Power Source Output Torque Control Based On Working Friction Characteristics Of Clutch

The clutch is an important component in the process of mechanical power transmission,which plays a very important role in the stable start of machinery and the transformation of working conditions.Clutch,from the friction plate separation and combination directly determines whether the clutch to participate in the power transmission,and in the clutch to participate in the process of power transmission will inevitably appear the relative motion between the main,from the friction plate and other conditions,so that the friction plate wear each other,reduce the service life of the friction plate.Based on this problem,the reverse research strategy is used in this paper,that is,according to the working state of the clutch,to constantly adjust the output torque of the power source,so as to make the clutch in the optimal working state and improve the acceleration performance of the machine.The main work of this paper is as follows:(1)The working friction characteristics of the clutch are analyzed.The relative motion state between the master and slave friction plates during the mechanical starting of the clutch is mainly studied.According to the force change of the driven friction plate and the speed difference between the master and slave parts,the working state of the clutch is divided into three stages.Firstly,the working characteristics of each stage are modeled and studied,and then the working characteristics of each stage are summarized.Finally,the working stages of the clutch are simulated and modeled in Simulink.(2)Research on BP Neural network for predicting engine output torque.Firstly,the structure and working principle of BP neural network are analyzed and studied.In this process,it is found that the activation function has a great influence on the accuracy of data prediction.Then,on the basis of previous studies on activation function,a new improved Xwish activation function is proposed.Finally,the curve fitting characteristics of the activation function and the traditional activation function in BP neural network are simulated and compared respectively.The results show that the performance of the activation function in data fitting is stronger than other activation functions.Therefore,the activation function is a relatively optimal choice in the control of power source output torque when constructing the neural network model of engine output torque prediction.(3)Modeling the power source output torque control environment components.Firstly,the torque control strategy in traditional hybrid vehicles is studied,and then a new hybrid model is built on this basis.Simulink / simscape,stateflow and other modules are used to model and analyze the power components of the hybrid vehicle.Finally,the BP neural network prediction model of engine output torque and the PID model of motor output torque under the new Xwish activation function are built respectively.(4)The power source output torque control model and comparison model of clutch working friction characteristics are constructed.According to the torque transfer relationship between the transmission components,the component models are connected and assembled.At the same time,the state transition module is inserted into the BP neural network model of engine output torque prediction and the PID model of drive motor output torque control.The hybrid vehicle model with torque controller and the vehicle model driven by pure engine are established respectively.The acceleration characteristics of the two under different loads and the energy loss caused by the relative motion of the friction plate are compared,which proves that the research has progressive significance.