| 引用本文: | 田书欣,周全,程浩忠,柳璐,路亮,江栗.基于鸽群优化算法的支持向量机在电力需求总量预测中的应用[J].电力自动化设备,2020,40(5): |
| TIAN Shuxin,ZHOU Quan,CHENG Haozhong,LIU Lu,LU Liang,JIANG Li.Application of pigeon-inspired optimization algorithm based SVM in total power demand forecasting[J].Electric Power Automation Equipment,2020,40(5): |
|
| 摘要: |
| 电力需求总量的科学预测是经济转型阶段电力系统规划与运行的重要依据。引入融合分段二次Lagrange插值函数的新型灰色关联理论，从经济发展、产业结构、用电环境以及居民生活4个方面分析社会经济新常态指标与电力需求总量之间的相关程度，筛选出影响电力需求增长率波动的关键因素；进而以电力需求总量及相关因素为训练数据集，利用融合莱维飞行特征的改进鸽群优化算法对支持向量机的参数进行优化，建立具有最佳参数、强泛化能力的电力需求总量预测模型。基于我国某电网区域电力需求历史实测数据的算例结果表明，所建模型具有更好的优化效率和预测精度。 |
| 关键词: 电力需求总量 新常态指标 鸽群优化算法 支持向量机 灰色关联 |
| DOI：10.16081/j.epae.202004016 |
| 分类号:TM715 |
| 基金项目:国家电网公司管理咨询项目(SGXN000GHWT17-00001) |
|
| Application of pigeon-inspired optimization algorithm based SVM in total power demand forecasting |
|
TIAN Shuxin1, ZHOU Quan2, CHENG Haozhong3, LIU Lu3, LU Liang2, JIANG Li2
|
|
1.College of Electric Power Engineering, Shanghai University of Electric Power, Shanghai 200090, China;2.Southwest Branch of State Grid Corporation of China, Chengdu 610041, China;3.Key Laboratory of Control of Power Transmission and Conversion, Ministry of Education, Department of Electrical Engineering, Shanghai Jiao Tong University, Shanghai 200240, China
|
| Abstract: |
| Scientific forecasting of total power demand is an important basis for power system planning and operation during the period of economy transition. A novel grey correlation theory combined with piecewise quadratic Lagrange interpolation function is introduced to analyze the correlation degree between social economic new normal indexes and total power demand from four aspects of economic development, industrial structure, power utilization environment and residential living, and the key factors that affecting the growth rate fluctuation of total power demand are screened. Further, the total power demand and its relevant factors are taken as training dataset, an improved pigeon-inspired optimization algorithm merging with Lévy flights is adopted to optimize the parameters of SVM(Support Vector Machine),and a forecasting model of total power demand with the best parameters and strong generalization capability is built. The case results based on the historical measured power demand data of a power grid region in China show that the proposed model has better optimization efficiency and forecasting accuracy. |
| Key words: total power demand new normal index pigeon-inspired optimization algorithm support vector machines grey correlation |
