Study On Security Control Of Electric Bus Power Battery System On Collision Conditions

With increasing popularity and deployment of new energy technologies, the number ofpure-electric bus has seen a dramatic in the public ground transport system. In the course oftrial operation, the exposed security issues have gained tremendous attention. Due to theinsulation failure or internal short circuit within the high-voltage battery pack, the vehiclegets fire causing serious harm to passengers and drivers on collision. Particularly, the sidecollision for the pure-electric bus would probably destroy the battery compartment locatedon the front panel, such as battery management systems and other electronic componentsdue to the deformation of the battery compartment door frame intrusion. As a result, it maybe possible to amass the cell excessive internal stress and incur single battery short circuitand even the electrolyte leakage. Therefore, the study on collision safety control system forhigh-voltage power battery is of great significance for improving the safety as well aspromoting mass adoption of pure-electric buses.Based on the analysis of the existing algorithms, a novel breaking control strategy forhigh-voltage battery pack is proposed in this thesis, while the relevant breaking controlsystem is developed. The primary goal of the study is to improve the security ofhigh-voltage battery pack on side collision by some intelligent identification algorithm.More specifics are shown as follows:Firstly, by interpreting the main tasks of the breaking control system, the underliningprinciple and constituent components, breaking schedule for single-box power battery packand battery power system are studied so as to determine the breaking strategies for thepower battery system, which lays firm foundation for the practical implementation ofcollision safety control system.Secondly, the acceleration signal is acquired by finite element model for the electricbus. At the same time, we utilize the wavelet de-noising method to extract features aboutnoise signal and simulate collision acceleration curve by computer. The acquisition ofsignal analysis identifies the occurrence of the collision and the severity of the collision. BPneural network is employed to predict the initial velocity of the collision, and then the state of the collision is further divided into three levels, i.e., mild collision, moderate collisionand severe collision. The severity of the collision is judged by fuzzy neural network, andthen the breaking controller decides whether or not to trigger the breaking signal.Finally, using Freescale microcontroller as the core IC, the embedded collision safetycontrol system for power battery system is built. According to the collision recognitionalgorithm and break control strategy, the software design is completed. Meanwhile, the dataacquisition test, execution devices’ response time test and collision simulation test arecompleted, and the data acquisition and storage capacity of control system, theeffectiveness real-time and reliability of collision identification are verified.