为降低换电站的运行费用,准确预测来车换电和利用分时电价差合理规划充电调度是核心环节。该文提出一种基于遗传算法优化的CatBoost模型,能够精确预测来车换电,并结合预测结果实现在高尖峰时期对开放充电口数的合理规划,同时引入光储系统进一步实现平均购电成本的降低。结果表明：通过SHAP摘要图和SHAP依赖图分析特征重要性和特征交互性,得到SOC、时间和距离3种特征对模型性能影响最大;GA-CatBoost在准确率、精确率、召回率和F1分数4种评估指标上均优于另外3种预测模型;基于预测结果,针对高尖峰时期的充电需求,制定多目标充电调度策略,在保障换电服务的同时显著降低换电站的平均购电成本;最后引入光储系统,仿真光储换一体站的分时电价工作模式,通过电力流动和购电成本的可视化分析,验证光储系统对电价优化的有效性,进一步减少换电站的电力采购成本。
Abstract
To reduce operational costs, accurate prediction of battery electric vehicle arrival time for battery swapping and strategic charging scheduling based on time-of-use electricity pricing are essential. This paper proposes a CatBoost model optimized by a genetic algorithm （GA）, which can accurately predicts vehicle arrivals for battery swapping. Based on the prediction results, the model enables optimal planning of charging ports during peak periods. Furthermore, an integrated photovoltaic and energy storage system is introduced to reduce the average electricity purchase cost. The results demonstrate that through SHAP summary plots and SHAP dependence plots, state-of-charge （SOC）, time, and distance are identified as the most influential features affecting model performance. The GA-CatBoost model outperforms three other prediction models in terms of accuracy, precision, recall, and F1 score. Based on the prediction outcomes, a multi-objective charging scheduling strategy is implemented for peak periods, significantly reducing the average electricity purchase cost while ensuring sufficient battery swapping services. Finally, with the introduction of a photovoltaic-storage system and simulation of the time-of-use pricing operation mode for an integrated system, the visual analysis of power flow and electricity cost validates the effectiveness of the system in optimizing electricity pricing, further reducing the station's electricity procurement costs.
关键词
电动汽车 /
光伏 /
储能 /
预测 /
电力负荷调度 /
换电站
Key words
battery electric vehicles /
photovoltaics /
energy storage /
prediction /
electric load dispatching /
battery swapping stations
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 孙波, 董浩, 王璐, 等. 计及电动汽车需求响应的风储混合系统日前联合优化调度模型[J]. 太阳能学报, 2021, 42(4): 107-114.
SUN B, DONG H, WANG L, et al.Joint optimal scheduling model for wind-storage hybrid system considering demand response of electric vehicle[J]. Acta energiae solaris sinica, 2021, 42(4): 107-114.
[2] 闫丽梅, 曾家威, 徐建军, 等. 含电动汽车和光伏的配电网演化动态博弈调度策略[J]. 太阳能学报, 2024, 45(5): 316-323.
YAN L M, ZENG J W, XU J J, et al.Scheduling strategy of distribution grid with electric vehicles and photolatic based on evolutionary dynamics game[J]. Acta energiae solaris sinica, 2024, 45(5): 316-323.
[3] SHALABY A A, SHAABAN M F, MOKHTAR M, et al.A dynamic optimal battery swapping mechanism for electric vehicles using an LSTM-based rolling horizon approach[J]. IEEE transactions on intelligent transportation systems, 2022, 23(9): 15218-15232.
[4] WANG S Y, CHEN A, WANG P X, et al.Short-term electric vehicle battery swapping demand prediction: deep learning methods[J]. Transportation research part D: transport and environment, 2023, 119: 103746.
[5] ZHANG F, LV H T, XING Q, et al.Deployment of battery-swapping stations: integrating travel chain simulation and multi-objective optimization for delivery electric micromobility vehicles[J]. Energy, 2024, 290: 130252.
[6] 王子奇, 侯思祖, 郭威. 考虑库存电池的光储换电站优化充电策略[J]. 电力自动化设备, 2022, 42(10): 142-148, 176.
WANG Z Q, HOU S Z, GUO W.Optimal charging strategy of PV-BESS-based battery swapping station considering inventory batteries[J]. Electric power automation equipment, 2022, 42(10): 142-148, 176.
[7] 何家智, 冯现大, 刘天琦. CatBoost算法结合Optuna框架预测砂土液化[J]. 济南大学学报(自然科学版), 2024, 38(4): 496-502.
HE J Z, FENG X D, LIU T Q.Sand liquefaction prediction by using CatBoost algorithm combined with Optuna framework[J]. Journal of University of Jinan (science and technology), 2024, 38(4): 496-502.
[8] HUANG G M, WU L F, MA X, et al.Evaluation of CatBoost method for prediction of reference evapotranspiration in humid regions[J]. Journal of hydrology, 2019, 574: 1029-1041.
[9] PROKHORENKOVA, LIUDMILA, et al.CatBoost: unbiased boosting with categorical features[J]. Advances in neural information processing systems, 2018, 31.
[10] JAIN, SHUBHAM. Introduction to genetic algorithm & their application in data science[J]. Analytics vadhya,(2017-07-05)[2024-01-06]. https://www.analyticsvidhya.com/blog/2017/07/introduction-to-genetic-algorithm/.
[11] 张强, 缪维跑, 刘青松, 等. 基于多目标遗传算法的垂直轴风力机专用翼型优化设计[J]. 太阳能学报, 2023, 44(4): 9-16.
ZHANG Q, MIAO W P, LIU Q S, et al.Optimal design of vertical axis wind turbine special airfoil based on multi-objective genetic algorithm[J]. Acta energiae solaris sinica, 2023, 44(4): 9-16.
[12] IKASARI D, IKASARI W, ANDIKA R.Implementation of haversine formula to determine the shortest path using web based application for a case study of high school zoning in depok[J]. American journal of software engineering and applications, 2021, 10(2): 19.
[13] CARRINGTON A M, FIEGUTH P W, QAZI H, et al.A new concordant partial AUC and partial c statistic for imbalanced data in the evaluation of machine learning algorithms[J]. BMC medical informatics and decision making, 2020, 20(1): 4.
基金
国家重点研发计划（2022YFB4201000）
PDF(3881 KB)
