Research On The Analysis Of Pipe Network Testing Data And Forecasting Techniques Based On The Theory Of Fuzzy Calculation

As a new means of transport for pipeline, it has insurmountable advantage compared with the road transport, the rail transport and so on in safe, efficient, low cost, easy to manage, centralized monitoring, etc. Pipeline’s safety and health are closely related to the businesses running, stability and development of social economic, and the security of national energy’s supply. Therefore, timely analysis and diagnosis for testing data of pipe network, pipeline failure mode and failure cause, predictions for the pipeline failure and response measures have a very important practical significance. So, in the safe operation of oil and gas pipelines, studies and risk assessment for the pipeline integrity is a crucial task.This paper use fuzzy computing theory to research the analysis models of pipeline inspection data and prediction. According to the characteristics about the quantitative and semi-quantitative of the inspection data and the needs of the inferential analysis on pipeline failure mode, firstly, this paper elaborates some theories and key technologies, such as fuzzy computing, fuzzy logic inference, fuzzy controller, fuzzy neural network, etc. Combined the data, the cases and the testing data of pipeline failure collected by the research group, it analysis a variety of the effects on the pipeline failure and abstract the data features. By analyzing the impact factors of pipeline damage and failure and the relationship between the factors, it establishes a metadata format of pipeline failure, data management model and data model of pipeline inspection. It uses the primary element analyzing methods and factor analysis to establish the screening and sensitivity analysis algorithms for influential factors of pipeline. Secondly, by fusing the fuzzy theory and neural network, and according to the fact that the pipeline failure factors are random, blurred and complex, it proposes a fuzzy neural network based on hybrid ant colony clustering algorithm, and this method will be used to diagnose the type of injury in the pipeline and pipe corrosion. Thirdly, according to the pipeline damage and corrosion for dynamic prediction, talking the fuzzy reasoning process neural network theory as basic theory, it constructs dynamic Prediction model based on improved particle swarm optimization.According to the results of all the research, it establishes an underlying database for management of pipeline data. It uses the development tools-ASP.NET to write computer programs for pipeline failure mode diagnosis and prediction algorithm, and put the algorithms into the process flow diagram and establishes a software system for the pipe network information management and warning analysis based on the process flow diagram. In the system, it can predict the pipeline failure in the visual environment rapidly based on real-time detection data. This provides the scientific basis for the risk assessment of the operators pipe network, also has important application prospect and practical value.