Study of Power Load Forecasting System Modeling With a Parallel Neural Network

• Main Authors: Ching-Chun Huang, 黃敬淳
• Other Authors: Ping-Teng Chang
• Format: Others
• Language: zh-TW
• Published: 2005
• Online Access: http://ndltd.ncl.edu.tw/handle/17154812385722107030

碩士 === 東海大學 === 工業工程與經營資訊學系 === 93 === The electric power is one of the important energy that is the foundation for the economic development of a country. For electric power business, accurate load forecasting plays an important role in economic scheduling of generating capacity, scheduling of fuel purchases, planning of energy transactions, and dispatching of generation units. Many methods have been developed for power load forecasting. Consequently, time series analysis, expert systems, grey system theory and artificial neural networks have been proposed for power load forecasting. Especially, artificial neural networks have been used widely. Not only too many input variables of power load forecasting need to be considered, but the traditional back-propagation network can’t adjust the input variables mutually with their relations, in this research, we developed a parallel neural network to forecast the power load. Then, we compared the actual power load with the results of load forecasting of the parallel neural network model, the back-propagation network model, the radial basis function network model and the general regression neural network model. According to the forecasting results, our parallel neural network models is more accurately than other methods, the mean absolute percentage error (MAPE) also reveals that our parallel neural network models perform better than other method.