Research On Adaptive Filtering Methods Of The Data Measured By Magnetic Flux Leakage Detectors

Both oil and natural gas are important energy for the national economic development and industrial production.On the current situation,and pipeline transportation has become the main way of oil and gas transportation in the sea.In recent years,the total mileage of oil and gas pipeline in our country has increased by thousands of kilometers,which provides a reliable transport channel for the transportation of the energy of oil and gas.To ensure safe pipeline operating,lengthen the service life of the pipeline,prevent leakage and environment pollution,according to the current regulations and international examples of submarine pipeline system at home and abroad,they are both necessary for submarine pipeline annual inspection and special inspection.Magnetic flux leakage detection is a main way of pipeline fault diagnosis.Research on adaptive filtering methods of the data measured by magnetic flux leakage detectors is the key research content on data preprocessing of the magnetic flux leakage detection,which lays a solid foundation for accurate and fast defect localization,feature extraction and life prediction,which has a great research value.In this paper,a set of intelligent adaptive filtering algorithms that realizes adaptive filtering processing of digital signals collected by the magnetic flux leakage inner detectors are designed,which has the function of fast and accurate data processing.The specific research contents are as follows:First,realize the baseline correction algorithm of the data measured by magnetic flux leakage inner detectors.According to the analysis and research on the problems that the single channel baseline drift and reference differences between sensor channels based on the magnetic flux leakage inner detectors,the two correction methods that are median correction and improved mean median correction are used to make the channel equalization and to unify the rules.Second,remove and compensate the abnormal data.In this paper,Artificial elimination of the abnormal data is realize d through the GUI interface of MATLAB,and intelligent eliminating of the abnormal data is realized by the threshold segmentation algorithm and the variable scale window interpolation fitting algorithm;According to the different situation of data missing,cubic spline interpolation is used to realize abnormal data points compensation or abnormal data points compensation are replaced by straight lines.Third,determine the evaluation criteria of the filtering effect.The theory of spectrum analysis are mainly told,and according to the data characteristics of the magnetic flux leakage detection,the spectral analysis is well applied to the process of the data analysis and processing.At the same time,in order to judge the quality of the filtering effect,a variety of subjective and objective evaluation methods is proposed.Fourth,realize the single channel adaptive filtering.The characteristics of the various kinds of filtering algorithms are concluded after comparing and analyzing the various kinds of filtering algorithms.Then combined with the characteristics of different signal types,single channel adaptive filtering algorithms including mix filtering algorithms of three kinds of signals,which are no defect signals,large defect signals and small fluctuation signals,are designed,which effect is better than the effect of a single filter algorithm.Fifth,realize the multi-channel adaptive filtering.noise in multi-channel will produce effect on each other,and multi-channel equalization is realized by comparison and analysis the degree of noise effects between adjacent channels.Multi-channel adaptive filtering algorithms including mix filtering algorithms of three kinds of regions,which are no defect regions,large defect regions and small fluctuation regions,are designed,which effect has further improved than the effect of single channel adaptive filtering algorithm.