Leak Identification Method of Water Supply Pipeline Based on Compressed Sensing and Least Squares Twin Support Vector Machine

As the scale of water supply pipelines continues to expand, pipeline leakage monitoring is entering the era of big data. Aiming at the problems of large data volume and information redundancy in traditional data collection methods, this article proposes a water supply pipeline leak identification method based on the combination of compressed sensing (CS) theory and least squares twin support vector machine (LSTSVM), which is called CS-LSTSVM. First, using the observation matrix to preserve the integrity of the information, the collected leakage signals are compressed and observed to obtain the observation dataset, thereby reducing the redundant information and volume of the data. Then, the corresponding feature information is extracted from the observation dataset to form a feature dataset. Finally, the feature dataset is sent into the LSTSVM recognition model to identify and classify the feature data through its excellent classification performance. The experimental results show that the proposed CS-LSTSVM method can greatly shorten the model training and testing time while maintaining a high leak identification accuracy, and the method has good robustness in pipeline leak detection. Among them, when the compression ratio (CR) is 50% and the observation matrix is a partial Fourier matrix, the recognition accuracy of the model reaches 98.56%, and the time consumption is reduced by 81.8%, which effectively improves the efficiency of water supply pipeline leak identification.

Automated summary

This article proposes a water supply pipeline leak identification method based on the combination of compressed sensing (CS) theory and least squares twin support vector machine (LSTSVM), called CS-LSTSVM. The method reduces redundant information and volume of data through compressed sensing, extracts feature information, and identifies leaks through the LSTSVM model, achieving efficient and accurate leak identification in water supply pipelines.