Computer-aided modelling template: Concept and application
Introduction
Computer-aided modelling is an important enabler in facing current and future challenges in product–process engineering. Modelling technology helps to reduce number of resource-demanding experiments and, at the same time, to deliver truly innovative solutions. Process model of reliable quality allows predicting process behaviour and based on this, to optimize the process and to improve understanding of the domain system.
However, the modelling technology also requires the necessary methods and tools to perform these functions. The construction of models is mainly manual, which is still not very effective. Manually developed models usually can be applied only for the specific problem for which it was developed, such models are usually not well documented and therefore difficult to maintain. Note also that currently, there is no systematic documentation approach as well as a knowledge representation system for model storage and retrieval, except for what is available in process simulators such as PROII™, ASPEN-PLUS™ and GPROMS™. Therefore, the reuse of existing modelling codes is minimal and every new modelling case requires full model development steps and involvement of expertise at a high level. Thus, there is a need for a system that would assist in the model development and use/reuse. This system should fulfil certain requirements. For example, it should allow generation of various types of models with corresponding degrees of details; it should provide expert help and guidance in the selection of model building blocks and in the generation of the model equations within a new building block; provide a knowledge representation structure for a model library together with model documentation; allow easy reuse and modification of existing models. This is in agreement with the recommendations of Foss et al. (1998), which emphasizes that a solid understanding of the model development process is required for the development of modelling tools, which in turn aims to improve the efficiency of the modelling process and the quality of the models. However, Foss et al. (1998) also notes the lack of implementation of modelling techniques in existing computer-aided modelling tools. More recently, Klatt and Marquardt (2009) have identified prospective improvements for the present modelling tools, such as multi-scale modelling features together with documentation, maintenance and reuse of models in an efficient and economical way. The benefits of the integration of modelling tools are highlighted in the study of Zhao et al. (2011), which shows the potential for generic, non-case-specific tools, while Kraus et al. (2012) emphasize the need for unification of modelling documentation. Heitzig (2011) points to the importance of the systematic framework for model development and the need for implementation of templates for modelling cases and creation of the template library.
Better application of the developed models can be achieved by reusing them through a specially developed modelling framework. This framework should increase reuse and exchange of models, provide proper documentations and support a variety of models. By using this modelling framework, the user will get a model structure, guidance and support during model development and model application and, therefore, efficiency of the modelling process will increase and the quality and reliability of the models will be improved (Mošat et al., 2013).
The objective of this study is to create modelling templates and combine them with other modelling tools within a modelling framework. The goal is to create a user-friendly system, which will make the model development process easier and faster and provide the way for unified and consistent model documentation. This idea of modelling template as part of a modelling framework is based on the work of Fedorova et al. (2013). Fig. 1 shows this modelling framework and its interaction with the modelling application template.
The computer-aided modelling framework consists of three parts: modelling and model analysis, model development and model template application. The central part-B defines the main pathway for model construction and analysis, which is based on a generic computer-aided tool ICAS-MoT (Heitzig et al., 2014). Part-C is related to model development, whose purpose is to help the user to generate the model equations with generic model building blocks (Jensen and Gani, 1999). Part-A, which is the focus of this paper, is concerned with the template-based modelling approach. The purpose of this approach is to enable modeller to create a general model (with all known phenomena and assumptions) for a given system, which can be used later to generate different versions of problem-specific models. To achieve this, the modeller is systematically guided through the steps of the different workflows and, at each step, the framework for template-based modelling identifies and integrates the required tools, database and library connections to guide the user through the corresponding workflow. This is explained further in Section 2.
The modelling framework includes a problem-specific model library, which contains developed and validated models with associated parameters, initial values and/or experimental data. All these models have been developed by using the modelling workflow (Heitzig et al., 2014) and this guarantees consistency of modelling information and documentation.
Depending on the modelling goals and needs, the modeller has the possibility to create a modelling case based on an existing problem-specific model from the library or an available model template or construct a new model by using the model development part of the modelling framework (see Part-C of Fig. 1).
Section 2 of the article presents the concept of template-based approach for modelling, whereas Section 3 describes the implementation of this approach as a software tool. Sections 4 Case study application – modelling template for fixed bed reactor, 5 Case study – oxidation of unsaturated fatty acid highlight the application of the template-based modelling through two case studies. The main conclusions are summarized in Section 6.
Section snippets
Template approach
A modelling template is a generic model on which other problem-specific models are based – in this way, it is similar to a superstructure of process alternatives – that is, it includes a collection of models for the process it represents. In the template, a model is decomposed into three sets of equation types: balance equations, constitutive relations and connection and/or conditional equations. This decomposition is based on the modelling hierarchy proposed by Cameron and Gani (2011). Within
Software implementation of modelling template
The template-based modelling approach is implemented as a software tool, providing easy and user-friendly interface for rapid and more efficient development and use of models. It provides an environment for creation of new templates or addition of new building blocks for existing templates and an environment for template use. It guides the user through the steps of the workflow for template-creation or use. It allows the user to impose changes to an existing template to create new templates
Case study application – modelling template for fixed bed reactor
This is the first of the two case studies highlighting the application of the modelling template-tool. This case study is related to the modelling of a catalytic membrane fixed bed reactor coupling dehydrogenation of ethylbenzene with hydrogenation of nitrobenzene. The reactor is composed of two compartments within a shell containing a bundle of hydrogenation tubes. On the shell side, dehydrogenation of ethylbenzene takes place producing styrene and hydrogen. Five side reactions also occur,
Case study – oxidation of unsaturated fatty acid
The second case study is related to oxidation of unsaturated fatty acids with hydrogen peroxide. The originally considered system includes two phases – oil and water, with phase transfer agent delivering oxidizer from the water phase to the oil phase, where reactions take place. The desired products of the process are short-chained acids, which are the result of cleavage of the unsaturated fatty acid.
This case study highlights the use of the template application for multiphase system with
Conclusions
Template-based modelling approach has been presented and implemented as a computer-aided software tool, which guides the user through workflow steps in order to generate a problem-specific model. The application of the tool has been highlighted with the two case studies, in which models are generated for different objectives.
The application of the tool was highlighted through the creation of a catalytic membrane fixed bed reactor template, where various features of model decomposition in the
Acknowledgements
This work is performed in the framework of the project ‘Model-Based Optimization & Control for Process-Intensification in Chemical and Biopharmaceutical Systems’ (OPTICO/G.A. No. 280813) funded by the European Commission. The contents of the publication reflect only the authors’ view.
References (19)
- et al.
Comparison of diffusion models in the modeling of a catalytic membrane fixed bed reactor coupling dehydrogenation of ethylbenzene with hydrogenation of nitrobenzene
Comput Chem Eng
(2012) - et al.
Kinetic investigations of the deactivation by coking of a noble metal catalyst in the catalytic hydrogenation of nitrobenzene using a catalytic wall reactor
Chem Eng Process
(1999) - et al.
Computer-aided modeling framework – a generic modeling template for catalytic membrane fixed bed reactors
Comput-Aided Chem Eng
(2013) - et al.
Computer-aided template for model reuse, development and maintenance
Comput-Aided Chem Eng
(2014) - et al.
Development of computer aided modelling templates for model re-use in chemical and biochemical process and product design: import and export of models
Comput-Aided Chem Eng
(2015) - et al.
A field study of the industrial modeling process
J Process Control
(1998) - et al.
A computer-aided framework for development, identification and management of physiologically-based pharmacokinetic models
Comput Chem Eng
(2014) - et al.
Perspectives for process systems engineering – personal views from academia and industry
Comput Chem Eng
(2009) - et al.
First principles pharmacokinetic modeling: a quantitative study on cyclosporin
Comput Chem Eng
(2013)
Cited by (23)
Neural network programming: Integrating first principles into machine learning models
2022, Computers and Chemical EngineeringModeling and simulation for design and analysis of membrane-based separation processes
2021, Computers and Chemical EngineeringProcess systems engineering – The generation next?
2021, Computers and Chemical EngineeringProREFD: Tool for Automated Computer-Aided Refrigerant Design, Analysis, and Verification
2021, Computer Aided Chemical EngineeringCitation Excerpt :Tool-1: Modelling Toolbox (MoT) is a mathematical modelling solver, which consists of in-house algebraic solvers, integrators, ordinary differential equation solvers and optimizers. MoT (Fedorova et al., 2015) is an equation-based simulation software that allows users to perform simulation processes without having to write any source code and has an option to convert models into COM-object s for use with external software. Tool-2: As shown in Fig 1, the Database Manager is a tool for searching and retrieving of the refrigerant properties (pure compounds and/or mixtures).
ProCACD: A computer-aided versatile tool for process control
2020, Computers and Chemical EngineeringCitation Excerpt :The modeling template guides the user to organize the model equations as balance equations, constitutive relations and connection and/or conditional equations, based on the modeling hierarchy proposed by Cameron and Gani (2011). Within each type of equations, multiple versions of model equations can be stored and retrieved by the use of binary variables to generate different problem-specific models (Fedorova et al., 2015). Fig. 3 highlights a generic model template where the four types of model equations and what they generally represent of the system are shown.
Optimization proposal for emulsions formulation considering a multiscale approach
2020, Chemical Engineering ScienceCitation Excerpt :It enables the generation and combination of elaborate building blocks (model templates) (Fedorova et al., 2015; Heintz et al., 2014), and the creation of a model template library (Kraus et al., 2012). Currently, these features can be found in commercial software only for process simulation like AspenPlusTM, gPROMSTM, and PROIITM, and in selected modeling and simulation tools such as MOSAIC (Klatt and Marquardt, 2009; Fedorova et al., 2015) and ICAS-MoT (Heitzig et al., 2014). Despite of the benefits that mathematical modeling and computer-aided tools offer, the development of specialized software for the design of chemical products has been limited.