Research Of Algorithmic Predictions On Tire Vertical Load And Wear Based On Tire Acceleration Signal

With the rapid development of China’s economy and society,vehicles have become an indispensable part of people’s lives.Tire is the only part of the contact with the road,which is closely related to the safety of the vehicle driving.Therefore,the status monitoring of tires has great practical significance.Intelligent tires are a solution that can meet the above needs and can be used to monitor the status of tires.At present,the condition monitoring functions of the more common intelligent tires on the market are mainly tire pressure monitoring and tire temperature monitoring.Other functions(such as friction coefficient monitoring)generally use indirect prediction to obtain the state information of the tire.In this paper,a simplified finite element model of a tire is first established,and the acceleration information of the center of the inner wall of the tire is extracted in the free rolling simulation to provide support for the subsequent research of the prediction algorithm.Then,according to these acceleration data,the prediction algorithms of tire vertical load and wear are researched respectively.The main work content is as follows:(1)Establishment of simplified finite element model for pneumatic tires.Starting from the tire cross-sectional view,a simplified tire finite element model is established,and the effectiveness of the finite element model is verified by radial stiffness and tire grounding imprint.The acceleration signal of the inner wall of the tire is obtained by explicit dynamic simulation in the free rolling state,and this information will be used in the subsequent research and development of various algorithms.(2)Acceleration data processing algorithm.The original acceleration data contains a lot of noise interference information,and the smoothing of the acceleration data is done with a digital filter.The tire rolling speed affects the number of sampling points per rolling cycle,so the parameters of the design filter are related to the tire rolling speed.After obtaining the smooth curve,the acceleration waveform is preliminarily processed to obtain the coordinates of some important feature points by designing algorithms such as peak identification,cycle segmentation,and feature point matching,which have a great correlation with the free rolling process of the tire.(3)Tire vertical load prediction algorithm.Starting from the characteristic points of the acceleration signal,the characteristic points related to the length of the contact patch are selected,and the length of the contact patch of the tire can be calculated according to the coordinates of these characteristic points.In the research,it was found that the length of the contact patch during tire rolling is directly related to the vertical load of the tire,but it is also affected by the tire pressure and rolling speed.Therefore,the relationship between the length of the contact patch and the three variables of load,tire pressure and rolling speed empirical formula.Substituting the length of the contact patch calculated from the acceleration signal into the empirical formula,the predicted value of the vertical load of the tire can be obtained.Verify the tire vertical load prediction algorithm with acceleration data.The relative error of the algorithm within the common load range generally does not exceed 10%.(4)Development of tire wear prediction algorithm.The relationship between wear and acceleration information is not obvious,so the method of neural network is adopted.Based on the acceleration characteristic points extracted in(2),further process the acceleration data to obtain several characteristic values.The characteristic values and the corresponding actual wear are used as training sets,and after training,a BP neural network model that can be used to predict tire wear is obtained.Use acceleration data to verify the tire wear prediction algorithm,the average absolute error of the algorithm within a common wear range is within 0.3mm.