An environment for heterogeneous model management in chemical process engineering

Modeling and simulation as a tool to support analyzing and improving chemical plants and processes is becoming more and more attractive. However, model development often is a lengthy task and requires an expert to obtain suitable results. Thus, in order to make further economic benefit from model-based activities, speeding up the model development process is an important objective. Fortunately, it becomes obvious that not all models have to be developed from scratch if a perspective is assumed that focuses on the lifecycle of an entity being modeled. Along this lifecycle, different models might be developed and all of them have some inherent dependencies because they all describe the same physical thing. Considering the lifecycle of a plant, for example, one recognizes that the models describing different modification stages of the plant are highly similar. Also, steady-state and dynamic models have some overlap such as the flowsheet structure of the plant. Further, data concerning physical property models and reaction kinetics is not only reusable, but should be kept consistent across various models. However, reusing those different parts is not that easy in practice. Models are developed by various employees of a company that sometimes change jobs or even the employer. Often, not only a single model is developed, but a variety thereof in order to take into consideration variations of some parameters or process alternatives. Further, those models are often developed using different tools to support various tasks. Tools and tasks range from simple mass balances and rough economic estimates using spreadsheets like MS Excel to dynamic simulation using e.g. gPROMS. Thus, it can be concluded, that model reuse leads to integration problems which must be tackled in order to leverage the efforts invested into model development. The first integration problem to solve is the combination of existing models in different formats and representations into a homogeneous library which is easy to browse or search. The second integration problem to address is to enable the combination of different models at simulation runtime so that they can be reused in their native simulation environment without reformulating or recoding them. The issues mentioned are exemplified with a case study from the production of polyamide6. capabilities of state-of-the-art tools with respect to model management and reuse are discussed and it is explained how the contributions of this thesis can improve the current situation. The architecture REALMS (Repository-based environment architecture for lifecycle modeling and simulation) of a heterogeneous modeling environment is then presented. It defines components suitable to solve the two problems described above, namely a model repository ROME (Repository for Open Modeling Environments) that permits storing, archiving, and assembling process models and a simulation platform CHEOPS which simplifies the integration of heterogeneous models at runtime. Both parts are integrated so that CHEOPS can execute heterogeneous experiments consisting of models stored in the repository ROME. A user interface based on a highly flexible component model is defined in order to enable the user to operate the overall system. The model repository ROME is based on a metamodel which is used to define a homogeneous view on the most relevant characteristics of chemical process models across various modeling tools. The metamodel is discussed as well as the mapping bewtween the metamodel and mathematical models presented in the polyamide6 scenario. The functionality realized with this metamodel and its implementation in an object-oriented database management system are then presented. The goal of the simulation platform CHEOPS is to run a simulation experiment which combines various simulators and coordinates them to achieve an overall solution of an experiment. Existing simulators are discussed with respect to their architectural properties and desired properties for an open simulation platform are stated. Based on these remarks, the architecture of the CHEOPS simulation platform and its realization are explained. Finally, the solution of the problem posed in the scenario presentation with the tools described in this work before conclusions are drawn.

Grantor dc:publisher

Publikationsserver der RWTH Aachen University

Year dc:date

2004

Author dc:creator

• von Wedel, Lars-Alexander

Contributors dc:contributor

• Marquardt, Wolfgang

• info:eu-repo/classification/ddc/660
• Polyamid 6
• Herstellung
• Prozessmodell
• Metadatenmodell
• Softwarelebenszyklus
• Softwarewerkzeug
• Softwarewiederverwendung
• Technische Chemie
• Modell Wiederverwendung
• Datenbank
• Simulation
• Integration CORBA

Statement dc:rights

• info:eu-repo/semantics/openAccess

Language dc:language

eng

Identifier

https://publications.rwth-aachen.de/search?p=id:%22RWTH-CONV-121258%22

OAI identifier oai:identifier

oai:publications.rwth-aachen.de:59477