基于IPSO-LSSVR算法的变电站工程造价预测方法  

Substation project cost prediction method based on improved particle swarm optimization-least squares support vector regression algorithm

作　　者：王林峰 刘云 亓彦珣[2] 周波 李洁 WANG Linfeng;LIU Yun;QI Yanxun;ZHOU Bo;LI Jie(School of Electrical and Electronic Engineering,Huazhong University of Science and Technology,Wuhan 430074,Hubei,China;Economic and Technological Research Institute,State Grid Hebei Electric Power Co.,Ltd.,Shijiazhuang 050001,Hebei,China;School of Electrical Engineering and Automation,Wuhan University,Wuhan 430072,Hubei,China;School of Energy,Power and Mechanical Engineering,North China Electric Power University,Beijing 102206,China;Hebei SECPT Computer Consulting Service Co.,Ltd.,Shijiazhuang 050081,Hebei,China)

机构地区：[1]华中科技大学电气与电子工程学院,湖北武汉430074 [2]河北省电力有限公司经济技术研究院,河北石家庄050001 [3]武汉大学电气与自动化学院,湖北武汉430072 [4]华北电力大学能源动力与机械工程学院,北京102206 [5]河北赛克普泰计算机咨询服务有限公司,河北石家庄050081

出　　处：《沈阳工业大学学报》2025年第2期168-175,共8页Journal of Shenyang University of Technology

基　　金：河北省自然科学基金重点项目(E2018210044);河北省教育厅科技项目(QN16214510D)。

摘　　要：【目的】电网建设项目中变电站工程造价的预测一直是影响项目成本管理的重要问题。然而,当前常用的变电站造价预测方法存在预测精度不足、计算效率低等问题,制约了预测模型在实际工程中的应用。为提高预测的准确性和计算效率,提出了一种基于改进的粒子群优化(IPSO)算法和最小二乘支持向量回归(LSSVR)算法的变电站工程造价预测方法。【方法】考虑到常规变电站与智能变电站在设备、技术和运维上的差异,通过分析这两类变电站的特点,对相关数据进行了有针对性的预处理,以去除噪声数据,填补缺失值,并将有效信息转换为特征向量,作为LSSVR模型的输入。为避免传统粒子群(PSO)算法易陷入局部最优解的问题,引入了一种混合调节策略,对PSO算法的惯性权重和学习因子进行优化,使得优化过程更加稳定并具备较强的全局搜索能力。通过该策略IPSO算法可以在全局搜索和局部搜索之间实现更好的平衡。利用IPSO算法优化LSSVR模型参数,并建立变电站工程造价预测模型。【结果】通过与其他预测模型进行比较分析得出结论,所提出的IPSO-LSSVR算法在预测精度上具有明显优势。具体来说,基于该模型的预测误差显著低于其他方法,可以将偏差控制在5%以内。改进后的粒子群优化算法能够有效避免陷入局部最优,确保了LSSVR模型在各种情况下都能提供较为准确的预测结果。【结论】基于IPSO优化LSSVR算法的变电站工程造价预测方法,克服了传统预测方法在预测精度和计算效率上的不足。在实际应用中,该方法能够为电网建设项目的成本管理提供更加准确的预测依据,从而有助于项目预算的合理制定和资源的有效配置。[Objective]The prediction of substation project cost in power grid construction projects has always been an important issue influencing project cost management.However,the currently commonly used substation cost prediction methods have problems such as insufficient prediction accuracy and low computational efficiency,which restricts the application of prediction models in actual projects.To improve the accuracy and computational efficiency of prediction,a substation project cost prediction method was proposed by combining the improved particle swarm optimization(IPSO)algorithm and least squares support vector regression(LSSVR)algorithm.[Methods]First,considering the differences in equipment,technology,and operation and maintenance between conventional substations and intelligent substations,the characteristics of these two types of substations were analyzed,and targeted preprocessing was performed on the relevant data to remove the noisy data,fill in the missing values,and convert valid information into feature vectors to be used as inputs of the LSSVR model.Next,to avoid the problem that the traditional particle swarm optimization(PSO)algorithm was prone to fall into the locally optimal solution,a hybrid adjustment strategy was introduced to optimize the inertia weights and learning factors of the PSO algorithm,which made the optimization process more stable and had a strong global search capability.With the help of this strategy,IPSO algorithm could achieve a better balance between global and local search.Finally,the IPSO algorithm was used to optimize the parameters of the LSSVR model,and a substation project cost prediction model was built.[Results]It is found from comparison with other prediction models that the proposed IPSO-LSSVR algorithm has significant advantages in prediction accuracy.Specifically,the prediction error of the model is significantly lower than those of other methods,and the deviation can be controlled within 5%.The IPSO algorithm can effectively avoid falling into local optima,which ensu

关 键 词：变电站 工程造价 造价预测 粒子群算法 最小二乘支持向量回归 预测精度 运算效率 混合调节策略 

分 类 号：TM769[电气工程—电力系统及自动化]