Research On The State Estimation And Management System Technology Of High-Power Li-ion Batteries

The energy consumption of the transportation field increases the global energy crisis and environmental pollution. Low-emission and energy-conservation electric vehicles have become the developing tendency of the energy transformation. With the advantages of high operating voltage, high energy density, long cycle life and non-pollution, the Li-ion batteries would become the mainstream pc wer source of future electric vehicle. How to improve the accuracy of state estimation and reliability of the battery management under complicated operating conditions have become key factors to improving the performance-price ratio of power battery pack and ensuring the safety of electric vehicles which are significant for the promotion of electric vehicles. Supported by the National High Technology Research and Development Program (2009AA11A113), my research focus on modeling and estimation of the state-of-charge (SOC); modelling and estimation of insulation between the battery pack and vehicle chassis; battery management system (BMS) scheme design. The main contents of this dissertation are listed as below:1. According to practical engineering problems, the robust estimation of SOC is discussed. To eliminate the disturbance of SOC estimation caused by the drift current noise in current sensors, a new working model of SOC estimation which takes the drift current as a state variable is proposed based on the combined model. Then, to suppress the parameter perturbations of this working model, the Unscented Particle Filter (UPF) method is applied to the synchronous estimation of both SOC and the drift current. This method effectively reduces the model errors introduced by the linearization process, and also enhances the suppression of the non-Gaussian noise. Robustness of this method is verified by comparing to the EKF and UKF methods on different perturbations of system parameters.2. Since the SOC-OCV curve is very flat on the steady-working platform of LiFePO4battery, an information fusion framework of SOC estimation based on the multi-source information fusion technology is proposed. This framework realizes the feature extraction and pattern classification of the charging and discharging process, and then optimizes the estimation models according to the specific pattern. The estimate model is switched to the best matched model to optimize the SOC estimation based on feature matching results at run time. By fusing the vehicle information and charge information based on the empirical knowledge and pattern classification, the key parameters of the battery system are estimated and adjusted online. Compared to the single estimation model, this method can increase the accuracy of the system state and parameter estimation, and also effectively restrains the jitter of SOC estimation on steady-working platforms. By utilizing the actual operation data, experiments and numerical simulations were conducted, and the superiority of this method is verified.3. To improve the accuracy of insulation detection between the high-power battery system and the chassis of electric vehicles under complicated operating conditions, a new insulation detection model which considers affections of battery internal resistances is built. Then a reliability algorithm (RA) based on this model is proposed. This algorithm realizes a reliability measurement of detecting dataset according to the change interval of the total voltage. Then a binary dataset which has the maximum reliability of calculating the system insulation resistance is selected. In order to restrain the random measurement noise, the moving average filter is employed. The experiments and numerical simulation results based on our test platform of electric vehicles are conducted to verify the superiority of the proposed method.4. A BMS scheme for high-power Li-ion batteries based on the Controller Area Network (CAN) bus communication technology is designed out of consideration of safety and reliability. The module voltage, temperature and other module information can be sampled by the battery management unit (BMU) of battery packs. The central control unit (CCU) carries out estimation, evaluation and decision missions based on information acquired through the internal CAN bus of BMS. Intelligent control and power optimization are realized by combining the vehicle CAN bus, the charge CAN bus and CCU. The software based on the real-time embedded operating system UCOS-II ensures the response performance of each interrupt and task, and also speeds up the progress of the development and debugging. This scheme provides superior scalability, configurability and portability. It has been commercially applied to more than a dozen of EV types.