| With the development of science, a large number of electric equipment surged into thepeople’s lives, consume of energy has been a problem which become particularly prominent.Electricity energy management system（EMS）has not only been used in metering andcontrolling of transmission and distribution systems of high voltage, but also concerned byclient for its effective and real-time metering and controlling. The Forecasting of electricitywill be the promise before the whole method which is used to saving the energy is given.This paper analysis the algorithm and the implement of the load forecasting, which isbased on the campus. Compare with the large grid load, load in low voltage has the feature ofstrong volatility and easily interference and so on. As the result, this paper established theforecasting model with the Fuzzy Neural Network (FNN) combined the advantages ofnetwork and fuzzy system. Firstly, according to the character of data, take the use of datapre-processing with the method comparing the data between the same sort of consume.Secondly, make the fuzzy processing of data such as outdoor temperature. Take thenormalized data as the input of the network. At last, get the data of forecasting through theFNN. Meanwhile, this paper made a comparison between different algorithms such as BP,RBF, FNN and Dynamic Fuzzy Neural Network (DFNN). Compare with single neuralnetwork such as BP and RBF, FNN become more accuracy and rapid. DFNN optimized thenumber of the fuzzy system in the FNN to reduce the influences of it.Load forecasting will be the part of the electric management system, so this article usedthe mixed programming between VC and MATLAB to make the using and promoting ofalgorithms more flexible. Meanwhile, the combination between VC and SQL Sever make thedata training become more convenience and optimize the model in real time. At last, the clientcould get the prediction of the load accurately from the model to make the further research ofsaving energy come true. |
PDF Full Text Request