Nowadays,electric vehicles(EVs)are increasingly equipped with advanced onboard devices capable of collecting and recording real-time charging data.The analysis of such data from a large-scale EV fleet plays a crucial ...Nowadays,electric vehicles(EVs)are increasingly equipped with advanced onboard devices capable of collecting and recording real-time charging data.The analysis of such data from a large-scale EV fleet plays a crucial role in supporting decision-making processes,particularly in the deployment of charging infrastructure and the formulation of EV-focused policies.Nevertheless,the challenges of collecting these data are significant,primarily due to privacy concerns and the high costs associated with data access.In response,this study introduces an innovative methodology for generating large-scale and diverse EV charging data,mirroring real-world patterns for cost-efficient and privacy-compliant use.Specifically,this approach combines Gibbs sampling and conditional density networks and was trained and validated using a real-world dataset consisting of approximately 1.65 million charging events from 3,777 battery EVs(BEVs)in Shanghai over a year.Results illustrate that the proposed model can effectively capture the underlying distribution of the original charging data,enabling the generation of synthetic samples that closely resemble real-world charging events.The approach is readily employed for data imputation and augmentation,and it can also help simulate future charging distributions by conditional generation based on anticipated development premises.展开更多
Unsupervised learning algorithms can effectively solve sample imbalance.To address battery consistency anomalies in new energy vehicles,we adopt a variety of unsupervised learning algorithms to evaluate and predict th...Unsupervised learning algorithms can effectively solve sample imbalance.To address battery consistency anomalies in new energy vehicles,we adopt a variety of unsupervised learning algorithms to evaluate and predict the battery consistency of three vehicles using charging fragment data from actual operating conditions.We extract battery-related features,such as the mean of maximum difference,standard deviation,and entropy of batteries and then apply principal component analysis to reduce the dimensionality and record the amount of preserved information.We then build models through a collection of unsupervised learning algorithms for the anomaly detection of cell consistency faults.We also determine whether unsupervised and supervised learning algorithms can address the battery consistency problem and document the parameter tuning process.In addition,we compare the prediction effectiveness of charging and discharging features modeled individually and in combination,determine the choice of charging and discharging features to be modeled in combination,and visualize the multidimensional data for fault detection.Experimental results show that the unsupervised learning algorithm is effective in visualizing and predicting vehicle core conformance faults,and can accurately predict faults in real time.The“distance+boxplot”algorithm shows the best performance with a prediction accuracy of 80%,a recall rate of 100%,and an F1 of 0.89.The proposed approach can be applied to monitor battery consistency faults in real time and reduce the possibility of disasters arising from consistency faults.展开更多
基金National Natural Science Foundation of China(72101153 and 72061127003)Shanghai Chenguang Program(21CGA72),Shanghai Eastern Scholar Program(QD2020057)+1 种基金Shanghai Frontiers Science Center of Artificial Intelligence and Deep Learning at NYU ShanghaiNYU Shanghai Doctoral Fellowships,and the NYU Shanghai Boost Fund.
文摘Nowadays,electric vehicles(EVs)are increasingly equipped with advanced onboard devices capable of collecting and recording real-time charging data.The analysis of such data from a large-scale EV fleet plays a crucial role in supporting decision-making processes,particularly in the deployment of charging infrastructure and the formulation of EV-focused policies.Nevertheless,the challenges of collecting these data are significant,primarily due to privacy concerns and the high costs associated with data access.In response,this study introduces an innovative methodology for generating large-scale and diverse EV charging data,mirroring real-world patterns for cost-efficient and privacy-compliant use.Specifically,this approach combines Gibbs sampling and conditional density networks and was trained and validated using a real-world dataset consisting of approximately 1.65 million charging events from 3,777 battery EVs(BEVs)in Shanghai over a year.Results illustrate that the proposed model can effectively capture the underlying distribution of the original charging data,enabling the generation of synthetic samples that closely resemble real-world charging events.The approach is readily employed for data imputation and augmentation,and it can also help simulate future charging distributions by conditional generation based on anticipated development premises.
文摘Unsupervised learning algorithms can effectively solve sample imbalance.To address battery consistency anomalies in new energy vehicles,we adopt a variety of unsupervised learning algorithms to evaluate and predict the battery consistency of three vehicles using charging fragment data from actual operating conditions.We extract battery-related features,such as the mean of maximum difference,standard deviation,and entropy of batteries and then apply principal component analysis to reduce the dimensionality and record the amount of preserved information.We then build models through a collection of unsupervised learning algorithms for the anomaly detection of cell consistency faults.We also determine whether unsupervised and supervised learning algorithms can address the battery consistency problem and document the parameter tuning process.In addition,we compare the prediction effectiveness of charging and discharging features modeled individually and in combination,determine the choice of charging and discharging features to be modeled in combination,and visualize the multidimensional data for fault detection.Experimental results show that the unsupervised learning algorithm is effective in visualizing and predicting vehicle core conformance faults,and can accurately predict faults in real time.The“distance+boxplot”algorithm shows the best performance with a prediction accuracy of 80%,a recall rate of 100%,and an F1 of 0.89.The proposed approach can be applied to monitor battery consistency faults in real time and reduce the possibility of disasters arising from consistency faults.
