Estimation Of Dynamic Battery SOC Using Fuzzy Arithmetic And The Realization On DSP

The development of the automotive industry has brought more and more convenience to people's everyday life, but with the appearance of harmonious development between human being and the nature, the designers concern not only the improvement of the performance but also the pressure that the vehicles bring to the environment and HEV appears just at this time. The energy of HEV is supplied by both the engine and the dynamic battery, these two parts works independently and can not be substituted by each other. Although HEV does not fully get rid of fuel engine it fetches up the shortage of the EV that the battery can not carry the expectant efficiency. In recently years the HVE industry has got a high speed development.Now electrochemistry battery still is one most important component of multienergy drivetrain no matter in parallel connection or in series drivetrain structure. HEV is like the common EV when it applies the brake: traction motor works in generating mode and the kinetic energy is transported to the battery by the traction motor. In complex multiple drivetrain structure the battery works as the main power supplier. So for the multiple powertrain structure and system control it is very important to know the dynamic characteristic of the battery.Now in the worldwide the research on HEV battery SOC is mainly divided into to different kinds: one is to research from the chemical character, the other is from the working parameter of the battery. The second does not care about the internal chemical activity, looking the battery as a dark box, but estimates the SOC by testing the outer feature while the battery works (voltage, current, temperature, etc).This method can avoid researching on the complicated chemical activities inside the battery and the outer feature is easy to measure. Many features are concerned with dynamic battery SOC, such as open circuit current, voltage, in circuit current, voltage, temperature, times it has been used an etc, so when using outer feature method to estimate SOC there are also many ways to adopt.By researching on the working features of lead acid battery it can be concluded that when the SOC is different the in circuit current and voltage will be different relevantly. It is to say that the in circuit current and voltage that measured while the battery is working can reflect the SOC. So the current SOC can be estimated if the relation between SOC and in circuit current and voltage. But from the analysis it can be learn that the relationbetween in circuit voltage, current and SOC is not linear, so it is hard to find a fit mathematic model to describe it. In addition, the discharging condition of dynamic battery that is used on HEV is different from common battery, mainly high pulse current with short time continuous little current. Because the running condition is different, the need for the energy is different. Because of these reasons the estimating method should have superior adaption capability: operation quantity be little, time delay be not to long and a suitable veracity. Considering all these factors fuzzy arithmetic is adopted.The basic researching course is to do enough experiments, depend on the fuzzy logic theory, simulate human fuzzy thinking method and finally carry out SOC estimation.Fuzzy estimating method can do real-time tracking of dynamic battery SOC, and need no precise complicated mathematic model. It needs only the data information that comes from the sensors and the professional information that comes from performance description system.The inputs of this system are in circuit current and voltage and the model is built based on two-in one-out Mamdani arithmetic. The discuss field and the function type is decided according to the experiment result. The voltage is divided into 9 fields and the current is divided into 6 fields, and the function type is the mixture of S type, Z type andΛtype. Then according to the relation between SOC and current and voltage in each discuss field, 54 rules between inputs and outputs is confirmed.The hardware of this system has 3 parts: data collection module, central data processing module and data transmission module. Data collection module takes charge in analog signal to digital signal transforming and sending the data to DSP. Central data processing module is the hard core of the system. It processes the useful data according to the arithmetic that designed in advance and sends the results to upper controller. The upper controller sends discharging and generating signal to the charger according to the data it gets from the central data processing unit. This system uses CAN bus to communicate because it has superior anti-jamming capability and in long distance communicating it has a good speed.According to the experiment it can be concluded that the SOC fuzzy estimating system in this thesis can estimate SOC in different status. Also it can do the test work online and does not affect the normal work of the battery. The error is in the allowable scope.This thesis concerns arithmetic and hardware. Because of the limitation of self competence, there may be some flaw of the design, hope to get your comment.