GT-SUITE

Controls Tutorials

VERSION 2016

by
Gamma Technologies

Copyright 2016 © Gamma Technologies LLC. All rights reserved.

All information contained in this manual is confidential and cannot be reproduced or
transmitted in any form or by any means, electronic or mechanical, for any purpose, without
the express written permission of Gamma Technologies LLC.
 GT Information

GT SUPPORT
• TELEPHONE: (630) 325-5848

• FAX: (630) 325-5849

• E-MAIL: [email protected]

• Web Address: www.gtisoft.com

• Address: 601 Oakmont Lane, Suite 220

Westmont, IL 60559
USA

Telephone Support Hours

8:00 A.M. to 5:30 P.M. Central Time Monday – Friday

 Table of Contents

TABLE OF CONTENTS
TUTORIAL 1: Defining and Using a PID Controller ........................................................................... 1
1.1 Introduction to PID Controllers (and alternative options) ............................................................... 1
1.2 Role about GT-SUITE Controllers .................................................................................................. 1
1.3 Controller Fundamentals ................................................................................................................. 2
1.3.1 Terminology ........................................................................................................................... 2
1.3.2 PID Controller Technical Background ................................................................................... 3
1.3.3 Controller and Plant Interaction ............................................................................................. 3
1.4 Methodology to Characterize the Plant and Build a Controller ...................................................... 5
1.4.1 Simulate the Step Function and the Plant's Response ............................................................ 5
1.4.2 Characterize the system by solving for K and τ ..................................................................... 5
1.4.3 Use K and τ to calculate the gains .......................................................................................... 6
1.4.4 Build the controller using the gains ........................................................................................ 6
1.4.5 Check the controller................................................................................................................ 6
1.5 Example: Target Boost Pressure by Wastegate Control .................................................................. 7
1.5.1 Simulate the Step Function and the Plant Response .............................................................. 7
1.5.2 Characterize the system by solving for K and τ ................................................................... 12
1.5.3 Use K and τ to calculate the gains ........................................................................................ 15
1.5.4 Build the controller using the gains. ..................................................................................... 15
1.5.5 Check the controller.............................................................................................................. 17
 Tutorial 1

TUTORIAL 1: Defining and Using a PID Controller

1.1 Introduction to PID Controllers (and alternative options)

A PID controller is used to control an input to a given system so that the output of that system reaches a
target value. There are many applications for a PID controller. A few simple examples include:

• controlling the wastegate diameter of a turbocharger to target a given boost pressure, turbine
RPM, or airflow rate
• controlling a throttle or injection rate to target engine speed or vehicle speed
• controlling a valve in a cooling system to target a coolant temperature

The generic nature of a PID controller--the fact that it can be used to target almost any quantity, also
means that effort is required to calibrate its "gains" so that the controller reaches the target:

1) In a stable manner--a steady output signal so that there is not oscillation around the target
2) In a prompt manner (i.e. reasonable time scale) so that excessive computational time is not required
waiting for the signal to move.

It can take time, and sometimes be difficult, to find a set of gains that reach the proper balance of
stability and prompt response; in fact, that is the purpose of this tutorial - to provide a means of finding
good gains. For this reason, GT-SUITE provides specialized templates, known as model-based
controllers, which can replace a PID controller for a few common engine control situations (see
"Targeting Controllers" group in the Controls tab of the template library), and several more are expected
to be added in the near future. In general, when a model-based controller is available, it should generally
be used instead of the PID because it will be both easier to use and more robust than a PID controller.

Examples of the model-based controllers include:

1) Throttle control for targeting part-load operation of SI engines
2) Injection rate control for targeting part-load operation of DI engines
3) EGR valve control for targeting and EGR fraction of and engine
4) Mass flow rate target control for orifices, throttles, and ball valves
5) Vehicle speed control (when using simple 'EngineState' engine)
6) Engine idle control (for integrated simulation with vehicle with simple 'EngineState' engine)

1.2 Role about GT-SUITE Controllers

It is common for "control engineers" (those engineers tasked with designing the control strategy for the
for an actual powertrain) to design controllers for tasks like the ones listed above using GT-SUITE, most
typically co-simulating with Simulink for the controls strategy. They typically focus a lot of attention on
simulating the system's behavior under the transient conditions, taking into account many details like
actuator and sensor rate limitations, interactions with other controllers, stability at all engine operating
conditions, safety considerations (i.e. acceleration limits), drivability, etc.

The purpose of the example in this tutorial is somewhat different…it is to create a "utility" for modeling.
This purpose is analogous to the many controllers that exist on laboratory software to target many things
such as engine speed, torque, EGR fraction, intercooler outlet temperature, etc. Most typically these
controllers are run at steady state or quasi-steady state, and designed strictly to aid the engineers working

Gamma Technologies © 2016

1
 Tutorial 1

the laboratory to run tests at particular conditions. These controllers are most typically not related to the
development efforts to make the engine control unit (ECU) strategy for the production engine/vehicle.

Another question that sometimes arises is the relationship between Simulink and GT-SUITE. Simulink is
the most common tool used in the automotive industry to develop ECU control strategies, and for several
good reasons. Simulink and GT-SUITE can readily be coupled together for controls studies. For more
information on co-simulation between GT-SUITE and Simulink please consult the Controls Coupling and
Real-Time manual (File\Manuals\Co-Simulation_And_UserCode\ControlsCouplingAndRealTime.pdf).

On the other hand, for the more modest objective of making simulation utility controllers it is more
convenient to use the controls library built into GT-SUITE. Reasons including the following:
1) GT-SUITE's library of controllers is often more convenient to use in GT-SUITE for such simple,
system-level tasks, especially in the case of post-processing and model building. (A key factor in
this is that Simulink does not have a "Case Setup" utility analogous to the one in GT-SUITE.)
2) Often times, the powertrain development engineers who need such controllers are not trained in
Simulink.
3) Software costs are minimized--to run GT-SUITE and Simulink coupled is not difficult, but it
does require a full license of both programs to run.
4) Support is available from Gamma Technologies for the combined powertrain and controls
system.

1.3 Controller Fundamentals

1.3.1 Terminology
System: The "system" refers to the simulation components in its entirety, including both the controller
and the plant.

Plant: The "plant" typically refers to the physical part of the system being modeled. In the example of
this tutorial, the engine, including its turbocharger, ambient, combustion conditions, injectors, etc., is the
plant.

Controller: The controller refers to the components that sense performance, make "decisions" according
to those results (typically through electronic means), and the actuators that in-turn operate on the plant.
For example, the wastegate is part of the plant, but a servo-motor that controls the wastegate position is
part of the controller.

Target Signal: The "target" value assigned to the controller. A well-designed controller will move the
actuator (such as injection rate, wastegate diameter, etc.) so that the plant's output (i.e. torque, BMEP,
manifold pressure, etc.) is forced to the specified target signal.

Response: The reaction of the plant to a change in an actuator input.

Characterize: The process of actuating the plant with some predefined inputs and studying the resultant
response of the plant in an attempt to describe its behavior. The plant can be thought of as a "black box",
and the goal is to find the coefficients to an equation, ideally a linear first order equation, that will
"characterize" the plant.

Gamma Technologies © 2016

2
 Tutorial 1

1.3.2 PID Controller Technical Background

GT-SUITE's control library includes the 'PIDController' template that is very convenient for building a
PID controller, although it is also possible to build one from scratch, using primitive control components,
such as gain, integral, transfer function, sum, product, etc.

Three values characterize a PID controller: Proportional gain, Integral gain, and Derivative gain. Please
refer to the documentation for 'PIDController' in the Reference Manual or on-line help for the exact
equations of the PID controller. The equations for a "true mathematical" PID controller are slightly
different than the governing equations presented in the documentation, but the GT-SUITE equations are
more commonly used in industry because of certain computational difficulties associated with true PID
mathematical differential equations.

A common practice in industry is to ignore the D (derivative) term of the PID controller. Derivative gain
is usually not necessary, and therefore its use creates an unnecessary complication when finding good
PID gains. This tutorial follows this practice by always setting the D term to 0.0, effectively making a PI
Controller. However, common practice is that it is still called a PID controller, despite the lack of a D
term.

The primary goal when building a PID controller is to find gains that will force the plant to the target
value as quickly as possible. In the controls field, the most common approach finding good gains is the
"trial-and-error" method, typically known as "controls calibration". Controls engineers use many
different utilities to automate the calibration process, and so it is typically an efficient method, especially
considering the complex interactions and behaviors that are exhibited by some engine and controller
combinations.

However, the controller is a time-based mechanism, and so the ability to run many trial-and-error tests
takes time—literally elapsed time. If the controlled event lasts a few seconds, then many controller
iterations can be run without consuming a great deal of time. However, if the event occurs very slowly,
the trial and error calibration method becomes much less attractive. Such is the case when running
simulations in GT-SUITE, because even though the events being modeled may occur over just a few
seconds in real time, the simulation may run significantly slower than real time, thus the computational
time can become very large. Therefore, an analytical method has been developed to find the P and I
gains.

1.3.3 Controller and Plant Interaction

The equations of a PI controller are perfectly complementary to a plant that behaves with a first-order
linear response. A first-order linear plant response is defined by the following reaction to a "step" input:

Gamma Technologies © 2016

3
 Tutorial 1

Step Input (X)

Actuator Input (X)

∆X

Time (t)

System Response (Y)

Y Response

∆Y

-1 0 1 2 3 4 5

Time (t)

Y = Yinitial + (K) (∆X) (1 – e-t/τ)

where: K = ∆Y / ∆X, also known as the output ratio

t = time
τ = time constant of the plant

Therefore, if the plant is assumed to have a linear first order response, then τ, ∆X, and ∆Y at some
operating condition are all that is needed to characterize the plant as a linear first order system. Once they
are known, complementary proportional and integral gains can be calculated analytically. This leads to
the following questions:
1) Does the plant (the engine) typically behave as a linear first-order system?
2) If so, how can τ, ∆X, and ∆Y be found?
3) Once the system has been characterized, how are the gains solved?

Fortunately, most of the engine's responses to inputs do behave closely enough to a first-order system for
the equations above to be useful. A diesel engine's boost pressure response to a change in wastegate
diameter is a very close match.

A spreadsheet (see $GTIHOME\V2016\xlsfiles\CONTROLS\controls.xls) and the information given in

this chapter are available to help with questions (2) and (3).

Gamma Technologies © 2016

4
 Tutorial 1

1.4 Methodology to Characterize the Plant and Build a Controller

The methodology to build the controller will be described in the given steps below, and then an example
of the methodology will be given. It is suggested that each step of this section be read concurrently with
the corresponding step of the example.

1.4.1 Simulate the Step Function and the Plant's Response

Run a simulation to determine the response of the plant to an input step. This is the easiest step to
understand, but also probably the easiest step in which a critical mistake can be made. Keep the
following in mind:
• Pick a nominal operating condition. Ideally the gains for the controller can be found at one
operating condition, and then used for all other operating conditions. Therefore, it is important
that non-controlled inputs such as engine speed be carefully chosen to represents the whole
operating range as much as possible. For example, for a wastegate controller, do not pick a very
low speed where the wastegate will hardly open - pick a higher engine speed where the wastegate
is active. Keep in mind what your ultimate objective is for the controller - if it is for a study at
2000 and 3000 RPM, then do the plant characterization at 2500 RPM.
• Carefully choose the two values of the input step so that the plant is "responsive" to the input,
and the inputs are "relevant". In a wastegate example, if the diameter is very small, the turbo
speed may exceed the turbo map, giving flawed simulation results. Also, if the wastegate is set
too big, the boost pressure prediction may be accurate and the plant may be responsive to the
wastegate change, but the operating condition will not be "relevant" because the wastegate would
never be open that far at low boost pressures.
• Save the transient output of step simulation, including filters, etc. Normally a 'SensorConn'
should be hooked to the item to be targeted, and make sure that its plot is turned on.
For all 'EngineCrankTrain' outputs such as torque, power, BMEP, speed, a 'SensorConn' is
connected to the 'EngineCrankTrain'. This is preferable to transient RLT data (i.e. 'RLTSensor')
because the sensor is updated every time that a cylinder fires, whereas the RLT variables only
update once per cycle.
When using a 'SensorConn' it is sometimes necessary to use a 'MovingAverage' part to filter the
oscillating signal. For example, when sensing boost pressure for a wastegate controller, the
pressure oscillations in the intake manifold should be averaged so that the wastegate diameter
does not also oscillate.
• Ensure that the simulation gets to steady-state, both before and after the step is imposed.

1.4.2 Characterize the system by solving for K and τ

In the "System Characterization" folder of the Excel control spreadsheet, specify the step input values
("prestep" and "poststep" respectively). However, care should be taken to specify the value that is most
linear for the system. For example, if flow is being controlled with a wastegate/throttle/valve, the flow
rate is approximately proportional to the effective cross-sectional area. Therefore, effective area, or
something proportional to it, such as a discharge coefficient or diameter-squared, is the best choice for
the control component, rather than diameter, throttle angle, valve lift, etc.

The spreadsheet also needs inputs for the steady state result, both before and after the step, and the time
at which the step is imposed. From this data, K will be calculated. The spreadsheet then has features that
can be used to easily find the time constant, τ.

Gamma Technologies © 2016

5
 Tutorial 1

1.4.3 Use K and τ to calculate the gains

The K and τ values will automatically be copied to the gains worksheet of the control spreadsheet. The
only inputs at this point are the settling time and damping ratio. Settling time is an "aggressiveness
factor" for the controller. If the plant had a perfect first-order behavior, it could be a very small value,
which is desirable because the controller will quickly force the system to steady state. However, the
controller can become unstable if the controller is too aggressive, especially if the plant does not have a
good fit in #2 above (unstable means that the controller and the output of the plant will continuously
bounce back and forth around the target, but never reach steady state). It is recommended that the settling
time be a value 2 to 3 times larger than the time constant, but longer when fit in #2 is poor. Under no
conditions should the settling time exceed 8 times the time constant. Finally, it is always recommended
that the damping ratio is set to 1.0.

1.4.4 Build the controller using the gains

Build the controller using the 'PIDController' template. Never limit the output of the 'PIDController'
downstream using other control components such as a 'Limiter' because a "wind-up" condition will occur.
Instead use the Limits within the controller, and if transient limits are necessary, use the 'PIDController'
input signal limits for this purpose.

The output from the 'PIDController' can then be connected to an 'ActuatorConn' to actuate the plant input.
If the input values specified in #2 above are not in a form consistent to the actuator (i.e. the inputs to the
spreadsheet were discharge coefficients, but the throttle is actuated by angle, or the inputs were diameter-
squared, but the wastegate is actuated by diameter), then use a 'Lookup1D' component to modify the
signal to the compatible input unit.

1.4.5 Check the controller

The system should be ready to go. Set a target for the controller and see how well it gets there. The best
way to monitor the controller is to select the "Performance Plot" from the 'PIDController' part's plot
options. If the system gets the steady-state value properly, but it is slower then desired, try making the
settling time in the spreadsheet less (see Step #3), and run again. If the output is unstable, then increase
the settling time, and re-apply the gains. There are also more detailed suggestions in the on-line
help/Reference Manual for 'PIDController' to modify its performance.

Gamma Technologies © 2016

6
 Tutorial 1

1.5 Example: Target Boost Pressure by Wastegate Control

This tutorial starts with the file %GTIHOME%\v2016\tutorials\Modeling_Applications\Controls\01-
PIDcontrol\PID-begin.gtm.

1.5.1 Simulate the Step Function and the Plant Response

An engine speed of 3000 RPM is a reasonable choice at which to build the PID controller. The original
example has a wastegate (WG) diameter = 12mm, which gives reasonable results close to the maximum
allowable turbine speed. Therefore, the WG needs to be studied in the region of 12 mm or greater for the
3000 RPM case, so that the map is not exceeded. Experience has shown that a step of only a few mm
typically results in a substantial change in boost pressure (especially for SI turbocharged engines).
Therefore, a modest step from 12 to 14 mm is chosen.

Open %GTIHOME%\v2016\tutorials\Modeling_Applications\Controls\01-PIDcontrol\PID-begin.gtm
and save it as PID-step.gtm. Edit the 'turb' object to impose a step at 1 second - Type "WG-Step" into the
object value field for Wastegate Diameter, then double-click on it to define it as a 'Profile Transient'
object. Fill in the Arrays folder as shown below to impose a step from 12 to 14mm. No changes are
needed in the Options folder.

Gamma Technologies © 2016

7
 Tutorial 1

When performing this procedure, make sure that the imposed step in WG diameter occurs after the
'Inertia Multiplier' attribute in the 'ShaftTurbo' object has returned to a value of 1.0. This can be checked
by opening the 'ProfilePeriod' object "inert-mult" from the 'tcshaft' object (edit 'tcshaft' and double click
"inert-mult"). In the Arrays folder, it can be seen that the "Inertia-Multiplier" returns to 1.0 after 18
cycles. Since 18 cycles = 0.72 seconds at 3000 RPM, the "Inertia-Multiplier" is indeed back at 1.0 after
1 second, the time of the imposed step. We may continue.

Gamma Technologies © 2016

8
 Tutorial 1

Go to Case Setup and turn off cases 2 and 3. Double check that RPM=3000 in Case 1.

Go to Run Setup. In the TimeControl folder, make Simulation Duration 100 cycles so that the simulation
will be able to run long enough to capture the tail of the response curve. Change the minimum number of
cycles to 30 (a value greater than the step change at 1 second=25 cycles) so that the simulation cannot
shut off until after the step function is imposed. You may also turn off Automatic Shut-off When Steady
State.

Gamma Technologies © 2016

9
 Tutorial 1

For this particular control situation, it is really the AVERAGE boost pressure that is to be targeted - it
would be impossible to make the pressure match the target at every crank-angle, due to the pressure
pulses from the engine; and even if we did, the wastegate would be constantly oscillating, which would
also be undesirable. Therefore, a 'MovingAverage' template will be used to smooth the signal before
feeding the signal into the controller. We want to include the 'MovingAverage' even in this step of
characterizing the system, since it contributes to the speed at which the system will respond to the input
step. Note: When using any of the model-based controllers mentioned in Section 1.1, it is not necessary
to pass the input signal through a 'MovingAverage' template because the controller will appropriately
condition the input signal by design.

Drag the 'MovingAverage' template from the "Controls" folder (Averaging and Filtering group) of the
template library into the PIDstep.gtm. Make a new object called "Boost-Smooth" for the moving average.
A window width of 1 cycle should be chosen. The other attribute values may be left as "def". Notice the
folder "Input/Output Signal Setup". Open that attribute folder and re-name the Input Signal Description
to "Intercooler Outlet Pressure (bar)" and the Output Signal Description to "Smoothed Intercooler Outlet
Pressure (bar)". This folder is available in a variety of controls templates, and allows the user to override
the signal labels, such that they assume more meaningful names. This is especially useful for very
complex models with many links, and can make complex model building and debugging significantly
easier.

Main Folder

Input/Output Signal Setup Folder

Gamma Technologies © 2016

10
 Tutorial 1

Let's assume that the boost pressure target is intended for the intercooler outlet - at the flowsplit part
named "icout." Drag the new 'Boost-Smooth' part on to the map, then link from the flowsplit to the
'MovingAverage'. A 'SensorConn' will automatically be placed in the middle, and you will be prompted
to choose the value to be sensed. Choose "Static Pressure". You will see that the Link ID for part Boost-
Smooth-1 is what was entered into the "Input Signal Description" attribute of the Boost-Smooth.

Then, turn on the plots in the 'MovingAverage' so that a record of the smoothed pressure is available.

Gamma Technologies © 2016

11
 Tutorial 1

Also, when the mouse is hovered over the link that was renamed, a tooltip textbox will appear which
displays the link name.

The model is ready. Run the simulation.

1.5.2 Characterize the system by solving for K and τ

Open the Excel spreadsheet %GTIHOME%\v2016\xlsfiles\CONTROLS\controls.xls. Resave it as
boost.xls. It asks for data about the step function simulation that we have just run. The first inputs to the
spreadsheet are the steady-state inputs and outputs before and after the step. The Pre-Step and Post-Step
Input Signals correspond to the input values of the wastegate: 12 and 14 mm degrees respectively.
However, WG controller will function better using the diameter2 because the flow rate through the
wastegate is approximately proportional to area, which is proportional to diameter2 (144 and 196 mm2, in
this case).

To get the steady-state outputs before and after the step, open PIDstep.gdx in GT-POST. View the output
plot of the 'MovingAverage' that we turned on in the previous step. (Compare this to the input plot to see
the purpose of the MovingAverage.) Then in the tree-view of GT-POST, expand the tree to find the data
set for the "Output". Right-click on it and select "View Data" to see the raw data in the plot. Find the
data just prior to 1.0 second (the step) and then at the simulation end. Note that your data may differ
slightly from the values shown below.

Gamma Technologies © 2016

12
 Tutorial 1

Enter the data just discussed into the boost.xls spreadsheet, as shown below. Since we are using
diameter2, enter 144 and 196 for the input signals. It also asks for the "Desired Number of Time
Constants". This is because we will be using the curve immediately above to "fit" the linear first-order
curve, thus allowing τ to be found. It is typically recommended that 1.5 times constant be chosen. The
Green-colored cell (B15) informs us that the value of Y (i.e. the sensed and filtered boost pressure) at
1.5τ is equal to 1.9756 bar in this case. Viewing the data again in GT-POST, find the time after the step
change when the pressure drops to the value shown in cell B15 and enter it into the spreadsheet.

Gamma Technologies © 2016

13
 Tutorial 1

Once the data has all been entered, the characterization is finished. Note the values for τ and K in the
"RESULTS" section of the spreadsheet. The fit of our equation using τ and K can be seen in the plot
below that was created in GT-POST by pasting the curve from the spreadsheet (just below the RESULTS
section) against the plot shown above by creating a new Report File (GU) in GT_POST. The range of the
plot may be shortened or lengthened by adjusting the time interval in cell B24.

Gamma Technologies © 2016

14
 Tutorial 1

1.5.3 Use K and τ to calculate the gains

Next task is to find the gains associated with τ and K. In the same spreadsheet, click on the "Gains
Worksheet" tab. The "Settling Time" input is generally recommended to be about 2-3 times the time
constant. However, the fit above is quite good, and experience with turbochargers has shown that a
shorter settling time of 15 to 20 engine cycles (0.6 to 0.8 seconds at 3000RPM) is usually a good choice.
The damping ratio should always equal 1. The gains are then calculated.

1.5.4 Build the controller using Gains Specified.

Resave PIDstep.gtm as PIDcontrol.gtm. Create a 'PIDController' object (found in the "Targeting
Controllers" section) called "Boost-Controller" using the gains calculated above. Specify a new
parameter called [Boost-Target] as the Target for the Input Signal value. Turn on (check) the "Display
Performance Monitor" attribute to get a monitor showing the input, target, and output signals of the
controller. For the minimum and maximum (in the "Limits" folder), remember that we are controlling the
diameter-squared. Let's assume a maximum wastegate diameter of 30 mm, and so the minimum is 0.0 and
the maximum is 302=900. Within the "Initialization" folder, initialize the wastegate to 0 and make the
Dwell Duration 3 cycles so that the controller is disabled for the first cycles when the simulation is not
yet initialized. The closer the initial output value is to the final value, the faster the controller will reach
its target, but in this case, the value will be set to 10 just to demonstrate the ability of the controller. No
changes are needed in the 'Convergence' or 'Solution Control' folders.

Gamma Technologies © 2016

15
 Tutorial 1

Even though we have chosen to control diameter-squared, the actuator option available on the 'turb' part
is "Wastegate Diameter (mm)". Use a 'MathFunction' template to perform a square-root operation on the
signal.

Gamma Technologies © 2016

16
 Tutorial 1

Some other details also need to be addressed:

• Edit the 'ShaftTurbo' component and set its "Inertia Multiplier" to 1.0. Because the time constant was
found when the inertia multiplier was 1.0, the controller should run with the same multiplier (it is
also no longer needed because the controller will "pull" the turbo speed to steady-state without the
artificially reduced turbo inertia).
• In the 'turb' object, enter any number (like 0.0) for the wastegate diameter, replacing the previous
reference to the 'ProfileTransient' object. The number will not be used, but the attribute cannot point
to a reference object when it is being overwritten by an 'ActuatorConn'.

Connect the control parts together as shown below. An 'ActuatorConn' will automatically be added to the
link between the 'MathFunction' and the 'turb' objects. Select "Wastegate Diameter" as the link input.

1.5.5 Check the controller

The control model is ready to run. In Case Setup, specify the [Boost-Target] as 2 bar for all three cases:
3000, 2500 and 2000 RPM. Turn all 3 cases on and run the model. The controller is successful because it
finds the target in a reasonable amount of time (i.e. efficiently) without excessive oscillations (i.e. good
stability). The results of the "Performance Plots" are shown here:

Gamma Technologies © 2016

17
 Tutorial 1

1.5.6 Build the controller using Gains Calculated.

The same exact controller can be built using the second set of radio-buttoned attributes titled "Gains
Calculated". The attributes "Slope, Time Constant (Tau), and Ratio of Settling Time to Time Constant"
can be used to fully-constrain the system. To do this, double-click on Boost-Controller, and instead
select the radio button for "Gains Calculated", as shown in the image below. Insert the Slope, Time
Constant, and Ratio of Settling Time to Time Constant that was calculated previously using the Excel
spreadsheet. Note that when this set of attributes is used, the damping ratio is always assumed to be 1.
Keep all other folders of the Boost-Controller the same, and re-run the simulation. The results of this
simulation should be identical as when using the "Gains Specified" radio button.

Gamma Technologies © 2016

18
 Tutorial 1

The benefit of using the "Gains Calculated" attributes is that the slope and time constant may be more
directly intuitive from some knowledge of the physical system itself. For this reason, it may be easier for
the user to simply make some educated estimates of the behavior of the system to achieve acceptable
controller behavior, rather than going through the calibration procedure for that particular system.

Gamma Technologies © 2016

19