A Fourier Space Basis Function Based 3D Modeling Method for Distributed Parameter Systems

A fourier space basis function based 3D fuzzy modeling method for distributed parameter systems is proposed in this paper. Unlike the traditional modeling method with model reduction, 3D fuzzy modeling integrates the time-space seperation and the time-space sysnsis into a 3D fuzzy model framework naturally. In this study, the construction of the 3D fuzzy model is of data-driven style. The antecedent parts of the 3D fuzzy rules is learned by Affinity Propagation Clustering algorithm. Fourier space basis function is considered as the concequent set of the 3D fuzzy rule, and its parameters are learned by the gradient descent algorithm. The proposed modeling method is successful applied to a simulated parabolic rod reactor system. In comparison with other modeling methods, the simulation results validate the superiority of the proposed modeling method.