Recomputing Causality Assignments on Lumped Process Models When Adding New Simplification Assumptions

Abstract

This paper presents a new algorithm for the resolution of over-constrained lumped process systems, where partial differential equations of a continuous time and space model of the system are reduced into ordinary differential equations with a finite number of parameters and where the model equations outnumber the unknown model variables. Our proposal is aimed at the study and improvement of the algorithm proposed by Hangos-Szerkenyi-Tuza. This new algorithm improves the computational cost and solves some of the internal problems of the aforementioned algorithm in its original formulation. The proposed algorithm is based on parameter relaxation that can be modified easily. It retains the necessary information of the lumped process system to reduce the time cost after introducing changes during the system formulation. It also allows adjustment of the system formulations that change its differential index between simulations.