Real-world Electric Vehicle Research Articles

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.

Battery State-of-Health Estimation Based on Incremental Capacity Analysis Method: Synthesizing From Cell-Level Test to Real-World Application

The incremental capacity analysis (ICA) method is widely used in battery state-of-health (SOH) estimation due to its high prediction accuracy and aging mechanism implications. However, realizing precise SOH metering for real-world electric vehicles (EVs) is still challenging, if not impossible, and comprehensive and large-scale laboratory tests necessitated are usually time-consuming and labor-intensive. This article proposes an enabling SOH estimation scheme based on the ICA method for real-world EVs. This is realized by combining an equivalent IC-value calculation for battery packs with cell-level battery tests while taking cell inconsistency into consideration. The effectiveness of the proposed method is verified using the datasets collected from both well-controlled laboratory tests and daily operating EVs. The results show that battery cells within a batter pack generally experience similar degradation routes, which means insignificant cell inconsistency development with aging, and the proposed method can realize an accurate pack-level SOH estimation both for laboratory battery packs and real-world EVs. By applying the proposed method, the root mean square errors (RMSEs) of battery SOH prediction for laboratory modules, packs, and an electric taxi are 0.00955, 0.02457, and 0.0204, respectively. This study presents a verified framework of applying the ICA-based method to realize pack-level battery SOH estimation based on cell-level tests.

State of Health Estimation of Lithium-Ion Batteries in Electric Vehicles Based on Regional Capacity and LGBM

Battery state of health (SOH) estimation is a prerequisite for battery health management and is vital for second-life utilization. Existing techniques implemented in well-controlled experimental conditions fail to reflect complex working conditions during actual vehicular operation. In this article, a novel SOH estimation method for battery systems in real-world electric vehicles (EVs) is presented by combing results of regional capacity calculation and a light gradient boosting machine (LGBM) model. The LGBM model is used to capture the relationship between battery degeneration and influential factors based on datasets from real-world EVs. The regional capacity, which is calculated through incremental capacity analysis with a Gaussian smoothing filter, is utilized to reflect the battery degradation level while ensuring high flexibility and applicability. Accumulated mileage, average charging current, average charging temperature, and start and end of SOC values are chosen as influential factors for model establishment. The effectiveness, complexity, superiority, and robustness of the proposed method are verified using data from real-world EVs. Results indicate accurate SOH estimation can be achieved with an average absolute error of only 0.89 Ah, where the MAPE and RMSE of the test vehicles are 2.049% and 1.153%, respectively.

Certifying Black-Box Policies With Stability for Nonlinear Control

Machine-learned black-box policies are ubiquitous for nonlinear control problems. Meanwhile, crude model information is often available for these problems from, e.g., linear approximations of nonlinear dynamics. We study the problem of certifying a black-box control policy with stability using model-based advice for nonlinear control on a single trajectory. We first show a general negative result that a naive convex combination of a black-box policy and a linear model-based policy can lead to instability, even if the two policies are both stabilizing. We then propose an <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">adaptive <inline-formula><tex-math notation="LaTeX">$\lambda$</tex-math></inline-formula>-confident policy</i> , with a coefficient <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\lambda$</tex-math></inline-formula> indicating the confidence in a black-box policy, and prove its stability. With bounded nonlinearity, in addition, we show that the adaptive <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\lambda$</tex-math></inline-formula> -confident policy achieves a bounded competitive ratio when a black-box policy is near-optimal. Finally, we propose an online learning approach to implement the adaptive <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\lambda$</tex-math></inline-formula> -confident policy and verify its efficacy in case studies about the Cart-Pole problem and a real-world electric vehicle (EV) charging problem with covariate shift due to COVID-19.

An Experimental Comparison of Physics-based and Machine-Learning-based Electric Vehicle Energy Consumption Estimation Methods

Electric vehicles (EVs) have gained attention in recent years due to their environmental friendliness and higher fuel efficiency. However, EV users may have concerns about their driving range. To alleviate such an anxiety, a pre-trip estimation of EV energy consumption can be helpful. There are two main approaches to predicting EV energy consumption: traditional model-based methods that use physical knowledge, and data-driven techniques that rely on machine learning methods. Although both types of methods show promise, little attention has been paid to experimentally compare their performance differences. To bridge this gap, this paper presents an experimental comparison study of three model-based and data-driven EV energy consumption estimation algorithms. Notably, real-world EV road-test datasets from urban driving are used for the comparative evaluation. Furthermore, this study offers a discussion of the pros and cons of each method, providing a guideline for algorithm improvement and selection.

State of health estimation for lithium-ion batteries in real-world electric vehicles

State of health estimation for lithium-ion batteries in real-world electric vehicles

Charging demand prediction in Beijing based on real-world electric vehicle data

The accurate estimation and prediction of charging demand play an essential role in charging infrastructure planning, power grid laying and efficient operations. In this paper, three-month real-world travel and charging records of 25,489 electric passenger vehicles in Beijing are utilized to quantitatively study the travel patterns and charging behaviors of electric vehicles and support charging demand prediction. A “multi-level & multi-dimension” travel and charging characteristic parameters extraction framework is proposed with >60 electric vehicle behavior characteristic parameters extracted from time, spatial, and energy dimensions. Then a two-step GMM and K-means-based clustering model is proposed to cluster the electric vehicle users into six categories with different usage habits, attributes, and characteristics. A trip-chain-simulation-based charging demand prediction model is further established to generate a week-period continuous activities simulation for specific category vehicles, in which the interdependences of travel and charging characteristic parameters are considered to make the trip chain simulation closer to reality. The city charging demands are aggregated from individual charging prediction results at a spatial resolution of 0.46 km and time resolution of 15 min quantitatively; the spatiotemporal aggregation results show that the proposed model can realize high accuracy prediction for real-world charging demands. The results are expected to provide the foundation for research on charging infrastructure planning and the impact of charging behaviors on the grid load.

Accurate State of Charge Estimation for Real-World Battery Systems Using a Novel Grid Search and Cross Validated Optimised LSTM Neural Network

State of charge (SOC) is one of the most important parameters in battery management systems, and the accurate and stable estimation of battery SOC for real-world electric vehicles remains a great challenge. This paper proposes a long short-term memory network based on grid search and cross-validation optimisation to estimate the SOC of real-world battery systems. The real-world electric vehicle data are divided into parking charging, travel charging, and finish charging cases. Meanwhile, the parameters associated with the SOC estimation under each operating condition are extracted by the Pearson correlation analysis. Moreover, the hyperparameters of the long short-term memory network are optimised by grid search and cross-validation to improve the accuracy of the model estimation. Moreover, the gaussian noise algorithm is used for data augmentation to improve the accuracy and robustness of SOC estimation under the working conditions of the small dataset. The results indicate that the absolute error of SOC estimation is within 4% for the small dataset and within 2% for the large dataset. More importantly, the robustness and effectiveness of the proposed method are validated based on operational data from three different real-world electric vehicles, and the mean square error of SOC estimation does not exceed 0.006. This paper aims to provide guidance for the SOC estimation of real-world electric vehicles.

Fault Diagnosis Method for Lithium-Ion Battery Packs in Real-World Electric Vehicles Based on K-Means and the Fréchet Algorithm.

Battery failure has traditionally been a major concern for electric vehicle (EV) safety, and early fault diagnosis will reduce many EV safety accidents. However, the short-circuit signal is generally very weak, so it is still a challenge to achieve a timely warning of battery failure. In this paper, an initial microfault diagnosis method is proposed for the data of electric vehicles in actual operation. First, a robust locally weighted regression data smoothing method is proposed that can effectively remove noisy data and retain fault characteristics. Second, an ordinary-least-squares-based voltage potential feature extraction method is proposed, which can effectively capture the small fault features of battery cells and achieve early warning. Third, a reference cell selection method based on K-means clustering is proposed, which can effectively reduce the false alarms caused by the inconsistency of each cell. Fourth, the Fréchet algorithm is introduced into the field of battery pack fault diagnosis and combined with thresholds for battery pack fault diagnosis and localization to accomplish the diagnosis and early warning of minor faults. Finally, the fault diagnosis method is validated by three actual running electric vehicles to verify the effectiveness, reliability, and robustness of the method.

Big data-driven decoupling framework enabling quantitative assessments of electric vehicle performance degradation

Electric vehicle (EV) performance in terms of the available driving range per charge and the energy consumption rate continuously degrades during its service life. Quantitative assessments of EV performance degradation play an important role in EV residual value analysis, battery management, and battery recycling. However, EV performance degradation is highly sensitive to both ambient temperature and battery aging states; coupled factors make its quantification challenging. Here, a novel big data-driven decoupling framework is proposed to investigate the partial relationships between EV performance degradation and each individual variable (e.g., temperature and total driving distances). The core innovation involves the decoupling process that can enable real-world and large-scale degradation assessments. The basic functionality of the decoupling is achieved by an iterative learning framework where different machine learning-based models can communicate with each other. It achieves the advantages of unsupervised training and high performance; the mean absolute error can be controlled less than 0.1 in the model validation of EV ranges. Its effectiveness is verified using different real-world EV datasets. By utilizing the framework, the changes in the range and energy consumption of EVs across 10 urban areas in China are assessed. The results show that the range and energy consumption rate of EVs are more greatly influenced by ambient temperature than by battery aging. Less consideration of variable decoupling may yield misleading results in EV performance analysis. Our proposed framework opens avenues for quantifying EV performance degradation via real-world EV data, which is critical to onboard and cloud-based EV research.

Battery state of charge estimation using temporal convolutional network based on electric vehicles operating data

Accurate estimation of state of charge (SOC) is crucial for battery management system (BMS). Since most of the existing estimation methods are based on laboratory data, the accuracy of battery SOC estimation in real driving scenarios is difficult to be guaranteed. This paper proposes a method to estimate battery SOC using temporal convolutional network based on real-world operating data of electric vehicles (EVs). In view of the characteristics of large volume and high dimensionality of real-world EVs data, attention mechanisms are added to the time steps and input feature dimensions of the data, respectively, to further improve the model estimation accuracy. Meanwhile, this paper proposes a new method to correct the SOC value provided by the cloud platform through developing an improved capacity decay model applicable to the real-world EVs data. Then the corrected SOC is used as the label for neural network training. Experiments are conducted using one-year EVs data as the testing dataset, and the results show that the proposed model performances well, with root mean square error (RMSE) <0.67 % in spring and 0.99 % in winter. In comparison with other neural network models commonly used to predict battery SOC, the results show that the proposed model predicts more stable and accurate results.

Comparative Study of Incremental Capacity Curve Determination Methods for Lithium-Ion Batteries Considering the Real-World Situation

The incremental capacity analysis (ICA) method is a promising method in battery state of health (SOH) estimation studies. The incremental capacity (IC) curve determination is one of the critical parts of the ICA method. However, the uncertain and incomplete charging conditions of real-world electric vehicles (EVs) significantly limit the IC curve determination. This article provides a thorough analysis of the practicality and limitations of four IC curve determination methods based on the datasets collected from real-world operating EVs with a comprehensive comparison scheme. The Lorentz function fitting method is improved by the differential evolution algorithm, breaking the limitation of fixed parameter boundary constraints. A novel PCHIP method is further proposed to determine the IC curve, with higher robustness to realistic uncertain and incomplete charging conditions. The proposed method is validated by real-world data from 40 EVs with low sampling frequency. The results show the extracted features from the IC curves, determined by the proposed method have a stronger correlation with the SOH, allowing the accurate SOH estimation with a 2% error. With less computational resources and sampling frequency requirements, this method shows great potential for the realistic battery management system implementation.

Electric Vehicle Velocity and Energy Consumption Predictions Using Transformer and Markov-Chain Monte Carlo

Although electric vehicles (EVs) are becoming a promising alternative for ground transportation, the issue of limited battery energy hinders its market penetration. Accurate prediction of electrical energy consumption along a given route would significantly relieve drivers’ anxieties and build their confidence in EVs. A high-fidelity energy estimation relies on an accurate velocity forecast. To this end, this article proposes a novel hybrid deterministic-stochastic methodology that utilizes inputs encompassing the route information, the driver’s characteristics, and the traffic flow’s uncertainties to predict the EV’s future velocity profile and energy consumption. The method comprises two components: a deterministic machine-learning-based transformer network and a stochastic Markov-chain Monte Carlo (MCMC) algorithm. Real-world EV data collected in four routes of different lengths and features are used to evaluate the method. The results demonstrate the enhanced performance of the approach in both velocity prediction and energy consumption estimation, compared to two popular baseline algorithms.

An Inconsistency Fault Diagnosis Method for Lithium-Ion Cells in the Battery Pack Based on Piecewise Dimensionality Reduction and Outlier Identification

Abstract The inconsistency of cells in the battery pack is one of the main causes of battery failure. In practical applications, the terminal voltage is an important parameter that is easy to obtain and can characterize the inconsistency of cells. In this paper, a fault diagnosis method based on piecewise dimensionality reduction and outlier identification is proposed according to the voltage inconsistency of cells in the battery pack. This method uses a piecewise aggregate approximation (PAA) algorithm with a shift factor to reduce the dimension of the cell voltage time series, after which a deletion mechanism is designed based on the clustering algorithm and outlier identification to calculate the clustering quality after deleting each cell, reflecting the deviate degree of each cell. In addition, a safety management strategy is designed based on the Z-score method, and an abnormality coefficient is set to evaluate the inconsistency of cells. The effectiveness of the proposed diagnosis method is verified by monitoring the voltage data of two real-world electric vehicles. The verification results show that the method can not only detect the inconsistency before the failure of the faulty cell in the battery pack in advance, but also reduce the risk of computational explosion caused by the voltage time series and accurately locate the faulty cell.

Data-driven battery state of health estimation based on interval capacity for real-world electric vehicles

State of health (SOH) estimation is critical to the safety of battery systems in real-world electric vehicles. Accurate battery health status is difficult to be measured during dynamic and robust vehicular operation conditions. This paper proposes a novel SOH estimation model based on Catboost and interval capacity during the charging process. A year-long operation dataset of an electric taxi is derived with all charging segments separated to construct the research dataset. The charging patterns are analyzed, and the segments with rich aging information are extracted, then a general aging feature of interval capacity is extracted by incremental capacity analysis. Furthermore, comparison with the other six machine learning methods is conducted, and five inputs are determined through Pearson correlation analysis, including start charging state of charge (SOC), end charging SOC, mileage, temperature of probe, and current. The results show the Catboost-based model achieves the best accuracy, with the mean absolute percentage error and root mean squared error limited within 2.74% and 1.12%, respectively. More importantly, a battery aging evaluation strategy and its further research plan is proposed for the application in real-world electric vehicles.

Battery Thermal Runaway Fault Prognosis in Electric Vehicles Based on Abnormal Heat Generation and Deep Learning Algorithms

Efficient battery thermal runaway prognosis is of great importance for ensuring safe operation of electric vehicles (EVs). This presents formidable challenges under widely varied and ever-changing driving conditions in real-world vehicular operations. In this article, an enabling thermal runaway prognosis model based on abnormal heat generation (AHG) is proposed by combining the long short-term memory neural network (LSTM) and the convolutional neural network (CNN). The memory cell of the LSTM is modified and the resultant modified LSTM-CNN serves to provide accurate battery temperature prediction. The principal component analysis is used to optimize the model input factors to improve prediction accuracy and to reduce computing time. A random adjacent optimization method is employed to automatically optimize the hyperparameters. Finally, a model-based scheme is presented to achieve AHG-based thermal runaway prognosis. Real-world EV operating data are used to verify the effectiveness and robustness of the proposed scheme. The verification results indicate that the presented scheme exhibits accurate 48-time-step battery temperature prediction with a mean-relative-error of 0.28% and can realize 27-min-ahead thermal runaway prognosis.

EVStationSIM: An end-to-end platform to identify and interpret similar clustering patterns of EV charging stations across multiple time slices

Transport electrification introduces new opportunities in supporting sustainable mobility. Fostering Electric Vehicle (EV) adoption integrates vehicle range and infrastructure deployment concerns. An understanding of EV charging patterns is crucial for optimizing charging infrastructure placement and managing costs. Clustering EV charging events has been useful for ensuring service consistency and increasing EV adoption. However, clustering presents challenges for practitioners when first selecting the appropriate hyper-parameter combination for an algorithm and later when assessing the quality of clustering results. In a clustering process, the ground truth data is normally not available for practitioners to validate different modeling decisions. Consequently, it is difficult to judge the effectiveness of the discovered patterns because there is no objective method to compare them. This work proposes an end-to-end platform prototype named “EVStationSIM” that allows for the creation of relative rankings of similar clustering results. The ultimate goal is to support users/practitioners by allowing them to identify and interpret similar clustering patterns of EV charging stations using multiple time slices. The performance of this proposed platform is demonstrated with a case study using real-world EV charging event data from charging station operators in New Brunswick, Canada. The case study illustrates how generated results can assist in downstream analytical tasks such as planning infrastructure allocation expansions.

Investigation on overcharge-caused thermal runaway of lithium-ion batteries in real-world electric vehicles

Overcharge is one of the most typical triggers leading to battery thermal runaway during real-worldvehicular operation. An in-depth exploration of the real-scenario-oriented failure mechanism of batteryovercharge is critical to safeguard the safe and durable operation of electric vehicles. With environmentalchanges and drivers’ random behaviors under real occasions considered, this paper demonstrates the multi-factor overcharge-to-thermal-runaway modelling theory based on electro-chemical thermal coupling analysis. A case study of a real-world electric vehicle that experienced an overcharge-causedthermal runaway accident is presented, and the pre-accident battery overcharge characteristics are dissected referring to the accident investigation reports. Furthermore, the multi-stage pre-accidentcharacterization is performed by dividing overcharge into three feature stages, such as charge, fully-charged instant, and overcharge. Various characterization parameters directly measurable and indirectly calculated are extracted, then a variable evaluation of the state of health degradation underovercharge is implemented. Moreover, a rudimentary probability model of overcharge-caused thermalrunaway is proposed combining all potential motivating factors, including environments and drivers’behaviors. This is the first of this kind to apply the actual accident case on overcharge-causedthermal runaway research, which can guide the battery overcharge protection to safeguard drivers andpassengers.

State-of-Health Estimation for LiFePO4 Battery System on Real-World Electric Vehicles Considering Aging Stage

Comprehensive and accurate estimation methods of the state of health (SOH) of battery systems play a significant role in online monitoring for the safe and reliable operation of electric vehicles (EVs). Most existing estimation techniques are established on top of data from well-controlled experimental environments and tend to focus on a single health feature, thus not practical for real-world EVs’ SOH monitoring throughout the life cycle. To address this problem, based on real-world EV operation data, a novel SOH estimation model is presented for LiFePO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">4</sub> battery systems in EVs with consideration of degradation mechanisms. The ohmic resistance and the peak value of the incremental capacity (IC) curve are extracted from a large number of electric buses as health features. The influence of temperature on the ohmic resistance is eliminated by exponential fitting, which transforms ohmic resistance into the relative change rate of ohmic resistance. With joint considerations of the two health features and the mechanism of battery aging, three stages of SOH degradation can be quantitively described. The test results show that the proposed model is able to track SOH degradation from 0 to over 300 000 km and reflect the SOH of the battery more comprehensively in the early and late life cycles of the battery compared with single-health-feature methods.

A Data-Driven Method for Battery Charging Capacity Abnormality Diagnosis in Electric Vehicle Applications

Enabling charging capacity abnormality diagnosis is essential for ensuring battery operation safety in electric vehicle (EV) applications. In this article, a data-driven method is proposed for battery charging capacity diagnosis based on massive real-world EV operating data. Using the charging rate, temperature, state of charge, and accumulated driving mileage as the inputs, a tree-based prediction model is developed with a polynomial feature combination used for model training. A statistics-based method is then used to diagnose battery charging capacity abnormity by analyzing the error distribution of large sets of data. The proposed tree-based prediction model is compared with other state-of-the-art methods and is shown to have the highest prediction accuracy. The holistic diagnosis scheme is verified using unseen data.