No abstract

With the research object of LiFePO 4 battery, this paper aims to correctly estimate the battery state of charge (SOC) by constructing a comprehensive SOC estimation strategy. Firstly, recursive least square (RLS) algorithm is adopted to realize online parameter identification of the equivalent battery model; and then an elaborate combination of RLS and Unscented Kalman Filter (UKF) is established, thus the battery model parameters used in UKF are actually obtained recursively by RLS; finally, SOC can be estimated by UKF. This strategy has an obvious adaptability due to the adoption of online parameter identification, so it is also called adaptive SOC estimation technique. Experimental results show that sometimes battery model parameters of different cells can be much different even though terminal voltages of these cells are very close or same when they are under resting state, and this inconsistency among LiFePO 4 batteries is captured by the RLS-UKF strategy presented in this paper; and of course battery SOC can also be correctly estimated by using the continuously updated model parameters.Keywords：LiFePO 4 battery; SOC estimation; online parameter identification; UKF IntroductionDuring recent several years, electric vehicle (EV) industry has been booming, and as the power source of the whole EV system, storage battery has drawn more and more attention.

Battery is the core equipment of the electrical vehicle (EV), battery life, and performance are critical to EV. Only motor variables are considered in most model predictive control (MPC) strategy in which battery indexes are neglected. A model predictive speed control (MPSC) strategy with adjustable C-rate of battery is proposed in this paper on EV with permanent magnet synchronous machine (PMSM). The rest capacity of battery is involved in the cost function, and the operating speed is related to the rest capacity of battery which adjusts the C-rate of the discharging currents further to protect battery. As the EV runs, the C-rate is decreased step by step according to the improved cost function, and the energy is saved to prolong mileage. The benefits of the proposed method including discharging capacities and operating mileage are corroborated by the simulation and repetitive experimental results under same conditions.

Summary This paper studies LiFePO4 (LFP) battery capacity fading diversity among different cells with same type and specification under same working states during their whole life cycle; and with consideration of this phenomenon, a novel battery state of health (SOH) estimation method with adaptability to capacity fading diversity is proposed. In order to cope with this capacity fading diversity, a machine learning structure involving a sparse auto‐encoder (SAE) and a backward propagation neural network (BPNN) is designed for battery SOH estimation. In this strategy, battery terminal voltage during the later stage of charging process is used as input of SAE; through the reconstruction of input signal, compressive feature of battery voltage is abstracted by SAE; then this compressive feature is used as the input signal of BPNN, and through nonlinear mapping of the neural network, battery SOH can be finally obtained. In this way, a relationship between the battery voltage information at its later charging stage and its SOH can be established. Verification tests show that this SAE‐BPNN based SOH estimation strategy possesses a good accuracy with adaptability to the capacity fading diversity and voltage differences among different battery cells, the SOH estimation error can be restrained within the range of ±5%, and it is also very convenient to adopt this method in real online battery management system (BMS).

Model predictive torque control with duty cycle control (MPTC-DCC) is widely used in motor drive systems because of its low torque ripple and good steady-state performance. However, the selection of the optimal voltage vector and the calculation of the duration are extremely dependent on the accuracy of the motor parameters. In view of this situation, A modified MPTC-DCC is proposed in this paper. According to the variation of error between the measured value and the predicted value, the motor parameters are calculated in real-time. Meanwhile, Model reference adaptive control (MRAC) is adopted in the speed loop to eliminate the disturbance caused by the ripple of real-time update parameters, through which the disturbance caused by parameter mismatch is suppressed effectively. The simulation and experiment are carried out on MATLAB / Simulink software and dSPACE experimental platform, which corroborate the principle analysis and the correctness of the method.