Fault detection method for gas turbine combined cycle unit operation state monitoring sensor

• Main Authors: ZHANG WENQIAN, JIAO DAOSHUN, SUN ZHENGBIAO, LI JUN, CHEN ZHIFENG, YUAN YUAN, WANG WENYI, ZHAI CHUNHUA, HUANG QING, SHI HUAREN, ZHOU JIAN, LIU JIASHU, GUI HONGBO, CAO DIANYAO, XI XINGUO, WANG KAIZHU, PAN HENAN, WANG QIANYUAN, YU JIABIN, YIN HAIQIANG, SU TONG, XIAO JUNFENG, LIAN XIAOLONG, ZHANG XUN, GU RUIXUAN
• Format: Patent
• Language: Chinese; English
• Published: 12.07.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

The invention discloses a fault detection method for a gas turbine combined cycle unit operation state monitoring sensor. The method comprises the following steps: integrating operation data of a gas turbine combined cycle unit at any time to form basic operation data; carrying out wavelet transformation on the basic operation data, and carrying out preliminary detection on a sensor fault according to a wavelet signal to obtain a suspected fault signal of the sensor; performing correlation analysis on the basic operation data, obtaining variables having strong correlation with suspected fault signals, constructing a signal characteristic variable matrix, and using the signal characteristic variable matrix as training data to train a signal prediction model; and outputting a sensor prediction signal by the trained signal prediction model, and performing sensor fault detection according to a residual error between the prediction signal and a sensor real-time signal. According to the method, the problem of model