Robust Optimization Of Control Parameters In DCT Shifting Process

The development of automatic transmission electronic control system and its application software is one of the core technologies of automatic.In the process of software development,the optimization and calibration of the control parameters in shifting process affect the comfort and the service life of parts.This paper relies on the project of Jilin Province Excellent Young Talents Fund "Research on off-line optimization method of control parameters in DCT shifting process considering uncertainties",in order to solve the problems of long calibration period,unclear relationship between control parameters and evaluation indexes,and the consistency of shifting quality throughout the product's life cycle.Taking the DCT shifting process as the research object,the principle analysis and modeling of the main oil pressure regulating subsystem,and the clutch pressure control subsystem are carried out,and the characteristics of the relevant valves were simulated and tested;Shift control strategies are established according to the principle of clutch switching control and hydraulic system.A DCT shifting process simulation platform based on AMEsim and Matlab is established.Based on the pressure curve of the DCT Power-on upshift process,the key points are extracted as the control parameters,and the evaluation indexes(vibration dose value,clutch energy density,shift duration)are established.The optimal Latin hypercube design method is used to establish a sample space where the input is the control parameter and the output is the evaluation index.Using RBF neural network to build an approximate model between the control parameters and the evaluation index,the model can reflect the relationship between input and output accurately.Based on the Sobol sensitivity analysis theory,calculate the main effect index and the total effect index of the shift process control parameters,and analyze the control parameters that have a greater impact on the evaluation index,which can be used for screening key parameters.Based on the an approximate model between the control parameters and the evaluation index,the multi-objective genetic algorithm is used to optimize the control parameters of the shifting process,which can greatly save the optimization time.Select the uncertain factors that have a greater impact on the shifting process through the sensitivity analysis method,including the working temperature of the hydraulic oil,the stiffness of the odd clutch spring,the idle stroke of the clutch,the clearance of plate in even clutch,and the spring stiffness of the solenoid valve.These uncertainty factors as noise factors are used for robustness optimization.Construct an approximate model connecting from shift process evaluation indicators to uncertainty factors and control parameters,and use SN ratio and the evaluation indicators as targets function to perform robustness optimization of the control parameters.Finally,the results of robust optimization and deterministic optimization are verified by hardware in the loop test.The results of robust optimization can resist the disturbance of uncertain factors better.Then,the feasibility of this research method is verified by the real vehicle test and the technical route can save engineer's calibration time,and can be used in other field of the parameters optimization and calibration of automobile.