Accounting for nonlinearity and periodicity in ozone model identification

This paper deals with the multi-model description of the ozone generation process in the Basse-Normandie region. The nonlinearity and periodicity have been taken into consideration. The nonlinear feature is accounted for resorting to the Takagi-Sugeno approach. This structure involves a set of local linear models (each one is valid for a certain range of operating conditions) and an interpolative mechanism that combines the outputs of the local models into a continuous global output. The periodicity feature consists in using the variable time to design three membership functions, where the day is devised into three equal intervals to construct three periodical models ( each one represents an interval of the day). The parameters of the global periodical model are estimated using the parametric identification approach and membership functions of the variable time to interpolate the three output of the periodical models to get the global output. The model thus obtained turned out to be satisfactory and currently used to build-up a predictor for this region.