Research On Fault Diagnosis Of Battery Sensors For Functional Safety

With the progress of the clean and efficient automobile industry worldwide,electric vehicles are gradually replacing traditional fuel vehicles in the field of road transportation.As a key component of electric vehicles,power battery packs work reliably and efficiently without the support of a Battery Management System(BMS).The BMS relies on the collected current,voltage,and temperature sensor signals to ensure that the battery works in the best state through identification and control strategies.If the sensor of the power battery pack fails,it may directly cause abnormal working conditions such as overcharging,over-discharging,and over-temperature of the battery,or even cause disasters.At the same time,the electronic devices included in modern automobiles have increased exponentially,which has led to a significant increase in the possibility of damage caused by electronic device failures.The promulgation of the second edition of the functional safety standard ISO 26262-2018,aimed at the functional safety of automotive electronic and electrical equipment,is urgently to be met by major automotive OEMs and component manufacturers.Therefore,the BMS must monitor the failure status of relevant sensors in real time in order to take corresponding safety measures in time to ensure that electric vehicles meet the requirements of ISO 26262 functional safety standards.Based on the ISO26262 functional safety standard,this paper conducted a systematic hazard analysis and risk assessment of the BMS,as well as the design of the safety mechanism,and then adopted a model-based fault diagnosis method to design relevant fault detection and isolation for the random hardware failure problem of the power battery pack sensor.Strategy.Mainly completed the following tasks:1.For ISO 26262 functional safety workflow,define related items of BMS,describe functions and boundaries;use Hazard Analysis and Risk Assessment(HARA)method to classify battery sensor functions For hazards caused by faults,determine the Automotive Safety Integration Level(ASIL)and safety goals,and derive functional safety requirements from the safety goals.Applying Fault Tree Analysis(FTA),Failure Mode and Effects Analysis(FMEA)to analyze BMS,and avoiding the risks caused by the unreasonable initial design of BMS;To solve the problem of random hardware failure of battery sensors,a model-based Troubleshooting methods.2.Discuss the principle of battery sensor fault diagnosis,compare the three methods of model-based fault diagnosis,analyze the diagnosis structure and suitable fields.Establish the first-order equivalent circuit model and lumped thermodynamic model of the battery,and realize the SIMULINK model of the battery electro-thermal including the thermodynamic characteristics;propose the space expression of different fault states of the battery sensor,and inject the battery model;apply the residual evaluation method based on the L2 norm to improve the fault Detection accuracy.3.For the needs of battery sensor diagnosis,based on the control and observation related theories,the fault diagnosis strategy based on Luenberger observer and sliding mode observer is used to achieve state space expression.For the non-linearity of the battery model,choose the battery typical working conditions.Ten operating conditions were linearized,and the results of linear and non-linear models were compared to verify the effectiveness of the linearized model.The sensor fault diagnosis strategy is designed based on the linearized battery model.The battery output(voltage,temperature)sensor residual generator is designed using the method based on Luenberger observer.The battery input is designed using the sliding mode observer method(current)sensor residual generator.In the system model,simulations are injected into faults,residuals are generated and evaluated,and faults are indicated.The results verify the effectiveness of the proposed random hardware failure diagnosis strategy for battery input and output sensors.