Gas Load Intelligent Prediction Research And Implementation

As the requirements of people's life quality for improving and the environmental protection needs, the use of natural gas that is a clean source of energy is vigorously popularized. While comprehensive constructing the gas pipe network to turn into "gas era" in our country, there exists some urgency problems to solve. For example, how to design the pipelines, planning the gas purchases and the storage, how to scheduling the peak and so on. These are the problems that the gas industrial development faces. Shanghai as a world financial center that China has advocated, the steady develop-ment of natural gas industry becomes especially important.This paper will make full analysis of Shanghai's gas load characteristics based on a comprehensive collection of historical gas load data of Shanghai. After studying the actual forecasting effect of some load forecasting methods which have been widely used to Shanghai, this paper puts forward a kind of gas load combination forecast method based on wavelet transform. This method is pretreatment using wavelet transform firstly. The raw data signal is decomposed into n layers under the action of wavelet decomposition so that the sub-sequences'periods are more apparent. Then forecast the each layer of wavelet sequence by single method which is in the comple-mentary model set (chapter 3 second quarter detail) respectively. We will get several predictive values after reconstructing. After calculating the drifts, get best weights of each prediction method, and can get the final prediction. The gas load fluctuates peri-odically in daily, monthly and yearly. And it presents the characteristics of superposi-tion cycles for there are many factors that influence the gas load such as seasonal al-ternate, holidays and other factors. The pretreatment method that using the wavelet decomposition is reasonable. The experimental results show that the effect of the model proposed in this paper is better than a single prediction model and general combinatorial models in precision and speed. In this way, it will take great applications and perspective.This task comes from projects of Shanghai normal university and we joint the project named"intellectualized research and development of Shanghai gas command system" with Shanghai Head-Union technology and innovation Co., LTD which is in charge of the research institution of computing technology of Shanghai. At present, we are developing the"Shanghai gas intelligence analysis prediction system". This paper shows the design and the programming in the next. Explore a joint programming method of VC and MATLAB. The method can reduce the burden of development work which is so heavy by realizing the MATLAB and VC development environment of integration, and enable to shorten the cycle of software programming. Because the development workload of this software is very big, the part of gas load forecasting is still in development. It haven't come true all yet. This is also the subsequent work of this subject. We hope it is able to be applied in practical use As soon as possible.