Research On Zero Drift Compensation Technology Of EPS Torque Sensor

Torque sensors play a crucial role in EPS(Electric Power Steering),indispensable in industry and automotive industry.Torque sensors are divided into contact and non-contact torque sensors,contact torque sensors work primarily through conductive slip rings,it is essential that wear and tear,causes zero drift of the torque sensor.Once a zero drift occurs,the sensor output voltage is not accurate,reduce sensor performance,in order to find out the reason of zero drift of torque sensor better,the structure and characteristics of potentiometer torque sensor are analyzed,it is concluded that fretting wear is the cause of zero shift.Avoid the effects of zero drift on the course of the car,software compensation is used to compensate the zero drift of the potentiometer torque sensor,to improve the performance of the sensor indicators.The neural network has strong learning ability and better solving nonlinear approximation problems,the BP neural network is used to compensate the zero drift of the potentiometer torque sensor.The potentiometer type torque sensor is tested on the total performance test platform of the tubular column,the wear test of potentiometer type torque sensor is carried out on the platform of wear test,compare the performance of the potentiometer torque sensors before and after wear,it shows that potentiometer type torque sensor has zero drift phenomenon.The main and sub voltage of the potentiometer before and after abrasion are collected,As input samples and expected outputs of the BP neural network,through MATLAB software,BP neural network algorithm is used to train learning,and the offset offset simulation curve is obtained.In order to further reflect the effect of zero drift compensation,the offset offset is programmed using the C language,enter into the EPS controller,experiments were conducted on an electric tube column assembly type performance test rig,the performance diagram after compensation is compared with the performance chart before compensation,by comparing the performance before and after compensation,we can see the effect of BP neural network algorithm in zero drift compensation.