A Graph‐Based Lithium‐Ion Battery Parameter Estimation Approach to Produce Diverse Synthetic Data

Abstract

Energy dense lithium-ion batteries are extensively used in all portable electronic devices and in electric vehicles as well. State-of-charge estimation of these batteries has been of considerable commercial interest as this key metric can be construed as the available range in electric vehicles. State-of-charge is also important to ascertain the remaining usage time in battery powered devices. In this paper a graph neural network-based approach is employed to estimate key battery parameters such as, voltage, battery capacity, etc. To the best of the authors' knowledge, this is the first paper to employ a graph-based approach to improve battery estimates. The pairwise interdependencies within the battery dataset are exploited to provide better battery estimates. The graph-based approach is compared with related statistical methods to highlight the effectiveness of this approach.

State of Charge Estimation for Lithium‑Ion Batteries Based on Extended Kalman Particle Filter and Orthogonal Optimized Battery Model

“…where X k and Y k are the system state variable and the system observation output variable, respectively. U k−1 is input state vector; W k−1 is the process noise, V k is the measurement noise, A k−1 , B k−1 is the coefficient matrix of the state variable, C k , D k is the coefficient matrix of the output variable, as shown in Equation (10):…”

“…[6] Accurate state of charge (SOC) estimation is essential for the charge/discharge control and thermal management of BMS. [7,8] Currently, SOC estimation methods can be divided into three categories: direct measurement method, [9] data-driven method, [10] and model-driven method. [11,12] The Ampere hour DOI: 10.1002/adts.202301022 (AH) method estimates SOC by integrating the charge/discharge current, but it relies on accurate initial SOC information and is susceptible to error accumulation.…”

“…Currently, SOC estimation methods can be divided into three categories: direct measurement method, [ 9 ] data‐driven method, [ 10 ] and model‐driven method. [ 11,12 ] The Ampere hour (AH) method estimates SOC by integrating the charge/discharge current, but it relies on accurate initial SOC information and is susceptible to error accumulation.…”