摘要
现有预测方法的预测结果与真实值存在差异,平均误差结果较大,为此提出基于并行粒子群算法的电力系统极限负荷预测方法。将原始负荷序列数据依次进行误差数据处理、归一化处理和标准化处理,使数据能够按照一定的比例缩放后存放在特定区间,从而得到无量纲数据。运用并行粒子群算法进行负荷预测,调整惯性权重增加全局与局部的搜索能力。计算适应度值并与最佳适应度值进行比较,位置更优时进行替换。当负荷预测结果满足预先设定的阈值时,短期负荷预测完成。实验结果表明,实验组的模型表现更好,预测结果与真实值更相符,计算的平均误差结果在0.32~0.36 MW,在规定误差范围内能够精准预测电力系统的极限负荷,具有更理想的预测结果。
There are differences between the predicted results of existing prediction methods and the actual values,and the average error results are relatively large.Therefore,a parallel particle swarm optimization algorithm based extreme load prediction for power systems is studied.Perform error data processing,normalization,and standardization on the original load series data,so that the data can be scaled to a certain proportion and stored in a specific interval,resulting in dimensionless data.Using parallel particle swarm optimization algorithm for load forecasting,adjusting inertia weights to increase global and local search capabilities.Calculate the fitness value and compare it with the optimal fitness value,and replace it if the position is better.When the load prediction results meet the pre-set threshold,short-term load forecasting is completed.The experimental results indicate that the model in the experimental group performs better and its predicted results are more consistent with the actual values.The average error results of the calculation are all between 0.32 and 0.36 MW,within the specified error range,which can accurately predict the ultimate load of the power system and have a more ideal prediction result.
作者
刘卓
许晨
叶攀
卢波
LIU Zhuo;XU Chen;YE Pan;LU Bo(State Grid Hubei Electric Power Co.,Ltd.,Huangshi Power Supply Company,Huangshi 435000,China)
出处
《通信电源技术》
2024年第1期41-43,共3页
Telecom Power Technology
关键词
并行粒子群
电力系统
负荷预测
parallel particle swarm optimization
power system
load prediction
