论文部分内容阅读
短期负荷预测主要指对未来几日或者几周时间内的电力电网负荷进行预测,目的是为发电厂的符合分配,机组启停以及设备检修和燃料能源供应计划提供指导。考虑到常规的SVR预测模型采用人工经验的方法对RBF核函数参数、不敏感系数和惩罚系数等参数进行选取,这样常规SVR算法的缺陷就是其性能会因为随机选取的参数而变得随机和不确定,因此本文使用人工鱼群优化算法对SVR参数选取进行优化。为了提高人工鱼群算法全局搜索能力,将全局最优的信息融入到人工鱼的觅食、聚群、追尾移动选择过程中。改进后的全局人工鱼群算法,能够避免传统的人工鱼群算法在对人工鱼移动方向选择时没有考虑全局信息而引起的收敛效率和收敛精度低等缺点,因此能够更加快速精确地搜索到全局最优解。最后,通过实验方法,对本文研究的短期电网负荷预测方法进行验证。结果表明,本文研究的短期负荷预测,预测精度较高,具有较好的工程应用价值。
Short-term load forecasting mainly refers to forecasting the electric grid load in the next few days or weeks in order to provide guidance for compliance with plant allocation, start and stop of units and equipment maintenance and fuel energy supply plans. Considering that the conventional SVR prediction model uses artificial experience to select the parameters of RBF kernel function, insensitivity coefficient and penalty coefficient, the disadvantage of the conventional SVR algorithm is that the performance of the SVR becomes random due to randomly selected parameters Therefore, this paper uses artificial fish swarm optimization algorithm to optimize SVR parameters selection. In order to improve the global search ability of artificial fish swarm algorithm, the global optimum information is integrated into the process of foraging, clustering and rear-end movement of artificial fish. The improved global artificial fish swarm algorithm can avoid the shortcoming of the traditional artificial fish swarm algorithm that the convergence efficiency and the convergence accuracy are not taken into account when the artificial fish swarm selection is not taken into account when selecting the moving direction of the artificial fish so that the overall situation can be searched more rapidly and accurately Optimal solution. Finally, the experimental method is used to verify the short-term grid load forecasting method. The results show that the short-term load forecasting in this paper has high prediction accuracy and good engineering application value.