Spatiotemporal modeling and estimation of temperature distribution in lithium ion batteries

Abstract

Electric vehicles (EVs) and hybrid electric vehicles (HEVs) have received great attention in both academic and industrial areas in recent years. Lithium-ion batteries are becoming increasingly popular for energy sources of EVs and HEVs due to their high specific energy and energy density. The safety, life, and performance of lithium ion batteries are all related to its thermal performance. A battery thermal management system that keeps the battery to work within an optimal temperature range is crucial in EV/HEV application. The battery thermal process is a typical distributed parameter system (DPS) which is spatiotemporal distributed. The online estimation of temperature distribution in vehicle battery systems is not easy as only few surface temperatures can be measured for the distributed parameter process. An accurate mathematical model that can be updated online is needed for real-time monitoring of the temperature distribution especially that cannot be measured. Modelling and online estimation of the temperature distribution in vehicle battery systems is not easy due to the following challenges and difficulties: • The thermal process described in nonlinear partial differential equations (PDE) is time/space coupled. Extensive computation is usually needed due to its infinite dimensional nature, which makes it difficult for direct online application. • Very few sensors can be used for online measurement of the spatiotemporal temperature process due to practical limitations. • The thermal dynamics vary significantly in different regions, particularly between the center region and the boundary region. This strong spatial nonlinearity will make traditional modeling methods difficult. • The chemical related behaviors in batteries are very sensitive to the external disturbances and variation of operational conditions, the developed model should be updated online for real time estimation. There is still an open area for online model based temperature distribution estimation of lithium ion battery systems. The main objective of this thesis is to develop a systematic method for modeling and estimation of spatiotemporal temperature distribution in vehicle battery systems. The developed method could be easily added to the existing battery management system. The estimated temperature distribution can be used both for battery thermal management and the battery state prediction and the health monitoring. The thermal process of lithium ion batteries is a typical nonlinear DPS which is spatiotemporal coupled. A reduced order model which is suitable for online application is needed for practical applications. The traditional Karhunen–Loèvev (KL) decomposition method is a global linear projection and reconstruction method that may not be optimal for nonlinear process. A local properties embedding based time/space separation approach is first proposed for the nonlinear thermal process studied. By incorporating the underlying local properties of the data, the proposed method can be more effective to nonlinear DPS than the traditional KL method. Due to the complex electrochemical reactions inside the battery and its sensitivity to the working environment, the thermal dynamics vary significantly in different regions, particularly between the center region and the surface region. This strong spatial nonlinearity will make these traditional modeling methods difficult. It is difficult to represent these spatiotemporal dynamics with a single physics-based model. A hybrid model is developed for spatiotemporal estimation of temperature distribution in lithium ion batteries. A simple but effective nominal model is first developed for real-time thermal management using the time/space separation method developed in the previous part. Subsequently, a data-based neural model is proposed to compensate the model-plant mismatch caused by both the spatial nonlinearity and other model uncertainties. Due to overwhelming complexity of the electrochemical related behaviors and internal structure inside the batteries, it is hard to obtain the accurate mathematical expression of heat generation based on the physical principal. A data based thermal model is proposed for lithium-ion batteries, where the uncertain chemical behaviors related heat generation term are approximated by extreme learning machine (ELM) based neural network. Both the uncertain parameters in the reduced order heat transfer model and the ELM model can be determined analytically in a linear way. This simple training process makes it superior for onboard application. The identified offline model should be updated online to adapt to the external disturbances and variation of working environments with few measured temperatures. An optimized sensor location based online estimation scheme for temperature distribution in lithium ion batteries is presented. Based on the two models developed in the previously parts, the physics based nominal model and the ELM based model, an adaptive observer is designed respectively. The sensor locations are designed offline by considering the effects of the sensor placement and the observer gain matrix simultaneously. With the designed observer, the measured current, voltage, and few temperature values, the temperature distribution of the whole battery can be estimated in real time. Numerical simulations demonstrated the effectives of the developed online model based temperature distribution estimation methods. The developed method is simple enough for online application. Only commonly measured signals in a typical vehicle battery management system, i.e. current, working voltage and a few temperature data, are needed for the proposed model. The developed model can be easily implemented in a model predictive control algorithm for effective mitigation of thermal runaway and hence be helpful in improving the safety of the battery. It can also be easily integrated into existing battery management systems for estimation of other temperature dependent battery states, such as state of health (SOH), capacity, and others.

Date of Award	16 Feb 2015
Original language	English
Awarding Institution	City University of Hong Kong
Supervisor	Hanxiong LI (Supervisor)