The Load Forecast And Distribution Of Natural Gas

The work of forecasting of gas load and distribution is one of the indicate which is measure of a gas company's modernization management,especially in our highly developed gas business. How to make use of existing data theories and methods to establish the appropriate scheduling system, forecasting model to improve resource utilization.has become a ignorable part of gas scheduling systemDecision support system is in order to improve decision making effect,by providing data, information and background information to definite policy goal and problem identification, establish decision-making modified.provide kinds of options and estimate the best.Through the repeated man-machine dialogue,wo can analysis,compare and judge for correct and effective decision-making.The system of forecasting and distribution of gas load is a decision support system.Through the forecasting of gas load of daily ,weekly,monthly and yearly,you can plan the production, implemente the gas supply and get the plan of distribution in accordance with relevant rules. Also,provinde a variety of information query interface for user.In this paper,firstly introduce the two model of neural network and support vector machines which use for data fitting and forecasting,and then use them to forecasting the gas load and compare with the the exponential smoothing method and the gray model.analysis the pros and cons,get the models characterstic.And further design the stucture and function module of city gas DSS. Mainly divided into the following modules: Module load forecasting, scheduling policy module, information query module. According to system functions it can forecast, and distribute according to relevant data, come to gas extraction projects and adjustment programs (gas is insufficient), and according to the nature of the various users come from deficiency of the distribution plan, that is. the number of users to stop gas supply and the time to stop the supplyso as to achieve the gas system of scientific decision-making and intelligent scheduling.