Executive Summary : | Electric vehicles (EVs) have become popular means of transportation in recent years and for the safe and reliable performance of EVs, it is necessary to monitor and control the performance of its Lithium-ion battery (LIB) packs and prevent battery operation outside the safe operating area. This is done by the means of battery management systems (BMS). The BMS makes use of monitoring algorithms for measuring the important states including state of charge (SoC), state of health (SoH) and state of available power (SoAP), of the batteries and ensures their safe operation. The dynamics of the LIBs can be captured more efficiently by means of equivalent circuit models (ECM) with finite number of RC elements. Higher the number of RC elements, higher is the accuracy of the state estimation. But this often results into higher computational cost. To eliminate this problem, this project proposal proposes exploring reduced-order modeling of the ECMs and then using those reduced-order models for the state of charge estimation. In addition, the SoC versus the terminal voltage measurement of an LIB is nonlinear, due to which the ECM is complex. Often, the nonlinearity is also tackled with high-order linear models. Model order reduction (MOR) is a promising technique that reduces the size and complexity of large-scale mathematical models while preserving their main characteristics. Reduced yet accurate models of higher-order systems can help in faster verification, control and optimization processes. In this proposal, the large-order ECMs will be handled using nonlinear model-order reduction techniques, one of which could be based on trajectory piecewise linear (TPWL) approximation proposed in the literature. Benefits of other MOR techniques will also be investigated. A comparative study will be performed on the reduced order models of ECMs obtained by different MOR variants. The best possible approximation will be then used for the estimation of SoC of the LIBs. |