3 and X4, respectively.

NY

(integer) is the number of dependent (Y) values associated with

each set of independent (X) values (1 NY 5)

is a double precision array dimensioned to NIND containing the

values of the independent variables for which interpolated Y
values are desired

is a double precision array of dimension NY containing

interpolated values of the dependent variables

INFO

is the calling TYPE's INFO array

(integer) is the corresponding line number to jump to if the

subprogram is present.

The data is read from logical unit LU at the start of the simulation. Thereafter, DynamicData
linearly interpolates for Y values given X values. If DynamicData is called with an X out of range

861

TRNSYS 16 Programmer's Guide

of the supplied data, then the closest specified value is used. DynamicData does not extrapolate
beyond the user supplied data.
The data supplied on the logical unit is read in free format. If NIND equals 4 and NX(4) is greater
than 1, then NX(4) values of the fourth independent variable X4 are read first. Similarly, if NIND
3 and NX(3) > 1, then a set of NX(3) values of the third independent variable, X3 are read. Next,
if NIND 2 and NX(2) > 1, then a set of NX(2) values of the second independent variable X2 are
read. Lastly, NX(1) values of the primary independent variable, X1, are read. The values of the
independent variables must be in ascending order, but need not be at regular intervals. The Y
values are read beginning with values corresponding to the smallest values of X1, X2, X3, and X4.
NX(1) sets of NY values of Y are read for each value of X2. NX(2) sets of NX(1) * NY values of Y
are read for each value of X3. In all, if 4 independent variables are employed, then NX(4) * NX(3)
* NX(2) * NX(1) * NY values of Y must be specified.
Example:
A user wishes to write a model for a water-to-water heat pump whose
performance depends upon the inlet flow stream temperatures to both the evaporator (Tevap)
and condenser (Tcond). DynamicData is to be used to evaluate both the capacity and COP. The
experimental data is shown in Table 8.4.4.4-1.
Table 8.4.4.4-1

Tcond = 20C

Tcond = 50C

Tevap (C)

Capacity (kJ/hr)

COP

Tevap (C)

Capacity (kJ/hr)

COP

10

35000

2.47

10

21000

1.73

20

41000

2.80

20

25000

1.96

30

49500

2.99

30

29000

2.09

The data is to be read from logical unit 10. Three values of Te and two values of Tc will be
supplied. The call to DynamicData within the calling Type routine might appear as
USE TrnsysFunctions
...
INGEGER :: NX(2),INFO(15),DUM
DOUBLE PRECIsion :: X(2),Y(2),CAP,COP,T_EVAP,T_COND
...
NX(1) = 3
!declares that there are 3 nx 1 values in the file.
NX(2) = 2
!declares that there are 2 nx 2 values in the file.
X(1)
= T_Evap
!sets x(1) to the local variable t_evap
X(2)
= T_Cond
!sets x(2) to the local variable t_cond

101

CALL DynamicData (10, 2, NX, 2, X, Y, INFO,*101)

CALL LINKCK(TYPE74, DynamicData,3,DUM)
IF (ErrorFound()) RETURN 1
CAP
COP
...

= Y(1) !sets the call result y(1) to local variable CAP.

= Y(2) ! sets the call result y(1) to local variable COP.

The syntax of the data file accessed through logical unit 10 would be as follows:

862

20.0 50.0

!Tcond values

10.0 20.0 30.0

!Tevap values

TRNSYS 16 Programmer's Guide

35000. 2.47

!Capacity and COP for Tcond=20C, and Tevap=10C

41000. 2.80

!Capacity and COP for Tcond =20C, and Tevap=20C

49500. 2.99

!Capacity and COP for Tcond =20C, and Tevap=30C

21000. 1.73

!Capacity and COP for Tcond =50C, and Tevap=10C

25000. 1.96

!Capacity and COP for Tcond =50C, and Tevap=20C

29000. 2.09

!Capacity and COP for Tcond =50C, and Tevap=30C

863

TRNSYS 16 Programmer's Guide

8.4.4.5.

ENCLOSURE (ENCL)

General Description
The utility routine ENCLOSURE (and its single precision companion ENCL) calculate view factors
between all surfaces of a rectangular parallelpiped. Up to 9 windows or doors may be located on
any of the 6 wall surfaces. ENCLOSURE is a double precision routine, meaning that its
arguments must be declared in the calling Type as double precision variables (as opposed to
REAL variables). The values returned by ENCLUSRE are also double precision. In TRNSYS
versions 15.x and below, ENCLOSURE was known simply as ENCL, which was a single
precision routine. ENCL was retained for backwards compatibility and every effort should be
made to use the double precision ENCLOSURE routine.
This routine is used by the TYPE 19 zone model. A call to ENCLOSURE is of the form
DOUBLE PRECISION :: SPAR(10),WPAR(10),FV
INTEGER :: INFO(15)
...
CALL ENCLOSURE (SPAR, WPAR, FV, INFO,*N)

where
SPAR

a double precision array of 10 values containing the 10 zone

geometry parameters described in the TYPE 19 documentation
(parameters 2-11 of geometry Mode 1).

WPAR

a double precision array dimensioned to 54, containing the

complete list of window or door geometry parameters described
in the TYPE 19 section for geometry Mode 1.

FV

a double precision array containing the view factors for all

surfaces as calculated by the subroutine.

INFO

the calling TYPE's INFO array.

is the corresponding line number to jump to if the subprogram is

present.

Subroutine ENCLOSURE outputs a table of the view factors between surfaces. The surface
numbers are as specified in the SPAR array.

Mathematical Description
Subroutine ENCL utilizes function subroutine VIEW to obtain view factors between continuous
rectangular surfaces. Reciprocity is also used to reduce the number of computations. In order to
determine view factors between surfaces that contain windows or doors, it is necessary to
perform some view factor algebra.
Consider the general case of m windows located on a surface i and n windows on a surface j as
shown in Figure 8.4.4.5-1.

864

TRNSYS 16 Programmer's Guide

jW
W i1

W i2

W i3

Wj2

j1

Wj3

W jn

W im

i
Figure 8.4.4.5-1

In general, the view factor between i and j is

Fi j = Fi ( j + W ji + ... + W jn ) Fi (W jk )
n

Eq. 8.4.4-9

k =1

where

Fi ( j +W j1 + ... +W jn ) =

(A

+ AW j1 + ... + AW jn F( j +W j1 + ...+W jn ) i
Ai

Eq. 8.4.4-10

F( j +W j1 + ...+W jn ) i = F( j +W j1 + ...+W jn ) ( i +Wi1 + ... +Wim ) F( j +W j1 + ...+W jn ) Wik

Eq. 8.4.4-11

k =1

Fi W jk =

AW jk F(W jk ) i

Eq. 8.4.4-12

Ai

FW jk i = FW jk ( i +Wi1 + ...+Wim ) FW jk Wil

Eq. 8.4.4-13

l =1

The variables Wi1 through Wim are the surface numbers associated with windows on wall i.
Likewise, Wj1 through Win refer to windows on wall j. Thus, (i+Wi1+...Win) and (j+Wj1+... Wjm),
each denote surfaces that are collections of individual surfaces. The variable A refers to the area
of a surface whose number is used as a subscription.

865

TRNSYS 16 Programmer's Guide

8.4.4.6.

FLUID_PROPS (FLUIDS)

The double precision utility routine called Fluid_Props and its single precision companion FLUIDS
are used to calculate the thermodynamic properties of various refrigerants based on correlations.
Arguments to the Fluid_Props routine should be declared in the calling Type as DOUBLE
PRECISION variables. Arguments to the FLUIDS routine should be declared in the calling Type
as REAL. With TRNSYS version 15.x and below, only the single precision version was available.
Since most kernel and standard Type variables were made DOUBLE PRECISION with the
release of TRNSYS version 16.0, every effort should be made to employ the double precision
version of this routine (Fluid_Props). Given two state properties, the Fluid_Props routine will
return the complete state of the refrigerant. A call to FLUIDS is of the form
INTEGER :: NREF,ITYPE,IFLAG
DOUBLE PRECISION :: PROP(9)
CHARACTER(len=2) :: UNITS
...
CALL FLUIDS (UNITS,PROP,NREF,ITYPE,IFLAGR,*N)

where
UNITS

(character) two character variable denoting unit system to be

employed by the fluids routine; SI for metric units and EN for
english units

PROP

double precision array holding the two known thermodynamic

properties of the refrigerant in question upon entry and the
complete state of the refrigerant upon return from the
Fluid_Props routine. This array must be dimensioned to 9 in the
routine that calls Fluid_Props. Each location in this array is
described below:

PROP(1)

temperature of refrigerant (F, C)

PROP(2)

pressure of refrigerant (psi, MPa)

PROP(3)

enthalpy of refrigerant (Btu/lbm, kJ/kg)

PROP(4)

entropy of refrigerant (Btu/lbm R, kJ/kg K)

PROP(5)

quality (0 < x < 1)

PROP(6)

specific volume (ft3/lbm, m3/kg)

PROP(7)

internal energy (Btu/lbm, kJ/kg)

PROP(8)

dynamic viscosity (lbm/ft hr, N s/m2)

PROP(9)

thermal conductivity (Btu/hr ft R, kJ/hr m K)

NREF

an integer variable denoting the refrigerant to be analyzed.

Available refrigerants are listed along with the required NREF
code number:
R-11
(11)

866

R-12

(12)

R-13

(13)

R-14

(14)

R-22

(22)

R-114

(114)

TRNSYS 16 Programmer's Guide

R-134A

(134)

R-500

(500)

R-502

(502)

AMMONIA

(717)

ITYPE

integer variable denoting which two properties are supplied as

known properties. ITYPE = 10* PROP indicator 1 + PROP
indicator 2 (12 to 98)

IFLAGR

an integer variable returning the error messages from the

FLUIDS call:
0 - no error found, calculations completed
1 - error found, calculations could not be completed

the corresponding line number to jump to if the subprogram is

present.

An example illustrating the use of the FLUID_PROPS routine is shown below. In the example,
two inputs (temperature and quality) are read and are placed in the PROP array. The Fluid_Props
routine is called. If the Fluid_Props routine is not found for some reason, the error handling
routine called LINKCK is called (refer to section 8.4.4.7 for more information). If Fluid_Props
returns, the example checks to make sure that no errors were reported by Fluid_Props by calling
an Access Function called ErrorFound( ) (refer to section 8.4.2.6 for more information). If errors
were detected, execution of the Type that called Fluid_Props stops immediately by returning
control to the calling program. If errors were not found then the PROP array has been filled. The
Second and third array spots are set to local variables and Type execution continues normally.

100

USE TrnsysFunctions
...
CHARACTER(len=2) :: UNITS
DOUBLE PRECISION :: PROP(9),XIN(2),TEMP,QUALITY,PRESSURE,ENTHALPY
INTEGER :: NREF
...
UNITS = SI !SI units are desired.
NREF = 12
!Refrigerant R12
...
TEMP = XIN(1)
QUALITY = XIN(2)
...
PROP(1) = TEMP
PROP(5) = QUALITY
CALL FLUID_PROPS(UNITS,PROP,12,15,IFLAG,*100)
CALL LINKCK('TYPE58','FLUID_PROPS',1,58)
IF (ErrorFound()) RETURN 1
...
PRESSURE=PROP(2)
ENTHALPY=PROP(3)

MATHEMATICAL DESCRIPTION
The correlations required to solve for the refrigerant properties are from the Engineering Equation
Solver program (F-chart software, 1994), a numerical solver employing thermodynamic
correlations from many different sources. Interested users should contact the reference for more
details on the correlations.
Enthalpies for ammonia and R134a are based on the reference state: enthalpy equal to zero at 40C.

867

TRNSYS 16 Programmer's Guide

Enthalpies for all other refrigerants are based on the reference state: enthalpy equal to zero at
0C
This subroutine checks for many improper inputs, such as qualities less than 0. or greater than 1.,
and the input of two properties that cannot be correct for one state. For these and other improper
inputs, the subroutine either prints a warning, corrects one of the inputs, and continues; or prints
an error and halts the simulation.
Subcooled properties are not available for refrigerants. If a subcooled state is specified, the
saturated liquid results will instead be provided.

868

TRNSYS 16 Programmer's Guide

8.4.4.7.

LINKCK

The LINKCK subroutine is a utility subroutine used for the detection and subsequent error
message printing of unlinked subroutines. With earlier version of TRNSYS (up to version 15.0)
the LINKCK routine was of vital importance because as a memory saving step TRNSYS was
often compiled to include only those kernel routines that were critical to a given simulation. With
advances in computing power and speed, this step became less and less important. LINKCK
remains as a safeguard against inadvertently unlinked routines but almost always, unlinked
routines will be caught during the compiling and linking processes. A call to LINKCK is of the
form:
CHARACTER(len=12) :: ENAME1,ENAME2
INTEGER :: ILINK,LNKTYP
...
CALL LINKCK(ENAME1,ENAME2,ILINK,LNKTYP,*N)

where
ENAME1

a 12-character variable identifying the subroutine from which the

call originated.

ENAME2

a 12-character variable identifying the subroutine that was called

ILINK

an integer indicating the steps to be taken by the LINKCK

program
=1
an unlinked subroutine has been found, generate an
error message and stop the program
=2
an unlinked subroutine has been found, generate a
warning but keep running
=3
an unlinked TYPE subroutine has been found, generate
an error message and stop the program
=4
warn the user that a subroutine requires use of an
external function which can not be link checked

LNKTYP

integer variable corresponding to the TYPE subroutine which

was not found in the TRNSYS executable

The LINKCK subroutine is provided to the users as a means to standardize the TRNSYS error
and warning messages associated with unlinked subroutines. The user should only call LINKCK
when an unlinked subroutine is detected or an external function is required.
The following example should clarify the use of the LINKCK subroutine. Refer to section 3.3.5 for
more details.
SUBROUTINE TYPE75(TIME,XIN,OUT,T,DTDT,PAR,INFO,ICNTRL,*)
...
CHARACTER(len=12) ENAME1,ENAME2
DOUBLE PRECISION :: SPAR(10),WPAR(10),FV
INTEGER :: ILINK,LNKTYP,INFO(15)
...
!CALL THE ENCLOSURE SUBROUTINE AND INFORM LINKCK IF NOT PRESENT
CALL ENCLOSURE(SPAR,WPAR,FV,INFO,*101)
!STOP THE PROGRAM IF ENCL IS NOT PRESENT
ILINK = 1
IDUM = 75
ENAME1 = TYPE75

ENAME2 = ENCLOSURE

869

TRNSYS 16 Programmer's Guide

CALL LINKCK(ENAME1,ENAME2,ILINK,IDUM)
101

870

!ENCL IS PRESENT - CONTINUE ON

CONTINUE
...

TRNSYS 16 Programmer's Guide

8.4.4.8.

MATRIX_INVERT (INVERT and DINVRT)

Subroutine Matrix_Invert is a double precision routine available to invert a matrix of up to 50 by

50 entries. In TRNSYS version 15.x and below, the subroutine had two forms; INVERT was a
single precision version and DINVRT was a double precision. With the release of TRNSYS 16.0,
the double precison routine was renamed and is recommended for use by all standard and user
written Types. INVERT and DINVRT were maintained for backwards compatibility. A call to
Matrix_Invert is of the form
INTEGER :: NRC,N,IFLAG
DOUBLE PRECISION :: A
...
CALL MATRIX_INVERT (NRC, N, A, IFLAG,*N)

where
NRC

(integer) is the column and row dimensions of the twodimensional A array as defined in the program that calls
Matrix_Invert.

(integer) the number of equations and unknowns associated with

the problem.

a double precision two-dimensional array (A(NRC, NRC)). Upon

calling Matrix_Invert A should contain the matrix to be inverted.
Matrix_Invert also returns the inverted matrix in the A array.

IFLAG

an integer flag that is set to 0 if the inversion proceeds correctly,

1 if the number equations and unknowns, N, exceeds 50, and 2
if the matrix is singular.

the corresponding line number to jump to if the subprogram is

present.

The inverted matrix is returned in the A array. The method used to invert the matrix is the GaussJordan reduction with maximum pivoting.

871

TRNSYS 16 Programmer's Guide

8.4.4.9.

MESSAGES

The Messages utility routine provides the Type programmer with a convenient method of
reporting error messages to the end user. When a Type catches a particular condition, it can
send a text string and information as to the severity of the condition to the Messages utility
subroutine. Messages will take care of logging the event, reporting it to both the simulation list
and log files, and stopping the simulation if appropriate.

INTERFACE
subroutine Messages(errorCode,message,severity,unitNo,typeNo)
integer :: errorCode
character (len=*) :: message, severity
integer :: unitNo, typeNo

(Note: Although the length of message is not explicitly declared, it should never exceed the
maxMessageLength global constant).

USAGE
ERRORCODE

errorCode is a standard TRNSYS error message number. Please

initializeErrorMessages() subroutine in Messages.f90 for existing messages.

refer

to

the

If you wish to generate a custom error message (as will most often be the case for user-written
Types), simply set ErrorCode to a value of 1, indicating the messages that the second argument
(message) contains the information to be printed.
MESSAGE

If you are generating a custom error message, this string variable will be printed in a standard
format by the Messages subroutine. It should consist of one line of text, shorter than
maxMessageLength. The example here below illustrates the use of Write to add run-time
information to the message.
If you are using a standard error message, the string text will be printed in addition to the
standard error message under the heading Reported Information. It is a method for providing
additional information about the error.
SEVERITY

severity is a string that indicates the severity of the message. Messages understands 4 levels of
severity:
"Notice": A notice is simply information that you would like the user to know
"Warning": The messages subroutine keeps track of how many "Warning" messages have
been generated. If the maximum allowable number of warnings (set by the LIMITS statement
in the input file) are exceeded, then the messages subroutine automatically generates a
"Fatal" error.
"Fatal": Designates a fatal error, i.e. an error that should stop the simulation. A call to
Messages with such an error should be followed by a "return 1" statement in the calling
Type. Note that TRNSYS will stop after going up the chain of "return" statements and giving

872

TRNSYS 16 Programmer's Guide

the chance to all subroutines to perform end-of-simulation operations. Further errors might
be generated during that process.
"Stop": The error will abort TRNSYS immediately without going up the chain of "return"
statements, by generating an exception in the DLL and returning control to TRNExe.exe.
This error type should only be used as a last resort, e.g. when returning from that error
would result in memory access violations or other exceptions. Exiting from a DLL by
generating an exception is generally considered as bad programming and users should only
use this technique if the cost of modifying the Type to handle the error properly is too high.
Note: severity is a Fortran literal string and can be enclosed in single or double
quotes. Messages will understand some case variants of the codes here above.
e.g. notice, Notice and NOTICE are accepted (other variants are not)

EXAMPLE
The following example illustrates calls to Messages with different severity levels by a Type for
which par(1)>10 is not acceptable (values <10.1 are just rounded to 10 and the simulation goes
on).
subroutine type200(time,xin,out,t,dtdt,par,info,iCntrl,*)

use TrnsysConstants

character (len=maxMessageLength) :: myMessage

integer info(15)

if ( par(1) < 10.0 ) then

write(myMessage,'("Par(1) = ",g," , which is OK")') par(1)
call Messages(-1,trim(myMessage),'Notice',info(1),info(2))
else if ( par(1) <= 10.1) then
call Messages(-1,'Par(1) has been rounded to 10','Warning',info(1),info(2))
else
call Messages(-1,'Par(1) is outside the acceptable range.',
&
'Fatal',info(1),info(2))
return 1
endif

Please note the use of RETURN 1 following the call to Messages with a severity of Fatal. When
the severity of the message is low (notice, message, or warning), the simulation should proceed;
thus no RETURN statement is required. In the case of a fatal message, however, the simulation
should be immediately stopped and the Type should return control to the TRNSYS kernel through
use of the RETURN 1 statement.

873

TRNSYS 16 Programmer's Guide

8.4.4.10. ParRead
This is an undocumented subroutine used internally by some standard Types. See the Source
code in ParRead.f90 for additional information.

874

TRNSYS 16 Programmer's Guide

8.4.4.11. PSYCHROMETRICS (PSYCH)

This utility routine is available to calculate various moist air properties based on ASHRAE
correlations and the Ideal Gas Law. A double precision version called PSYCHROMETRICS and a
legacy single precision version called PSYCH are available. It is recommended that all user
written routines employ the double precision version and that all non integer arguments to the
Psychrometrics routine be declared as DOUBLE PRECISION variables in the calling Type. A call
to Psychrometrics is of the form
DOUBLE PRECISION :: TIME,PSYDAT
INTEGER :: INFO(15),IUNITS,MODE,WBMODE,EMODE,STATUS
...
CALL PSYCHROMETRICS(TIME,INFO,IUNITS,MODE,WBMODE,PSYDAT,EMODE,STATUS)

where
TIME

(double precision) This is the standard TIME used in TRNSYS.

It is used in the PSYCHROMETRICS subroutine only to identify
the time at which warnings or errors occur.

INFO

This is the standard integer INFO array used in TRNSYS (see

Section 8.4.3). It is used in the PSYCHROMETRICS subroutine
to identify the unit and type number of the component which calls
the subroutine, which will be printed during a simulation if
warnings or errors occur.

IUNITS

an integer variable equal to 1 or 2 identifying the units desired for

the properties. IUNITS = 1 results in properties in SI units.
IUNITS = 2 results in properties in English units.

MODE

an integer variable from 1 through 6. Three properties are

needed to specify the moist air state. The MODE identifies
which two properties besides pressure are need to be input,
leaving the other properties to be calculated. The modes are as
follows:
1:

input dry bulb and wet bulb temperatures (PSYDAT(2)

and PSYDAT(3))

2:

input dry bulb temperature and relative humidity

(PSYDAT(2) and PSYDAT(4))

3:

input dry bulb and dew point temperatures (PSYDAT(2)

and PSYDAT(5))

4:

input dry bulb temperature

(PSYDAT(2) and PYSDAT(6))

5:

input dry bulb temperature and enthalpy (PSYDAT(2)

and PSYDAT(7))

6:

input humidity ratio and enthalpy (PSYDAT(6) and

PSYDAT(7)) . If saturation conditions occur, enthalpy is
reset to saturation enthalpy at the given value of
humidity ratio.

7:

input humidity ratio and enthalpy (PSYDAT(6) and

PSYDAT(7)) . If saturation conditions occur, humidity
ratio is reset to saturation humidity ratio at the given

and

humidity

ratio

875

TRNSYS 16 Programmer's Guide

value of enthalpy. 8:
input relative humidity
enthalpy (PSYDAT(6) and PSYDAT(4)).

and

WBMODE

an integer variable equal to 0 or 1. If WBMODE = 0, the wet bulb

temperature will not be calculated in MODEs 2 through 6. If
WBMODE = 1, the wet bulb temperature is calculated. At PATM
= 1.0 and temperatures common to HVAC applications, a
correlation (1) is used to find the wet bulb temperature given
enthalpy. For other pressures and temperatures, a Newton's
iterative method is used with the equations given by ASHRAE (2)
to calculate the wet bulb temperature. If the wet bulb
temperature is not needed, WBMODE should be set to 0 to
decrease the computational effort. If WBMODE = 0, the wet bulb
temperature is set to the dry bulb temperature. PSYDAT an
array containing the properties of the moist air. This array must
be dimensioned to 9 in the routine that calls
PSYCHROMETRICS. Each location in this array is described
below:

PSYDAT(1)

(double precision) The total system pressure in atmospheres.

The subroutine prints a warning if PATM is over 5 atmospheres,
which is considered here to be an upper limit for using ideal gas
relations for air and water mixtures.

PSYDAT(2)

(double precision) dry bulb temperature (C).

PSYDAT(3)

(double precision) wet bulb temperature (C).

PSYDAT(4)

(double precision) relative humidity (fraction).

PSYDAT(5)

(double precision) dew point temperature (C).

PSYDAT(6)

(double precision) humidity ratio (kg water/kg dry air).

PSYDAT(7)

(double precision) enthalpy (kJ/kg dry air).

PSYDAT(8)

(double precision) density of the air water mixture (kg/m3)

PSYDAT(9)

(double precision) density of the air portion of the mixture (kg dry
air/m3)

EMODE

an integer variable equal to 0, 1, or 2 specifying how warnings

are handled. If EMODE=0, no warnings will be printed. If
EMODE=1, only one warning for each condition will be printed
throughout the simulation. If EMODE=2, warnings will be printed
every time step. (EMODE=2 is useful for program and simulation
development, but it can produce an excessive amount of output).

STATUS

an integer variable equal to 0 through 13 identifying the warning

condition that occurred within subroutine PSYCH. 0 indicates no
warnings. For a complete listing of warnings 1 through 13 the
subroutine source code should be reviewed.

the corresponding line number to jump to if the subprogram is

present.

MATHEMATICAL DESCRIPTION
In determining a number of moist air properties the water vapor saturation pressure (PWS) is
required. The correlations (ASHRAE, 2001) used for PWS are accurate over the temperature
range of 100C to 200C. A warning is printed if moist air states occur outside this temperature
range.

876

TRNSYS 16 Programmer's Guide

The correlations for the dew point temperature (Ashrae, 2001) are accurate over the temperature
range of -60C to 70C. A warning is printed if moist air states occur outside this dew point
range.
Enthalpies are based on the following reference states:
air enthalpy is zero at 0C
liquid water enthalpy is zero at 0C
This subroutine checks for many improper inputs, such as relative humidities less than 0. or
greater than 1., dew point temperatures greater than the dry bulb temperature, and the input of
two properties that cannot be correct for one state (such as a humidity ratio greater than
saturation humidity ratio for dry air at a given dry bulb temperature). For these and other improper
inputs, the subroutine prints a warning, corrects one of the inputs and continues, or prints an error
and halts the simulation.

877

TRNSYS 16 Programmer's Guide

8.4.4.12. RCHECK
The RCHECK subroutine was added to TRNSYS 14 to provide input-output mismatch checking
for all components, especially those that do not have the benefit of INPUT/OUTPUT unit
connection checking as part of an interface program such as IISiBat or the TRNSYS Simulation
Studio. A call to RCHECK is of the form:
INTEGER ::INFO(15)
CHARACTER(len=3) :: YCHECK(:),OCHECK(:)
...
CALL RCHECK(INFO,YCHECK,OCHECK)

where
INFO

This is the standard INFO array used in TRNSYS

(see Section 8.4.3). It is used in the PSYCH subroutine to
identify the unit and type number of the component which calls
the subroutine, which will be printed during a simulation if
warnings or errors occur.

YCHECK

A 3-character array containing the expected input types for the

component.

OCHECK

A 3-character array containing the output types for the

component

The RCHECK array is passed the input and output types for each component in the simulation
and stores these variables in arrays with the same structure as the XIN and OUT arrays. When
the TRNSYS processor checks for input-output mismatches, the arrays are inspected to ensure
that the variable types for an input-output connection are consistent.
A user formulating a new component should utilize the input/output checking feature of TRNSYS
for two reasons: input/output checking ensures that TRNSYS connections are correctly defined,
and output information such as output type and output units are carried with the output and are
able to be processed by the TYPE 25 printer and the TYPE 57 unit conversion routine.
The first step in input/output checking is to characterize the inputs and outputs using Table
3.4.12.1. Users wishing to create additional input-output types or add a unit conversion to an
existing input-output type should modify the file UNITS.LAB, keeping the same format and style
as the original. Refer to the TYPE 57 unit conversion routine for more details on modifying this
file. The input-output types should be stored in the YCHECK and OCHECK arrays respectively.
The final step in using the input/output checking is to call the RCHECK array at the proper time,
with the proper arrays. The RCHECK routine should be called during the first iteration in most
cases (INFO(7)=-1 see Section 8.4.3) and after the call to the TYPECK subroutine (see Section
3.4.1).

878

TRNSYS 16 Programmer's Guide

The following example should help clarify the use of the RCHECK subroutine:
SUBROUTINE TYPE75(TIME,XIN,OUT,T,DTDT,PAR,INFO,ICNTRL,*)
!3 INPUTS AND 4 OUTPUTS FOR THIS TYPE
!INPUTS: TEMPERATURE IN (F)
!
MASS FLOWRATE IN (LBM/HR)
!
CONTROL SIGNAL (0 OR 1)
!OUTPUTS: TEMPERATURE OUT (C)
!
MASS FLOWRATE OUT (KG/HR)
!
HEAT TRANSFER RATE (KJ/HR)
!
HEAT TRANSFER RATE (W)
...
CHARACTER(len=3) :: YCHECK(3),OCHECK(4)
INTEGER :: INFO(15)
...
DATA YCHECK/TE2,MF3,CF1/
DATA OCHECK/TE1,MF1,PW1,PW2/
...
IF (INFO(7).EQ.-1) THEN
...
CALL RCHECK(INFO,YCHECK,OCHECK)
...
ENDIF
Table 8.4.4.12-1: TEMPERATURE

VAR. TYPE #

VAR. UNITS

VAR.TYPE

MULT. FACTOR

ADD. FACTOR

1
2
3
4

C
F
K
R

TE1
TE2
TE3
TE4

1.0
1.8
1.0
1.8

0
32
273.15
492

Table 8.4.4.12-2: LENGTH

VAR. TYPE #

VAR. UNITS

VAR. TYPE

MULT. FACTOR

ADD. FACTOR

1
2
3
4
5
6

m
cm
km
in
ft
miles

LE1
LE2
LE3
LE4
LE5
LE6

1
100
1000
39.3701
3.28084
6.21371 E-04

0
0
0
0
0
0

Table 8.4.4.12-3: AREA

VAR. TYPE #

VAR. UNITS

VAR. TYPE

MULT. FACTOR

ADD. FACTOR

1
2
3
4
5
6

m2
cm2
km2
in2
ft2
mi2

AR1
AR2
AR3
AR4
AR5
AR6

1
1 E+04
1 E-06
1550
10.7639
3.86102 E-07

0
0
0
0
0
0

Table 8.4.4.12-4: VOLUME

VAR. TYPE #

VAR. UNITS

VAR. TYPE

MULT. FACTOR

ADD. FACTOR

1
2
3
4
5
6

m3
l
ml
in3
ft3
gal

VL1
VL2
VL3
VL4
VL5
VL6

1
1000
1 E+06
6.10237 E+04
35.3147
264.172

0
0
0
0
0
0

879

TRNSYS 16 Programmer's Guide

Table 8.4.4.12-5: SPECIFIC VOLUME

VAR. TYPE #

VAR. UNITS

VAR. TYPE

MULT. FACTOR

ADD. FACTOR

1
2
3
4

m3/kg
l/kg
ft3/lbm
in3/lbm

SV1
SV2
SV3
SV4

1
1000
16.0185
2.76799 E+04

0
0
0
0

Table 8.4.4.12-6: VELOCITY

VAR. TYPE #

VAR. UNITS

VAR. TYPE

MULT. FACTOR

ADD. FACTOR

1
2
3
4
5

m/s
km/hr
ft/s
ft/min
mph

VE1
VE2
VE3
VE4
VE5

1.0
3.6
3.28084
196.85
2.23694

0
0
0
0
0

Table 8.4.4.12-7: MASS

VAR. TYPE #

VAR. UNITS

VAR. TYPE

MULT. FACTOR

ADD. FACTOR

1
2
3
4
5

kg
g
lbm
ounces
ton

MA1
MA2
MA3
MA4
MA5

1
1000
2.20462
35.274
1.10231 E-03

0
0
0
0
0

Table 8.4.4.12-8: DENSITY

VAR. TYPE #

VAR. UNITS

VAR. TYPE

MULT. FACTOR

ADD. FACTOR

1
2
3
4

kg/m3
kg/l
lbm/ft3
lbm/gal

DN1
DN2
DN3
DN4

1.0
0.001
6.2428 E-02
8.3454 E-03

0
0
0
0

Table 8.4.4.12-9: FORCE

VAR. TYPE #

VAR. UNITS

VAR. TYPE

MULT. FACTOR

ADD. FACTOR

1
2
3

N
lbf
ounce

FR1
FR2
FR3

1.0
0.224809
3.59694

0
0
0

Table 8.4.4.12-10: PRESSURE

VAR. TYPE #

VAR. UNITS

VAR. TYPE

MULT. FACTOR

ADD. FACTOR

1
2
3
4
5
6
7
8

BAR
kPa
Pa
ATM
psi
lbf/ft2
in. H2O
in. Hg

PR1
PR2
PR3
PR4
PR5
PR6
PR7
PR8

1
100
1 E+05
0.986923
14.5038
2.08854 E+03
401.463
29.53

0
0
0
0
0
0
0
0

Table 8.4.4.12-11: ENERGY

VAR. TYPE #

VAR. UNITS

VAR. TYPE

MULT. FACTOR

ADD. FACTOR

1
2
3
4

kJ
kWh
Cal
ft-lbf

EN1
EN2
EN3
EN4

1
2.77778 E-04
238.846
737.562

0
0
0
0

880

TRNSYS 16 Programmer's Guide

5
6

hp-hr
BTU

EN5
EN6

3.72506 E-04
0.947817

0
0

Table 8.4.4.12-12: POWER

VAR. TYPE #

VAR. UNITS

VAR. TYPE

MULT. FACTOR

ADD. FACTOR

1
2
3
4
5
6
7

kJ/hr
W
kW
hp
BTU/hr
BTU/min
Tons

PW1
PW2
PW3
PW4
PW5
PW6
PW7

1
0.277778
2.77778 E-04
3.72505 E-04
0.947817
1.57969 E-02
7.89847 E-05

0
0
0
0
0
0
0

Table 8.4.4.12-13: SPECIFIC ENERGY

VAR. TYPE #

VAR. UNITS

VAR. TYPE

MULT. FACTOR

ADD. FACTOR

1
2
3

kJ/kg
BTU/lbm
ft-lbf/lbm

SE1
SE2
SE3

1
0.429923
334.553

0
0
0

Table 8.4.4.12-14: SPECIFIC HEAT

VAR. TYPE #

VAR. UNITS

VAR. TYPE

MULT. FACTOR

ADD. FACTOR

1
2
3

kJ/kg-K
W-hr/kg-K
BTU/lbm-R

CP1
CP2
CP3

1
0.277778
0.238846

0
0
0

Table 8.4.4.12-15: FLOW RATE

VAR. TYPE #

VAR. UNITS

VAR. TYPE

MULT. FACTOR

ADD. FACTOR

1
2
3
4

kg/hr
kg/s
lbm/hr
lbm/s

MF1
MF2
MF3
MF4

1
2.77778 E-04
2.20462
6.12395 E-04

0
0
0
0

Table 8.4.4.12-16: VOLUMETRIC FLOW RATE

VAR. TYPE #

VAR. UNITS

VAR. TYPE

MULT. FACTOR

ADD. FACTOR

1
2
3
4
5
6
7

m3/hr
m3/s
l/hr
l/s
ft3/s
ft3/hr
gpm

VF1
VF2
VF3
VF4
VF5
VF6
VF7

1
2.77778 E-04
1000
0.277778
9.80958 E-03
35.3144
4.40286

0
0
0
0
0
0
0

Table 8.4.4.12-17: FLUX

VAR. TYPE #

VAR. UNITS

VAR. TYPE

MULT. FACTOR

ADD. FACTOR

1
2
3

kJ/hr-m2
W/m2
BTU/hr-ft2

IR1
IR2
IR3

1
0.277778
8.8055 E-02

0
0
0

Table 8.4.4.12-18: THERMAL CONDUCTIVITY

VAR. TYPE #

VAR. UNITS

VAR. TYPE

MULT. FACTOR

ADD. FACTOR

1
2
3

kJ/hr-m-K
W/m-K
BTU/hr-ft-R

KT1
KT2
KT3

1
0.277778
0.160497

0
0
0

881

TRNSYS 16 Programmer's Guide

Table 8.4.4.12-19: HEAT TRANSFER COEFFICIENTS

VAR. TYPE #

VAR. UNITS

VAR. TYPE

MULT. FACTOR

ADD. FACTOR

1
2
3

kJ/hr-m2-K
W/m2-K
BTU/hr-ft2-R

HT1
HT2
HT3

1
0.277778
4.89194 E-02

0
0
0

Table 8.4.4.12-20: DYNAMIC VISCOSITY

VAR. TYPE #

VAR. UNITS

VAR. TYPE

MULT. FACTOR

ADD. FACTOR

1
2
3
4
5
6

N-s/m2
kg/m-s
poise
lbf-s/ft2
lbf-hr/ft2
lbm/ft-hr

VS1
VS2
VS3
VS4
VS5
VS6

1
1
10
2.08854 E-02
5.80151 E-06
2419.08

0
0
0
0
0
0

Table 8.4.4.12-21: KINEMATIC VISCOSITY

VAR. TYPE #

VAR. UNITS

VAR. TYPE

MULT. FACTOR

ADD. FACTOR

1
2
3
4

m2/s
m2/hr
ft2/s
ft2/hr

KV1
KV2
KV3
KV4

1
3600
10.7639
3.87501 E+04

0
0
0
0

Table 8.4.4.12-22: MISCELLANEOUS

VAR. TYPE
Dimensionless

DM1

Degrees

DG1

Percentage

PC1

Month

MN1

Day

DY1

Hour

TD1

Control function

CF1

In user-written routines that have output variable types dependent on input variable types, such
as a statistics or integrator routine, the call to RCHECK must come on the second iteration of the
first time step, after a call to the EQUATER subroutine. The EQUATER subroutine will determine
the output connected to the unknown input and provide the appropriate input type. A call to the
EQUATER subroutine has the following form:
CHARACTER(len=3) :: TRUTYP,INDES,ODES
DOUBLE PRECISION :: X1,A1,X2,A2
INTEGER :: IU,INNO, OUTU,OUTNO,ICK1,ICK2
CALL EQUATER(IU,INNO,INDES,TRUTYP,OUTU,OUTNO,ODES,X1,A1,X2,A2,ICK1,ICK2)

where

882

IU

the unit number of the user-written component routine calling the

EQUATER subroutine (integer variable)

INNO

the input number with the undetermined input type (integer

variable)

TRNSYS 16 Programmer's Guide

INDES

the desired input variable type - not used in this capacity, set to
NAV for not available (3-character variable)

TRUTYP

a 3 character array returning the output variable type connected

to this input (this is the desired result for this application) (3
character variable)

OUTU

the unit number of the output connected to the unknown input

(not used in this capacity) (integer variable)

OUTNO

the output number of the OUTU unit connected to the unknown

input (not used in this capacity) (integer variable)

ODES

the desired output variable type for this unknown input - not used
in this capacity (3-character variable)

X1

the multiplication factor for the INDES variable - not used in this
capacity (double precision value)

A1

the addition factor for the INDES variable - not used in this
capacity (double precision value)

X2

the multiplication factor for the OUTDES variable - not used in

this capacity (double precision value)

A2

the addition factor for the OUTDES variable - not used in this
capacity (double precision value)

ICK1

error flag for calculation (1 = error found) (integer variable)

ICK2

error flag for calculation (1 = error found) (integer variable)

Users attempting to use the input/output checking algorithm in this capacity should refer to the
TYPE 55 Periodic Integrator subroutine and the EQUATER subroutine source code for more
details.

883

TRNSYS 16 Programmer's Guide

8.4.4.13. REWIND
The Rewind Utility is designed to rewind a data file whose end has been reached through multiple
read statements. It is able to then skip a specified number of steps from the beginning of the file
and leave a pointer ready to read the first non skipped line. A call to Rewind should have the
following form:
INTEGER :: LU,SKIP,IU,IT,IE
...
CALL REWIND(LU,SKIP,IU,IT,IE)

where
LU

(integer) the Fortran logical unit number ASSIGNed to the data

file that is to be rewound.

SKIP

(integer) the number of lines that should be skipped in the data

file once it is rewound.

IU

(integer) the unit number of the Type calling the Rewind routine.

IT

(integer) the type number of the Type calling the Rewind routine.

IE

(integer) an error code that is returned by the Rewind routine and

is set to zero if no errors were encountered while rewinding the
data file or skipping lines and which is set to a vale of 1 if errors
were encountered.

EXAMPLE
The following example shows the usage of the Rewind routine and a possible method for
checking whether errors occurred during the rewind process.
INTEGER :: LU,SKIP,IU,IT,IE,INFO(15)
...
LU = JFIX(PAR(1)+0.1)
SKIP = 10
IU = INFO(1)
IT = INFO(2)
CALL REWIND(LU,SKIP,IU,IT,IE)
IF (IE.GT.1) THEN
RETURN 1
ENDIF
...

884

TRNSYS 16 Programmer's Guide

8.4.4.14. SOLCEL
SOLCEL Models silicon cell I-V characteristics
Contributor:

Professor Don L. Evans, Arizona State University. Tempe, AZ 85281

(provided as-is. Check the source code for additional information)

The SOLCEL subroutine is called by Types 49 and 50 and uses a set of empirical values to
model the cells. Data for Spectrolab type cells, provided by Professor Evans, is "built-in" and is
used by default if no other data are supplied by the user. Users may supply their own data by
adding an extra parameter to the list of TYPE 49 or TYPE 50 in modes 5-8. This extra parameter
is interpreted as the logical unit number of the file containing the cell characteristics. The data
are read using 9F10.0 format and should be ordered as follows; (default values are listed after the
definition of the parameter):
First record:
ICELL

Cell type indicator, 0 for Spectrolab and Solarix Types, 1 for RCA types (peak
power tracking only for RCA); (0).

CELLPF

Ratio of cell area to absorber area; if ICELL = 1, the numerical value is ignored,
but a number must be present; (0.8).

Second record:
TL

A low temperature reference ( C); (43)

TH

A high temperature reference ( C); (175)

CL

A low reference concentration; (6)

CH

A high reference concentration; (20)

VOCLL

Open circuit voltage of a cell at temperature TL and concentration CL (volts);

(0.625)

ISCLL

Short circuit current of the cell at temperature TL and concentration CL (amps);

(0.649)

ISCLH

Short circuit current of the cell at temperature TL and concentration CH (amps);

(0.625)

ISCHL

Short circuit current of the cell at temperature TH and concentration CL(amps);

(2.16)

ISCHH

Short circuit current of the cell at temperature TH and concentration CH(amps);

(2.16)

Third record (Not used if ICELL = 1):

ACELL

Gross cell area (m2); (0.0004)

EG

Cell band gap at 0 Kelvin; (14000)

AOMIN

Value of Ao. at concentration of 0; (1)

DAODC

Slope of the Ao. vs concentration curve; (0.018)

RSH
RSMIN

Shunt resistance (ohms); (10000)

Value of series resistance when it becomes essentially independent of
concentration (ohms); (0.01)

885

TRNSYS 16 Programmer's Guide

886

CONTR

Minimum concentration for which the series resistance can be considered to be

equal to RSMIN; (10)

DRSDC

Slope of the series resistance vs concentration curve below a concentration

CONTR; (-0.005)

NSER

Number of cells to be wired in series in the array (10)

TRNSYS 16 Programmer's Guide

8.4.4.15. STEAM_PROPS (STEAM)

The double precision STEAM_PROPS routine and its single precision counterpart STEAM are
used to calculate the thermodynamic properties of steam based on correlations from the National
Bureau of Standards. Given two state properties, the STEAM_PROPS routine will return the
complete state of the fluid. A call to STEAM_PROPS is of the form
CHARACTER(len=2) :: UNITS
DOUBLE PRECISION :: PROP(7)
INTEGER :: ITYPE,IERR
...
CALL STEAM_PROPS(UNITS,PROP,ITYPE,IERR,*N)

where
UNITS

two character variable denoting unit system to be employed by

the fluids routine; SI for metric units and EN for english units

PROP

double precision array holding the two known thermodynamic

properties of steam upon entry and the complete state upon
return from the STEAM_PROPS routine. This array must be
dimensioned to 7 in the routine that calls STEAM_PROPS. Each
location in this array is described below:

PROP(1)

(double precision) temperature of the steam (F, C)

PROP(2)

(double precision) pressure of the steam (PSI, MPa)

PROP(3)

(double precision) enthalpy of the steam (BTU/lbm, kJ/kg)

PROP(4)

(double precision) entropy of the steam (BTU/lbm R, kJ/kg K)

PROP(5)

(double precision) quality (0 < x < 1)

PROP(6)

(double precision) specific volume (ft3/lbm, m3/kg)

PROP(7)

(double precision) internal energy (BTU/lbm, kJ/kg)

ITYPE

integer variable denoting which two properties are supplied as

known properties. ITYPE = 10* PROP indicator 1 + PROP
indicator 2 (12 to 76)

IERR

an integer variable returning the error messages from the

STEAM_PROPS call:
= 0 - no error found, calculations completed
> 0 - error found, calculations could not be completed

the corresponding line number to jump to if the subprogram is

present

An example illustrating the use of the STEAM_PROPS routine is shown with the following
considerations: SI units are desired, and Temperature (PROP(1)) and Quality (PROP(5)) are
known.

USE TrnsysFunctions
...
CHARACTER(len=2) :: UNITS
DOUBLE PRECISION :: PROP(7),XIN(2),TEMP,QUALITY,PRESSURE,ENTHALPY
...

887

TRNSYS 16 Programmer's Guide

100

UNITS = SI !SI units are desired.

...
TEMP = XIN(1)
QUALITY = XIN(2)
...
PROP(1) = TEMP
PROP(5) = QUALITY
CALL STEAM_PROPS(UNITS,PROP,INFO(1),INFO(2),IFLAG,*100)
CALL LINKCK('TYPE58','STEAM_PROPS',1,58)
IF (ErrorFound()) RETURN 1
...
PRESSURE=PROP(2)
ENTHALPY=PROP(3)

MATHEMATICAL DESCRIPTION
The correlations required to solve for the steam properties are from the EES program (8), a
numerical solver employing thermodynamic correlations from many different sources. Interested
users should contact the reference for more details on the correlations.
Enthalpies for steam are based on the following reference state: enthalpy equal to zero at 0C
This subroutine checks for much improper input, such as qualities less than 0. or greater than 1.,
and the input of two properties that cannot be correct for one state. For these and other improper
inputs, the subroutine prints a warning, corrects one of the inputs, and continues; or prints an
error and halts the simulation.
Subcooled properties are not available for steam. If a subcooled state is specified, the saturated
liquid results will instead be provided.

888

TRNSYS 16 Programmer's Guide

8.4.4.16. Storage of Data Between Time Steps

Since a single Type subroutine may be used for several UNITS in one simulation, care must be
taken when storing values that are to be used from one time step to the next. If a tank model, for
instance, requires the initial tank temperature throughout the simulation, simply setting a variable
TZERO to the tank temperature when INF0(7) = -1 will not work. When two tanks are used, the
local variable TZERO will be set twice. Throughout the simulation, it will contain the initial
temperature of the second tank. As a result, the first tank will not be modeled correctly.
One solution to this problem is to use the OUT array. Since each unit is allocated a minimum of
20 outputs, there is often enough space to store a few extra values which are not actual outputs.
The number of outputs can be increased beyond 20 by using INFO(6) (see section 8.4.3.2).
However, for large storage requirements (greater than 10 or so), it is recommended that a series
of calls to utility routine be used instead. With the release of TRNSYS 16, these utility routines
replace the S array and the STORE common block as the recommended solution to data
storage by components.
The first of the three utility routines reserves a requested number of spots in the global data
structure. The second routine retrieves the values from the global storage structure that
correspond to the current Type. The third sets new values in the global structure for the current
Type. The following section provides general information about the manner in which the three
storage related utility routines are called. An example of using storage in a Type follows.
SETSTORAGESIZE

INTERFACE
subroutine setStorageSize(n,INFO)
integer :: n, INFO(15)

DESCRIPTION
The setStorageSize subroutine should be called when INFO(7) = -1. This will tell the storage data
structure how many spots it needs to reserve for your component. The subroutine has two
arguments:
n: The integer number of storage spots required by this component.
INFO: The standard INFO array is sent so that the storage data structures know what unit
and type number are requesting storage.
GETSTORAGEVARS

INTERFACE
subroutine getStorageVars(storageArray,m,INFO)
integer :: m, INFO(15)
double precision :: storageArray(*)

DESCRIPTION
When the user wants to obtain a previously stored value (usually at the first call of a new time
step, i.e. when INFO(7) = 0), a call is made to the getStorageVars subroutine. The subroutine
takes the following arguments:

889

TRNSYS 16 Programmer's Guide

StorageArray: a locally defined double precision array dimensioned to n (see setStorageSize

above) that will contain the stored variables that are being retrieved from storage.
m: The integer number of variables that are to be retrieved. This argument allows the user to
retrieve only a subset of the stored variables. For example, if only the 5th stored variable is
desired, setting m to 5 will retrieve stored variables 1 through 5 and place them into spots 1
through 5 of storageArray
INFO: the standard INFO array is sent so that the storage data structures know what unit
and type number are requesting storage
SETSTORAGEVARS

INTERFACE
subroutine setStorageVars(storageArray,m,INFO)
integer :: m, INFO(15)
double precision :: storageArray(*)

DESCRIPTION
When the user wants to set storage at the end of a time step, a call is made to the
setStorageVars subroutine when INFO(13) > 0. Much like the getStorageVars subroutine, the
setStorageVars subroutine takes the following arguments:
storageArray: a locally defined double precision array dimensioned to n (see setStorageSize
above) that will contain the stored variables that are being retrieved from storage.
m: the integer number of variables that are to be stored. This argument allows the user to
store only a subset of an array that is used in the Type.
INFO: the standard INFO array is sent so that the storage data structures know what unit
and type number are requesting storage.

STORAGE EXAMPLE
A tank model which uses global storage to store the initial tank temperature might contain the
following lines:
...
C PERFORM FIRST CALL MANIPULATIONS
IF (INFO(7).EQ.-1) THEN
...
!reserve space in the double precision storage structure
StorageSize = 1
CALL setStorageSize(StorageSize,INFO)
...
!return to the calling program
RETURN 1
ENDIF
C PERFORM INITIAL TIME STEP MANIPULATIONS
IF (TIME.LT.(TIME0+DELT/2.)) THEN
...
!set initial values of variables in the double precision storage structure
Stored(1) = T
CALL setStorageVars(Stored, StorageSize,INFO)
...
!return to the calling program
RETURN 1

890

TRNSYS 16 Programmer's Guide

ENDIF
...
CALL getStorageVars(Stored,StorageSize,INFO)
TZERO = Stored(1)
...

The above example is exceedingly simplistic. The result of it is that no matter what is done to the
local variable TZERO later in the Type, it will always be reset to the initial value of T each time the
Type is called anew. A much more common situation arises when not only do we wish to use an
initial value in calculations but at the end of the time step, we wish to set the final calculated value
as the initial value for the next time step. In TRNSYS version 15.x and before, there was no way
for a Type to know when it was being called for the last time in a given time step. With the release
of TRNSYS version 16, a post convergence call to all Types was added, simplifying this
process. In the following modification to the above example, TZERO is set to an initial value.
TFINAL is calculated based on TZERO. At the end of each time step, the converged value of
TFINAL is set as the value of TZERO for the next time step.
...
C PERFORM POST CONVERGENCE MANIPULATIONS
IF (INFO(13).GT. 1) THEN
CALL getStorageVars(Stored,StorageSize,INFO) !recall the storage values
Stored(1)=Stored(2)
!update the TZERO spot with the value stored
!in the TFINAL spot.
CALL setStorageVars(Stored,StorageSize,INFO) !set the new storage values
ENDIF
C PERFORM FIRST CALL MANIPULATIONS
IF (INFO(7).EQ.-1) THEN
...
!reserve space in the double precision storage structure
StorageSize = 2
CALL setStorageSize(StorageSize,INFO)
...
!return to the calling program
RETURN 1
ENDIF
C PERFORM INITIAL TIME STEP MANIPULATIONS
IF (TIME.LT.(TIME0+DELT/2.)) THEN
...
!set initial values of variables in the double precision storage structure
Stored(1) = T
Stored(2) = T
CALL setStorageVars(Stored, StorageSize,INFO)
...
!return to the calling program
RETURN 1
ENDIF
...
!retrieve the stored value of the initial temperature for this time step.
CALL getStorageVars(Stored,StorageSize,INFO)
TZERO = Stored(1)
...
!calculate TFINAL based on the initial temperature.
TFINAL = TZERO + ...
...
!update the second storage spot with the calculated value of TFINAL
Stored(2) = TFINAL
CALL setStorageVars(Stored,StorageSize,INFO)
...

891

TRNSYS 16 Programmer's Guide

8.4.4.17. TABLE_COEFS (TABLE)

The double precision subroutine TABLE_COEFS and its single precision companion TABLE
return transfer function coefficients corresponding to the standard walls, partitions, floors, and
ceilings given in the ASHRAE Handbook of Fundamentals. Double Precision indicates that the
arguments to the TABLE_COEFS routine must be declared in the calling Type as DOUBLE
PRECISON instead of as REAL variables. The single precision version (called TABLE) was the
only version available in TRNSYS version 15.x and below. It was retained for backwards
compatibility but every effort should be made to use the double precision version:
TABLE_COEFS. The TABLE_COEFS routine is used by the TYPE 19 Zone Model. A call to
TABLE_COEFS is of the form
INTEGER :: ITYPE,NTABLE, IU,NBS,NCS,NDS
DOUBLE PRECISION :: B,C,D
...
CALL TABLE_COEFS(ITYPE,NTABLE,IU,NBS,NCS,NDS,B,C,D,*N)

where
ITYPE

an integer variable equal to 1, 2, or 3 denoting which type of

surface is desired. ITYPE = 1 corresponds to roof coefficients,
ITYPE = 2 is for exterior walls, and ITYPE = 3 denotes interior
partitions, floors and ceilings. The tables containing the surface
descriptions can be found in the documentation for TYPE 19;
both in this manual and in the Component Description manual.

NTABLE

(integer) number of wall, roof, etc., from the table specified by

ITYPE.

NBS

(integer) the number of b coefficients for the desired wall as

returned by the subroutine.

NCS

(integer) the number of c coefficients for the desired wall as

returned by the subroutine.

NDS

(integer) the number of d coefficients for the desired wall as

returned by the subroutine.

a double precision array dimensioned to 10 that contains the b

coefficients. These are the coefficients for the current and
previous sol-air temperatures (i.e. bo to bNBS-1).

a double precision array dimensioned to 10 that contains the c

coefficients. These are the coefficients for the current and
previous room temperatures (i.e. co to cNCS-1).

a double precision array dimensioned to 10 that contains the d

coefficients. These are the coefficients for the previous heat
fluxes (d1 to dNDS).

the corresponding line number to jump to if the subprogram is

present.

Subroutine Table_Coefs reads the transfer function coefficient from the file ASHRAE.COF
accessed through FORTRAN logical unit 8. This logical unit can be changed by modifying a data
statement within the subroutine. After the coefficients have been read for a particular wall, the
coefficients are stored internally within the subroutine for use on subsequent calls to
Table_Coefs.

892

TRNSYS 16 Programmer's Guide

Note: Transfer function coefficients in the file ASHRAE.COF are different from
those actually found in the ASHRAE handbooks. Because the TYPE 19 singlezone model handles the inside radiative heat transfer separately, the inside
radiative resistance is not included in the ASHRAE.COF transfer coefficients.
Thus while the file ASHRAE.COF is well-suited for use with TYPE 19 and other
models which calculate radiative heat transfer separately, it must not be
assumed that this file matches the ASHRAE handbook values.

893

TRNSYS 16 Programmer's Guide

8.4.4.18. TAU_ALPHA (TALF)

GENERAL DESCRIPTION
The double precision function Tau_Alpha and its single precision counterpart TALF calculate the
transmittance-absorptance product () of a solar collector as a function of angle of incidence of
radiation and collector construction. The transmittance of one or more sheets of glass or plastic
can be obtained by specifying an absorptance of 1. Tau_Alpha is a double precision routine,
meaning that all but its first argument must be declared in the calling Type as double precision
variables (as opposed to REAL variables). The value returned by Tau_Alpha is also double
precision. In TRNSYS versions 15.x and below, Tau_Alpha was known simply as TALF, which
was a single precision routine. TALF was retained for backwards compatibility and every effort
should be made to use the double precision Tau_Alpha routine.
A call to function TAU_ALPHA is of the form
DOUBLE PRECISION :: TAUALF,TAU_ALPHA,THETA,XKL,REFRIN,ALPHA,RHOD
INTEGER :: N
TAUALF = TALF(N,THETA,XKL,REFRIN,ALPHA,RHOD)

The function arguments are interpreted as follows:

N

(integer) number of covers

THETA

(double precision) angle of incidence (degrees)

XKL

(double precision) product of cover thickness and extinction

coefficient

REFRIN

(double precision) index of refraction

ALPHA

(double precision) absorber surface absorptance

RHOD

(double precision) reflectance of the N covers to diffuse radiation

If the value of RHOD is not known, then TALF can calculate it: When the sixth argument in the
function reference is less than zero, the reflectance of the N covers to diffuse radiation is
determined. The variable used as the sixth argument in the call to TALF is then set to this value
so that on subsequent calls to TALF the value of RHOD may be supplied.
Due to FORTRAN considerations, functions can not be checked for unlinked subroutines. Use
the LINKCK subroutine to warn the user that this function must be linked to the TRNSYS
executable.

MATHEMATICAL DESCRIPTION
The function subprogram Tau_Alpha calculates the transmittance-absorptance product () of a
solar collector for a given angle of incident radiation. The collector is described by the number of
glazings, N, each with refractive index ng and extinction length kL, and the absorptance of the
collector plate . The model uses Fresnel's equation for the specular reflectance at a planar
interface (assuming unpolarized incident radiation) and accounts for multiple reflections. The
transmittance for diffuse radiation is approximated as the transmittance for specular radiation at
an incident angle of 60. Fresnel's equations for reflectance at a planar interface for
perpendicular and parallel polarized radiation are the following:

894

TRNSYS 16 Programmer's Guide

perpendicular

1 =

sin 2 (1 2 )
sin 2 (1 + 2 )

parallel

2 =

Eq. 8.4.4-14

tan 2 (1 2 )
tan 2 (1 + 2 )

Eq. 8.4.4-15

The angles 1 and 2 are related by Snell's Law with the index of refraction for air of unity.

sin 2 =

sin 1
ng

Eq. 8.4.4-16

The transmittance and reflectance of a single glazing for each component of polarization,
including absorption and multiple reflections within the glazings, can be derived to give the
following [1,2]:

Ti ,1 =

a (1 i ) 2
2
2
1 a i

a 2 (1 i ) 2
Ri ,1 = i 1 +
2
2
1 a i

Eq. 8.4.4-17

Eq. 8.4.4-18

The internal transmittance, a, is a function of the extinction length and the transmission angle.

kL

cos 2

a = exp

Eq. 8.4.4-19

The transmittance and reflectance of N parallel plates, of the same material and thickness, can
be determined by the following recursive formulae.

Ti ,1Ti , N 1

Ti , N =
Ri , N

1 Ri ,1Ri , N 1
Ri ,1Ti , N
= Ri , N 1 +
1 Ri ,1Ri , N 1

Eq. 8.4.4-20
Eq. 8.4.4-21

The transmittance-absorptance product of the collector plate and glazing assembly can then be
calculated as

( ) =

(T

1, N

+ T2, N )

2(1 (1 ) RD )

where RD is the reflectance of the N covers to diffuse radiation.

reflectance for specular radiation at 60 incident angle.

Eq. 8.4.4-22

RD is calculated as the

895

TRNSYS 16 Programmer's Guide

8.4.4.19. TYPECK
The purpose of TYPECK is to perform some simple checks on the TRNSYS input file to ensure
the proper number of INPUTS, PARAMETERS, and DERIVATIVES have been specified for the
given component, and to terminate the simulation after compilation of the component input file
with appropriate error messages if errors are found. The form of a TYPECK call is:
INTEGER :: IOPT,INFO(15),NI,NP,ND
...
CALL TYPECK(IOPT,INFO,NI,NP,ND)

where
IOPT-

896

indicates to TYPECK what kind of check to perform:

<0

the compilation will be terminated immediately after

printing an explanatory error message.

>0

the simulation will be terminate after all UNITS have

been called once or 20 error messages have been
issued.

1 or -1

check the number of user furnished INPUTS,

PARAMETERS, and DERIVATIVES from the INFO
array against the values of NI, NP, and ND. An error
message will be printed only if TYPECK finds an
error.

2 or -2

TYPE routine has found unspecified error and wants

TYPECK to flag it.

3 or -3

TYPE routine has found a bad INPUT or number of

INPUTS and wants TYPECK to flag it.

4 or -4

TYPE routine has found a bad PARAMETER or

number of PARAMETERS and wants TYPECK to
flag it.

5 or -5

TYPE routine has found a bad DERIVATIVE or

number of DERIVATIVES and wants TYPECK to
flag it.

TYPE routine has determined that the maximum

number of UNITS of that TYPE has been exceeded
and wants TYPECK to flag it and stop the
simulation.

INFO-

TYPE's INFO array containing pertinent bookkeeping

data on the UNIT (see Section 8.4.3).

NI-

correct number of INPUTS

NP-

correct number of PARAMETERS

ND-

correct number of DERIVATIVES

TRNSYS 16 Programmer's Guide

8.4.4.20. VIEW_FACTORS (VIEW)

GENERAL DESCRIPTION
The double precision function View_Factors (and its single precision companion VIEW) calculate
the view factor between two planer rectangles of any orientation. For common orientations,
analytical expressions are used. With A call to View_Factors is of the form:
F21 = VIEW(ITYPE,A,B,C,D,E,F,G,H,X3,Y3,Z3,NA1,NA2)

The integer variable ITYPE defines the type of orientation between the two surfaces. There are
five possibilities. The arguments A, B, C, D, E, F, G, H, X3, Y3, and Z3 (all double precision) are
dimensions or coordinates that depend upon the orientation chosen. Not all of these variables
are used for each orientation. The possible orientations and required dimensions are shown in
Figures 8.6.7.16-1 to 8.6.7.16-5. The VIEW factor from surface 2 to 1 is determined by the
routine. The integer arguments NA1 and NA2 are the number of differential areas for each
surface, when numerical integration is used to determine VIEW factors. NA1 is only used for
ITYPE equal to 3, 4, or 5. NA2 is only for ITYPE equal to 5.

2
1
B

C
Figure 8.4.4.20-1:ITYPE=1; parallel identical rectangles.

A
1
B
2
C
Figure 8.4.4.20-2: ITYPE=2; perpendicular rectangles with a common edge.

G
1

2
A
C
Figure 8.4.4.20-3: ITYPE=3; nonidentical parallel rectangles.

897

TRNSYS 16 Programmer's Guide

A
2
B

F
1

C
Figure 8.4.4.20-4: ITYPE=4; perpendicular rectangles without a common edge.

(X2, Y2, Z2)

(X1, Y1, Z1)

(X3, Y3, Z3)

Z
Y
2
(0, 0, 0)

B
X

Figure 8.4.4.20-5: ITYPE=5; two rectangular surfaces of any orientation (X1=C, Y1=D, Z1=E, X2=F,
Y2=G, Z2=H from subroutine call statement).

MATHEMATICAL DESCRIPTION:

ITYPE = 1
The view factor from surface two to one of Figure 8.4.4.20-1 is determined with the following
expression taken from Reference 1.

2 (1 + X 2 )(1 + Y 2 )
F21 =
ln
XY 1 + X 2 + Y 2

1/ 2

+ 1 + Y 2 tan 1

X
1+ Y 2
Eq. 8.4.4-23

+ Y 1 + X 2 tan 1

Y
1+ X2

X tan 1 X Y tan 1 Y

where,
X = A/C
and
Y = B/C

ITYPE = 2

898

TRNSYS 16 Programmer's Guide

From Reference 1, the view factor from surface two to one of Figure 8.4.4.20-2 is

F21 =

1
1
1
Y tan 1 + X tan 1
Y
X
Y

)(

1 1+ Y 2 1+ X 2
+ ln
4 1+ X 2 + Y 2

) Y (1 + X
(1 + Y )(X

X2 +Y2

) X (1 + X
) (1 + X )(X
Y2

+Y 2
2
+Y 2

X 2 + Y 2 tan 1
2

)
)

+Y 2
2
+Y 2

X2

Eq. 8.4.4-24

where,
X = B/A
and
Y = C/A

ITYPE = 3
The VIEW factor from surface two to one of Figure 8.6.7.16-3 is found by numerical integration
using the following expression:

d 1 2

F21 =

dA1

Eq. 8.4.4-25

A1

A2

where Fd1-2 is the view factor from an elemental area dA1 on surface 1 to surface 2 and A1 and
A2 are the areas of surfaces 1 and 2, respectively. The number of differential areas used in the
integration, NA1, is specified as an argument in the call to VIEW.
The view factor Fd1-2 is determined by breaking surface 2 into 4 sections such that the normal to
the center of the element passes through a common corner of each section of surface 2.
From Reference 1,
4

Fd 1 2 =
i =1

Xi
Yi
Yi
Xi
1
tan 1
tan 1
+
2
2
2
2 1 + X 2
1 + Xi
1 + Yi
1 + Yi
i

Eq. 8.4.4-26

where

Xi =

Ai
B
and Yi = i
C
C

and Ai and Bi are dimensions of each section on surface 2, measured in the same directions as A
and B.

899

TRNSYS 16 Programmer's Guide

ITYPE = 4
The view factor from surface 2 to 1 of Figure 8.4.4.20-4 is also determined using Equation
8.6.7.16-3 The view factor Fd1-2 is found by breaking surface 4 into 2 sections such that the
normal to a common corner of the sections splits the element in half. From Reference 1,
2

Fd 1 2 =
i =1

1 1 1
tan

Yi
2

Yi
2

X i + Yi

tan 1

2
2
X i + Yi
1

Eq. 8.4.4-27

where

Xi =

A
C'
and Yi =
Bi
Bi

Bi is the dimension of each section measured in the direction of B. C' is the distance from the
edge of surface 2 to the center of the element on surface one.

ITYPE = 5
For two rectangular surfaces of arbitrary orientation as shown in Figure 8.4.4.20-5, the view factor
from 2 to 1 is:

F21 =

1
A2

cos 1 cos 2
dA2 dA1
S 2
A1 A2

Eq. 8.4.4-28

where S is the distance between the centers of elemental areas on each surface, 1 and 2 are
angles between the normals to each element and the line connecting their centers, and A1 and
A2 are the areas of surfaces 1 and 2. Equation (8.6.7.16-6) is solved by numerical integration.
The number of elemental areas used in the integration procedure, NA1 and NA2, are specified in
the argument list in the call to VIEW.
The distance S is

S=

( X d 2 X d 1 )2 + (Yd 1 Yd 2 )2 + (Z d 2 Z d 1 )2

Eq. 8.4.4-29

where (Xd1,Yd1,Zd1) and (Xd2,Yd2,Zd2) are the coordinates of the elemental areas on surfaces
1 and 2, respectively. For Surface 2,

cos 2 =

For Surface 1,

8100

Z d1 Z d 2
S

Eq. 8.4.4-30

TRNSYS 16 Programmer's Guide

v v
SN
cos 1 =
N S

Eq. 8.4.4-31

v
S is the vector originating at the center of the elemental area on surface 1 and ending at
v
the center of the second element on surface 2 and N is a normal vector to the elemental area on

where

surface 1.
In general,

v v

v
v
v
v
S = sxi + s y j + sz k
v
v
v
v
N = nx i + n y j + nz k

Eq. 8.4.4-32
Eq. 8.4.4-33

where i , j , and k are unit vectors in the x, y, and z directions, respectively. Substitution of
Equations (8.6.7.16-10) and (8.6.7.16-11) into Equation (8.6.7.16-9) gives

cos =

S x nx + S y n y + S z nz
2

Sx + S y + Sz

nx + n y + nz

Eq. 8.4.4-34

where,
Sx = Xd2 - Xd1
Sy = Yd2 - Yd1
Sz = Zd2 - Zd1

Eq. 8.4.4-35
Eq. 8.4.4-36
Eq. 8.4.4-37

If nz is arbitrarily chosen to be 1, then

nx =

K n yY 1 Z1

X1
X2 X2

K 1
Z1 Z 2
+
X1 X1

ny =
X2
Y1
Y2
X1

Eq. 8.4.4-38

Eq. 8.4.4-39

where,

X3

Z1 Z 3 Y 2

X
1

K=
X
2

1
Y 3
X 1

X2 X2
X3

Y 1
Z1 Z 2 Y 3
Y 1
X1 X1
X1

X 3 X 3
X2
Y 1 1
Y 1
Y 2
X1
X 1
X1

Eq. 8.4.4-40

8101

TRNSYS 16 Programmer's Guide

The coordinates X1, X2, X3, Y1, Y2, Y3, Z1, Z2, Z3 are as defined in Figure 8.4.4.20-5.

8102

TRNSYS 16 Programmer's Guide

8.4.5.

Solving Differential Equations

TRNSYS versions 11.1 and higher provide two methods for solving differential equations which
may occur in component models: an approximate analytical solution using the subroutine
Differential_Eqn (see Section 8.4.4.3), and a numerical solution using one of three numerical
methods (see Manual 07-TRNEdit). The analytical method is recommended whenever practical it
is possible to write the differential equation in the following general form:

dT
= aT + b
dt

Eq. 8.4.5-1

The numerical method generally requires shorter time steps and more computation for
comparable accuracy and numerical stability. Users intending to utilize the Powells method
TRNSYS solver (SOLVER 1) should use the Differential_Eqn subroutine whenever possible.
Numerical solutions to differential equations, when required can be obtained through the DTDT
and T arrays in the subroutine calling sequence. To use a very simple example, if the equation

dT
Q& i = C 1
dt

Eq. 8.4.5-2

were part of the mathematical model and the DERIVATIVE of the temperature T1 with respect to

& and a constant value C, it would appear

time t were to be integrated for a known heat flow rate Q
in the subroutine in a manner similar to that shown below.
.
.
DTDT(1) = QDOT / C
.
.
OUT(2) = T(1)
.
.

DTDT(1) is the differential dT1/dt, and T(1) is the result of integrating DTDT(1). In this example
T1 is also desired to be the second OUTPUT of the sequentially ordered OUTPUT list. Note that
the subscripts of OUT and T do not have to be the same, but that T(n) will always be the result of
integrating DTDT(n).
A DERIVATIVES control statement followed by an initial values line must be included among the
component control statements for any model which uses the DTDT array. The presence of a
DERIVATIVES command signals TRNSYS to integrate derivatives placed in the DTDT array and
to return results in the T array, repeating the procedure at each timestep until either the
appropriate error TOLERANCE or the iteration LIMIT is reached. The initial values of the
dependent variables (i.e., T(n)) are available to the component subroutine, if needed, through the
T array at the first call of the simulation (i.e., when INF0(7) = - 1; see Section 8.4.3).

8103

TRNSYS 16 Programmer's Guide

8.4.6.

Use of the Alternate Return (RETURN 1)

TRNSYS checks to make sure that all routines required by the TRNSYS input file are included in
the TRNSYS executable file by utilizing the FORTRAN alternate return. In previous TRNSYS
versions run with compilers allowing unresolved externals, unlinked subroutines were not
checked and were often the source of strange errors. The TRNSYS routines and all user-written
routines that call other subroutines should include an alternate return. The following example
should help the user understand the workings of the alternate return.
C* CALLING PROGRAM **
CALL TYPE2(TIME,XIN,OUT,T,DTDT,PAR,INFO,ICNTRL,*101)
100
WRITE(*,*) ERROR - TYPE 2 IS NOT LINKED !
CALL MYSTOP(1001)
RETURN 1
101
CONTINUE
.
.
C* CALLED SUBROUTINE **
SUBROUTINE TYPE2(TIME,XIN,OUT,T,DTDT,PAR,INFO,ICNTRL,*)
.
RETURN 1

The calling program will call the TYPE 2 subroutine and return to line 101 when the subroutine
calculations are completed due to the RETURN 1 line in the TYPE 2 subroutine. If the RETURN
1 is missing from the TYPE 2 subroutine or the subroutine is not linked, the program will return to
line 100 of the calling program and the error message will be flagged.
All user written routines should contain an asterisk (*) at the end of the subroutine designation
and RETURN 1s every place a RETURN is desired. Failure to do so will cause the unlinked
subroutine error message to be displayed and the program to terminate.
All user-written routines that call other subprograms should also include the alternate return
feature to check for unlinked subprograms. To activate the alternate return:
1)

Call the subprogram with an asterisk-line number combination at the end of the
subprogram designation.

2)

Specify the line number given in the step above as a CONTINUE line.

3)

Between the subprogram call and the CONTINUE statement, place an error
message and a:
CALL MESSAGES(-1,Error Message Text,fatal,INFO(1),INFO(2))

command - a call may instead be placed to the LINKCK subroutine (see Section
8.4.4.7).
4)

8104

Make sure all RETURNs in the subprogram are replaced with a RETURN 1
statement. Refer to the example given or a FORTRAN manual for more details.

TRNSYS 16 Programmer's Guide

8.4.7.

Use of the ICNTRL Array

The ICNTRL array, which was new to version 14, conveys to the component subroutines (TYPE1,
TYPE2, etc.) and to the TRNSYS equation processor, the states of the controller variables. With
the release of TRNSYS 14, a new control strategy was designed to eliminate the problems
caused by previous TRNSYS controllers sticking in order to solve the system of equations (see
Manual 07-TRNEdit). With TRNSYS 14, the TRNSYS executive program, rather than the
component models, directly controlled discrete variables. At the start of each time step, the
values of all discrete variables are known. TRNSYS 14 (and subsequent versions) force these
values to remain at their current state until a converged solution to the equations is obtained or an
iteration limit is encountered. During these calculations, the component models calculate the
desired value of the discrete variables but they do not actually change the settings. After a
converged solution is obtained, or the maximum number of iterations is exceeded, TRNSYS
compares the current discrete variable values with the desired values. If they do not differ and a
converged solution has been obtained, the calculations are completed for the time step.
Otherwise, TRNSYS changes the values of the discrete variables to the desired setting and
repeats this process, taking care to not repeat the calculations with the same set of discrete
variables used previously.
To incorporate the new control strategy into user-written component routines, the ICNTRL and
INFO arrays must be set up correctly. INFO(11) should specify the number of discrete control
variables in the component routine. ICNTRL(2*(i-1)+1) and ICNTRL(2i) will contain the control
state at the previous iteration for controller i in the component model. (controlled by the TRNSYS
processor) and the desired control state at this iteration respectively.
The ICNTRL array is an integer array and must therefore pass and receive only integer values.
The assumption in TRNSYS is that an ICNTRL value of 1 implies the control signal was enabled
while an ICNTRL value of 0 denotes a disabled control signal.
For example, a user wishes to describe an on/off controller with hysteresis effects (control signal
depends on previous control signal). The component formulation would be of the following format:
UDB=PAR(1)
! UPPER DEAD BAND TEMP. DIFFERENCE
LDB=PAR(1)
! LOWER DEAD BAND TEMP. DIFFERENCE
.
INFO(11)=1
! 1 CONTROL VARIABLE)
.
.
TH=XIN(1)
! UPPER INPUT TEMPERATURE
TL=XIN(2)
! LOWER INPUT TEMPERATURE
.
.
STATEOLD=ICNTRL(1) ! CONTROL STATE AT PREVIOUS SOLUTION
STATENEW=STATEOLD
IF(STATEOLD.EQ.1) THEN
IF((TH-TL).LT.LDB) STATENEW=0
ELSE
IF(TH-TL).GT.UDB)) STATENEW=1
ENDIF
ICNTRL(2)=STATENEW
.

At every iteration, the value of ICNTRL(1) holds the controller state at the previous solution while
ICNTRL(2) holds the desired controller state at this iteration. The TRNSYS processor will check
to see if ICNTRL(1) is equal to ICNTRL(2) at convergence. If the values are not equal at

8105

TRNSYS 16 Programmer's Guide

convergence, ICNTRL(1) will be set to the desired controller state at convergence (ICNTRL(2))
and the process will be repeated.
IMPORTANT!! If a new component makes use of the ICNTRL array, then SOLVER 1 must be
used in any TRNSYS input file that contains that component.

8106

TRNSYS 16 Programmer's Guide

8.4.8.

Converting Types from earlier TRNSYS Versions

As programming standards have evolved, so too have the conventions required by TRNSYS
Types. The following sections provide the user with information on the steps required to convert
Types written for earlier versions of TRNSYS to later versions.

8.4.8.1.

Converting components from version 13.x to version

14.1

TRNSYS routines written for version 13.1 or earlier have to be modified for use with TRNSYS
version 14.1 and higher. The steps are summarized below for convenience - refer to the sections
listed for full details.
1) Change the TYPE specification from:
CALL TYPEn(TIME,XIN,OUT,T,DTDT,PAR,INFO)

to:
CALL TYPEn(TIME,XIN,OUT,T,DTDT,PAR,INFO,ICNTRL,*)

2) Replace all RETURN statements in the subroutine with RETURN 1

Failure to replace all returns will result in an unlinked subroutine error. (see Section
8.4.6)
3) Change control algorithms to accommodate the Powells Method control strategy
(Optional) (see Section 8.4.7)
4) Change the XIN and OUT arrays to DOUBLE PRECISION variables. (see Section
8.3.2)
5) Change the INFO array dimension from 10 places to 15. (see Section 8.4.3)
6) Define the input and output variable types in the YCHECK and OCHECK arrays and
call the RCHECK subroutine to perform input/output checking. (see Section 8.4.4.12)
7) All calls to the following routines should be modified to include the link checking
option: DATA,ENCL,TABLE,INVERT,DINVRT,FIT,DFIT and PSYCH. (see Section
8.4.4.7)

8.4.8.2.

Converting components from version 14.1 to version

14.2

TRNSYS routines written for version 14.1 have to be modified for use with TRNSYS 14.2 for
Windows. The Windows operating system does not deal well with two common features of
TRNSYS version 14.1 subroutines. The steps are summarized below for convenience.
1)

Change any statements that contain a STOP statement from:

STOP

to the following if the Stop occurred in the primary Type subroutine:

8107

TRNSYS 16 Programmer's Guide

CALL MYSTOP(1001)
RETURN 1

or to the following if the Stop occurred in a secondary subroutine:

CALL MYSTOP(1001)
RETURN

The number 1001 can be changed to another error number if so desired.

2)
Windows does not allow any write-to-screen statements. Any WRITE or PRINT
statements that print to the screen such as:
WRITE (*,*) There is an error in this type.

need to be replaced with a statement that writes the text to the listing file.
WRITE(LUW,*) There is an error in this type

Note: In most cases, writing to the list file requires the inclusion of the common block that
contains the value for the logical unit. Include the following line:
COMMON /LUNITS/ LUR, LUW, IFORM, LUK

8.4.8.3.

Converting components from version 14.2 to version

15.x

There are no modifications necessary to a TRNSYS 14.2 Type in order for it to be used with
TRNSYS version 15.x.

8.4.8.4.

Converting components from version 15.x to version

16.x

The steps involved with converting TRNSYS components written for TRNSYS version 15.x to the
TRNSYS 16.x coding standard are included in section 8.2.1.3 of this manual.

8108

TRNSYS 16 Programmer's Guide

8.4.9.

Instructions to rebuild TRNDll.dll with Compaq

Visual Fortran 6.6

The following sections give detailed information on the compiler settings used to rebuild
TRNDll.dll. They are for reference only. To use the default workspace, launch CVF 6.6, go to
"File/Open workspace" and select %TRNSYS16%\Compilers\Cvf66\Cvf66.dsw. You do not have
to modify any setting to recompile and debug TRNSYS if you use the project included in the
distribution. The instructions hereunder are for reference only. They explain how this workspace
was created and give all modifications to default settings. You will also need to use the
instructions here below if you want to use the open-source version of Type 56 (see section
8.4.9.4).
Supported Compiler version: 6.6b (Note: there has been a 6.6c update but it appears to be
unstable and it introduces some "features" that are incompatible with many existing programs. It
was pulled back by HP-Compaq, then posted again. We advise against using it)

8.4.9.1.

Cvf66 Workspace

In CVF, a Project is a set of files and settings that are used to build an application (.exe or .dll). A
Workspace is a container for one or more projects. We will first create an empty Workspace:
Go to File/New, select the "Workspace" Tab
Select "Blank Workspace"
Place in %TRNSYS16%\Compilers
Workspace name: Cvf66

8.4.9.2.

TRNDll project

The TRNSYS kernel and the standard components are compiled and linked into one DLL,
TRNDll.dll. To create the project:
Go to File/New, select the "Project" Tab
Project type: select "Win32 Dynamic Link Library"
Select "Add to current workspace"
Project name: TRNDll
The folder name should read %TRNSYS16%\Compilers\Cvf66\TRNDll
Click OK
Select "Empty DLL"
Click Finish, then OK

8.4.9.3.

Project settings

Go to Projects/Settings
In the left-hand window, select the "TRNDll" project
In the "Settings for" drop-down list box showing "Win32 Debug", select "All configurations".
We will now choose the settings that are common to the Debug and Release configurations.

8109

TRNSYS 16 Programmer's Guide

There are numerous settings here controlled by a series of tabs at the top. Below are the
important settings for each of these tabs. Each tab can have several "Categories" controlled by a
drop-down list, so there are many settings involved. Only the settings that must be changed from
their default values are mentioned.
Table 8.4.9.3-1: TRNDll Settings CVF 6.6, All Configurations

Tab
Debug

Drop-down

Value

Executable for debug

session

..\..\..\Exe\TRNExe.exe

Program arguments

You can optionally enter the name of the deck file

that you wish to debug here, to bypass the file/open
dialog in TRNExe.exe

General

General
Compatibility

Fortran

Parameter

External
Procedures

Floating
Point

Runtime

Generate Source
Browse Information
Predefined
Preprocessor Symbols
Powerstation 4.0
compatibility options
Arguments Passing
Convention
String Length Argument
Passing
Floating point exception
handling
Extend precision of
Single precision
constants
Generate traceback
information
Runtime Error checking

Checked
TRNSYS_MULTI_DLL, TRNSYS_WIN32
Check "Libraries" and "Other Run-time Behavior"
C, by Reference
After All Args
0
Checked
Checked
Check all boxes

In the "Settings for" drop-down list box, select "Win32 Debug". We will now choose the settings
that only apply to the Debug configuration. Here again, only the settings that must be changed
from their default values are mentioned.
Table 8.4.9.3-2: TRNDll Settings CVF 6.6, Win32 Debug

Tab

Drop-down

Parameter

Value

Link

Input

Ignore Libraries

libcd.lib, msvcrtd.lib

Post-build description

Post-build: Copying Debug TRNDll.dll and

TRNDll.lib to the Exe folder

Post-build command(s)

copy "Debug\TRNDll.dll" "..\..\..\Exe\TRNDll.dll"

copy "Debug\TRNDll.lib" "..\..\..\Exe\TRNDll.lib"

Postbuild
step

In the "Settings for" drop-down list box, select "Win32 Release". We will now choose the settings
that only apply to the Release configuration. Here again, only the settings that must be changed
from their default values are mentioned.
Table 8.4.9.3-3: TRNDll Settings CVF 6.6, Win32 Release

Tab

Drop-down

Parameter

Value

Link

Input

Ignore Libraries

libcd.lib, msvcrtd.lib

Post-build description

Post-build: Copying Release TRNDll.dll and

TRNDll.lib to the Exe folder

Post-

8110

TRNSYS 16 Programmer's Guide

build
step

Post-build command(s)

copy "Release\TRNDll.dll" "..\..\..\Exe\TRNDll.dll"

copy "Release\TRNDll.lib" "..\..\..\Exe\TRNDll.lib"

Notes:
The ignored libraries must change depending on the libraries used in the precompiled files.
Currently Type56 is distributed in Release mode, i.e. linked against the release version of
the standard libraries. Consequently the "debug" version must be ignored (libcd.lib and
msvcrtd.lib). If a Debug version of Type56 was linked to the DLL, the "non-debug" version of
default libraries should be ignored (libc.lib and msvcrt.lib).
The post-build step makes sure the required files are copied to the Exe folder without adding
unnecessary compiler files to that folder and making sure a copy of those files stays in the
Debug or Release directory

8.4.9.4.

Project files

On the left-hand window, go to "TRNDll files". You can delete the "Resource Files" and "Header
Files" folders, which will not be used. Go to the "Source file" folder. Right-click on it and choose
"New folder". Create the "Kernel" folder. Right click on this folder and choose "Add files to folder".
Browse to %TRNSYS16%\SourceCode\Kernel and select all .for and .f90 files. Click OK to add
the files to the project.
Go back to the "Source file" folder. Right-click on it and choose "New folder". Create the "Types"
folder. Right click on this folder and choose "Add files to folder". Browse to
%TRNSYS16%\SourceCode\Types and select all .for and .f90 files (if any). Click OK to add the
files to the project.

TYPE 56 MULTIZONE BUILDING

You can either use the Open source version (which does not include all Type56 features) OR use
the precompiled .OBJ files. To use the precompiled files:
Go to the "TRNDll Files" root folder and create the "Precompiled" folder.
Select this new folder, right-click and choose "Add files to folder".
Browse to %TRNSYS16%\SourceCode\Type56\ and choose a subdirectory according to
your compiler (i.e. Cvf66 here). Select all .obj and .lib files (if any) and add them to the
project.
To use the Open Source version:
Go to "TRNDll Files" , "Source Files" and then "Types". Right-click, select "Add files to
folder".
Browse to Browse to %TRNSYS16%\SourceCode\Type56\OpenSource and select all .for
and .f90 (if any) files.

8.4.9.5.

Building the TRNSYS DLL (TRNDll.dll)

Make sure the right project (TRNDll) is active in Project/Set active project, in case you have
several projects.

Select the appropriate configuration (Debug or Release) in Build/Set Active Configuration

Go to Build/Rebuild all

8111

TRNSYS 16 Programmer's Guide

Note: You need to do a Rebuild all before compiling any file that uses TRNSYS data modules. If
you try to compile a routine using the data modules before performing a "Rebuild" you will get an
message saying "error in opening module file" and the compilation will fail. The "Rebuild"
operation compiles the modules and creates the module files (.mod) before compiling the files
that require them.

8.4.9.6.

Debugging TRNSYS

Debugging TRNSYS is easy using the Visual Studio: Just press F5 to start TRNSYS. If you left
"Project Settings/Debug/General/Program Arguments" empty, an "open file" dialog box will show
up and you can select the deck file from there.
Before starting the debugger, you can set/remove breakpoints anywhere in the code using the
F12 key. You can then step through the code one instruction at a time using the F10 key (F11 to
enter called subroutines or functions).
Hint: To force the compiler to stop when an error such as an "out of bounds"
array occurs, you have to set it to stop on exceptions. To do this, launch the
debugger (F5). While it is stopped in the File/open dialog box, come back to the
debugger and go to Debug/Exceptions. Select all exceptions in the lower text
field, select "Stop always" for the action and click on "change". CVF has a bug in
the display but if you then click OK and come back at the same place you will see
that your changes have been made. From now on, CVF will stop when such
exceptions are generated and it will indicate the line that caused the problem.

8.4.9.7.

Known issues with CVF 6.6

ERRORS ABOUT IFPORT.MOD, IFWIN.MOD, ETC. (FILE NOT FOUND)

This is a bug in CVF but it does not have any effect on TRNSYS. The "if*.mod" files are only used
when compiling and linking with the Intel Visual Fortran, but CVF checks for their existence (in the
wrong location) before even checking that it is going to use them. So you can safely ignore those
messages.

CVF UNNECESSARILY RECOMPILING SOME ROUTINES WHEN REBUILDING TRNDLL.DLL

CVF has a known issue that causes it to recompile some files even though they have not been
changed. This usually does not happen immediately but after using the project for some time you
may notice that some files are recompiled without a valid reason. To fix that issue, you can try the
following:
In CVF, with the right project and configuration selected, do a Build/Clean
In Windows explorer, browse to %TRNSYS16%\Compilers\Cvf66\TRNDll\Debug and delete
all
files
that
may
have
been
left
over.
Do
the
same
for
the
%TRNSYS16%\Compilers\Cvf66\TRNDll\Debug folder
Then do a Build/Update all dependencies
Then do a Rebuild all
This should solve the problem for a while. Moving the project from one location to another one or
from one computer to another one is known to cause the problem to reappear.

8112

TRNSYS 16 Programmer's Guide

HOW TO "FIX" A PROJECT THAT CRASHES WHEN YOU TRY DEBUGGING TRNSYS
Occasionally, some project files get corrupt and CVF crashes as soon as you try to debug
TRNSYS. This is a known compiler issue. The following fix has been found to work in most
cases:
Delete ALL files in the %TRNSYS16%\Compilers\Cvf66 directory except for Cvf66.dsw and
the subdirectories (only delete FILES, e.g. Cvf66.opt)
Delete ALL files in the %TRNSYS16%\Compilers\Cvf66\TRNDll directory except for
TRNDll.dsp and the subdirectories (only delete FILES, e.g. TRNDll.plg)
Delete
EVERYTHING
in
%TRNSYS16%\Compilers\Cvf66\TRNDll\Debug
%TRNSYS16%\Compilers\Cvf66\TRNDll\Release

and

Then reopen Cvf66.dsw. Some settings like breakpoints will be lost but the important project
settings are still there. The choice of debugging EXE is lost and you must re-enter
..\..\..\Exe\TRNExe.exe in Project settings / Debug / General / Executable for debug session.
If you are still unable to debug TRNSYS, please reinstall a fresh copy of the project from your
original installation media or recreate the workspace after deleting all existing files, according to
the instructions here above, after deleting all existing files in %TRNSYS16%\Compilers\Cvf66.

8113

TRNSYS 16 Programmer's Guide

8.4.10. Instructions to add a new project to the CVF66

workspace
This section explains how to add a new project to the CVF66 workspace provided with
TRNSYS 16 in order to create a library of user components.
Note: You can use the "Export as Fortran" command in the TRNSYS Studio to
generate a code template and a CVF project template with the settings that are
listed in this section. The instructions here below are given for reference only and
for users who want to control more advanced options. Please refer to the
"Getting Started" TRNSYS Manual and to the TRNSYS Studio manual for more
information on the "Export as Fortran" command.

8.4.10.1. Cvf66 workspace

The workspace file is %TRNSYS16%\Compilers\Cvf66\Cvf66.dsw. In Compaq Visual Fortran, a
workspace is essentially a collection of Projects, each project being a set of files that allow you to
build an executable program or a DLL when linked together. When you open the workspace, you
will see a list of included projects on the left (by default only the TRNDll project is present).

8.4.10.2. New Project

Go to File/New, select the "Project" Tab
Project type: select "Win32 Dynamic Link Library"
Select "Add to current workspace"
Choose a Project name, e.g. MyDll.dll
The folder name should read %TRNSYS16%\Compilers\Cvf66\MyDll
Click OK
Select "Empty DLL"
Click Finish, then OK
This should add the MyDll project to the Workspace explorer (docked window on the left side of
the screen by default).

8.4.10.3. Project settings

Go to Projects/Settings and in the left-hand window, select the "MyDll" project (or right-click
on the MyDll Project in the workspace explorer and select "Settings")
In the "Settings for" drop-down list box showing "Win32 Debug", select "All configurations".
We will now choose the settings that are common to the Debug and Release configurations.
There are numerous settings here controlled by a series of tabs at the top. Below are the
important settings for each of these tabs. Each tab can have several "Categories" controlled by a
drop-down list, so there are many settings involved. Only the settings that must be changed from
their default values are mentioned.
Table 8.4.10.3-1: New DLL Settings CVF 6.6, All Configurations

8114

TRNSYS 16 Programmer's Guide

Tab

Drop-down

General
Debug

Parameter

Value

Output files

..\..\..\Userlib

Executable for debug

session

..\..\..\Exe\TRNExe.exe

Program arguments

You can optionally enter the name of the deck file

that you wish to debug here, to bypass the file/open
dialog in TRNExe.exe

General

General
Compatibility
External
Procedures

Fortran
Floating
Point

Runtime

Generate Source
Browse Information
Powerstation 4.0
compatibility options
Arguments Passing
Convention
String Length Argument
Passing
Floating point exception
handling
Extend precision of
Single precision
constants
Generate traceback
information
Runtime Error checking

Checked
Check "Libraries" and "Other Run-time Behavior"
C, by Reference
After All Args
0
Checked
Checked
Check all boxes

In the "Settings for" drop-down list box, select "Win32 Debug". We will now choose the settings
that only apply to the Debug configuration. Here again, only the settings that must be changed
from their default values are mentioned.
Table 8.4.10.3-2: New DLL Settings CVF 6.6, Win32 Debug

Tab

Drop-down

Parameter

Value

Fortran

Preprocessor

Include and Use Paths

..\TrnDll\Debug

Link

Input

Ignore Libraries

See note hereunder

In the "Settings for" drop-down list box, select "Win32 Release". We will now choose the settings
that only apply to the Release configuration. Here again, only the settings that must be changed
from their default values are mentioned.
Table 8.4.10.3-3: New DLL Settings CVF 6.6, Win32 Release

Tab

Drop-down

Parameter

Value

Fortran

Preprocessor

Include and Use

Paths

..\TrnDll\Release

Link

Input

Ignore Libraries

See note hereunder

Notes:
The ignored libraries must be adapted to the precompiled object files or libraries that you link
to the DLL. The linker will tell you which libraries cause conflicts. The most frequent ones are
libc.lib and msvcrt.lib, as well as their debug versions, libcd.lib and msvcrtd.lib.

8.4.10.4. Project files

On the left-hand window, go to "TRNDll files". You can delete the "Resource Files" and "Header
Files" folders, which will not be used.

8115

TRNSYS 16 Programmer's Guide

Go to the "Source file" folder. Right click on this folder and choose "Add files to folder". Browse to
your source files (Types and possibly the utility routines they require) and add them.
Add the "TRNDll.lib" file that you will find in %TRNSYS16%\Exe to the project (e.g. in a
"Libraries" folder that you create). This will allow your DLL to use all the symbols exported by
TRNSYS.

8.4.10.5. Building the new DLL

Make sure the right project (MyDll) is active in Project/Set active project, in case you have
several projects.
Select the appropriate configuration (Debug or Release) in Build/Set Active Configuration
Go to Build/Rebuild all
MyDll.dll should be created in %TRNSYS16%\UserLib. You can debug components in this DLL
by making the project active in Project/Set active project and launching the Exe used for
Debugging (TRNExe.exe) by pressing F5.

8.4.10.6. Troubleshooting multiple DLL's

SHARING LOGICAL UNITS (FILES) BETWEEN DLL'S
Typical symptoms:
You assign a file in the TRNSYS input file and that file seems to be inaccessible to user-Types
compiled in an external DLL.
An "Inquire" statement tells you that the Logical Unit is not associated to a file
Attempts to write to / read from the file create "fort.nnn" (where nnn is the logical unit)
Solution:
If a TRNSYS Type uses a file that is opened and closed by the kernel (i.e. a file managed through
an "ASSIGN" statement in the deck, or an "External File" in a Studio proforma), the logical unit is
associated to the actual file in TRNDll. If a Type that accesses the file is compiled into a separate
DLL in UserLib:
Both DLL's must be linked to the DLL form of the Fortran run-time libraries. Use
"/libs:dll" in the command line, or set "Project Settings\Fortran\Libraries" to "Single-threaded
DLL" in the Visual Studio ("Debug Single-threaded DLL" for debug configuration).
Furthermore, all DLL's must be compiled in the same mode (Debug or Release). Note
that by default TRNSYS is shipped with a Release TRNDll.dll, but a Debug DLL is available
in "\Compilers\Cvf66\TRNDll\Debug".

8116

TRNSYS 16 Programmer's Guide

8.4.11. Instructions to rebuild TRNDll.dll with Intel

Visual Fortran 8.1
The following sections give detailed information on the compiler settings used to rebuild
TRNDll.dll. They are for reference only. To use the "solution" (Visual Studio .Net equivalent of the
Visual Studio 6.0 "workpace"), launch Microsoft Visual Studio .Net, go to "File/Open Solution" and
select the followinf file: %TRNSYS16%\Compilers\Ivf81-Mvs2003\Ivf81-Mvs2003.sln. You do not
have to modify any setting to recompile and debug TRNSYS if you use the project included in the
distribution. The instructions hereunder are for reference only. They explain how this solution was
created and give all modifications to default settings. You will also need to use the instructions
here below if you want to use the open-source version of Type 56 (see section 8.4.11.4).
Supported Compiler version: 8.1.019 or above. The latest tested version is identified by the
following Package ID: w_fc_p_8.1.019. You can find out the Package ID you are using by typing
"ifort" at the command prompt (you may have to add the path to ifort.exe to your PATH
environmental variable or to type the full path at the command prompt, by default: " C:\Program
Files\Intel\Fortran\compiler80\IA32\BIN\ifort.exe".
Note: The solution distributed with TRNSYS 16 is created using Microsoft Visual
Studio 2003. Unfortunately, Microsoft changed the solution file format between
Visual Studio .Net 2002 and 2003, so the solution is not usable using MVS .Net
2002. However, the distributed projects for Intel Visual Fortran can easily be
imported in a newly created solution if you use MVS .Net 2002. All the important
settings are in those projects, not in the solution.

8.4.11.1. Ivf81-Mvs2003 Solution

In Intel Visual Fortran, a project is set of files and settings that are used to build an application
(.exe or .dll). A Solution is a container for multiple projects. The solution itself can be built, which
builds all of the contained projects in a specified order (which can be adjusted by the user).
Go to File/New, select the "Blank Solution" option
In "name", type in "Ivf81-Mvs2003" (without the quotes)
For "Location", browse to %TRNSYS16%\Compilers
Click OK. This
Mvs2003.sln

will

create

the

file

%TRNSYS16%\Compilers\Ivf81-Mvs2003\Ivf81-

8.4.11.2. TRNDll project

The TRNSYS kernel and the standard components are compiled and linked into one DLL,
TRNDll.dll. To create the project:
Go to File / New / Project
Under "Project Types", select "Intel Fortran Project"
Under "Templates", select "Dynamic-link Library"
Type in "TRNDll" for the name
Check the "Add to solution" radio button. The Location will be adapted automatically and
should say: %TRNSYS16%\Compilers\Ivf81-Mvs2003

8117

TRNSYS 16 Programmer's Guide

Click OK.
This will launch the "Fortran Dynamic Link Library Wizard".
Select "DLL Settings" under "Overview", in the left frame, and choose "Empty Project" as
the DLL Type.
Click "Finish".
This
will
create
the
following
Project
file
with
%TRNSYS16%\Compilers\Ivf81-Mvs2003\TRNDll\TRNDll.vfproj

default

settings:

8.4.11.3. Project settings

In the Solution Explorer (right-hand window by default), right-click on the TRNDll project and
select "Properties"
In the "configuration" drop-down list box showing "Active (Debug)", select "All configurations". We
will now choose the settings that are common to the Debug and Release configurations.
There are numerous settings here controlled by a tree structure on the left. Below are the
important settings for each of these tabs. Each tree branch can have several "Categories", so
there are many settings involved. Only the settings that must be changed from their default
values are mentioned.
Table 8.4.11.3-1: TRNDll Settings IVF 8.1, All Configurations

Branch

Category

Parameter

Value

General
Data
Floating
Point

Predefined Preprocessor Symbols

Local Variable Storage
Extend precision of Single
precision constants
Check Array and String bounds
Check Edit Descriptor Data Type
Check Edit Descriptor Data Size
Check for Actual arguments using
temporary storage

TRNSYS_MULTI_DLL; TRNSYS_WIN32
All variables SAVE

Fortran
Run-time

Yes
Yes
Yes
Yes
Yes

In the "configuration" drop-down list box, select "Debug". We will now choose the settings that
only apply to the Debug configuration. Here again, only the settings that must be changed from
their default values are mentioned.
Table 8.4.11.3-2: TRNDll Settings IVF 8.1, Win32 Debug

Branch

Category

Debugging

Parameter

Value

Command

..\..\..\Exe\TRNExe.exe
You can optionally enter the name of the deck file
that you wish to debug here, to bypass the file/open
dialog in TRNExe.exe

Command arguments

Fortran

Libraries

Build
events

Post-build
Event

Runtime library
Command Line
Description

Debug Single-threaded DLL

copy "Debug\TRNDll.dll" "..\..\..\Exe\TRNDll.dll"
copy "Debug\TRNDll.lib" "..\..\..\Exe\TRNDll.lib"
Post-build: Copying TRNDll.dll and TRNDll.lib to the
Exe folder

In the "configuration" drop-down list box, select "Release". We will now choose the settings that
only apply to the Release configuration. Here again, only the settings that must be changed from
their default values are mentioned.

8118

TRNSYS 16 Programmer's Guide

Table 8.4.11.3-3: TRNDll Settings IVF 8.1, Win32 Release

Branch

Category

Parameter

Value

Fortran

Libraries

Runtime library

Single-threaded DLL

Build
events

Post-build
Event

Command Line
Description

copy "Release\TRNDll.dll" "..\..\..\Exe\TRNDll.dll"

copy "Release\TRNDll.lib" "..\..\..\Exe\TRNDll.lib"
Copying TRNDll.dll and TRNDll.lib to the Exe folder

Notes:
The post-build step makes sure the required files are copied to the Exe folder without adding
unnecessary compiler files to that folder and making sure a copy of those files stays in the
Debug or Release directory

8.4.11.4. Project files

In the Solution Explorer, you can optionally remove the "Header Files" and "Resource Files"
folders under "TRNDll", since they are not used (right-click on the folder and select
"Remove").
Add Kernel routines:
Right-Click on "Source Files" under the "TRNDll" project.
Select "Add", then "New Folder".
Create a folder and rename it to "Kernel".
Right-click on that newly-created folder and select "Add", then "Add Existing Item".
Browse to %TRNSYS16%\SourceCode\Kernel and select all .for and .f90 files.
Click OK to add the files to the project.
Add Standard Types:
Go back to the "Source file" folder.
Right-click on it and choose "New folder".
Create the "Types" folder.
Right-click on this folder and choose "Add", then "Add Existing Item".
Browse to %TRNSYS16%\SourceCode\Types and select all .for and .f90 files (if any).
Click OK to add the files to the project.

TYPE 56 MULTIZONE BUILDING

You can either use the Open source version (which does not include all Type56 features) OR use
the precompiled .OBJ files. To use the precompiled files:
Go to the "TRNDll / Source Files" root folder and create the "Precompiled" folder.
Select that folder, right-click and choose "Add", then "Add Existing Item".
Browse to %TRNSYS16%\SourceCode\Type56\ and choose a subdirectory according to
your compiler (i.e. Ivf81 here). Select all .obj and .lib files (if any) and add them to the project
(Tip: change the filter in "File Type" to "all files") .
To use the Open Source version:
Go to "TRNDll", "Source Files" and then "Types". Right-click, select "Add", then "Add
Existing Item".
Browse to Browse to %TRNSYS16%\SourceCode\Type56\OpenSource and select all .for
and .f90 (if any) files.

8119

TRNSYS 16 Programmer's Guide

8.4.11.5. Building TRNDll.dll

Make sure the right project (TRNDll) is active (its name should be in bold characters in the
Solution Explorer). If not, make it active (Select, right-click, choose "Set as StartUp Project").
Select the appropriate configuration (Debug or Release) in Build/Configuration Manager
Go to Build/Rebuild Solution
Note: You need to do a Rebuild before compiling any file that uses TRNSYS data modules. If you
try to compile a routine using the data modules before performing a "Rebuild" you will get an
message saying "error in opening module file" and the compilation will fail. The "Rebuild"
operation compiles the modules and creates the module files (.mod) before compiling the files
that require them.

8.4.11.6. Known issues with IVF 8.1

LINKER WARNINGS
You will get the two following warnings when building TRNDll:
LNK4049: locally defined symbol xxx imported
LNK4217: locally defined symbol xxx imported in function yyy
You can safely ignore those warnings.

8120

TRNSYS 16 Programmer's Guide

8.5. References
Engineering Equation Solver, F-Chart Software, 4406 Fox Bluff Road, Middleton WI, 1994.
www.fchart.com
ASHRAE. 2001. Handbook of Fundamentals. American Society of Heating Refrigeration, and Air
Conditioning Engineers, Atlanta, Ga.

8121