EP-4741846-A1 - METHOD AND SYSTEM FOR PARAMETER IDENTIFICATION IN ELECTROCHEMICAL MODELS OF BATTERIES

EP4741846A1EP 4741846 A1EP4741846 A1EP 4741846A1EP-4741846-A1

Abstract

A method and system for identifying parameters of an electrochemical model for characterization and/or operational management of batteries. Battery performance data is collected under various conditions, applying an electrochemical model to simulate battery chemistry and degradation. The model parameters are segmented into subsets related to specific conditions and aging effects. An optimization framework then estimates each subset of parameters by resolving individual optimization problems based on the performance data. These estimated parameters are used to refine the electrochemical model, which is subsequently employed for further characterization and management of batteries.

Inventors

COUTO MENDONCA, Luis Daniel
MULDER, Grietus
PONNETTE, Rafaël Frank L.

Assignees

Vito NV

Dates

Publication Date: 20260513
Application Date: 20241106

Claims (15)

A method for identifying parameters of an electrochemical model for characterization and/or operational management of batteries, the method comprising: obtaining measurement performance data of a battery under various battery operating conditions; employing an electrochemical model to simulate battery cell electrochemistry and degradation mechanisms; dividing model parameters of the electrochemical battery model into multiple subsets related to specific battery operating conditions and aging phenomena; utilizing an optimization-based framework that is configured to estimate each subset of model parameters by solving separate optimization problems using measurement performance data, wherein each optimization problem estimates a subset of parameters relevant under the specific operating conditions corresponding to the measurement performance data; adjusting the electrochemical battery model with the estimated parameters; and using the updated electrochemical battery model for characterization and/or operational management of batteries.
The method according to claim 1, wherein the model parameters are divided into subsets corresponding to different stages and conditions of battery use and aging, and each subset is estimated using measurement data relevant to that mechanism.
The method according to claim 1 or 2, wherein a multi-step identification procedure is performed including: identifying a set of initial parameters of a new battery cell, the initial parameters including equilibrium parameters that characterize the battery's behavior under pseudo-equilibrium-conditions and dynamic parameters that characterize the battery's behavior under dynamic operational conditions; utilizing the identified initial parameters as a basis for modeling battery aging, wherein the modeling includes: incorporating calendar aging parameters into the model, which account for degradation effects due to the passage of time; and incorporating cycle aging parameters into the model, which account for degradation effects due to operational use; wherein the initial parameters serve as foundational configurations for the battery model, and adjustments are made to these configurations based on the calendar aging parameters and cycle aging parameters to model and/or predict the battery's degradation over time.
The method according to any one of the preceding claims, wherein the parameter identification comprises: determining, using low-current and/or open-circuit potential measurement performance data, a low-current or open-circuit potential, OCP, function of the battery electrodes to establish baseline equilibrium parameters; estimating, using dynamic measurement performance data including charging and discharging at various rates, dynamic parameters that indicate the battery's response under operational conditions; estimating, using data from aging tests where batteries are held at specific states of charge without cycling, calendar aging parameters that indicate calendar aging; estimating, using data from measurements in which the battery is subjected to repeated charge-discharge cycles under controlled conditions, cycle aging parameters that indicate cycle aging; applying an optimization algorithm to refine all identified parameters ensuring the model predictions align closely with the observed data; and combining all identified parameters into a comprehensive battery model for predicting the battery's behavior over its expected lifespan under various usage scenarios.
The method according to claim 4, wherein OCP functions per electrode are determined by measuring voltage response of the battery electrodes under low-current or zero-current conditions and fitting the data to obtain functional representations of the OCPs, wherein the OCP functions are determined by fitting measured OCP data to a set of basis functions selected from exponentials and hyperbolic functions.
The method according to claim 3, 4 or 5, wherein estimating equilibrium parameters involves solving a nonlinear least squares optimization problem that minimize the difference between measured and modeled voltage under low-current conditions.
The method according to any one of the preceding claims 3-6, wherein estimating dynamic parameters involves solving an optimization problem that minimize the difference between measured and modeled voltage during charge/discharge cycles at different current rates, to determine solid-phase diffusion time coefficients, reaction rate constants for positive and negative electrodes, and/or a lumped resistance parameter.
The method according to any one of the preceding claims 4-7, wherein estimating calendar aging parameters involves modeling capacity fade over time under storage conditions and solving optimization problems to fit the model to the measured capacity loss data, in order to determine at least parameters associated with solid electrolyte interphase (SEI) layer growth.
The method according to claim 8, wherein the calendar aging parameters further include parameters that are function of the state of charge (SOC) so as to account for SOC-dependent degradation.
The method according to any one of the preceding claims 4-9, wherein estimating cycle aging parameters involves modeling capacity fade due to cycling-induced degradation mechanisms and solving optimization problems to fit the model to the measured capacity loss data during cycling, to determine parameters associated with at least surface cracking propagation in the electrode materials of the battery.
The method according to any preceding claims, wherein the battery model is a physics-based electrochemical model, comprising: a single-particle model (SPM) for simulating the electrochemical interactions within the battery under various operational states; an extension of the single-particle model to include degradation mechanisms represented by solid-electrolyte interphase (SEI) layer growth for calendar aging and surface cracking propagation for cycle aging, whereby: the SEI layer growth is modeled to reflect degradation due to chemical and physical changes occurring over time as a result of environmental conditions and state of charge; the surface cracking propagation is modeled to account for mechanical degradation effects resulting from the cyclic stress and strain experienced during operational use of the battery; and wherein the integrated SEI layer growth and surface cracking propagation mechanisms are adjusted based on the initial equilibrium and dynamic parameters.
The method according to any one of the preceding claims, wherein discretization of the model equations is performed using spectral methods.
A system for identifying parameters of an electrochemical model that includes aging mechanisms for characterization and/or operational management of batteries, the system comprising: one or more processors; memory storing instructions that, when executed by the one or more processors, perform the method according to claim 1-12.
A method for operating a lithium-ion battery within a defined safe operating area, determined based on the electrochemical model parameters identified according to the method according to any one of the preceding claims 1-12.
A system for parameter identification in an electrochemical model of batteries, the system comprising: a battery testing apparatus configured to simulate operational conditions and collect measurement performance data under various battery operating conditions; at least one processing unit that is configured to perform the method according to any one of the preceding claims 1-12, using the measurement performance data collected by the battery testing apparatus, to identify parameters of an electrochemical model that includes aging mechanisms for characterization and/or operational management of batteries.

Description

FIELD OF THE INVENTION The invention relates to a method and system for identifying parameters of an electrochemical model for characterization and/or operational management of batteries. The invention also relates to a method for operating a lithium-ion battery within a defined safe operating area, determined based on the electrochemical model parameters identified. The invention also encompasses the field of systems and methods for collecting and processing battery performance data under a variety of operating conditions to facilitate the accurate modeling and management of batteries Furthermore, the invention relates to a system for parameter identification in an electrochemical model of batteries, the system comprising: a battery testing apparatus configured to simulate operational conditions and collect measurement performance data under various battery operating conditions so as to collect data for identifying parameters of an electrochemical model of the battery being tested. Furthermore, the invention relates to methods and systems for analyzing and managing the performance and operational characteristics of batteries. Further, the invention is also concerned with various aspects of battery management, including parameter identification in battery models, battery characterization, operational management, and age-related degradation modeling. Additionally, the invention relates to a computer program product. BACKGROUND TO THE INVENTION Battery technology is widely used for various applications. For instance, lithium-ion batteries are increasingly utilized across a wide range of modern applications, from portable consumer electronics and electric vehicles, to large-scale energy storage systems. These batteries are favored for their high energy density, long life cycle, and relative stability. Yet, despite these advantages, the effective management and maintenance of batteries, such as for instance lithium-ion batteries, remain challenging due to the complexity and variability in their aging processes as well as slight variations in parameters amongst manufactured cells. A primary issue within the current state of the art is the inefficiency in predicting the longevity and performance deterioration of lithium-ion batteries under different use conditions over time. Degradation of these batteries not only diminishes their capacity but also affects other performance metrics, impacting safety and reliability considerations for end-users. The degradation in lithium-ion batteries stems from a multitude of interconnected physical and chemical processes, making it difficult to predict and manage. Sources of degradation include but are not limited to, the growth of the solid-electrolyte interphase (SEI) layer, electrode delamination, corrosion of current collectors, and other phenomena associated with the loss of lithium inventory and active material. Different batteries may experience these degradation mechanisms to varying extents depending on their specific chemical compositions, physical cell architecture and operational usage patterns. Several models exist to predict the life-time and deterioration of batteries. Existing prediction models can be broadly classified into empirical or semi-empirical models, physics-based models, equivalent circuit models, and data-driven models. While physics-based models offer detailed insights into the internal processes of a battery, they tend to be computationally demanding. Empirical and semi-empirical models, although faster and simpler, often lack the depth of understanding needed to accurately predict battery life under varying conditions. Data-driven models require substantial data and are prone to inaccuracies when used outside the specific conditions under which they were developed. Current technologies often struggle with flexibility and require extensive recalibration as operational or environmental conditions change. For instance, many models do not seamlessly account for the complexity of interdependent degradation mechanisms. This inability to accurately account for, predict, and manage battery degradation under multifarious conditions leads to significant inefficiencies. Batteries may be prematurely discarded or replaced, incurring unnecessary costs and environmental waste, or alternatively, they may be used beyond their safe operational limits, posing safety risks. Hence, there is a clear and present need for an improved and more practical approach in the modeling and prediction of battery degradation mechanisms that comprehensively address the nuanced interplay of aging factors under varying operational conditions. Such an approach would ideally reduce predictive errors, extend battery life, reduce costs, and increase the safety and efficiency of battery-dependent technologies. Unfortunately, the inclusion of equations in the battery model for various degradation mechanisms under multifarious conditions poses a big challenge with respect to the validation o