Result: Online Impedance Identification using Pseudo-EIS Excitation on Self-Reconfigurable Batteries

Advanced energy storage devices known as Self-Reconfigurable Batteries (SRBs) have the ability to dynamically change their configuration in order to follow variable power and energy requirements. In contrast to conventional batteries, SRBs can instantly switch the connections between the cells in order to modify the electrical architecture, opening new possibilities for battery control and efficiency maximization. CEA, which is working on this technology since 2012, has started in 2020 a collaboration with Schaeffler (former Vitesco Technology) to study the adaption of this technology in an automotive solution. Considering the specific features, requirements and constraints of the application domain, the benefits and drawbacks in terms of performance and cost have been assessed. To ensure optimal performance and safety, it is crucial in a Battery Electric Vehicle (BEV) that the Battery Management System (BMS) delivers accurate internal state estimators, such as the State-of-Charge, State-of-Health or State-of-Power. In the current state of technology, these estimators rely on Electric Equivalent Circuit (ECM) models that are often defined in laboratory tests in the Beginning of Life (BoL) of the battery. To ensure that these estimators remain accurate throughout the life cycle of the BEV, a delicate updating task of the model’s parameters must be performed. One of the ways to identify these parameters is to realize an Electro-chemical Impedance Spectroscopy (EIS), which is an image of the cell complex impedance. The trend in the automotive industry is to have dedicated circuitry that can generate the excitation signal. But, since SRB technology has the capability to control the current in the cells in a discrete way (bypass or serial in the power path), it is possible for a given group of cells to define a switching pattern that forms a Pseudo Random Binary Sequence (PRBS). This signal mimics a sinusoidal excitation without dedicated circuitry and can do so during the normal operation of the battery.The study examines the quality of the identified models with respect to a known reference, and validates the approach on aged cell data to demonstrate that the proposed method offers accurate tracking of the battery model parameters throughout the degradation process.