A Research On Natural Gas Load Forecasting And Peak Shaving Methods

In order to improve energy structure and environment pollution, natural gas, as a clear energy source, had a widely demand year by year, the supply of natural gas become harder than before. There are lots of difficulty problem in an engineering, such as how to forecast and plan the consumption of natural gas, ensure steady supply, enhance the gas supply scheme’s reliability and economic and so on The purpose of this article is to solve these.The article mainly study natural gas load forecast, the way of peak shaving and regulator station’s optimization. In the first part author uses correlation analysis to find out the main influence factor, forecasts short-term load by regression analytical method and artificial neural network model, with the help of SPSS and MATLAB software, the results show that forecast load is agreement with true load basically, Then, base on the characteristic of natural gas middle-term load, take a city winter peak gas consumption stage as the example, carries on the middle-term load forecast with the BP artificial neural networks forecast model, forecast error within 10 percent, which means this method has directive function in future.Secondly discuss natural gas long-term load forecast, consider natural gas user’s characteristic and rank, combined with the change of energy structure, purposes new user classification method at the first time. This article establishes long-term load forecast model based on classify user for residential users, commercial users, automotive gas users, heating users and industry users, based on classify user’s characteristic and their forecast methods. Compare with reality year load the mean error is 9.3 percent, belongs to high accuracy forecast, which suggests that classify forecast method suitable for long-term load forecast.In the last part, author do some research with gas peaking scheme and regulator station’s optimization, and put forward the concept of years cost per gas peaking, engineers and technicians can do economic analysis with it, through compare different peak-shaving methods and calculate peak shaving load of a city, the paper aims to get a suitable gas peak shaving method. Finally set up a mathematical model, the purpose of this model is to get the best quantity and work radius of regulator station, and author give an example to confirm that this model has practical value.