Real-world Electric Vehicle Research Articles

Collision-Caused thermal runaway investigation of li-ion battery in Real-World electric vehicles

Due to the limitations of static and non-realistic load conditions in laboratory environments, existing battery thermal runaway mechanism/characterization studies are unable to realistically restore the operating conditions and extract thermal runaway characteristics. In this paper, the actual thermal runaway historical data of collision electric vehicles are used. In-depth data mining and characterization of historical data from different perspectives are performed for the three months before the accident. Several feature parameters and outliers are extracted in the paper to explore the characteristics and mechanisms of thermal runaway due to collisions in the real world. In addition, based on the anomaly fluctuation analysis, a modified overall Shannon entropy is proposed to infer the exact collision time and location. The results show that the diagnostic results match the thermal runaway location in the accident report. In this paper, we used actual data to count the anomalous information before the occurrence of thermal runaway, and utilized the corresponding evaluation indexes to diagnose the anomalous features in the early stage of the accident 91 days in advance. This is the first study of thermal runaway of batteries caused by collision based on actual accident cases and real vehicle operation data.

Voltage fault diagnosis and misdiagnosis analysis of battery systems using the modified Shannon entropy in real-world electric vehicles

Power battery systems fault diagnosis of electric vehicles (EVs) is a key technical approach to ensure the safety and reliable operation of the vehicle, of which avoiding misdiagnosis not only can reduce the driver's safety concerns, but also is a reflection of the reliability of the diagnostic method. This paper first proposes a modified Shannon entropy-based battery fault diagnosis method for identifying cells with abnormal voltage fluctuations in battery systems, and the method is implemented online by calculating the Shannon entropy of the voltage sequence in a moving time window. Then, the defined sensitivity factor (SF) can provide an efficient and accurate assessment of the extent of abnormal voltage fluctuations. Further, to improve the fault diagnosis accuracy of the method, we validated the diagnostic model by using battery operation data from a battery monitoring cloud platform with a sampling frequency of 0.1 Hz and eventually found a large number of false alarm cells in the diagnostic results. Based on this, we have analysed the fault diagnosis mechanisms of the model and thus obtained the causes of misdiagnosis. Finally, a solution to reduce misdiagnosis is proposed and its effectiveness is verified. To further test its comprehensive performance, we compared the accuracy of the method before and after optimization based on normal vehicle data, and the results showed a 81.9 % reduction in the average relative misdiagnosis rate of the model, thus greatly improving the reliability of this fault diagnosis strategy, creating conditions for the online application of the method and providing ideas for the analysis and improvement of the accuracy rate of other fault diagnosis methods.

Robust Operation Scheme of EV Charging Facility With Uncertain User Behavior

With the widespread of electric vehicles (EV), the necessity of installing EV charging facility has rapidly increased. Since operating EV charging facility are costly and complex, choosing a proper power operation scheme is critical in implementing a cost-effective charging service. Therefore, this paper proposes an efficient power operation scheme for EV charging facility by jointly analyzing not only monetary issues with various power rate policies and battery wear-out costs of the energy storage system (ESS) but also the uncertainty issues caused by EV users' behavior under different charging requirements. By considering the issues, this paper analyzes and formulates the problem with robust optimization techniques. By using the real-world EV operation datasets, the effectiveness of the proposed scheme is analyzed in terms of daily peak power and total cost. In addition, by applying the proposed scheme in an actual power system, it is shown that the proposed scheme is able to reduce the overall cost by <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$20.5\%$</tex-math></inline-formula> and reduce peak power by <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$17.3\%$</tex-math></inline-formula> compared to other benchmark models.

Multi-forward-step battery voltage prediction for real-world electric vehicles using gated recurrent units

Accurate battery voltage prediction performs a critical role in electric vehicles, especially safety risks prognosis in future periods. This paper proposes a multi-forward-step battery voltage prediction method using gated recurrent units. Actual vehicle operation data and weather data are used together for model training. A two-step procedure including data cleaning and correlation analysis is conducted to improve training efficiency and prediction accuracy. Furthermore, the effects of different model parameters on the prediction results are also investigated to obtain an optimal set of model parameters. Additionally, the model is tested with the data of the second year. The results proved that the model not only has a high prediction accuracy, but also has a strong stability and robustness. On this basis, a research plan for real-time multi-forward-step battery voltage prediction is further developed, whereby identifying occurrence of battery failure and mitigate the risk of thermal runaway by voltage predicting.

Towards Nearly Zero-Energy Buildings: Smart Energy Management of Vehicle-to-Building (V2B) Strategy and Renewable Energy Sources

As countries worldwide strive for net-zero emissions, electric vehicles (EVs) have gained significant momentum. EVs, particularly through Vehicle-to-Everything (V2X) technology, including Vehicle-to-Building (V2B), are recognized as key enablers for achieving nearly-zero energy buildings (NZEB). This research presents an advanced Smart Energy Management System (SEMS) with a novel V2B strategy, offering precise one-hour resolution to reduce grid reliance and optimize peak load management. The SEMS coordinates building energy and EV charging while harnessing the potential of renewable energy resources and battery storage systems. By carefully scheduling EV charging and discharging activities, it maximizes the efficient use of available energy resources. Importantly, the algorithm integrates real-world EV parking patterns, accounting for factors such as fleet composition, parking durations, and charging priorities. Economic evaluation encompasses not only direct EV battery degradation costs but also external expenses associated with environmental impacts. Simulation outcomes underscore a remarkable reduction in building grid power demand (65%) and a substantial decrease in carbon emissions (64%). The PV+V2B Scenario emerges as the preferred choice, striking an optimal balance between initial investment costs, energy savings, and carbon footprint reductions. This study holds great academic and practical significance, contributing to the advancement of smart, sustainable, and economically viable net-zero cities.

Analysis and key findings from real-world electric vehicle field data

Analysis and key findings from real-world electric vehicle field data

Personalized Velocity and Energy Prediction for Electric Vehicles With Road Features in Consideration

Electric vehicles (EVs) seem to be an eminent alternative for ground transportation. Yet, human drivers may suffer from EV’s range anxiety, which is engendered by EV’s limited battery capacity and sparse charging stations, particularly in rural areas. A reliable prediction of EV battery energy consumption for the intended route before traveling can alleviate this feeling of unease. Fundamentally, an accurate speed trajectory forecast lays the foundation for a dependable energy consumption prediction. This paper originates a data-driven paradigm to predict an EV’s speed considering various road features and individuals’ driving characteristics. Two modified Transformer neural networks, which outperform the traditional recurrent neural networks, are exploited to extract information from input features and predict vehicular acceleration or velocity profile. Additionally, a novel energy consumption model is suggested to pioneeringly take the tire slip ratio into account and utilize a neural-network-predicted powertrain efficiency map. By design, the proposed algorithm can offer EV speed and energy consumption estimations prior to the start of a trip. Experimental data collected in real-world EV test drives on four different routes are employed to validate the method. Compared to a baseline approach, the proposed scheme yields a superior accuracy on both velocity and energy consumption predictions for EVs.

Assessing the value of information for electric vehicle charging strategies at office buildings

Smart charging strategies for electric vehicles (EVs) require as input information such as energy requirement and dwell time. In practice, that information is often not available. However, estimations may be obtained from historical charging behavior. This paper investigates the added value of historical information for EV charging strategies based on real-world EV charging data collected in an office building parking lot with 125 chargers. Furthermore, it provides valuable insights into EV charging behavior at office building parking lots, based on a statistical analysis of the data. The added value of data availability in EV charging strategies for day-ahead planning and real-time operation of office building parking lots is assessed via a set of quality metrics that measure user satisfaction and impact on the local grid. Offline charging strategies under various degrees of available information are validated by comparing their performance with the real-time operation of the parking lot. Results show a power peak reduction of more than 50% using historical data and simple estimations of arrival times, dwell times, and energy requirement. A trade-off between power peaks and service quality (on average 4.4kWh energy not served) is observed. It was found that knowledge of individual average energy provides higher added value compared to knowing individual average dwell time in both offline planning and real-time operation of the parking lot.

A Battery Capacity Estimation Framework Combining Hybrid Deep Neural Network and Regional Capacity Calculation Based on Real-World Operating Data

Efficient battery capacity estimation is of utmost importance for safe and reliable operations of electric vehicles (EVs). This paper proposes a battery capacity estimation framework based on real-world EV operating data collected from forty electric buses of the same model operating in two cities. First, a reference capacity calculation method is presented by combining the Coulomb counting method with the incremental capacity analysis method. Then the impacts of temperature, current, and State-of-Charge on battery degradation are quantitatively analyzed. Using the historical probability distributions as battery health features, a hybrid deep neural network model that combines a convolutional neural network with a fully-connected neural network is proposed for battery capacity estimation. The validation results show that the proposed model outperforms the state-of-the-art methods and reaches a mean absolute percentage error of 2.79 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\%$</tex-math></inline-formula> , while maintaining low computational cost.

Battery safety issue detection in real-world electric vehicles by integrated modeling and voltage abnormality

Detecting battery safety issues is essential to ensure safe and reliable operation of electric vehicles (EVs). This paper proposes an enabling battery safety issue detection method for real-world EVs through integrated battery modeling and voltage abnormality detection. Firstly, a battery voltage abnormality degree that is adaptive to different battery types and working conditions is defined. Then an integrated battery model is developed by combining an electrochemical model, an equivalent circuit model (ECM), and a data-driven model to evaluate the normal voltage. To ensure normality of input current, a current processing model is presented. The performance of the proposed scheme is examined under random loading profiles using operating data collected from real-world EVs. The results show that the integrated battery model can precisely predict normal battery terminal voltage, with mean-squared-errors of 1.034e−4 V2, 7.221e−5 V2, and 4.612e−5 V2 for driving, quick charging, and slow charging, respectively. The accuracy in classifying faulty and normal batteries is verified based on the operating data collected from 20 EVs.

Digital Twin-Based Remaining Driving Range Prediction for Connected Electric Vehicles

&lt;div&gt;Electric vehicles (EVs) suffer from long charging time and inconvenient charging due to limited charging stations, which are the main causes of drivers’ range anxiety. Real-time and accurate driving range prediction can help drivers plan journeys, alleviate range anxiety, and promote EV development. However, predicting the EV driving range is challenging due to different weather, road conditions, driver habits, and limited available data. To address this issue, this article proposes a novel digital twin-based driving range prediction method. First, a one-year real-world EV dataset in Beijing is utilized. Detailed feature selection is conducted for the dataset, and six key features are extracted: battery SOC, consumed battery SOC, battery total voltage, battery maximum cell voltage, battery minimum cell voltage, and mileage already driven. Then, a random forest method is used to train the EV driving range prediction model using the features described earlier. Four prediction models with different adopted features are trained, respectively. Finally, the sliding window algorithm is proposed for the input of random forest to investigate its impact on prediction accuracy in the four prediction models, and different window sizes are evaluated. Results show that the sliding window algorithm can significantly improve the prediction model with only SOC as input, while it can deteriorate other models with more features. The most accurate model taking all six features as inputs provides 89.8% data that has an accuracy of over 80%, while data proportion of the prediction model without past energy consumption is only 31.8%.&lt;/div&gt;

Hybrid deep learning model for efficient state of charge estimation of Li-ion batteries in electric vehicles

State of charge (SoC) estimation is critical for the safe and efficient operation of electric vehicles (EVs). This work proposes a hybrid multi-layer deep neural network (HMDNN)-based approach for SoC estimation in EVs. This HMDNN uses Mountain Gazelle Optimizer (MGO) as a training algorithm for the deep neural network. Our method leverages the intrinsic relationship between the SoC and the voltage/current measurements of the EV battery to accurately estimate the SoC in real time. We evaluate our approach on a large dataset of real-world EV charging data and demonstrate its effectiveness in comparison to traditional SoC estimation methods. Four diverse Li-ion battery datasets of electric vehicles are employed which are the dynamic stress test (DST), Beijing dynamic stress test (BJDST), federal urban driving schedule (FUDS), and highway driving schedule (US06) with different temperatures of 0oC,25oC,45oC. The comparison is made with Mayfly Optimization Algorithm based DNN, Particle Swarm Optimization based DNN and Back-Propagation based DNN. The evaluation indices used are normalized mean square error (NMSE), root mean square error (RMSE), mean absolute error (MAE), and relative error (RE). The proposed algorithm achieves 0.1% NMSE and 0.3% RMSE on average on all datasets, which validates the effective performance of the proposed model. The results show that the proposed neural network-based approach can achieve higher accuracy and faster convergence than existing methods. This can enable more efficient EV operation and improved battery life.

Battery Safety Risk Assessment in Real-World Electric Vehicles Based on Abnormal Internal Resistance Using Proposed Robust Estimation Method and Hybrid Neural Networks

Battery safety issue is developing as one of the main hinders restricting the further application of real-world electric vehicles (EVs). Internal resistance (IR) is one of the important parameters to reflect battery safety, because bigger abnormal IR will cause more heat generation and make the battery easier to cross the critical condition of thermal runway. Safety risk assessment based on abnormal IR can locate this kind unsafe batteries and ensure the safe operation of EVs. Existing studies mostly focus on IR estimation in laboratory, yet is insufficient in real-world abnormal IR detection, which is influenced by random conditions and coupled factors. To cope with this issue, a method is proposed to detect unsafe battery, thereby predicting the thermal runaway. The method can be divided into three parts, i.e., IR estimation, normal IR prediction, and IR evolution law construction and safety risk assessment. First, we analyze the problems of battery IR estimation in real-world scenarios, and propose a robust method to estimate the IR only based on sparse voltage and current. Second, a novel hybrid neural network model is designed and trained to predict normal IRs inputted by temperature, mileage, and state-of-charge. The model combines the advantages of different neural network structures to improve the performance. A double-layer grid searching method is proposed to optimize hyperparameters of the hybrid neural network, aiming to achieve the best accuracy under constructed neural network structure. Finally, the safety boundary indicated by IR is formed by statistic method and the IR evolution law is constructed, then a strategy is proposed to make residual evaluation for safety risk assessment. All the proposed methods are driven by EV historical data to reflect real-world conditions. The results showcase that the method is effective in the cloud for the small IR prediction error with the mean-square-error, mean-relative-error, and max error of 2.45*e-4, 1.78%, and 0.15. In addition, the method can periodically classify the unsafe and normal batteries with 98.6% accuracy for the EVs with the same specification.

Multi-dimension statistical analysis and selection of safety-representing features for battery pack in real-world electric vehicles

Battery safety issues have been raising anxiety of electric vehicle (EV) consumers and threatening to the occupant’s life. Many battery safety management/monitoring methods and features have been reported, but which features can better represent battery safety in real-world EVs has been rarely discussed and compared in the unified level. In this study, the multi-feature and multi-dimension statistical analysis for battery pack safety in numerous real-world electric EVs is deployed. Firstly, an EV state distinction scheme is proposed to cope with the different characteristics of EV driving and EV charging, and twenty statistical features are constructed to extract the characteristics of battery voltages. Then, a scheme is proposed to select the feature which can effectively represent battery safety in real-world EVs. A multiple iteration Gaussian mixture model is proposed in the scheme to construct the actual feature distribution and cope with the random sensor error and data outlier. Finally, a model is constructed to analyze the statistical dimensions of selected battery safety-representing feature, and statistical analysis from different seasons, mileages, state-of-charges, and EV states is deployed. Results showcase that the selected feature is competent to distinguish between normal and unsafe batteries in real-world EV packs of different conditions. Various interesting conclusions based on the selected feature are also summarized to explore some common laws of battery safety in real-world EV application and support for the establishment of battery safety management method.

Recognition of Driving Behavior in Electric Vehicle’s Li-Ion Battery Aging

In the foreseeable future, electric vehicles (EVs) will play a key role in the decarbonization of transport systems. Replacing vehicles powered by internal combustion engines (ICEs) with electric ones reduces the amount of carbon dioxide (CO2) being released into the atmosphere on a daily basis. The Achilles heel of electrical transportation lies in the car battery management system (BMS) that brings challenges to lithium-ion (Li-ion) battery optimization in finding the trade-off between driving and battery health in both the long- and short-term use. In order to optimize the state-of-health (SOH) of the EV battery, this study focuses on a review of the common Li-ion battery aging process and behavior detection methods. To implement the driving behavior approaches, a study of the public dataset produced by real-world EVs is also provided. This research clarifies the specific battery aging process and factors brought on by EVs. According to the battery aging factors, the unclear meaning of driving behavior is also clarified in an understandable manner. This work concludes by highlighting some challenges to be researched in the future to encourage the industry in this area.

Data-Driven Distributionally Robust Electric Vehicle Balancing for Autonomous Mobility-on-Demand Systems Under Demand and Supply Uncertainties

Electric vehicles (EVs) are being rapidly adopted due to their economic and societal benefits. Autonomous mobility-on-demand (AMoD) systems also embrace this trend. However, the long charging time and high recharging frequency of EVs pose challenges to efficiently managing EV AMoD systems. The complicated dynamic charging and mobility process of EV AMoD systems makes the demand and supply uncertainties significant when designing vehicle balancing algorithms. In this work, we design a data-driven distributionally robust optimization (DRO) approach to balance EVs for both the mobility service and the charging process. The optimization goal is to minimize the worst-case expected cost under both passenger mobility demand uncertainties and EV supply uncertainties. We then propose a novel distributional uncertainty sets construction algorithm that guarantees the produced parameters are contained in desired confidence regions with a given probability. To solve the proposed DRO AMoD EV balancing problem, we derive an equivalent computationally tractable convex optimization problem. Based on real-world EV data of a taxi system, we show that with our solution the average total balancing cost is reduced by 14.49%, and the average mobility fairness and charging fairness are improved by 15.78% and 34.51%, respectively, compared to solutions that do not consider uncertainties.

Machine learning-based state of health prediction for battery systems in real-world electric vehicles

Carbon dioxide emission reduction is a significant benefit of electric vehicles. State of health prediction is essential to ensure the safety and reliability of the battery system, which is a key part of electric vehicles. This paper proposes a novel state of health prediction strategy based on ohmic internal resistance and long short-term memory networks. Driving conditions are extracted based on the current and speed of real-world vehicles. Combining the equivalent circuit model and Kalman filter, the parameter identification of the power battery system is performed. Ohmic internal resistance is selected as the health state characterization parameter. According to the ohmic internal resistance results obtained from the identification, Pearson correlation analysis is used to obtain the parameters with the highest correlation to the ohmic internal resistance. These parameters are used as input samples for the long short-term memory network to get the evaluation value of health status. The model is trained and tested using data from different operating conditions and vehicles. The root mean square errors of the model outputs are below 0.02 Ω, demonstrating the method's effectiveness. The proposed method is expected to have considerable application in advanced battery management systems.

A state of health estimation framework based on real-world electric vehicles operating data

The accurate estimation of battery state of health (SOH) is essential to alleviate mileage anxiety for the drivers of electric vehicles (EVs) and achieve echelon use of the battery. The data collected in real-world reflects complex operating conditions, which make the study of real-world EVs possible. In this paper, a framework for the SOH estimation of real-world EVs considering parameters optimization is proposed. First, the battery's capacity is calculated utilizing the deep charging process, and the non-linear trend of capacity decay is described by Generalized Additive Models (GAM). Then, the internal and external health features characterizing battery aging are extracted from the EVs data. The quality of the health features is evaluated by correlation analysis, and the parameters of the health features are selected with Bayesian optimization. Finally, the PSO-ELM model for SOH estimation is established by optimizing the weights and biases of extreme learning machine (ELM) by particle swarm optimization (PSO). The experimental results show that the framework proposed is able to accurately estimate the SOH of the real-world vehicle battery pack with a maximum error of less than 0.4 %.

Thermal fault prognosis of lithium-ion batteries in real-world electric vehicles using self-attention mechanism networks

Battery temperature is directly related to various thermal faults and thermal runaway of a battery system in electric vehicles. Its prediction is critical to the safety and durability of the battery system. This paper presents a novel clustering-based data partitioning method and a partitioned all-season coverage model established by the spiral self-attention neural network to perform real-time multi-forward-step temperature prediction for battery systems. A yearlong real-world operational dataset of an electric taxi is deployed. The proposed series-temperature method can generate a generic and non-interfering model for all temperature probes. The results demonstrate that the partitioned training model has excellent online prediction performance for non-differentiated real-world vehicle operation in all climates. The prognosis feasibility and reliability of the model for various temperature anomalies are also validated, providing a reliable tool for battery safety by determining the advent of thermal faults and mitigating the risk of thermal runaway. Furthermore, the flexible and adjustable predicted horizons could provide sufficient processing time for predicting and preventing battery anomaly/fault. This method's superiority, stability, flexibility, and robustness are verified under different hyperparameters and operating patterns.

Timely Thermal Runaway Prognosis for Battery Systems in Real-World Electric Vehicles Based on Temperature Abnormality

Hundreds of thermal-runaway-induced battery fire accidents have been occurring to real-world electric vehicles (EVs) in recent years, exposing life to danger and causing property losses. Timely and fast battery thermal runaway prognosis is essential but restricted by limited parameters and complex influencing factors during real-world operation of EVs, i.e., environment, driving behavior, and weather. To cope with the issue, several data-driven methods are combined, and the thermal runaway prognosis is realized by two steps, i.e., temperature prediction by the modified extreme gradient boosting (XGBoost) and then abnormality detection by the principal component analysis (PCA) and density-based spatial clustering of applications with noise (DBSCAN). The XGBoost is modified and trained by data of real-world EVs to couple the influencing factors during the real-world operation of EVs. For parameter optimization, the “pretraining and adjacent grid optimizing method” (P-AGOM) and the “adjacent grid optimizing method” (AGOM) are proposed to achieve locally optimal hyperparameters for XGBoost and DBSCAN. Verified results showcase that the XGBoost-PCA-DBSCAN achieves accurate 5-min-forward temperature prediction, and the mean square errors (mses) of four seasons are only 0.0729, 0.0594, 0.0747, and 0.0523, respectively. By modification of XGBoost, the mse of temperature prediction is reduced by 31.2%. In addition, the 35-min-forward thermal runaway prognosis by the XGBoost-PCA-DBSCAN will provide the driver sufficient response time to minimize the loss of life and property.