An adaptive modeling method for time-varying distributed parameter processes with curing process applications

Abstract

Curing processes are nonlinear distributed parameter systems (DPS) with time-varying spatiotemporal dynamics. However, existing data-driven modeling methods have only considered time-varying dynamics in the time direction and paid less attention to those in the spatial direction. This has led to poor modeling accuracy for nonlinear DPS with time-varying spatiotemporal dynamics. In this paper, we propose an adaptive modeling method to estimate the distribution model for this kind of DPS. An adaptive time/space separation is first developed to decompose the time/space coupling dynamics. Time-varying spatial basis functions are then constructed, which can represent time-varying dynamics in the spatial direction. An adaptive T–S fuzzy modeling method is further developed for online learning of unknown dynamics derived from the data. This modeling can adapt to real-time spatiotemporal variation after the time/space synthesis since it utilizes time-varying spatiotemporal dynamics. Finally, curing experiments successfully test and demonstrate the effectiveness of the proposed method.