Fuzzy logic (FL) controllers (FLCs) ensure .stable arid robust, control of multi-input multi-output, (MIMO) nonlinear industrial processes with no reliable models. To respond to the plant, changes and the requirements of industrial implementation, a design approach is suggested for computationally simple MIMO FL supervisor-based adaptive FLCs (SAFLCs) suitable for use by programmable logic controllers (PLCs). The SAFLC consists of a main model-free MIMO FLO and a MIMO FL supervisor (FLS) for' keeping desired system performance by on-line adaption of the FLC’s scaling factors. The real-time plant control by the empirically designed FLC provides data for the derivation via genetic algorithms (GAs) of a transfer matrix-based Takagi-Sugeno-Kang (TSK) plant model and for its validation. The TSK plant model enables computational system sensitivity analysis for the design of the optimal structure MIMO FLS. The SAFLC is approximated to a simple PLC feasible parallel distributed compensation (PDC) using GAs and real-time plant, control data used also for the PDC validation. The Lyapunov SAFLC system stability is studied via the TSK-PDC system representation and linear matrix inequalities. The designed SAFLC real-time control of coupled levels reduces the system settling time, overshoot and coupling in comparison to the FLC control.
|Number of pages
|International Journal of Innovative Computing, Information and Control
|Published - Apr 2017