Battery Management System Design And Parameters Estimation Method Research

The exacerbation of energy crisis and environmental pollution make the research of electric vehicles an important direction of auto research. The focus should be put on enhancing the function and cycle life of the power source of electric vehicles─the battery pack. Therefore, the key point of ensuring the security and effective operation of electric vehicles lies in building a comprehensive battery management system to realize a real-time detection of the battery information and achieve a dynamic management of the charging and discharging process of the battery. To address the above mentioned problems,this article provides a designation of electric vehicle power management system based on the embedded real-time operating system. The system is able to detect the voltage, current and temperature of the battery and apply algorithms which derivated from the Kalman filtering algorithm to estimate the SOC, SOH on the basis of nonlinear state space model. Then the system can effectively control the charging and discharging process of the battery pack and guarantee its safety and long-lasting effective operation. The main research contents of this article are listed below:1. We have designed a battery management system based upon the distributed wireless sensor networkThe voltage, temperature and other key information of the battery modules located in different parts of the vehicle can be detected through the use of the network. After the detection, the system will automatically transmit the acquired information to the central Management unit, CMU, which are based on UC/OSII embedded operating system. CMU achieves in the functions of data collection, parameters estimation, real-time controlling, real-time displaying and is enabled with the ability of reasonably adjusting and dealing with different situations. The distributed wireless sensor network enhanced the expandability of the system and provided simple configuration process. The UC/OSII real-time management system promised a concurrent processing of each real-time task. Many BMS products derived from this system have been practically used on several commercial vehicles.2. SOC estimation of the lithium battery (a) To solve the problem of multi-string battery pack SOC estimation, the system has built a Vmin state-space model on the ground of single battery model after fully considering the bucket effects of in-series battery pack, and has extended the kalman filtering algorithm on the basis of the Vmin state-space module. Thus realized a recursive SOC estimation of the battery pack. The validation of the reliability of the data algorithms under different working conditions and comparison between conventional algorithms showed that the Vmin-EKF can give an accurate estimation of the SOC and exhibit a strong inhabitation of the initial state error and the current input noise.(b) Lithium battery model input variables─the constant current value drift make it a difficult job to give an accurate SOC estimation. To eliminate the above mentioned problem, we can treat the current input drift as the parameters of the model, and use the joint-ekf algorithm to actualize the synchronous estimation of both the SOC and the current input noise. Compared with similar algorithms, this algorithm has the advantage of giving better estimation results using less complicated process. It's very suitable to be used in the embedded battery management system.3. SOH estimation of the lithium batteryAfter analyzing the factors influencing the SOH of the battery and defining the two state variables that reflects the SOH─the internal resistance of the battery and the battery rated capacity, a reduction Dual-EKF algorithm is proposed for state variables estimation. This way of calculation breaks the coupling solution process and calculates each part respectively. The approach reduces the complexity of the algorithm, but also has a high precision, the simulation results demonstrate these advantages.