(588b) ICAS-MoT, a Computer-Aided Modeling Tool

Model-based techniques play a role of increasing importance in product-process design across a number of industries because they are indispensible in the development of improved products and innovative processes. Model development, analysis and solution are basic prerequisites to apply these techniques but can be very expertise- and time-demanding tasks.

This motivates the development of a general computer-aided modeling tool targeted to increase the efficiency of the process of model development, analysis, application and storage. This is achieved by

1. providing a general structure of the modeling process and combining the required steps to fulfill the desired tasks in work-flows which can be followed by the modeler and

2. combining the required features and numerical methods throughout the modeling process in the same computer-aided modeling framework.

With respect to point 1, a general work-flow for model development and analysis as well as model application in simulation and optimization problems have been derived and tested on case studies and examples of problems reported by others. This work-flow forms the general structure of the modeling tool and its architecture.

Based on this, key features and methods needed for the different steps of the work-flow have been identified and incorporated in the computer-aided modeling tool. One important issue, for example is the model documentation in order to facilitate later re-use of the models. In this context the modeler needs to provide the modeling goal, assumptions, phenomena in the system, etc. Furthermore, support needs to be offered in constructing and providing the actual model equations. Methods for model analysis, like degree of freedom analysis, construction and analysis of the incidence matrix, optimal ordering of equations and deriving an appropriate solution strategy are incorporated into the tool. For model application the modeling tool needs to combine solvers for algebraic equations, ordinary differential equations, partial differential equations and all combinations of these equation types. For the solution of dynamic systems the solution of the steady state model and the evaluation of the stability of the possible steady states using the eigenvalue properties is important. The eigenvalues, moreover allow conclusions on the stiffness of the system as well as the potential for model simplification. In order to solve optimization problems the tool needs to provide an optimizer. Further numerical methods of importance, especially for model identification, are sensitivity analysis, identifiability analysis and uncertainty analysis. In addition, it is desirable that the modeling tool allows an automated report generation.

The above mentioned features and structure have been implemented in the modeling tool ICAS-MoT which will be presented. The different features and numerical methods will be highlighted by showing the analysis and solution of a non-trivial modeling problem.