Models Hierarchy

Abstract

This chapter presents a hierarchical approach to model SOFC behavior during both steady and transient conditions. The recourse to such an approach is motivated by the high computational intensity characterizing optimization algorithms, especially those aiming at large-scale design and definition of on-field applicable control and diagnosis strategies for SOFC-based systems, either destined to stationary generation or transport applications (Rizzoni et al. in Modeling, simulation, and concept design for hybrid-electric medium-size military trucks, pp. 1–12, 2005). In principle, this issue may be satisfactorily addressed by exclusively using black-box/lumped models of the system under development. Nevertheless, against such choice are major drawbacks, such as the much extended data sets required for identification and validation of black-box models, together with the need of running new experiments whenever system specifications change. It was demonstrated, with regard to internal combustion engine modeling (Arsie et al. in A hierarchical system of models for the optimal design of control strategies in spark ignition automotive engines, pp. 473–488, 1999), that the best compromise between accuracy, experimental costs, computational time, and flexibility is achieved by using a mixed modeling approach, with white, gray- and black-box models integrated within a hierarchical structure. This approach can be usefully extended to flexible SOFC systems, especially because of the high costs to be faced to run transient experiments and the high variety of fuel cell designs, which are under current investigation.