摘要
针对电力工程投资精益化管理的需求,文中结合电力公司多年积累的工程数据资源进行了智能化数据挖掘算法的研究,提出一种基于电力工程设备材料历史价格和价格波动因素的工程数据预测方法。该方法基于支持向量机(SVM)、灰狼优化(GWO)和差分进化算法原理展开,其中,SVM的参数识别基于结构风险最小原则,避免了算法迭代过程中的过拟合现象;GWO算法引入了动态进化算子和非线性收敛因子,从而减小陷入局部最优解的可能性。在算法仿真时,重点关注了电网投资中的常用设备和材料,从不同的角度考虑设备材料价格的影响因素。数值实验结果表明,该算法在进行电网投资需求预测时,平均相对误差为2.86%,相比于GA-BP算法提升了3.56%。
Aiming at the needs of lean management of power engineering investment,this paper studies the intelligent data mining algorithm combined with the engineering data resources accumulated by power companies for many years.An engineering data prediction method based on historical price and price fluctuation factors of power engineering equipment and materials is proposed,which is based on Support Vector Machine(SVM),Grey Wolf Optimization(GWO)and differential evolution algorithm.The parameter identification of SVM is based on the principle of minimum structural risk,which avoids over fitting in the iterative process;GWO algorithm introduces dynamic evolution operator and nonlinear convergence factor,which reduces the possibility of falling into local optimal solution.In the algorithm simulation,we focus on the common equipment and materials in power grid investment,and consider the influence factors of equipment and materials price from different angles.The results of numerical experiments show that the average relative error of the algorithm is 2.86%,which is improved by 3.56%compared with GABP algorithm.
作者
姚日权
费英群
丁云峰
王胜华
田林
YAO Riquan;FEI Yingqun;DING Yunfeng;WANG Shenghua;TIAN Lin(Huzhou Power Supply Company,State Grid Zhejiang Electric Power Co.,Ltd.,Huzhou 313000,China)
出处
《电子设计工程》
2022年第7期63-67,共5页
Electronic Design Engineering
基金
国网浙江省电力有限公司湖州供电公司资本性项目标准成本研究项目(5500-202006151A-0-1-02)。
关键词
数据挖掘
电网工程
SVM
GWO
差分进化
data mining
power grid engineering
SVM
GWO
differential evolution
