机构地区: 广西大学电气工程学院
出 处: 《电力系统及其自动化学报》 2007年第1期63-67,共5页
摘 要: 针对短期负荷预测的特点,提出基于粒子群(PSO)优化的模糊神经网络短期负荷预测模型。将PSO与模糊优选人工神经网络进行融合,在对模糊优选神经网络训练中采取PSO算法和梯度下降算法相结合的方法,充分发挥PSO全局寻优的能力和梯度下降局部细致搜索优势。对广西某地区进行短期负荷预测,并与实际值进行比较分析,结果表明这一模型应用于短期负荷预测能获得较高的预测精度,是一种行之有效的短期负荷预测方法。 According to the trait of power load, this paper proposes a PSO based fuzzy neural network model for short-term load forecasting. The model makes full use of the global optimization of PSO and local accurate searching of BP. Practical example indicates that the application of PSO-FNN to short-term load forecasting is feasible and effective, which can obtain more accurate result than conventional methods.
领 域: [电气工程]