Black-Box Fuzzy Identification of a Nonlinear Hydrogen Fuel Cell Model

Abstract

A fuzzy  identification of the dynamical system  model is developed upon a data generated by a software simulator of a hydrogen fuel cell. The data presents a black box  model, just composed by inputs and outputs, carry no  additional information, and showing a strong nonlinear behavior. The choice for a fuzzy identification is based on the data features, and the malleability of the mathematical fuzzy  technique. This approach allows to accomplish the objectives of the research, among which, the validation of the method for it used in other industrial problems.  The dynamic system identification process is performed using a fuzzy clustering through  the Gustafson and Kessel algorithm, and a Takagi Sugeno fuzzy inference method. Validation tests are performed  in terms of the 4-fold technique, confirming the lack of the data over-training. These  results make the fuzzy approach looks as a promising tool for black-box identification  of non linear dynamic systems.