Experimental investigation of evolving cloud-based fuzzy control of a pilot thermal exchanger under a decentralized framework

Abstract

Relying on the Robust Evolving Cloud-based Control (RECCo) protocol, a decentralized evolving fuzzy control scheme is presented in this paper for a strongly interacting thermal exchanger process and for the first time experimentally investigated on the real pilot plant. Loop pairing information is first derived based on experimental analysis of the process dynamics. Accordingly, independent RECCo-controlled loops of system temperature are arranged in functional blocks without any prior decoupling or process modelling. The evolvable fuzzy controller structure is learned concurrently relying on data streams only. Simulation and experiments are carried out to support the design procedure and verify the control performance. The results show reasonable tracking performance under different scenarios.