A Novel Enhanced SOC Estimation Method for Lithium-Ion Battery Cells Using Cluster-Based Lstm Models and Centroid Proximity Selection

TY - UNPB

T1 - A Novel Enhanced SOC Estimation Method for Lithium-Ion Battery Cells Using Cluster-Based Lstm Models and Centroid Proximity Selection

AU - Al-Alawi, Mohammed

AU - Jaddoa, Ali

AU - Cugley, James

AU - Hassanin, Hany

PY - 2024/3/13

Y1 - 2024/3/13

N2 - In line with the global mission in achieving the net zero target through deployment of renewable energy technologies and electrifying the transportation sector; precise and adaptable State of Charge (SOC) estimation for Lithium-ion batteries has emerged as a critical need. This study introduces an innovative SOC estimation method for Lithium-ion batteries, featuring a novel Cluster-Based Learning Model (CBLM) that integrates K-Means and Fuzzy C-Means clustering with the predictive capabilities of Long Short-Term Memory (LSTM) networks. A key innovation of this research is the application of the CBLM to SOC estimation, paired with a dynamic SOC estimation process enabled by a centroid proximity selection mechanism. This approach significantly enhances the precision and adaptability of SOC predictions under varying operational conditions. Comprehensive evaluations demonstrate the model's superior performance, with reductions in Root Mean Square Error (RMSE) to as low as 0.65% and Mean Absolute Error (MAE) to 0.51%, reducing state-of-art benchmark model errors by margins of 61.8% and 68.5% respectively. Additionally, the maximum error exhibited a notable decrease, emphasizing the model's reliability in worst-case-scenarios. These findings set new benchmarks in battery management systems and expands the potential of machine learning applications in energy storage, offering an advanced tool for real-time SOC estimation with enhanced accuracy and reliability.

AB - In line with the global mission in achieving the net zero target through deployment of renewable energy technologies and electrifying the transportation sector; precise and adaptable State of Charge (SOC) estimation for Lithium-ion batteries has emerged as a critical need. This study introduces an innovative SOC estimation method for Lithium-ion batteries, featuring a novel Cluster-Based Learning Model (CBLM) that integrates K-Means and Fuzzy C-Means clustering with the predictive capabilities of Long Short-Term Memory (LSTM) networks. A key innovation of this research is the application of the CBLM to SOC estimation, paired with a dynamic SOC estimation process enabled by a centroid proximity selection mechanism. This approach significantly enhances the precision and adaptability of SOC predictions under varying operational conditions. Comprehensive evaluations demonstrate the model's superior performance, with reductions in Root Mean Square Error (RMSE) to as low as 0.65% and Mean Absolute Error (MAE) to 0.51%, reducing state-of-art benchmark model errors by margins of 61.8% and 68.5% respectively. Additionally, the maximum error exhibited a notable decrease, emphasizing the model's reliability in worst-case-scenarios. These findings set new benchmarks in battery management systems and expands the potential of machine learning applications in energy storage, offering an advanced tool for real-time SOC estimation with enhanced accuracy and reliability.

KW - State of Charge (SOC) Estimation

KW - Lithium-Ion Batteries

KW - LSTM

KW - Battery Management Systems (BMS)

KW - Dynamic SOC Estimation

KW - Cluster-Based Learning Model

U2 - 10.2139/ssrn.4758431

DO - 10.2139/ssrn.4758431

M3 - Preprint

T3 - EST-D-24-02589

BT - A Novel Enhanced SOC Estimation Method for Lithium-Ion Battery Cells Using Cluster-Based Lstm Models and Centroid Proximity Selection

ER -