Battery Lifetime Prediction Research Articles

Machine Learning for Battery Quality Classification and Lifetime Prediction Using Formation Data

Machine Learning for Battery Quality Classification and Lifetime Prediction Using Formation Data

Battery Lifetime Prediction and Degradation Analysis with Machine Learning Method

Due to environmental pollution and climate crisis caused by fossil fuels, eco-friendly alternative energy sources are gaining attention. Particularly, electric vehicles (EVs) are being recognized as the next-generation means of transportation. The most crucial technologies in electric vehicles are lithium-ion batteries and Battery Management Systems (BMS). However, the extended lifespan of lithium-ion batteries leads to prolonged assessment periods, making it challenging to measure and improve performance over time. Additionally, the operation of batteries under various conditions makes it difficult to measure state of charge (SOC) and state of health (SOH) in real-time, introducing many variables. With recent advancements in predictive technology through machine learning, accurate predictions can be made without complex physical knowledge, leveraging high-quality data.In this study, machine learning techniques were applied using charge/discharge data from lithium-ion LiNi0.5Mn0.3Co0.2O2 (NMC532)/graphite pouch cell batteries. The values measured during the initial charge/discharge cycles were converted into variables that could assess battery degradation modes. Machine learning algorithms were then applied, followed by model tuning to reduce overfitting and enhance accuracy. This research contributes to real-time battery management system data using machine learning technology to predict the remaining useful life (RUL) of lithium-ion batteries with high accuracy, using only initial cycles. This approach enhances the efficiency of BMS data in real-world scenarios and contributes to the overall time efficiency of battery research. Acknowledgement This research was supported by National R&D Program through the National Research Foundation of Korea (NRF) funded by Ministry of Science and ICT(NRF-2021K1A4A8A01079455).

Towards Rational Design of Battery Formation Protocols: From Electrochemical Modeling to Factory Deployment

In battery manufacturing, the formation process is both paramount and problematic. It is paramount since all batteries undergo formation charging and aging steps to build a resilient solid electrolyte interphase (SEI) and to screen for defects. It is problematic because the formation process is expensive to operate, remains a major source of factory energy demand, requires larger factory footprints, and takes an order of magnitude longer than nearly every other manufacturing step. Despite the centrality of the formation process in battery manufacturing, steps taken to optimizing formation protocols remain ad-hoc in the absence of design principles and physical models.We highlight recent progress in developing a principled understanding of the SEI formation process applied towards industrial battery formation process (i.e. in the context of a full cell). We review a reduced-order electrochemical model of the formation process that captures first cycle charge dynamics and subsequent cycling and aging degradation behavior [1]. The model predicts measurable quantities such as first cycle efficiency (FCE) and irreversible thickness growth under multiple formation protocols. We discuss opportunities to integrate the electrochemical formation model with differential voltage analysis (DVA) to visualize the connection between first cycle charge dynamics from the perspective of shifting electrode stoichiometries [2] and comment on applications towards physics-informed battery lifetime prediction [3]. We finally share recent data elucidating the effect of formation temperature, pressure, and protocol on cycle life.

Attention Mechanism-Based Neural Network for Prediction of Battery Cycle Life in the Presence of Missing Data

As batteries become widespread applications across various domains, the prediction of battery cycle life has attracted increasing attention. However, the intricate internal mechanisms of batteries pose challenges to achieving accurate battery lifetime prediction, and the inherent patterns within temporal data from battery experiments are often elusive. Meanwhile, the commonality of missing data in real-world battery usage further complicates accurate lifetime prediction. To address these issues, this article develops a self-attention-based neural network (NN) to precisely forecast battery cycle life, leveraging an attention mechanism that proficiently manages time-series data without the need for recurrent frameworks and adeptly handles the data-missing scenarios. Furthermore, a two-stage training approach is adopted, where certain network hyperparameters are fine-tuned in a sequential manner to enhance training efficacy. The results show that the proposed self-attention-based NN approach not only achieves superior predictive precision compared with the benchmarks including Elastic Net and CNN-LSTM but also maintains resilience against missing-data scenarios, ensuring reliable battery lifetime predictions. This work highlights the superior performance of utilizing attention mechanism for battery cycle life prognostics.

Early battery lifetime prediction based on statistical health features and box-cox transformation

Early battery lifetime prediction is important for both safety reasons and battery development. It predicts battery lifetime before it degrades significantly and has the advantage of being low-cost, time-saving, and providing timely feedback. However, due to the nonlinear degradation behavior and limited data in the early stage, early battery lifetime prediction is difficult. In this paper, statistical health features are extracted from temperature and voltage data, and a statistical transformation method is proposed to enhance the feature's importance and help increase the prediction accuracy of battery lifetime. First, candidate health features are constructed by analyzing temperature and voltage data, and Box-Cox transformation (BCT) is used to enhance their linear correlation with battery lifetime to help with feature selection. The Pearson correlation coefficients of extracted temperature and voltage features with battery lifetime reach −0.73 and − 0.90 respectively. Early battery lifetime prediction based on the first 100 cycles of data is then conducted using Gaussian process regression with a root-mean-square-error (RMSE) of 112.14 cycles. Comparison experiments show that Gaussian process regression outperforms the other four common machine learning methods in both BCT and non-BCT situations. Moreover, adding temperature feature and BCT have a positive effect. When they are not used, RMSEs would increase by 13.8 % and at most 209 %, respectively. At last, prediction based on fewer data (50 cycles) is tested and gained acceptable result of an RMSE of 149.68 cycles.

Investigating explainable transfer learning for battery lifetime prediction under state transitions

Battery lifetime prediction at early cycles is crucial for researchers and manufacturers to examine product quality and promote technology development. Machine learning has been widely utilized to construct data-driven solutions for high-accuracy predictions. However, the internal mechanisms of batteries are sensitive to many factors, such as charging/discharging protocols, manufacturing/storage conditions, and usage patterns. These factors will induce state transitions, thereby decreasing the prediction accuracy of data-driven approaches. Transfer learning is a promising technique that overcomes this difficulty and achieves accurate predictions by jointly utilizing information from various sources. Hence, we develop two transfer learning methods, Bayesian Model Fusion and Weighted Orthogonal Matching Pursuit, to strategically combine prior knowledge with limited information from the target dataset to achieve superior prediction performance. From our results, our transfer learning methods reduce root-mean-squared error by 41% through adapting to the target domain. Furthermore, the transfer learning strategies identify the variations of impactful features across different sets of batteries and therefore disentangle the battery degradation mechanisms and the root cause of state transitions from the perspective of data mining. These findings suggest that the transfer learning strategies proposed in our work are capable of acquiring knowledge across multiple data sources for solving specialized issues.

Attention towards chemistry agnostic and explainable battery lifetime prediction

Predicting and monitoring battery life early and across chemistries is a significant challenge due to the plethora of degradation paths, form factors, and electrochemical testing protocols. Existing models typically translate poorly across different electrode, electrolyte, and additive materials, mostly require a fixed number of cycles, and are limited to a single discharge protocol. Here, an attention-based recurrent algorithm for neural analysis (ARCANA) architecture is developed and trained on an ultra-large, proprietary dataset from BASF and a large Li-ion dataset gathered from literature across the globe. ARCANA generalizes well across this diverse set of chemistries, electrolyte formulations, battery designs, and cycling protocols and thus allows for an extraction of data-driven knowledge of the degradation mechanisms. The model’s adaptability is further demonstrated through fine-tuning on Na-ion batteries. ARCANA advances the frontier of large-scale time series models in analytical chemistry beyond textual data and holds the potential to significantly accelerate discovery-oriented battery research endeavors.

Semi-supervised learning for explainable few-shot battery lifetime prediction

Semi-supervised learning for explainable few-shot battery lifetime prediction

Status, challenges, and promises of data‐driven battery lifetime prediction under cyber‐physical system context

AbstractEnergy storage is playing an increasingly important role in the modern world as sustainability is becoming a critical issue. Within this domain, rechargeable battery is gaining significant popularity as it has been adopted to serve as the power supplier in a broad range of application scenarios, such as cyber‐physical system (CPS), due to multiple advantages. On the other hand, battery inspection and management solutions have been constructed based on the CPS architecture in order to guarantee the quality, reliability and safety of rechargeable batteries. In specific, lifetime prediction is extensively studied in recent research as it can help assess the quality and health status to facilitate the manufacturing and maintenance. Due to the aforementioned importance, the authors aim to conduct a comprehensive survey on the data‐driven techniques for battery lifetime prediction, including their current status, challenges and promises. In contrast to existing literature, the battery lifetime prediction methods are studied under CPS context in this survey. Hence, the authors focus on the algorithms for lifetime prediction as well as the engineering frameworks that enable the data acquisition and deployment of prediction models in CPS systems. Through this survey, the authors intend to investigate both academic and practical values in the domain of battery lifetime prediction to benefit both researchers and practitioners.

Robust battery lifetime prediction with noisy measurements via total-least-squares regression

—Machine learning technologies have gained significant popularity in rechargeable battery research in recent years, and have been extensively adopted to construct data-driven solutions to tackle multiple challenges for energy storage in embedded computing systems. An important application in this area is the machine learning-based battery lifetime prediction, which formulates regression models to estimate the remaining lifetimes of batteries given the measurement data collected from the testing process. Due to the non-idealities in practical operations, these measurements are usually impacted by various types of interference, thereby involving noise on both input variables and regression labels. Therefore, existing works that focus solely on minimizing the regression error on the labels cannot adequately adapt to the practical scenarios with noisy variables. To address this issue, this study adopts total least squares (TLS) to construct a regression model that achieves superior regression accuracy by simultaneously optimizing the estimation of both variables and labels. Furthermore, due to the expensive cost for collecting battery cycling data, the number of labeled data samples used for predictive modeling is often limited. It, in turn, can easily lead to overfitting, especially for TLS, which has a relatively larger set of problem unknowns to solve. To tackle this difficulty, the TLS method is investigated conjoined with stepwise feature selection in this work. Our numerical experiments based on public datasets for commercial Lithium-Ion batteries demonstrate that the proposed method can effectively reduce the modeling error by up to 11.95 %, compared against the classic baselines with consideration of noisy measurements.

Data-driven rapid lifetime prediction method for lithium-ion batteries under diverse fast charging protocols

Accurate lithium-ion battery lifetime prediction is essential for equipment maintenance and safety assurance in practical applications. The influence of different charge protocols on the lifetime varies greatly, which raises tremendous challenges for rapid lifetime prediction. In this paper, a comprehensive data-driven rapid lifetime prediction method for batteries under diverse fast charging protocols is proposed. It straightforwardly establishes the relationship between the charging conditions and the impacts on battery lifetime. A deep neural network model with the Bayesian optimization algorithm for hyperparameter search (BOA-DNN) and three traditional shallow machine learning algorithms consisting of support vector machine (SVM), Gaussian process regression (GPR), and extreme gradient boosting (XGBoost) are employed. The verification is performed on an authoritative and popular dataset with 69 types of distinct fast charging conditions. The eventual results indicate that the proposed approach has attained reliable performance of battery lifetime degradation trajectory and battery end-of-life (EOL) prediction, in which the BOA-DNN model outperforms SVM, GPR, and XGBoost. Especially in battery EOL prediction, the prediction mean absolute percentage error of the BOA-DNN model is below 3.58%.

Early prediction of battery lifetime based on graphical features and convolutional neural networks

Accurate lifetime prediction of lithium-ion batteries in the early cycles is critical for timely failure warning and effective quality grading. Convolutional neural network (CNN), with excellent performance in feature extraction, has gained increasingly attentions in battery prognostics. However, since degradation test normally takes years to complete, employing end-to-end CNNs directly for battery lifetime prediction is impractical due to the limited number of available training samples and the scarcity of features in the early cycles. Instead of directly feeding the raw data, in this work, we propose to use graphical features for early lifetime prediction. Three feature curves, including capacity-voltage curve, incremental capacity curve, and capacity difference curve are used to construct graphical features. Specifically, the incremental capacity curve and capacity difference curve are derived from capacity-voltage curve, aiming to extract more information from both intra-cycle and inter-cycle perspectives. The evolution patterns of these feature curves over the initial 100 cycles show evident correlations with battery lifetime, and are termed as the graphical features. The three graphical features, after some proper transformation, are stacked into a three-channel image before feeding to the CNN model. Five classical CNNs, with different structures and key parameters, are investigated for battery lifetime prediction. Comparative experiments are conducted to study the influence of different feature combinations, voltage segments, and discharge cycles on the prediction performance. Experimental results demonstrate that simple CNNs with only a few convolutional layers can achieve satisfying prediction results. Additionally, networks with rectified linear unit are shown to outperform those with other activation functions.

Accelerated Prediction and Analysis of Lithium-Ion Battery Lifetime Using Efficient Electrochemical Models

Lithium-ion batteries degrade over their lifetime due to several parasitic reactions that occur in the cell. These reactions lead to loss of lithium inventory, loss of active material, mechanical and thermal damage to the cell’s component. Predicting cell aging and capacity fade is crucial for battery management systems (BMS) to monitor and control the battery's state of health. Physics-based models are valuable in battery management systems (BMS) to monitor the battery’s state of health and predict battery failure. BMS algorithms require fast codes that can predict and estimate battery parameters in real-time and control the battery’s performance under different loads.1 Common reduced-order models, such as the SPM (Single Particle Model) consider approximations that may not be valid as the cell ages. The Tanks-in-Series model is a model developed by Subramaniam et al., where the governing equations from the full-order ‘pseudo-2-D’ (p2D) model are volume-averaged over each cell region.2 This ensures that the physics captured by the mass and charge conservation equations are maintained without loss of accuracy while providing simulation solutions at a few milliseconds. The Tanks-in-Series model is ideal for predicting cell aging due to its reduced computation speed and predictability beyond other reduced-order models.This work combines the Tanks-in-Series model with governing equations for several degradation mechanisms, including the growth of the SEI layer, lithium plating, particle cracking, etc. A comprehensive analysis of capacity fade predictions is shown, with comparisons to predictions from the P2D model. References V. Ramadesigan et al., J. Electrochem. Soc., 159, R31–R45 (2012).A. Subramaniam et al., J. Electrochem. Soc., 167, 013534 (2020).

Broad Transfer Learning Network based Li-ion battery lifetime prediction model

The Broad Transfer Learning Network (BTLN) model uses only one-third of the parameters as the common neural network does to achieve the similar prediction performance that a Multi-Layer Perceptron (MLP) model has. This brings considerable benefits to AI-related studies that highly rely on computational power. This study also proposed a feature mapping technique that applied a linear transformation to the original features to improve the performance of various learning models. This study combines both broad learning and transfer learning techniques. Data augmentation is used to expand the training dataset. It proves that, under certain conditions, the model with a broader network can perform better. The broad network structure can act as an effective feature extractor. The transform learning algorithm can increase the training efficiency due to decreased trainable parameters. The performance improvement of neural network models is particularly remarkable. The performance improves by 3.5% without any change to the model architecture. In this paper, the performance of the BTLN model proposed based on Root Mean Square Error (RMSE) improves by 18.5%. Compared to common neural network models, the training parameters and overall parameters are lowered to 14.57% and 36.28%, respectively.

Lithium-Sulfur battery lifetime prediction, a reliability approach

The remaining useful life (RUL) phenomenon of lithium sulfur batteries is characterized by being a nonlinear phenomenon that implies difficulties to determine it; performance tests to evaluate and predict the RUL involve long times and a large sample data that, sometimes, is limited for an adequate characterization due to insufficient data. The present work addresses the problem through a reliability analysis considering a highly censored reduced sample, which allows characterizing is reliability key indicators and RUL through the Weibull distribution and the reduced bias adjustment (RBA) method, considering uncertainty in the estimation. The results show that the method is capable to predict RUL of lithium-sulfur batteries, as well as the usefulness of reliability indices.

Semi-Supervised Self-Learning-Based Lifetime Prediction for Batteries

Accurate and reliable degradation and lifetime prediction for lithium-ion batteries is the main challenge for smart prognostic and health management. This paper proposes a novel semi-supervised self-learning method for battery lifetime prediction. Firstly, three health indicators are extracted from the partial capacity-voltage curve. Secondly, the capacity estimation model and lifetime prediction model are built using data from three randomly selected batteries in the source domain. Then, the health indicators are used to reconstruct the historical capacities to provide pseudo values for self-training of the lifetime model. Finally, the self-trained lifetime model is used to predict future degradation. The uncertainty expression is also included to provide the probabilistic prediction of future capacities. Different application scenarios are considered in the verification. The mean lifetime prediction error is less than 23 cycles with only three known checkpoints for batteries aging under different profiles. Predictions for different battery types show that the errors are less than 50 cycles with relative errors less than 4.1% for long lifespan batteries, and less than 20 cycles with relative errors less than 5.21% for short lifespan batteries. This paper guides proper solutions for lifetime prediction when the labeled capacities in the real world are limited.

Learning model for battery lifetime prediction of LoRa sensors in additive manufacturing

Today, an innovative leap for wireless sensor networks, leading to the realization of novel and intelligent industrial measurement systems, is represented by the requirements arising from the Industry 4.0 and Industrial Internet of Things (IIoT) paradigms. In fact, unprecedented challenges to measurement capabilities are being faced, with the ever-increasing need to collect reliable yet accurate data from mobile, battery-powered nodes over potentially large areas. Therefore, optimizing energy consumption and predicting battery life are key issues that need to be accurately addressed in such IoT-based measurement systems. This is the case for the additive manufacturing application considered in this work, where smart battery-powered sensors embedded in manufactured artifacts need to reliably transmit their measured data to better control production and final use, despite being physically inaccessible. A Low Power Wide Area Network (LPWAN), and in particular LoRaWAN (Long Range WAN), represents a promising solution to ensure sensor connectivity in the aforementioned scenario, being optimized to minimize energy consumption while guaranteeing long-range operation and low- cost deployment. In the presented application, LoRa equipped sensors are embedded in artifacts to monitor a set of meaningful parameters throughout their lifetime. In this context, once the sensors are embedded, they are inaccessible, and their only power source is the originally installed battery. Therefore, in this paper, the battery lifetime prediction and estimation problems are thoroughly investigated. For this purpose, an innovative model based on an Artificial Neural Network (ANN) is proposed, developed starting from the discharge curve of lithium-thionyl chloride batteries used in the additive manufacturing application. The results of experimental campaigns carried out on real sensors were compared with those of the model and used to tune it appropriately. The results obtained are encouraging and pave the way for interesting future developments.

Comparative Analysis of Battery Cycle Life Early Prediction Using Machine Learning Pipeline

Lithium-Ion battery system is one of the most critical but expensive components for both electric vehicles and stationary energy storage applications. In this regard, accurate and reliable early prediction of battery lifetime is important for optimizing life cycle management of batteries from cradle to grave. In particular, accurate aging diagnostics and prognostics is crucial for ensuring longevity, performance, safety, uptime, productivity, and profitability over a battery's lifetime. However, current state-of-art methods do not provide satisfactory prediction performance (lack of uncertainty quantification) using early degradation data. In the present work, to produce the best model for both battery cycle life point prediction and range prediction (i.e., confidence intervals or prediction intervals), a pipeline-based approach is proposed, in which a full 33-feature set is generated manually based on battery degradation knowledge, and then used to learn the best model among five machine learning (ML) models that have been reported in the battery lifetime prediction literature, and two quantile regression models for battery cycle life prediction. The calibration and sharpness property of battery cycle life range prediction is properly evaluated by their coverage probability and width respectively. The experimental results show that the gradient boosting regression tree model provides the best point prediction performance, while the quantile regression forest model provides the best range prediction performance with both full 33-feature set and the MIT 6-feature set.

Battery aging mode identification across NMC compositions and designs using machine learning

•A machine-learning framework for aging mode identification in Li-ion batteries •Aging modes are classified using cycle-by-cycle electrochemical features •Battery design, chemistry, and use condition determine predominant aging modes •Understanding aging phenomena creates opportunities for building more robust batteries A comprehensive understanding of lithium-ion battery (LiB) lifespan is the key to designing durable batteries and optimizing use protocols. Although battery lifetime prediction methods are flourishing, diagnosis of the root causes of aging and degradation have not yet been well developed nor studied for a broad mixture of designs and use cases. Here, we create a machine-learning (ML)-based framework that distinguishes aging modes using multiple electrochemical signatures recorded cycle-by-cycle. The predominant aging behaviors include a combination of loss of active materials in cathode (LAMPE) and a loss of Li inventory (LLI) in Li plating or solid electrolyte interphase (SEI) formation, manifested from 44 batteries representing two cathode chemistries, two electrode loadings, and five charging rates. The aging mode classification accuracy is 86% using features within the first 50 cycles and increases to 88% beyond 225 cycles. The same features can quantify the percentage of end-of-life LAMPE with only 4.3% of error. A comprehensive understanding of lithium-ion battery (LiB) lifespan is the key to designing durable batteries and optimizing use protocols. Although battery lifetime prediction methods are flourishing, diagnosis of the root causes of aging and degradation have not yet been well developed nor studied for a broad mixture of designs and use cases. Here, we create a machine-learning (ML)-based framework that distinguishes aging modes using multiple electrochemical signatures recorded cycle-by-cycle. The predominant aging behaviors include a combination of loss of active materials in cathode (LAMPE) and a loss of Li inventory (LLI) in Li plating or solid electrolyte interphase (SEI) formation, manifested from 44 batteries representing two cathode chemistries, two electrode loadings, and five charging rates. The aging mode classification accuracy is 86% using features within the first 50 cycles and increases to 88% beyond 225 cycles. The same features can quantify the percentage of end-of-life LAMPE with only 4.3% of error.

Battery Lifetime Prediction via Neural Networks with Discharge Capacity and State of Health

The market share of electric vehicles (EVs) has grown exponentially in recent years to reduce air pollution and greenhouse gas emissions. The principal part of an EV is the energy storage system, which is usually the batteries. Thus, the accurate estimation of the remaining useful life (RUL) of the batteries, for an optimal health management and a decision-making policy, still remains a challenge for automakers. In this paper, the problem of battery RUL prediction is studied from a new perspective. Unlike other estimation strategies existing in the literature, the proposed technique uses an intelligent prediction of the lifespan of lithium–iron–phosphate (LFP) batteries via a modified version of neural networks. It uses a data-driven life estimation approach and optimization method and does not require any prior comprehension and initialization of any parameters of the battery model. To validate and verify the proposed technique, we use LFP battery data sets, and the experimental results showed that the proposed methodology can well learn the characteristic relationship of battery discharge capacities as well as its state of health (SOH), where the battery life cycle changes as the battery ages with time and cycles.