摘要
在分析电力负荷曲线特性的基础上,将日负荷分成工作日和非工作日,并着重考虑温度对负荷曲线特性的影响,将BP算法和模拟退火(SA)算法相结合,对某电网的日负荷数据进行实际计算,发现考虑预测日类型和温度等因素后,负荷预测精度有很大提高.
Based on the analysis of the load's character, the data is differentiated according to the type day. At the same time, the influence of temperature on the load's character is considered. This paper proposed a neural network with BP and SA algorithm, which combines the property of BP with the property of SA algorithm. The numerical tests show that the accuracy will be improved after considering the influence of the day type and temperature.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2004年第9期1544-1547,共4页
Journal of Shanghai Jiaotong University
关键词
短期负荷预测
神经网络
模拟退火
Backpropagation
Neural networks
Numerical methods
Simulated annealing
Time series analysis
