论文部分内容阅读
小水电大多为无调节能力的径流式电站,汛期与大中型水电挤占输电通道,导致水电弃水及地区窝电现象日趋严重,因此尽可能准确地掌握小水电发电能力、制定合理的大小水电协调调度计划愈显重要。对此提出了基于模糊聚类和BP神经网络相结合的小水电短期发电能力FC-BP预测方法,将训练样本根据历史运行数据分类,建立相应的BP网络,对待测样本识别归类,预测小水电装机日利用小时数,并将该方法应用于云南省盈江县和云龙县小水电短期发电能力预测中。结果表明,FC-BP预测方法较传统ANN模型预测精度有所提高,且泛化能力更强。
Most small-scale hydropower plants are non-regulated radial-flow hydropower stations. Hydropower stations and large and medium hydropower plants occupy power transmission channels during the flood season, resulting in water and electricity discarding and worsening of power generation in the area. Therefore, it is imperative to have a clear grasp of the hydropower generating capacity and make reasonable hydropower coordination Scheduling is even more important. Based on the combination of fuzzy clustering and BP neural network, this paper proposes FC-BP prediction method for short-term generation capacity of SHP. Based on historical operation data, the training samples are classified according to the historical operating data to establish the corresponding BP network. Hydropower installed capacity utilization hours, and the method is applied to Yingjiang County, Yunnan Province and Yunlong County, short-term power generation capacity forecast. The results show that the prediction accuracy of FC-BP prediction method is higher than that of traditional ANN model, and its generalization ability is stronger.