Application of Neural Networks to Optimal Dispatch of Hydrothermal Coordination

• Main Authors: Jou, Chuan-Ling, 周川凌
• Other Authors: Huang, Shyh-Jier
• Format: Others
• Language: zh-TW
• Published: 1997
• Online Access: http://ndltd.ncl.edu.tw/handle/95341726098513697899

碩士 === 義守大學 === 電機工程研究所 === 85 === The schedule of generating systems depended merely on the operators'' experience in the past. The disadvantage lies not only on wasting generation cost, but also on damages of generating units because of frequent operation. Since the characteristics of various generating units in a power plant are quite different, the main task concerned by the power company is to reduce generation costs by effective dispatch of each generator output in a generation system. This paper applies neutral networks and dynamic programming to find the optimum hydrothermal generation output. By this method, it can reduce computation time and storage requirement. Furthermore, it is suitable for various types of load curves. When the neural network training data base is built and well-trained, the optimum generation output can be solved in a short time. This paper offers an example of solving optimum dispatch. It can be extended to solve the industry automation.