Intelligent Built-in Test Design of Controller Module By Improved Biologically Inspired Neural Network

Abstract

Built-in test (BIT) technology is widely employed in heavy-duty gas turbine control systems for fault recognition. However, it is difficult to obtain an excellent fault diagnostic ability by using the conventional BIT technology, and the false alarm rate is high. In this paper, a design of intelligent BIT based on improved biologically inspired neural network (BINN) is proposed to reduce false alarm. Firstly, massive historical measurement data of controller module is collected and used as training dataset and test dataset. Secondly, intelligent BIT based on improved BINN is designed to deal with the issue of module state identification and reduce false alarm rate. Finally, the effectiveness of proposed approach is validated by the given extensive numerical simulation results and experimental results.