SOH Prediction in Li-ion Battery Energy Storage System in Power Energy Network

The prediction of the State of Health (SOH) of Li-ion batteries is crucial for the system safety and stability of the entire energy network. In this paper, we analyse the role of Li-ion batteries as balancing batteries in the communication-energy-transportation network, which are key nodes for energy exchange. These batteries have different states of health and are constantly in a state of bi-directional energy transfer through charging and discharging. They also require coordinated charging and discharging in the network. Due to these differences, the degradation rates of the different balancing batteries vary, making it necessary to monitor the SOH of each battery in real time. To address the problem of numerous nodes and large computational requirements in the network, we propose a method based on the Transformer architecture for accurate and fast estimation, which can alleviate the communication pressure of the energy network layer. In this method, we select the voltage change rate as the only key health indicator and perform correlation analysis while the battery is in a constant current CC charging mode. According to the test results on NASA public battery data set, we continuously collect 3, 5, 10 and 20 voltage change rate values as inputs to the model and validate 1–4 layer transformer models. The model with 20 voltage change rate values as inputs and 4 layers has the best prediction performance, with a Mean Absolute Error (MAE) as low as 4.85%, while the Root Mean Square Error (RMSE) is 6.02%. In addition, the proposed method can process the time series data of multiple network nodes in real time and in parallel, with high prediction accuracy and stable performance, which makes it suitable for widespread use in power energy networks.