| At present, it is paid more attention for the environmental and energy problems. The fuel oil vehicle is denounced by its high energy consumption and the high pollution. Electric vehicles have been developed for it's benefit of zero emission and low noise. Many countries have regarded electric vehicles as main aim of automobile industry. Energy management system is the key technology of electric vehicle. A good energy management system is based on knowing the exact state-of-charge (SOC). In this respect, estimating the state-of-charge is an essential element to the development of electric vehicles.This paper does research on lead-acid battery. First, describes the electrochemistry characteristic of lead-acid battery, point out the factors of affecting residual capacity(RC) and the difficulties of estimating stage-of-charge. Then, the most common techniques for estimating the SOC of electric vehicle's battery in the world are introduced. And the disadvantages of these methods are described. Based on these, this paper raise a new approach to estimate SOC based on the Electric quantity balanced model. A multi-layer LM neural network is structured to forecast the loss capacity on charge or discharge. A rapid learning algorithm that is realized model structure identification is also proposed. Using the neural network parallel computation, from the study characteristic, it enhances the accuracy on the forecasts. This method has accurate determination, and did not need too much data about battery.In addition, this paper describes the design of system's software and hardware. The platform is based on the C8051F005 MCU. Carries on certain precision to the accumulator cell operational parameter the survey, And it transmits through the serial port to the database in the personal computer, Thus obtains the parsing algorithm data. Some experiments was done by the system, and it gets some data .It establish the model of the battery by the data. And the model is validated.
|
PDF Full Text Request