Research And Implementation Of Ultra-Short Term Load Forecasting Based On Web

Power system short-term load forecasting (STLF) has important function in the reliable, security, and economical operation of power system. It is one of the important routines for power dispatch departments and critical component of Electricity Market Operation System. It is widely used in the dispatching and operation planning of power systems. Accurate load forecasting will improve operation security and stability and save the generation cost. With the establishment and development of the power market, STLF has been one of the significant symbols for the power system modern management.In recent years, the power corporations, especially automatic generating control (AGC) and dynamic economic dispatch put forward urgent demand for ultra-short term load forecasting, i.e. predict the future load data from several minutes to 1 hour. It plays an important role in the electricity market in the long run. We can find the research and development of ultra-short term load forecasting will be significant and practical value. Studying theory and algorithms of STLF and employing an effective and applicable STLF system are becoming an important task. The algorithms of STLF and its application software system are studied elementarily in this paper.As an important module of EMS, the ultra-short term load forecasting attracts attention more and more. This dissertation introduces constituent, characteristic and category of electric load, then analyzes the factors which influences the precision of the load forecasting. At the same time the paper studies the methods of data preprocessing and error computation. After compare the merits and shortcomings for the common methods of load forecasting, this paper establishes two models——ANN and WNN(wavelet neural network) to carry out ultra-short term load forecasting work based on the history load data of Taiyuan power grid. By using MATLAB 7.0 software, a three layers network was used to set up the model, and in the process of training and study continuously introduces BP algorithm to revise the weight value of ANN. Substituting wavelet function for sigmoid function in neural network to form WNN. Wavelet decomposition and reconstruction can be realized by the Mallat algorithm of Multi-Resolution Analysis (MRA) in wavelet transform theory. After selecting proper mother wavelet and scaling function, the load series was decomposed into one low-frequency and some high-frequency sub-series in the wavelet domain and the problem of load feature extraction was solved. Using different ANN models independently forecasting each sub-series, adding the forecasting results of all sub-series, then the whole forecasted load data series are obtained. The representative examples prove that this algorithm for ultra-short term load forecasting has high precision of distinguish, strong capability of popularization, easy training and high currency, it can satisfy the requirements in practical application.Along with the development of the information-based progress of power system, the development and applications of the power enterprise information system are gradually becoming hot point in the trade. According to the practical demands of electric department, a whole STLF management system based on Web for district power networks is successfully developed. The system integrating with Dispatching Automation System has advantage of real-time, economy and practicality. As the grid becomes more and more complex, it is discovered that the traditional Client/Server (C/S) mode means inefficient, misgovernment, difficult maintenance and expensive development outlay. But the integration of production, management, dispatching and command is very necessary, and informational, exoteric and data sharing is very important in the power system. So Browser/ Server (B/S) distributed frame is used here in the system. MS SQL Server with safety and stability is employed as background database platform. Core program and GUI are developed by C#, which is an OOP and visual programming tool. In the process of designing software completely embodies the characteristic and superiority of .NET. Establishing a resource sharing platform, the program provides a new way to organize the graphic information so as to be easily shared and modified through the Web. The functions realized in this system are as follows: load forecasting, load query, analysis and statistics, data assess and print, dispatching information management, etc. The product is open, interactive and expanding well, and it can increase running efficiency and management level of the electric power enterprise.