摘要
为了准确预测未来新能源汽车充电负荷,提前为城市电网的调度以及充电站储能设备的建立做准备,采用支持向量回归与粒子群优化算法构建一种基于小样本的新能源汽车保有量预测模型,并在此基础上利用改进蒙特卡洛模拟法对江苏省未来某日新能源汽车充电负荷进行预测。首先对数据进行预处理时,通过灰色关联度分析法筛选出与保有量关联度较大的几个关键影响因素,然后对回归模型超参数进行寻优,最后通过改进蒙特卡洛法让模型自主判断用户每日充电次数并进行充电行为模拟以得到日充电负荷预测曲线,为防止峰值叠加和城市电力系统规划管理提供理论支持。
On the purpuse of accurately forecasting the load of new energy vehicles'charging,preparing for the dispatching of urban power grid and the establishment of energy storage equipment of charging station in advance.A new energy vehicles'ownership forecast model based on small sample is constructed by Support Vector Regressio and Particle Swarm Optimization.On this basis,the daily load in future of new energy vehicles'charging in Province Jiangsu will be forecasted by modified Monte Carlo method.Initiallly,the key influential factors that are closely related to the number of ownership are filtered by using the grey relational degree analysis method in the data preprocessing.Secondly,the hyperparameters of regression model will be optimized.Eventually,the daily load forecast curve of new energy vehicles'charging can be obtained by modified Monte Carlo method which can make the model independently choose the charging times and simulate the charging behavior,provides theoretical support for preventing peak superposition and urban power system planning and management.
作者
殷铭
YIN Ming(School of Electrical and Information Engineering,Anhui University of Science and Technology,Huainan 232001,China)
出处
《电工技术》
2024年第24期44-49,共6页
Electric Engineering
关键词
负荷预测
支持向量回归
粒子群优化
保有量预测
蒙特卡洛
load forecast
Support Vector Regression
Particle Swarm Optimization
ownership forecast
Monte Carlo
