Data Analysis of Magnetic Flux Leakage Detection Based on Multi-Source Information Fusion

Abstract

For the analysis of the magnetic flux leakage detection data in pipelines, a single information source data analysis method is used to determine the pipeline characteristics with uncertainty. A multi-source information fusion data analysis technology is proposed. This paper makes full use of the information collected by the multi-source sensors of the magnetic leakage internal detector, and adopts distributed and centralized multi-source information fusion analysis technology. First, pre-analyze and judge the information data of the auxiliary sensors (speed, pressure, temperature) of the internal magnetic flux leakage detector. Then, the data of the main sensor, ID / OD sensor, axial mileage sensor, and circumferential clock sensor of the magnetic flux leakage detector are analyzed separately. Finally, the RBF neural network + least squares support vector machine (LSSVM)fusion analysis technology is adopted to realize the fusion analysis of multi-source information. The results show that this method can effectively improve the quality and reliability of data analysis compared with traditional single information source data analysis.