Research On Long-Term Electric Load Forecasting Model And Its Database Software System

Electric load forecasting is the forward work for power system planning. The accuracy of the forecasting results is significant for the rationality, scientificity and economy of the final planning scheme and it also enables us to maintain the stable and economical operation of power system.There are many random and fuzzy factors in load forecasting and the forecasting results are deeply affected by many factors, such as load characteristic, industrial structure, climate condition, etc. So it's very important to dig into the internal law of historical load data and reveal the main factors affecting load forecasting. The constructing of long-term electric load forecasting model which can reflect district load characteristic and the application of load forecasting database software can change the traditional rule of thumb in power system planning.This thesis details the fundamental principles and their respective features of many kinds of load forecasting model, and then selects regression analysis, deterministic time series analysis, auto regression moving average and grey model as the basic algorithms of the forecasting software. In the system design of the software, this thesis adopts the popular B/S (browse/Server) mode. Taking full use of the advantages of the B/S mode, any planners can use the familiar IE browser software to access or manage the historical load data from different places through different access mode, such as LAN, WAN, Internet/Intranet etc. In the realization of this software, this thesis uses the ASP (active server page), JavaScript and VBScript as the developing language to design the WEB user interface, and selects the Microsoft ACCESS database to store, manage and enquiry the historical load data. Moreover, to exchange data between WEB pages and ACCESS database, this thesis uses the ADO component to visit the ACCESS database from the WEB pages. In the implementation of load forecasting algorithms, this thesis takes full use of the powerful numerical calculation function of MATLAB and finishes the programming of each algorithm using MATAB language. Furthermore, using the MATLAB Web Server as the calculation server, this thesis realizes the data exchange between WEB pages and MATLAB calculation engine.Utilizing the historical load and power sales amount data of Shapingba power supply bureau in Chongqing Electric Power Company, this thesis carries out load forecasting to verify the WEB based database software of long term load forecasting, and the forcasting results show its excellent applicability.