Capacity State-of-Health Estimation of Electric Vehicle Batteries Using Machine Learning and Impedance Measurements

Abstract

With the increasing adoption of electric vehicles (EVs) by the general public, a lot of research is being conducted in Li-ion battery-related topics, where state-of-health (SoH) estimation has a prominent role. Accurate knowledge of this parameter is essential for efficient and safe EV operation. In this work, machine-learning techniques are applied to estimate this parameter in EV applications and in diverse scenarios. After thoroughly analysing cell ageing in different storage conditions, a novel approach based on impedance data is developed for SoH estimation. A fully-connected feed-forward neural network (FC-FNN) is employed to estimate the battery’s maximum available capacity from a small set of impedance measurements. The method was tested for estimation in long-term scenarios and for diverse degradation procedures with data from real EV batteries. High accuracy was obtained in all situations, with a mean absolute error as low as 0.9%. Thus, the proposed algorithm constitutes a powerful and viable solution for fast and accurate SoH estimation in real-world battery management systems.

Artificial Intelligence Approaches for Advanced Battery Management System in Electric Vehicle Applications: A Statistical Analysis towards Future Research Opportunities

“…High accuracy of SOH prediction was obtained under varying conditions with coefficients of determination between 0.896 and 0.992. Barragán-Moreno et al [88] applied the BPNN model to predict the maximum capacity of EV batteries. The performance of the proposed model was verified using diverse degradation data from real EV batteries.…”

Cloud and Edge Computing for Smart Management of Power Electronic Converter Fleets: A Key Connective Fabric to Enable the Green Transition

“…Despite the progressive battery price reduction, they still represent a major part of the cost in stationary storage applications and EVs. Moreover, performance degradation under heavy usage is a considerable disadvantage [34]. Particularly, precise information about the current health status of batteries and predictions about their remaining useful lifetimes are a key to optimal battery asset management-but a still unresolved scientific challenge.…”

“…Our works bring improvements to the current landscape [34], [39], [40]. First, even though our test subjects are stationary batteries, our state-of-health inference approaches are designed so that they can be applied to either mobile or stationary systems, using both test-and operational-based data.…”

“…2) In contrast, EV batteries work as isolated systems, and therefore, asset coordination is not a major issue nowadays. In this case, edge/ cloud symbiosis can enhance battery performance by providing state-of-health estimations that do not require inconvenient time-consuming tests [34]. In particular, for any given batterybased application, we can reduce data throughput from megabytes to kilobytes (see the "Battery Degradation Analysis and Forecast" section) by implementing load collectives and damage models locally.…”

Skin layer classification by feedforward neural network in bioelectrical impedance spectroscopy

“…The studies of tissue disorders based on BIS and MLA illustrate the ability of MLA skin layer classification. For impedance measurements, Feedforward Neural Network (FNN) is appropriate because it can model complex relationships between inputs and outputs, making it a powerful tool for high-accuracy classification tasks that can be performed quickly with less computational power and can use parallel computation to increase the algorithm’s efficiency [ 14 ]. The skin layer classification based on impedance measurement, which focuses specifically on predicting source indicator o k , is a preliminary step to skin dielectric characteristics diagnosis.…”