1. The pipeline leak detection algorithm based on D-S evidence theory and signal fusion mechanism
- Author
-
Wenhao Xie, Yuan Liu, Xiaoyan Wang, and JunTao Wang
- Subjects
Pipeline leakage detection ,Information fusion ,D-S evidence theory ,BPA ,Discount factor ,Engineering (General). Civil engineering (General) ,TA1-2040 - Abstract
In this paper, a pipeline leakage detection algorithm based on information fusion of pressure and flow is proposed, and its core work is the construction of BPA based on discount factors. The wavelet packet decomposition is carried out for the original signals, and the processed signals are used to train different SVM classifiers to achieve the first identification results. For the samples whose initial classification results are not completely consistent, BPA is calculated according to the confusion matrixes of the initial SVM classifiers. In this paper, static discount factors and dynamic discount factors are constructed using different methods, and dynamic discount factors are modified when decisions fail. Then comprehensive discount factors are constructed based on the combination of different static discount factors and dynamic discount factors. The Shafer discount rule is used to modify BPA. Finally, D-S evidence theory is used to fuse the BPA results of all classifiers under different discount combinations. Experiments show that this algorithm can effectively use the correlation of evidences to reasonably fuse the decision results of multiple sensors, overcome the problem that the accuracy of a single sensor is not high enough for leakage identification, and improve the identification accuracy of pipeline leakage.
- Published
- 2023
- Full Text
- View/download PDF