Read the full-text of this article for free

Title: A time series neural network-based early warning system for thermal power station

Authors: Chen Yang; Hengyao Pan

Addresses: School of Energy and Power Engineering, Chongqing University, Chongqing, 400044, China ' School of Energy and Power Engineering, Chongqing University, Chongqing, 400044, China

Abstract: Real-time monitoring and early warning are critical for power system stability and security as thermal power units become more crucial in power generation. This paper reports a time series neural network-based thermal power unit condition warning mechanism. First, time series analysis simulates the dynamic changes of the units to capture their long-term trend and time-varying characteristics over operation using multi-dimensional thermal power unit operation data. Time-series data is processed by long short-term memory (LSTM), which also learns features for highly precise defect prediction. Furthermore, designed in this work is a multi-dimensional performance assessment system based on fault detection rate (FDR), early warning lead time (EWLT), false alarm rate (FAR), and diagnostic accuracy (DA) to entirely assess the proposed method. Tests reveal that the time series neural network-based thermal power unit status warning system can let thermal power units run reliably.

Keywords: thermal power unit; condition monitoring; fault warning; time series analysis; long short-term memory; LSTM.

DOI: 10.1504/IJICT.2025.144650

International Journal of Information and Communication Technology, 2025 Vol.26 No.4, pp.106 - 122

Received: 31 Dec 2024
Accepted: 14 Jan 2025
Published online: 25 Feb 2025 *