Research on Sequential Prediction Model for Gas Pipeline Pressure Value Based on Spatial Correlation

Abstract

Pipeline failure is the most common condition in gas pipeline networks. It is directly reflected in the abnormal pressure values in pipelines. To solve this problem, it is very important to scientifically predict the abnormal pressure of the gas network in a certain period of time in the future. This paper presents a sequential prediction model for gas pipeline pressure using Gradient Boosting Regression Tree as its primary function, which is termed as sequential GBRT(SGBRT). In SGBRT, a series of GBRTs are connected sequentially to extend the lead time of forecast, each of them taking a forecast value from an immediate preceding GBRT as input. The output of each GBRT is modified by adding an expected value of error so that the residual variance of the forecast series is minimized. In addition, through the spatial correlation analysis of the pipeline network, the study excavates the dependence between pressure value of different parts of the pipeline, so that the input features of the model are extended. The experimental results show that the prediction model is correct and effective, which improves the prediction accuracy of gas pipeline pressure value.