Search

The key to battery management systems (BMS) is an accurate and real-time prediction on State of Charge (SOC) of the power battery. The methods of estimating SOC of power battery were analyzed. The gray neural network model was introduced into the field of SOC estimation. The model is used to establish the relationship of SOC, discharge current and rebound voltage. The model can be used to estimate SOC of battery on line by detecting the values of rebound voltage and discharge current during the discharge process. The model of SOC estimation is proved feasible by comparing the experimental data with calculated data. [ABSTRACT FROM AUTHOR]

The lithium ion battery is considered as the experimental object, and its discharge characteristics are studied. A model of the battery in different charge-states is established by a tool of neural network while battery's rebound voltage, temperature and load are set as input parameters. The validity of the model is tested based on the experimental data. The accuracy, adaptability and stability of the SOC in this model is validated in a variety of the working conditions, and the accuracy of the model is demonstrated to be higher than 5%. [ABSTRACT FROM AUTHOR]

To study the ignition characteristics of semiconductor bridges (SCB) for automobile airbag igniters, a model was proposed to compute the electrical characteristics of SCB based on the structural characteristics, resistance properties, and the working process of the bridge. Numerical simulation was conducted to analyze the voltage and current trends at two ends of bridges with various geometric dimensions and those with identical geometric dimensions and oxide layers of different thicknesses. Voltage-current curves for bridges of three different dimensions were obtained through numerical simulation. Bridges of larger geometric dimensions were found to require higher voltage to reach the melting point, which means higher energy input is required. [ABSTRACT FROM AUTHOR]