Research And Implementation On Data Analysis System Of Gas Pipeline

As continuous development of pipeline technology, pipeline has become the main way for gas transportation. It also influenced pipeline inspection technology developed so quickly in the past 30 years. Magnetic Flux-leakage inspector is one kind of advanced equipment for collecting the data of pipeline running status. Magnetic Flux-leakage inspector only can detect the different thickness of metal pipe, but there isn't any exact mapping rule for matching signals and defects, so how to recognize pipeline defects is difficult. Traditional recognition is done manually by human experts. According to decades'experience, human experts collected a lot of information and summarized some rules on how to match them. Based on these useful rules, we started our research and implementation -- Data analysis system of gas pipeline.Generally the data file generated from pipeline inspector is so complex and huge, so the first task is to find out an effective solution on how to read and show original data in graphics by researching on mass data. Avarage differences, amplitude differencese, half-baked data generated by mechanism failure is universal for original data. In order to implement the automic analysis, deficient data should be adjusted and repaired first. By study on data mining, system adopt a good mechanism on how to find out the hidden rules. These hidden rules can help to rebuild the ideal data file. Because there are a lot of uncertain factors in trick recognition, experts often get the different result when handle the uncertained area. Vague mathematics and detailed solutions is studied for trick recognition. Because the position information is so important and should be included in final analysis report, this dissertation introduces two kinds of localizer usually used in inspectors and analyses their data structures and finally shows the algorithm on data matching.By studying on this project, finally Data Analysis System of Gas Pipeline was finished successfully. According to the comparation with manual analysis results of history projects, discrimination rate of pipeline defects can get 93% in the new system.