A Study On Load Forecasting In Jingxi Natural Gas Pipeline

As an kind of environmental friendly energy, natural gas will occupy the leading position of energy consumption in the future from the current development trend of energy consumption. Facing the rapid development of natural gas industry, natural gas pipeline industry will also obtain new opportunities and challenges, however, natural gas pipeline network planning and operation management, must rely on accurate load forecasting, especially in the natural gas supply system between upstream and downstream signed "take-or-pay contract", the gas load or demand forecasting has become more important, it relates to the natural gas supply system security, reliability and natural gas company economic benefits and so on. Therefore, the natural gas load forecasting have a very important significance to improve the supply of natural gas load forecasting system management level, ensure the safety supply, improve economy of the supply system.First of all, the various basic theories and the latest research methods about forecast of natural gas demand is summarized systematically in this paper, then analysis the law of use of natural gas and related factors of Jingxi natural gas pipeline, based on the prediction theories and characteristics of the exponential smoothing forecasting method, the gray forecasting method and the neural network method, constructed the double exponential smoothing model, theGM(1,1) gray prediction model and BP neural network model respectively, forecasted natural gas consumption demand situation in Jingxi pipeline from2011to2015. Then, constructed the non-optimal and optimal combination forecasting model based on the three models above, used the entropy method and the sum of error square minimum method, forecasted natural gas consumption demand situation in Jingxi pipeline from2011to2015. The results showed combination model has higher precision, satisfied the accuracy requirements of the practical application of engineering.Based on the yearly load forecasting, considering the characteristics of the monthly load forecasting, proposed monthly load forecasting model based on BP neural network, complete the prediction of the model calculation with the MATLAB software. The results show that the model has a good predictive accuracy and stability, there are good prospects of application in the monthly natural gas load forecasting.