基于L1-SSA-SVM模型的城镇燃气短期负荷预测  

Short term load forecasting of urban gas based on L1-SSA-SVM model

作　　者：肖荣鸽[1] 夏海平 李雨泽 刘鸿嘉 张青松 XIAO Rongge;XIA Haiping;LI Yuze;LIU Hongjia;ZHANG Qingsong(School of Petroleum Engineering,Xi’an Shiyou University,Xi’an 710065,China)

机构地区：[1]西安石油大学石油工程学院,陕西西安710065

出　　处：《区域供热》2024年第6期150-158,共9页District Heating

摘　　要：精准及时的燃气负荷预测对于充分调配燃气资源,保障居民用气问题是十分重要的,然而燃气负荷本身具有非线性的特点,故想要建成固定模拟机理十分困难。为了找到更加精准的燃气负荷预测模型,提出了一种利用L1(Least absolute shrinkage and selection operator,Lasso Regression)范数特征选择分析关联度,基于优化麻雀算法SSA(Sparrow Search Algorithm,ISSA)优化支持向量机(support vector machines,SVM)的燃气负荷预测模型。利用L1范数特征选择在燃气负荷相关的11个影响因素中选择,通过分析不同影响因素之间的关联度,让其中一部分影响因素的系数缩小,产生一个稀疏权值矩阵,剔除关联度相对较小的影响因素。将关联度较高的影响因素作为SVM的输入,再对支持向量机模型的惩罚因子c及核函数参数g进行优化,建立L1-SSA-SVM模型预测城镇燃气负荷,并验证其准确性和有效性。结果表明:所提出模型MAPE为0.65%,远低于传统的支持向量机模型以及传统麻雀搜索算法优化的支持向量机模型,同时,使用L1范数特征选择对影响燃气负荷预测的因素进行筛选能有效地提高所建立模型的预测精准度,文中所提出的L1-SSA-SVM模型具有十分广泛的适用性。Accurate and timely gas load prediction is crucial for fully allocating gas resources and ensuring residential gas consumption.However,gas loads themselves have non-linear characteristics,making it very difficult to establish a fixed simulation mechanism.In order to find a more accurate gas load prediction model,a gas load prediction model using L1 norm feature selection analysis correlation and support vector machine(SVM)optimization based on the Sparrow Optimization Algorithm(ISSA)is proposed.Using L1 norm feature selection to select among the 11 influencing factors related to gas load,by analyzing the correlation degree between different influencing factors,reducing the coefficients of some influencing factors,a sparse weight matrix is generated to eliminate the relatively small correlation factors,and the high correlation factors are used as inputs for SVM(Support Vector Machines,SVM).Then,ISSA(Improving Sparrow Search Algorithm,ISSA)is used to optimize the penalty factor c and kernel function parameter g of the support vector machine model,establish the ISSA-SVM model to predict urban gas load,and verify its accuracy and effectiveness.The results show that the proposed model has a MAPE of 0.65%,which is much lower than traditional support vector machine models and support vector machine models optimized by traditional sparrow search algorithms.

关 键 词：燃气负荷预测 麻雀搜索算法 L1范数特征选择 支持向量机模型 

分 类 号：TU996.3[建筑科学—供热、供燃气、通风及空调工程]