The Design And Implementation Of Power Load Forecasting Based On Net

With the rapid development of economy, power "bottleneck" problem isprominent, parts appear periodically, inherent power supply tense or lack ofelectricity, which bring serious impact area of regional economic development. AsManagement department of power grid, on the change trend of the consumption ofelectricity is at an effective prediction for electric power production developmentplanning is one of the most important content. Accurate, timely to forecast the powerload, from a security point of view can improve the security and stability of powersystem, can tell from the production to provide to reduce cost, reduce excess capacity.Place, in particular, more to solve the problem of and coordinated proportion ofelectricity, electricity consumption forecast is generated for this issue. How toaccurately forecast the power demand has always been a need to be solved by thedepartment of power system, a key problem.Electricity the power the historical data of the prediction is actually based onregion analysis, extract the because of the rule of data, how to establish a reasonablemathematical model for a process. For now, a lot of kinds of forecasting methods inthis field have been proposed, but there is always to predict the content is notcomprehensive, the relatively single forecasting model, and most dependent onhuman, the real-time effect is not good, the precision is not high. In order to changethis situation, this set of prediction system was designed using the modular way ofthinking. The thesis firstly analyzes the power consumption forecasting method ofbackground and significance, methods for the forecast of electricity consumption inthe power system at home and abroad are summarized, the recent development ofprediction software for electricity market development present situation has carriedon the summary, and this paper introduces the present situation of the intelligentdecision-making system. On the basis of these work, summarizes the current powersystem units or departments and determine the advantages of using the software, theresearch target of this article is determined.Secondly, for the study of the theory of the consumption forecast method in the power system has carried on the detailed introduction, and for the function of theload forecasting model and performance details are given.Prediction method is more, in order to adapt to the current electricityconsumption and the parameters of the involved more complex characteristics suchas specific load, adopt the method of comparing a variety of prediction methods. Thelast is chosen from among several algorithms optimal algorithm. For day forecastmodel, adopt the way of optimizing combination to improve the accuracy ofprediction, are less than8%, for medium and long-term prediction model, theprediction relative error less than5%.In this paper, the designed system has good man-machine interface, practical,easy to operate. The software has good engineering application, the promotion ofsignificance.