Research on extraction of ultrasonic signal features of heat transfer fouling

Fouling in heat exchange leads to the decrease of heat exchange efficiency and potential safety hazards in the heat exchange process. When using ultrasonic time-domain reflectometry to detect fouling in small-diameter heat exchange tubes, the problems of signal aliasing and noise interference will occur. The above problems, this paper with the method of particle swarm optimization combined with matching pursuit of ultrasound probe received signal sparse decomposition, and matching degree of its signal sparse decomposition atomic library is studied in detail, the influence of construction mode of atom library on signal sparse feature extraction results is analyzed through finite element simulation signal of heat transfer fouling. In this paper, the effect of sparse decomposition feature extraction is examined by reconstructing the heat exchange tube fouling detection signal to verify that the adopted sparse decomposition method has good signal processing capability.