Using Fuzzy Neural Network with Genetic Algorithm for Forecasting Electric Load and Noise of Power Line Communication

• Main Authors: CHIA-HUI KAO, 高嘉輝
• Other Authors: Yuang-Shung Lee
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/92611157571184720857

碩士 === 輔仁大學 === 電子工程學系 === 94 === Electricity is widely applied in many aspects of modern life. Precise forecasting of electricity consumption may not only reduce operational and maintenance cost for power companies but also enhance the reliability of power systems, as well as avoid shortage of supply that causes damage and inconvenience to customers. The power line was not only transferring the electricity but also to act as the medium of communication. Through the convenient of power line networks which could improve the question of last mile. But the power line would affected by variable impedance, attenuation of distance and interference. Owing to the variably characteristic of power line, the noise question always was an important issue. In this paper, the proposed gradient descent neural network assist with the fuzzy theorem and the reduced form genetic algorithm are adopted to training our forecasting network and model. Those networks are simulated by MATLAB program. And the training data profile is acquired from the home. Through compare those simulation results, would expect to have some contribution in the research field.