Research on identifying drainage pipeline blockage based on multi-feature fusion

In order to solve the difficulties of detecting partial blockage in urban drainage pipeline and to determine the degree of blocking, a novel method based on multi-feature fusion technique of recognizing pipe blockage is proposed in this paper. Firstly, the acoustic response signal collected from a section of working sewer pipe is decomposed by 3 levels wavelet packet decomposition, and the high-energy wavelet packet nodes are selected to reconstruct the signal to establish feature components. Then the characteristics of wavelet energy entropy, approximate entropy and fractal box dimension of the feature components are extracted respectively, so that the classification feature sets can be constructed. Finally, the particle swarm optimization algorithm is used to optimize the parameters of the SVM classifier to identify the blockage fault signal. The results from the experiments have shown that the method can not only effectively identify the different degrees of blocking failure, but also eliminate the impact of the lateral connection from fault identification. As a result, the accuracy rate of pipeline blockage identification is improved, and the method also provided a research foundation for early fault detection of working pipeline.