SoftwareX 12 (2020) 100572

Contents lists available at ScienceDirect

SoftwareX
journal homepage: www.elsevier.com/locate/softx

Original software publication

Dyssol—An open-source flowsheet simulation framework for

particulate materials
∗
Vasyl Skorych , Maksym Dosta, Stefan Heinrich
Institute of Solids Process Engineering and Particle Technology, Hamburg University of Technology, Denickestrasse 15, 21073 Hamburg, Germany

article info a b s t r a c t

Article history: Dyssol is a modelling framework for the dynamic flowsheet simulation of processes designed for
Received 11 May 2020 handling of particulate materials. Main distinctive features of this software are the comprehensive
Received in revised form 16 July 2020 description of multidimensionally distributed particulate materials, the application of transformation
Accepted 16 July 2020
matrices and the use of sequential-modular simulation approach. This cross-platform system can be
Keywords: easily extended with new models, applied for calculation of large datasets and coupled to the external
Flowsheet simulation programme packages.
Dynamic modelling © 2020 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license
Multidimensional distributed parameters (http://creativecommons.org/licenses/by/4.0/).
Granular materials
Solids processes

Code metadata

Current code version v0.10.0.2

Permanent link to code/repository used of this code version https://github.com/ElsevierSoftwareX/SOFTX_2020_208
Code Ocean compute capsule -
Legal Code Licence BSD-3-Clause
Code versioning system used git
Software code languages, tools, and services used C++, Qt, HDF5, zlib, SUNDIALS, KISS FFT, Inno Setup.
Compilation requirements, operating environments & dependencies Windows, Visual Studio 14 2015, Qt 5.11.3 msvc2015 / Qt 5. 11.3 msvc2015_64,
Qt Visual Studio Tools for Visual Studio 2015, Git
If available Link to developer documentation/manual https://github.com/FlowsheetSimulation/Dyssol-
open/tree/master/Documentation
Support email for questions [email protected]

Software metadata

Current software version v0.10.0.2

Permanent link to executables of this version https://github.com/FlowsheetSimulation/Dyssol-open/releases/latest
Legal Software Licence BSD-3-Clause
Computing platforms/Operating Systems Microsoft Windows
Installation requirements & dependencies Visual C++ Redistributable for Visual Studio 2015
If available, link to user manual–if formally published include a reference to https://github.com/FlowsheetSimulation/Dyssol-
the publication in the reference list open/tree/master/Documentation
Support email for questions [email protected]

1. Motivation and significance

∗ Corresponding author. Complex production processes are used for the intermedi-
E-mail address: [email protected] (V. Skorych). ate handling or for the final manufacturing of the numerous

https://doi.org/10.1016/j.softx.2020.100572
2352-7110/© 2020 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
2 V. Skorych, M. Dosta and S. Heinrich / SoftwareX 12 (2020) 100572

products, from the processing of the iron ore and ending with the etc.) and is divided into a set of intervals according to the dis-
production of pharmaceutical compounds. They often consist of cretization grid. A material is defined as a set of values, where
several transformation steps occurring in different apparatuses, each one describes the mass fraction of particles with specific
which are connected into a single chain by material or energy combination of properties. An important aspect here is the correct
streams. These complex structures are referred to as flowsheets. and efficient handling of these large datasets. In Dyssol, all mate-
Flowsheet simulation frameworks can be effectively used to de- rial streams are described with a sparse data format implemented
scribe the behaviour of such processes, optimize them, develop based on the hierarchical tree topology [16]. This significantly
control strategies or to design new products. reduces the size of stored data and the amount of calculations.
In flowsheet simulations, each apparatus or transformation To describe changes in particle distributions, transformation ma-
step is described using separate units that are connected by trices are applied. They allow to avoid the loss of information
material streams, whereby the output of one unit can be con- about the secondary distributions and significantly extend the
nected to the input of another. From a computational point of applicability of units [9].
view, material streams are a general construct that describes the The reader is referred to our previous works to obtain more
flows of mass, energy or information and is applied to transfer detailed information about the simulation algorithm and anal-
them between units. Depending on the input streams and the ysis of its efficiency [8,16], the importance and treatment of
current state of the unit, the transformations occurring in the distributed parameters [9], and the validation of the modelling
unit are calculated, which leads to changes in the output streams. system [8].
Such flexible and extendable approach makes it easy to define
various process configurations, modify process parameters, and 2.2. Software architecture
numerically investigate their behaviour.
There are numerous commercial flowsheet simulation tools The Dyssol framework was developed using an object-oriented
like AspenPlus [1], gPROMS Formulated Products [2], ChemCad approach. Main software components can be divided into mod-
[3], etc. However, most of these programmes have been de- ules, which are schematically shown in Fig. 1.
veloped with a focus on liquid–vapour systems with simpli- The core of the system is the Flowsheet and Simulator modules.
fied description of the solid phase. As a result, a comprehensive The first one describes the structure of the simulated process and
numerical description of the particulate material cannot be ef- all the parameters of the included models. The latter one is re-
fectively performed in them. Due to various aspects, such as sponsible for the execution of the main computational algorithm
the description of the solid phase by interdependent multidi- and for the data transfer between units. Flowsheet contains a set
mensional distributed parameters, modelling of solids processes of Units, material Streams and information on the topology of the
requires specific treatment [4–6]. simulated process. All units have access to the library of built-in
In this contribution, we present a new open-source simulation Equations solvers.
framework Dyssol, which has been especially developed with a The most important of the Auxiliary modules are the following.
focus on modelling solids processes. This framework has been Material database that organizes access to materials and their
developed based on our previous research [7–9], in the scope properties and allows manipulating them using graphical user
of the DynSimFP project [10]—a priority research programme interface (GUI). Constant, temperature- and pressure-dependent
financed by German Research Foundation (DFG) from 2013 until parameters are supported, as well as properties describing the
2020. interaction of two different materials, such as interface tension.
For more details, see [16]. The Models manager module acts as an
2. Software description interface between the simulation system and units and solvers
from external libraries. Finally, the Data storage subsystem orga-
2.1. Methods nizes writing and reading of data on a hard drive and a runtime
data caching.
A sequential-modular approach [11,12] is used to perform Models of solvers and units are organized as shared libraries
calculations in Dyssol. Compared to the equation-oriented (simul- (*.dll or *.so). They can be developed separately from Dyssol
taneous) approach, modular simulation has a set of advantages and dynamically linked at runtime. To simplify the creation of
that play a decisive role, especially for treatment of solids pro- new models, the simulation system is equipped with a Models
cesses. The most important is that each single unit can be solved creator tool. It includes templates of new modules, examples of
separately using its own calculation strategy. This feature plays their implementation, and provides access to the programming
a crucial role in the case of particulate materials, where single interfaces (API) necessary to develop a fully-functional units or
process steps are often described with complex mathematical solvers.
models of heterogeneous nature. For example, coupling of several The software provides a simplified Graphical user interface to
units described with multidimensional population balance mod- configure units and streams connections and to perform initial
els (PBM) into one equation set can already make the application results analysis. Moreover, there exist Command-line interface,
of the equation-oriented approach unfeasible. which can be used to run simulations in batch mode. More
To perform data transfer between separate models and to detailed description of each module of the simulation system can
reach convergence, the waveform relaxation approach is applied. be found in [8] and [16].
This method can be effectively used for modular simulations
on the scale of the entire flowsheet and on the scale of indi- 2.3. Implementational aspects
vidual process units [13]. Furthermore, to improve overall effi-
ciency, several convergence methods (direct substitution, Weg- Dyssol is implemented with C++ programming language using
stein’s [14] and Steffensen’s [15] approaches) and several extrap- several external libraries and tools. The graphical user interface
olation methods (nearest neighbour, linear and cubic spline) are is based on the Qt library and Qt software framework [17]. The
implemented [16]. storage subsystem uses the hierarchical data format HDF5 and
In Dyssol, granular materials are described by discretized mul- the HDF5 library [18]. This ensures portability of the simulation
tidimensional interdependent distributed parameters. Each di- framework and allows the use of external packages to analyse and
mension corresponds to a specific particle property (size, shape, edit Dyssol files. Built-in equation solvers are implemented using
 V. Skorych, M. Dosta and S. Heinrich / SoftwareX 12 (2020) 100572 3

Fig. 1. Simplified representation of the architecture of the simulation system Dyssol.

the IDA [19] and KINSOL [20] solvers from the SUNDIALS [21] Table 1
package. The KISS FFT library [22] is utilized to perform a fast Main model parameters used for milling process.

Fourier transform, needed in some units and solvers. The Dyssol Inlet 1 Inlet 2
installation package for the Windows platform is assembled using Mass flow 0.1 kg/s until 4 h: 0 kg/s
the Inno Setup [23] utility. The Models creator tool is based on after 4 h: 0.3 kg/s
µsize 200 mm 3
350 mm3
a preconfigured solution for Visual Studio 2015, which includes
σsize 20 mm3 30 mm3
projects with required API libraries and templates of units. µmoisture 0.2 0.4
The main version of the programme is developed for Win- σmoisture 0.04 0.02
dows, where the graphical and command-line versions of the Screen 1 Screen 2
user interface are available. Microsoft Visual C++ 14.0 is used
Separation diameter 268.1 mm3 143.8 mm3
to compile and build Dyssol for Windows. For Linux, only the Separation sharpness until 2 h: 10 until 2 h: 10
command-line version exists, which is build using g++-7 com- after 2 h: 3.5 after 2 h: 3.5
piler. Mill 1 Mill 2
Hold-up mass 200 kg 100 kg
2.4. Software functionalities xlim 520 mm3 610 mm3
k 9.5 5.5
Dyssol allows to investigate the dynamics of complex pro- φ 0.5 0.9
duction processes applying extended description of granular ma- ψ 4.1 4.5
ϑ 2.5 3.2
terials with multidimensional distributed parameters and their
advanced treatment with the help of transformation matrices.
Solid, liquid and vapour phases are described by a set of overall
(mass flow, temperature, pressure), concentrated (phase fraction) – Improving performance by introducing advanced paralleliza-
and distributed (compound fraction) parameters. For the solid tion techniques.
phase, it is possible to extend the description with additional
dependent distributed parameters, such as porosity, moisture 3. Illustrative example
content or form factor.
Available standardized interfaces for implementing new mod- To illustrate the usage of Dyssol for complex processes, we
els and solvers allow one to expand the system with new mod- analyse the dynamics of the sieve-mill circuit. The flowsheet
ules. The current installation package (version v0.10) includes a structure is shown in Fig. 2 (left) and main model parameters are
set of the following steady-state and dynamic models: agglom- listed in Table 1. Screens are calculated according to the steady-
erator [24], bunker, crushers (e.g. [9,25]), cyclone [26], granula- state model of Molerus and Hoffmann [28]. Mills are dynamic
tor [27], heat exchanger, mixer, screens [28–31], splitter, time units, developed in [9], which apply Austin & Luckie [34] selection
delay. In addition to the solvers from the SUNDIALS package, Dys- S (x) and Austin [35] breakage B (x, y) functions in form
sol contains three solvers of agglomeration population balance ( )k
x
equations: cell average [32], fixed pivot [33] and FFT-based [24]. S (x) = , (1)
It is worth mentioning that the current version has some xlim
( )ψ−2 ( )ϑ−2
limitations compared to other commercially available products. φ·ψ x ϑ x
Therefore, in the next releases of Dyssol, the following challenges B (x, y) = · + (1 − b1) · · , (2)
y y y y
are to be addressed:
where x and y are volumes of particles, xlim is the maximum
– Improving GUI for visual configuration of flowsheets and particle size, and k, φ , ψ and ϑ are model parameters. Both units
extended results analysis. implement the approach with transformation matrices [9].
– Extension of the materials database by connecting to exist- Solid material is described by two interdependent distribu-
ing databanks and incorporating tools for thermodynamic tions: 200 classes of particle volumes ranging from 0 to 500 mm3
calculations. and moisture content on interval 0 to 50% divided into 50 classes.
– Adding the ability to handle multi-component particles. All distributed parameters are described with Gaussian function
– Extension of the model library with new relevant units. with mean value µ and standard deviation σ . The first inlet
– Simplification of the way of expanding the system with new continuously supplies small dry particles during the whole pro-
models. cess. The second inlet is switched on 4 h after the start of the
4 V. Skorych, M. Dosta and S. Heinrich / SoftwareX 12 (2020) 100572

Fig. 2. Flowsheet of the milling process (left) and initial material distributions in the inlets and the mills’ holdups (right).

simulation and adds large wet particles to the system. Initial Totally 8896 time points are generated, which is 5.4 ms per time
holdup material in both mills has the same distribution as Inlet point. All measurements were made on the computer equipped
1. Additional dynamics in the process arise due to a change in with CPU i7-7700K. More information on Dyssol performance and
the operating mode of screens, in which the separation sharpness the effect of various parameters can be found in [8].
changes after 2 h of process time. During the simulation, it is as-
sumed that there is no drying and no moisture exchange between 4. Impact
the particles.
Fig. 3 shows the dynamic response of the process to changes in The main advantages and distinctive features of the Dyssol
process conditions. The steady state is almost achieved after two framework are:
hours. At this time point, the parameters of screens are modified
– Dynamic simulation of complex process structures, consist-
and, as a consequence, the material mass flows and the particle
ing of interconnected heterogeneous units.
sizes change rapidly. However, after an additional hour, a new
– Application of the sequential-modular approach that simpli-
steady state is reached. Finally, at the time point of 4 h, the
fies extension of the system.
second inlet begins to supply large particles to the system. As a
– Utilization of the waveform relaxation method supplemen-
result, parameters of the material change again, and the process
ted by convergence and extrapolation methods for dynamic
stabilizes within the next hour.
calculations of process structures with recycling streams.
The time-dependent change of particle distributions is shown
– Description of the solid phase with multidimensional inter-
in Fig. 4. The size distribution responds to the variations in
dependent distributed parameters.
the process parameters, and also changes with time due to the
– Proper handling of granular materials using the transforma-
operation of the mills. One can see large disturbances of particle
tion matrix approach.
sizes at time points 0, 2 h and 4 h, when parameters of the process
– Possibility to efficiently process interdependent distributed
change rapidly, followed by a gradual stabilization. At the same
parameters with fine grids and a large number of classes due
time, changes in the dependent distributed parameter are caused
to dynamic data caching and special data formats.
only by the additional supply of a new material at time point 4 h.
– General applicability for various dynamic and steady state
For the rest of the time, the moisture content remains constant.
processes.
The simulation results demonstrate that the approach with
– Standardized interfaces for extending the system with new
transformation matrices implemented in Dyssol allows to avoid
units and solvers.
the loss of information about secondary distributed parameters
– High modularity and expandability of the system due to its
that are not explicitly included in the model equations. Despite
modular structure.
the fact that the screens and mills were designed considering only
– Open-source BSD licencing.
the particle size distribution, the moisture content was implicitly
handled and correctly calculated. Several research projects were already carried out using the Dys-
The calculation of this flowsheet with two-dimensional distri- sol system. It was applied to numerically analyse the dynamics
bution of material parameters (total 200 × 50 = 10000 classes) of a chemical looping combustion [36,37]. To simulate the pro-
until the first steady state takes around 1146 s. A total of 26011 cess, dynamic models of a circulating fluidized bed, a bubbling
time points is generated in 8 material flows, which corresponds to fluidized bed and a syphon, and a steady-state cyclone model
44.1 ms per time point. Calculating the case but only with a one- were developed. In this project, several simulation studies of the
dimensional distribution (200 classes by size), takes about 48 s. transient behaviour of the process were conducted using Dyssol.
 V. Skorych, M. Dosta and S. Heinrich / SoftwareX 12 (2020) 100572 5

Fig. 3. Response of the milling circuit to variation of process parameters.

Fig. 4. Dynamic change in particle size (left) and moisture content (right) during simulation for the product and mills’ holdups.

It was found that all developed models are generally applicable of more complex stirred tank reactor models for precipitation
in a specific parameters range. The performed numerical calcu- processes were developed and simulated in Dyssol.
lations showed good agreement between the simulation results In [24], a new approach to solving a one-dimensional PBM
and experimental data. for agglomeration process was proposed and investigated. The
Dyssol was also used to study the processes of particle for- new method assumes using of the separable approximation of the
mation during precipitation crystallization [38]. A confined im- agglomeration kernel and the subsequent fast Fourier transform
pinging jet mixer model was developed by combining a newly to speed up the calculations. Using Dyssol, the new method was
proposed limiting time scale approach with a population balance studied and compared with the well-known fixed pivot tech-
solver. This model was investigated in a combined framework nique, and its applicability to dynamic modelling of continuous
consisting of Dyssol, MATLAB, and PHREEQC [39] to study the ef- agglomeration processes with complex structures was analysed.
fect of supersaturation and mixing on the precipitation processes. Simulation studies have shown that the new method can provide
Based on the data obtained during the work, several versions a more effective solution if certain restrictions are met.
6 V. Skorych, M. Dosta and S. Heinrich / SoftwareX 12 (2020) 100572

In [40], a generic flowsheet model for precipitation of spar- [8] Skorych V, Dosta M, Hartge E-U, Heinrich S. Novel system for
ingly soluble salts was developed and implemented. Dyssol, cou- dynamic flowsheet simulation of solids processes. Powder Technol
2017;314:665–79. http://dx.doi.org/10.1016/j.powtec.2017.01.061.
pled with the PHREEQC hydrochemistry software, was used as
[9] Skorych V, Das N, Dosta M, Kumar J, Heinrich S. Application of transfor-
a model development environment and as a simulation tool. To mation matrices to the solution of population balance equations. Processes
verify the model, a comparison was made between the modelling 2019;7(8). http://dx.doi.org/10.3390/pr7080535.
and experimental precipitation of several compounds. Experi- [10] DFG-Priority programme SPP 1679. 2020, http://dynsim-fp.de/en/home/,
mental and model studies have shown good agreement between [Accessed 16 April 2020].
[11] Sargent RWH, Westerberg AW. SPEED-UP In chemical engineering design.
the results for both simple unidirectional circuits and complex
Trans Inst Chem Eng 1964;42:190. http://dx.doi.org/10.13140/2.1.2585.
schemes with recirculation of material. 6001.
New approaches to formulating transformation matrices from [12] Hillestad M, Hertzberg T. Dynamic simulation of chemical engineer-
population balance equations (PBE) for grinding and agglomera- ing systems by the sequential modular approach. Comput Chem Eng
tion processes were proposed in [9]. The finite volume method for 1986;10:377–88. http://dx.doi.org/10.1016/0098-1354(86)87008-9.
[13] Dosta M. Modular-based simulation of single process units. Chem Eng
spatial discretization and the second-order Runge–Kutta method
Technol 2019;42(3). http://dx.doi.org/10.1002/ceat.201800671.
to obtain the complete discretized form of the PBE were used [14] Wegstein JH. Accelerating convergence of iterative processes. Commun
to derive transformation laws. Applying the proposed approach, ACM 1958;1(6):9–13. http://dx.doi.org/10.1145/368861.368871.
dynamic units of a mill and an agglomerator were developed. [15] Mathews JH, Fink KD. Numerical methods using MATLAB. 3rd ed. Prentice
Using Dyssol, the resulting models were numerically investigated Hall; 1999, p. 90–100.
[16] Skorych V, Buchholz M, Dosta M, Heinrich S. Dyssol – a framework for dy-
applied to an exemplary production process.
namic simulation of interconnected solids processes. In: Heinrich S, editor.
Since several years, the Dyssol software is intensively used in Dynamic flowsheet simulation of solids processes. Springer International
the educational process at Hamburg University of Technology. It Publishing; 2020, http://dx.doi.org/10.1007/978-3-030-45168-4.
is applied to study the basics of steady-state and dynamic mod- [17] Qt – Cross-platform software development for embedded & desktop. 2020,
elling as part of the courses ‘‘Particle Technology’’ and ‘‘Multiscale https://www.qt.io/, [Accessed 11 May 2020].
[18] The HDF5⃝ R
library & file format. 2020, http://hdfgroup.org/solutions/hdf5/,
Simulation of Granular Materials’’.
[Accessed 16 April 2020].
[19] Hindmarsh AC, Serban R, Collier A. User Documentation for IDA V3.1.0
5. Conclusions (SUNDIALS V3.1.0). Center for Applied Scientic Computing Lawrence
Livermore National Laboratory; 2017.
An open source platform for flowsheet simulations of complex [20] Collier AM, Hindmarsh AC, Serban R, Woodward CS. User Documenta-
tion for KINSOL V3.1.0 (SUNDAILS V3.1.0). Center for Applied Scientific
dynamic processes involving particulate materials, named Dyssol,
Computing Lawrence Livermore National Laboratory; 2017.
has been presented. Important features of the software are the [21] SUNDIALS: SUite of Nonlinear and DIfferential/ALgebraic Equation Solvers.
high flexibility, customizability and extensibility of the system, 2020, http://computing.llnl.gov/projects/sundials, [Accessed 16 April 2020].
which provides the widest application scope. [22] A Fast Fourier Transform (FFT) library KISS FFT. 2020, http://github.com/
Thanks to the advanced description of the solid phase using mborgerding/kissfft, [Accessed 16 April 2020].
[23] Inno Setup. 2020, http://jrsoftware.org/isinfo.php, [Accessed 16 April 2020].
multidimensional distributed parameters, as well as the pro-
[24] Skorych V, Dosta M, Hartge E-U, Heinrich S, Ahrens R, Le Borne S. Inves-
posed methods for their handling utilizing transformation ma- tigation of an FFT based solver applied to dynamic flowsheet simulation
trices, this system provides ample opportunities for studying of agglomeration processes. Adv Powder Technol 2019;30(3):555–64. http:
various processes involving granular materials. //dx.doi.org/10.1016/j.apt.2018.12.007.
[25] Bond FC. Crushing and grinding calculations. Part I. Br Chem Eng
1961;6(6):378–85.
Declaration of competing interest
[26] Muschelknautz E. Die Berechnung von Zyklonabscheidern für Gase. Chem
Ing. 1972;44(1/2):63–71. http://dx.doi.org/10.1002/cite.330440112.
The authors declare that they have no known competing finan- [27] Heinrich S, Peglow M, Ihlow M, Henneberg M, Mörl L. Analysis of the start
cial interests or personal relationships that could have appeared up process in continuous fluidized bed spray granulation by population
to influence the work reported in this paper. balance modelling. Chem Eng Sci 2002;57:4369–90. http://dx.doi.org/10.
1016/S0009-2509(02)00352-4.
[28] Molerus O, Hoffmann H. Darstellung von windsichtertrennkurven durch
Acknowledgement ein stochastisches modell. Chem Ing Tech 1969;41(5+6):340–4. http://dx.
doi.org/10.1002/cite.330410523.
This work was supported by the German Research Foundation [29] Plitt LR. The analysis of solid–solid separations in classifiers. CIM Bull
(DFG) via Priority Programme SPP 1679 ‘‘Dynamic simulation of 1971;64(708):42–7.
[30] Hennig M, Teipel U. Stationäre Siebklassierung. Chem Ing Tech
interconnected solids processes’’. 2016;88:911–8. http://dx.doi.org/10.1002/cite.201500161.
[31] Radichkov R, Müller T, Kienle A, Heinrich S, Peglow M, Mörl L. A numerical
References bifurcation analysis of continuous fluidized bed spray granulation with
external product classification. Chem Eng Process 2006;45:826–37. http:
[1] Aspen plus. 2020, http://www.aspentech.com/en/products/engineering/ //dx.doi.org/10.1016/j.cep.2006.02.003.
aspen-plus, [Accessed 16 April 2020]. [32] Kumar J, Peglow M, Warnecke G, Heinrich S. An efficient numerical
[2] PSE Products – gPROMS FormulatedProducts. 2020, http://psenterprise. technique for solving population balance equation involving aggregation,
com/products/gproms/formulatedproducts, [Accessed 16 April 2020]. breakage, growth and nucleation. Powder Technol 2008;182(1):81–104.
[3] CHEMCAD – Chemical engineering simulation software by chemstations. http://dx.doi.org/10.1016/j.powtec.2007.05.028.
2020, www.chemstations.com/CHEMCAD/, [Accessed 16 April 2020]. [33] Kumar S, Ramkrishna D. On the solution of population balance equa-
[4] Rossiter AP, Douglas JM. Design and optimization of solids processes. Part tions by discretization – I. A fixed pivot technique. Chem Eng Sci
1. A hierarchical decision procedure for process synthesis of solids systems. 1996;51(8):1311–32. http://dx.doi.org/10.1016/0009-2509(96)88489-2.
Chem Eng Res Des 1986;64:175–95. [34] Austin LG, Luckie PT. The estimation of non-normalized breakage
[5] Hartge EU, Pogodda M, Reimers C, Schwier D, Gruhn G, Werther J. distribution parameters from batch grinding tests. Powder Technol
Flowsheet simulation of solids processes. KONA 2006;24. http://dx.doi.org/ 1971;5(72):267–71. http://dx.doi.org/10.1016/0032-5910(72)80030-5.
10.14356/kona.2006017. [35] Austin LG. A discussion of equations for the analysis of batch grinding data.
[6] Dosta M, Litster J, Heinrich S. Flowsheet simulation of solids processes: Powder Technol 1999;106:71–7. http://dx.doi.org/10.1016/S0032-5910(99)
Current status and future trends. Adv Powder Technol 2020;31(3):947–53. 00047-9.
http://dx.doi.org/10.1016/j.apt.2019.12.015. [36] Haus J, Hartge E-U, Heinrich S, Werther J. Dynamic flowsheet simulation
[7] Dosta M, Heinrich S, Werther J. Fluidized bed spray granulation: Analysis of of gas and solids flows in a system of coupled fluidized bed reactors
the system behaviour by means of dynamic flowsheet simulation. Powder for chemical looping combustion. Powder Technol 2017;316:628–40. http:
Technol 2010;204:71–82. http://dx.doi.org/10.1016/j.powtec.2010.07.018. //dx.doi.org/10.1016/j.powtec.2016.12.022.
 V. Skorych, M. Dosta and S. Heinrich / SoftwareX 12 (2020) 100572 7

[37] Haus J, Hartge E-U, Heinrich S, Werther J. Dynamic flowsheet simulation [39] PHREEQC Version 3. 2020, http://usgs.gov/software/phreeqc-version-3,
for chemical looping combustion of methane. Int J Greenhouse Gas Control [Accessed 27 April 2020].
2018;72:26–37. http://dx.doi.org/10.1016/j.ijggc.2018.03.004. [40] Rehage H, Scherer S, Kind M. A steady-state precipitation model for flow-
[38] Metzger L. Process simulation of technical precipitation process – The sheet simulation and its application. Comput Chem Eng 2019;128:524–37.
influence of mixing. Karlsruhe: KIT Scientific Publishing; 2018, http://dx. http://dx.doi.org/10.1016/j.compchemeng.2019.06.030.
doi.org/10.5445/KSP/1000076187.