A Design of TSK-Based ELM for Prediction of Electrical Power in Combined Cycle Power Plant

Abstract

This paper is concerned with the prediction of full load electrical power output of a base load operated Combined Cycle Power Plant (CCPP) based on Takai-Sugeno-Kang (TSK)-based Extreme Learning Machine (ELM). Here TSK-based ELM is designed by a systematic approach to producing automatic fuzzy if-then rules, while the conventional ELM is designed without knowledge information. The design of TSK-ELM consists of two main steps. In the first step, an initial randomly partition matrix is generated and cluster centers for random clustering are estimated. These centers are used to determine the premise part of fuzzy rules. Next, the linear parameters of the TSK fuzzy type in consequent part are estimated using the Least Squares Estimate (LSE) method. The experiments were performed on prediction of electrical power in CCPP by the presented TSK-ELM. The input variables include hourly average ambient variables temperature, ambient pressure, relative humidity and exhaust vacuum. The output variable is used to predict the net hourly electrical energy output. The experimental results revealed that the presented TSK-ELM showed good performance in compared to the original ELM.