基于模型-数据融合方法的锂离子电池储能系统惯性支撑持续功率边界在线估计

Abstract Lithium-ion battery energy storage system (BESS) demonstrates great potential to provide inertia support to the power grid. The balance between the efficient inertia support and secure operation of battery is challenging, which requires accurate estimation of battery output boundary, especially in online working conditions. However, the existing methods for assessing the output power boundary of battery usually ignore the special inertia-supporting output profile and the requirement for online application, limiting the accuracy and efficiency. This paper proposes a novel online estimation method of inertia-supporting sustaining power boundary (SPB) of BESS based on model-data fusion method (MDFM). First, a series of experiments are conducted to investigate the impedance characteristics of battery under inertia-supporting condition, based on which a negative resistor-based equivalent circuit model (ECM) is developed to involve the nonlinear solid-phase diffusion effects of battery. Recognizing the nonlinear impact of state of charge (SOC) and discharge current rate on the negative impedance, a support vector machine (SVM) is applied to model the negative impedance, where the experimental results are input as the training data. Then, an MDFM-based method is proposed for online parameter estimation of the improved ECM, where the negative impedance is estimated by the SVM in real-time. Based on the ECM, the inertia-supporting SPB of BESS, constrained by the cut-off voltage, SOC and maximum current thresholds , is estimated online by a multi-constraint-based method. Finally, experiments are conducted to validate the MDFM-based ECM estimation method and the multi-constraint-based online SPB estimation method. Compared to conventional peak power estimation methods, the proposed method significantly improves the accuracy of BESS's output boundary assessment in an online manner.