Prediction Method And Experimental Study On Tire Noise And Optimization Design For Low-noise Tire

Tire noise is one of the most important factors reflecting tire performance. The performance and control of tire noise are important research fields in tire industry and urban transportation. It is significant to promote technological progress of the tire industry and improve the urban environment.Tire noise test is the necessary means which establishes and validates noise prediction model. The optimization design of low noise is based on tire noise prediction. The optimization control of tire noise is an important part in the design and development of new products. In this paper the main research works include the methods of tire noise prediction, tire noise experiments and optimization design of low-noise tire. The laboratory test system was designed and a complete test method of tire noise was proposed. According to the noise experiments a new prediction method of tire noise was developed and the prediction model of tire noise was established based on BP neural networks. Finally objective optimization design of low noise for tread patterns was carried out.The following aspects are the main research works and innovation:(1) A novel prediction method of tire noise was proposed based on BP neural networks. According to the complexity of acoustic mechanism of the tire and the diversity of factors influencing the tire noise, essentially the tire sound is a complex nonlinear process. Based on analysis of the existing prediction methods of tire noise, a prediction model of tire noise was studied employing artificial neural networks, and then the nonlinear mapping relationship was established between influencing factor of tire noise and noise performance. An interactive design and verification modeling method was proposed based on neural network applying to the most widely used BP neural networks.(2) In order to obtain learning samples of neural network model, LDR test system for the noise prediction model was designed based on the analysis of the current testing methods of tire noise. The main components and key technologies of test system were discussed. And also hardware and software of the test system were analyzed and designed. (3) A test method and specification of tire noise were proposed based on the LDR method. In order to get representative test data as the learning samples, adopting the orthogonal experimental methods to design the test program, the universal and representative noise test data are obtained based on a limited number of tests. In the paper the noise of smooth tires and patterns tires was analyzed and compared.(4) A new prediction model of tire noise based on BP neural network was established and the predicting simulation was carried out on the different types of tire noise. A method was proposed to extract structural parameters of the tire tread patterns as the main input of the model. The method can be used to classify tread patterns, extract parameters of the patterns features and achieve the parameterization of the irregular and complex patterns structure.(5) With the basic idea of simulated annealing algorithm, ASAGA method was proposed based on improved adaptive genetic algorithm. For the problems of premature and local optimal solutions during the low noise optimization of the tire applying traditional GA, the method which is integrated into the advantages of global optimization of SA algorithm improves GA and meet the goal of optimization design of low noise in terms of optimization efficiency and results. In the paper, considering the tread patterns as the optimal object, the implementation of tire noise optimization design was discussed. And also numerical simulation and analysis of tire example were carried out.