Neuro-Fuzzy Adaptive PID Control of Thermoelectric Module for Metal Hydride Reactor

Abstract

This paper presents a neuro-fuzzy proportional integral derivative (PID) control technique for improving thermodynamic performance of a metal hydride (MH) reactor via heating/cooling effects generated by a thermoelectric module. The thermal behavior of the MH reactor coupled with a thermoelectric module is numerically studied by mathematical representations of genuine practical applications. It is found that the integrated system has strong nonlinearity owing to thermal characteristics. To obtain the desired performances of the MH reactor, a neuro-fuzzy PID control is used in real-time implementation. A non-linear optimization of a back-propagation technique is applied for fine-tuning the parameters of the neuro-fuzzy PID controller. The simulated results show the effectiveness of the proposed technique compared to conventional PID control.