Based On The Analysis Of The Model Is Built And The Level In The Pipeline Corrosion In The Diagnosis Of Applied Research

With the development of our national economy, to raise the level of science andtechnology, demand for energy increased significantly, while the pipe transportation hasbecome the most efficient energy transport, convenient and save one of the mode oftransportation. But the increase of oil and gas pipeline, pipe service fixed number of year ofgrowth, laying pipes and the environment is different, so that tube corrosion accidentshappen repeatedly to people’s lives and property safety brought great hidden dangers.Therefore to early detection and early pipeline replacement can avoid the happening ofcalamity. The traditional pipe failure mode has a lot of uncertainty diagnosis method, notvery good locate corrosion pipe of the specific area. Analytical hierarchy process is usingqualitative and quantitative methods to solve the decision making problems, so that it wouldbe using analytic hierarchy process to the pipeline failure mode diagnosis has the very goodadaptability, and has the important theoretical and practical application value.Papers to the pipeline detection by data and fault type of inductive analysis, usinganalytic hierarchy process, the solution of the problem into decision-making, target layerand rule layer, In decision-making target layer，based on a large number of decision-makingpipeline detection data, failure mode of the pipes to the classified analysis; in the rule layer，based on the level of analysis using intelligent modeling method constructing mode，thetesting data of the pipes to the analysis, forecast; in the schematic layer, according to thetesting data pipe to determine the pipeline corrosion level, that is dangerous, serious,moderate or mild. In view of the actual application of model building method to have strongflexibility and intelligent requirements, the paper constructs the fuzzy hierarchy analyticalmethod, and fuzzy comprehensive evaluation method, the support vector machine andcascade neural network diagnosis model, can be stratified and influence factors to realizerelationship pipeline corrosion situation analysis, in the information processing mechanismof the pipeline corrosion and the factors affecting the corrosion degree of diagnosis hasbetter adaptability.Design and development for pipeline detection data analysis software system, theactual data processing got good, the result shows that the establishment of the paperdiagnostic methods and processing the feasibility and effectiveness of the model.