基于自组织映射(SOM)神经网络对水电站发电排序问题进行了研究.采用耗水率、枯期贡献率、水库可调库容、水库来水等指标作为水电站发电排序的依据,将其作为SOM神经网络的输入,对发电站进行聚类分析,在聚类结果上依据耗水率、枯期贡献率制定出具体的发电排序方案.算例分析结果表明,应用该方法进行发电排序科学合理、实用可行,能够满足节能发电调度的需要.%The generation sorting of hydropower stations is studied based on Self-organizing Map (SOM) neural network. Taking water consumption rate, distribution ratio in dry period, regulating storage and inflow as the basis of sorting and the inputs of SOM neural network, the cluster analysis on hydropower stations is made, and based on the results, the generation sorting can be obtained according to water consumption rate and distribution ratio in dry period. The case study shows that the method is reasonable and practical, and can meet the demand of energy-saving generation dispatching.
展开▼
