Abstract:Li-ion batteries (LIBs) are becoming ubiquitous in portable, stationary, and electrical energy storage applications that power various devices, from cell phones to the zero-emission plane. Market research suggests the $26.9 billion-a-year (2018) worldwide LIB market could reach $107 billion-a-year by 2025. Taking electric vehicles (EVs) for example, according to the Electric Drive Transportation Association (EDTA), the number of EV sales in the United States market has increased from 345 vehicles in 2010 to 331,617 in 2019, with a total cumulative number of 1.44 million vehicles over the ten-year sales period. This trend has also been observed in the global market as EVs' sales reached 2.1 million worldwide in 2019, boosting the stock to 7.2 million. Although the COVID-19 pandemic dramatically affected the global EV market and prompted the market to plummet over the year 2020 relative to 2019. The global EVs stock projections are expected to increase from 7.2 million to 14 million in 2025 and 25 million in 2030, accounting for 10% of global passenger vehicle sales in 2025 and 28% in 2030. Despite rapidly skidding into the EV era - a technical fait accompli- significant obstacles to vehicle electrification and adoption, especially the safety of their batteries, continue to threaten that future. The risk of anomaly generation and evolution (e.g., thermal runaway and exploration) is currently a major safety concern and challenge in applications powered by LIBs. The traditional first-principle simulations and comprehensive experiments are time-consuming or less effective in detecting and predicting the above problems induced by a synergistic effect of multiple degradation mechanisms. The above challenge eventually necessitates predictive or prognostic capability for Prognostics and Health Monitoring (PHM) under a complexly hostile working environment.This proposal will break away from existing practices, and instead, for the first time, the proposed data-driven prognosis (DDP) requires the assumption of the existence of a conservation principle but does not require a priori knowledge of the key mechanisms and boundary conditions. Alternatively, the exact form of the conservation functional is only specified locally, in the neighborhood of each observation point (and is assumed to be piecewise quadratic in the current work), whereas its global form remains unknown. The conservation principle is then defined as the minimization of system curvature at each observation point. The DDP methodology is based on the local curvature of the mathematical manifold representing the Euler-Lagrange governing equation of a system. The curvature represents a derivative of the total energy with respect to sensed variables, and instability is anticipated whenever the curvature exceeds a critical threshold, denoted by the inverse of the local length-scale of the manifold (or its extrinsic curvature). The length-scale, in turn, is calculated by solving the first moment of the state-space description (conservation of linear momentum) based on two consecutive time instants of measurements. In another sense, the DDP technique can estimate a mass normalized Energy Exchange Amplitude (EEA) that also represents a scaled “entropy rise” of the system. After acquiring the input data from the literatures, they were fed into DDP in a continuous order separated by each time step. At each time step, two consecutive time instances were considered that are sufficient to make the prognostication. The results of the simulation using the DDP are shown below. First the progression of PDI which shows the percentage of points that have reached different path dependence index (PDI) markers as shown in Figure 1 (a). It signifies the instants when the battery shows local instabilities. As shown in the figure, battery starts to show local instability from time instant 20 and beyond, highlighting the instability's progression. The chain length (CL) values and residual curvature information were analyzed as well and shown in Figure 1 (b) along with other criterias. As illustrated in the Figure, at time instant 10, the system crosses the critical chain length threshold and intensifies around time instance 20. Furthermore, the residual curvature (RC) value exceeds the critical threshold at time instance 19 and 21. The moment when all three criterias were met, the battery is expected to fail which is time instance 21 as shown in the Figure. Figure 1

Data-Driven Prognosis of Multiscale and Multiphysics Complex System Anomalies: Its Application to Lithium-ion Batteries Failure Detection

A Physics-Informed Hybrid Data-Driven Approach with Generative Electrode-Level Features for Lithium-Ion Battery Health Prognostics

A Physics-Informed Hybrid Multitask Learning for Lithium-Ion Battery Full-Life Aging Estimation at Early Lifetime

Battery safety: Machine learning-based prognostics

Battery prognostics and health management from a machine learning perspective

Physics-based Prognostics of Lithium-Ion Battery Using Non-Linear Least Squares with Dynamic Bounds

The State of Lithium-Ion Battery Health Prognostics in the CPS Era

Developing an online data-driven approach for prognostics and health management of lithium-ion batteries

Hybrid physics-based and data-driven modeling with calibrated uncertainty for lithium-ion battery degradation diagnosis and prognosis

An Integrated Probabilistic Approach to Lithium-Ion Battery Remaining Useful Life Estimation

Specialized deep neural networks for battery health prognostics: Opportunities and challenges

A review on prognostics and health management (PHM) methods of lithium-ion batteries

Probabilistic Machine Learning for Battery Health Diagnostics and Prognostics—review and Perspectives

Insights and reviews on battery lifetime prediction from research to practice

Lithium-Ion Battery Prognostics through Reinforcement Learning Based on Entropy Measures

A Migration-Based Method for Non-Invasive Revelation of Microscopic Degradation Mechanisms and Health Prognosis of Lithium-Ion Batteries

Battery Lifetime Prognostics

Data-driven health estimation and lifetime prediction of lithium-ion batteries: A review

(Digital Presentation) Data-Driven Prognosis of Lithium-Ion Batteries Thermal Runaway Early Warning and Detection

A Hybrid Prognostic Approach for Remaining Useful Life Prediction of Lithium-Ion Batteries

Lithium battery prognostics and health management for electric vehicle application – A perspective review