Study On Noise During AMT Shifting Process

With the development of auto-transmission technology and people's pursuit of power and comfort, more and more attention has been paid to shift quality. A perfect shifting process should be smooth, quiet, harmonious, natural and so on. Today, shift quality has become to be an important branch in the field of auto-transmission. The content of shift quality research includes (1) comfort index, such as shift jerk caused by the change of transmission ratio and shift noise caused by the fluctuation of engine speed and vibration; (2) fuel and emission index which put environment and resources into consideration; (3) part-longevity index which will determined by the maximal load on the parts of transmission system during shifting process; (4) manipulatory index which reflects in what extent the responding speed satisfies the driver's intent. This thesis, supported by"Research on the Dynamic System of Shift Quality Synthetic Evaluation"-National Science Foundation of China (50505014), focuses mainly on the research of shift noise.As an important part of shift quality, it is very significant to research on shift quality. The aim of this paper, based on AMT car, is to find out the influence to shift noise during shifting process by each parameter. AMT realized its shift automation on the basis of traditional transmission and dry clutch. In the shifting process, the influence to noise is mainly caused by the variety of engine's and gearbox's parameters. And engine is the largest source of vehicle noise, mainly including combustion noise, mechanism noise, inflation noise, emission noise and fan noise. Although inflation noise and emission are the two of the largest engine noise sources, as to the noise during shifting process that the thesis focuses on, the mechanism noise caused by the variation of engine speed and torque should be paid more attention. The gearbox noise is considered to be caused by these factors: the variety of the mesh stiffness, transmitting errors, the mesh impact of gears, the node impact caused by friction and the vibration of gearbox. Due to the change of transmitting ratio and gear-mesh in the shifting process, the mesh stiffness correspondingly changes and thus the load upon gears and the periodical impact also change greatly. Therefore, the noise of the whole gear system varies.In shifting process, the change of transmitting ratio causes the fluctuation of engine speed and torque, and thus shift noise occurs. Because it is a very complex process, to simulate this process by computer is quite an intricate task and the result may not be practicable. An experimental research method based on real vehicle is applied in this thesis. By using sound pressure fluctuation as a reflection to the variation of noise during shifting process, analyzing the noise in time domain with the variation of each impact factor, it is obviously shown that the four indices, variation of engine speed, variation of engine torque, clutch engaging speed and shift time, impact the variation of noise most. In this thesis, those impact factors have been researched and in what extent those factors impact the noise individually has been analyzed. It will establish a theoretical basis to construct shift quality evaluation system.Because of the strong correlation among those impact factors, it is not easy to construct a mathematical model to describe. If using traditional statistical method, the workload is too heavy and the result is probably not satisfactory. Neural network has its unique advantage dealing with this kind of task with its self-study ability, error toleration ability and non-line approach ability. This thesis has constructed a network model on the relation between the four impact factors and the noise variation by using BP network. After trained with a large number of experimental data, the error of this model can be controlled in an acceptable span. By testing this model by several additional experimental data, it is proven that this model can reflect the function relation between the four parameters and the noise fluctuation. Then, based on this network model, the value-test method can be used to analyze the four parameters'impacting extent. After it, a conclusion is drawn that, during the shifting process from 1st gear to 2nd gear, the four parameters with the sequence of the impact ability to noise fluctuation are: variation of engine speed, variation of engine torque, clutch engaging speed and shift time.In the end of this thesis, a series of comparison tests have been made to validate the conclusion above. The testing car starts up with 1st gear, and accelerates to the highest gear. Through changing the throttle opening and the shift rule, a series of data in different work conditions can be obtained. After comparing and analyzing the four indices in different work condition and the noise variation, the conclusion mentioned above can be validated.