Adaptive Neural Network-Based PID Controller Design for Output Temperature Regulation of a Gas Micro-Turbine

Efficient regulation of Exhaust Gas Temperature (EGT) in Micro-Gas Turbines (MGTS) is paramount for maintaining operational stability and preventing thermal degradation. However, the inherent nonlinearities and significant thermal time delays of MGTs pose substantial challenges for conventional linear controllers. This paper proposes an Adaptive Neural Network-Based PID (ANN-PID) control strategy designed for real-time EGT regulation. In this framework, an artificial neural network acts as an intelligent supervisor, dynamically tuning the PID gains through a gradient-descent-based learning law to compensate for system uncertainties. To validate the proposed method, a comprehensive simulation study was conducted, comparing the ANN-PID performance against a standard industrial PID controller. The results demonstrate that while the conventional PID suffers from sluggish recovery and poor disturbance rejection, the proposed adaptive controller significantly enhances the transient response, reducing the settling time by approximately ۶۵% and providing robust rejection of external load perturbations. These findings underscore the effectiveness of integrating neuro-learning capabilities with classical control architectures for high-precision thermal management in distributed power generation systems.

کلیدواژه ها:

Adaptive PID control ، artificial neural network ، exhaust temperature control ، micro gas turbine