Research on sparse decomposition processing of ultrasonic signals of heat exchanger fouling

Ultrasonic time-domain reflectometry has the advantages of real-time and accurate detection of fouling in heat exchange tubes, but owing to the complex detection environment, the received ultrasonic is noisy, and there are problems such as mutual interference between echoes. This article presents a new method for sparse decomposition based on selecting the atomic dictionary with a high degree of matching. The imperialist competitive algorithm is used to optimize the atomic search process of the matching pursuit. Additionally, the global optimization ability of the imperialist competitive algorithm is improved by improving the moving mode and adding the update strategy. This way, problems such as low matching accuracy and slow searching speed of sparse decomposition are improved. This article reports on the characteristic reconstruction experiment run on the ultrasonic detection signal for heat transfer fouling to test the signal processing effect of the improved sparse decomposition method.