Multi-scale Spatiotemporal Modeling and Intelligent Control for BMS

Description

An effective temperature management system is crucial to the optimal operation of the batteryand its health monitoring. Estimation of the internal temperature distribution of the cell is verydifficult due to its multi-scale spatiotemporal dynamics and no sensing within the cell. Effectivetemperature management of the battery pack is also difficult due to lack of direct spatial controlfunction.This project aims to design a multi-scale spatiotemporal modelling method for internaltemperature distribution of the cell/pack and a spatiotemporal model based management scheme.This intelligent management scheme could be easily added to the existing battery managementsystem. A multi-phase approach is innovatively proposed to integrate modeling and controlthrough real-time experiment, physical simulation, and analytical modeling.‧First, the temperature distribution in the cell will be modelled in multi-scale from particlereaction to heat generation. An approximate model based iterative optimization is proposedin first time to iteratively estimate the model parameters using experimental data.‧Second, the analytical model of the cell temperature distribution can be constructed under thespace-time separation, and trained with data from the physical simulation. The deep learningis proposed in first time for sensor reduction. With the help of sensors in the pack, theanalytical model of the cell will be updated during the real-time operation and graduallybecomes a feasible operational model for state prediction and health monitoring of the battery.‧Finally, with the peltier plates added in proper positions and the air flow control mechanism,a spatiotemporal fuzzy control scheme will be designed for optimal management of thetemperature field in the battery pack.