Research On Real-time Optimization And Decision Of Gearshift Point For Stepped Automatic Transmission

Automatic transmission is a significant symbol of the development level of the automobile industry.It is the main embodiment of the international competitiveness of automotive core technology and has an important share in the automotive industry.In the 21 st century,low carbon,environmental protection,comfort and safety have become the theme of the development of the world automobile industry.With the continuous advancement of traditional automotive technology and the development of various new energy vehicles,the requirements for automatic control and matching technology of automotive drivetrains are getting higher,and automatic transmission control technology is becoming increasingly complex.Gearshift decision is the core content of the automotive automatic control theory,that is,to select a reasonable shifting timing according to the working condition,which directly affects the dynamic performance,fuel economy and driving maneuverability,etc.In view of look-up table type shift schedule based on steady-state conditions suffering from many issues,such as long development cycle,high cost,and unable in real-time optimal trade-off between vehicle dynamic and fuel economy,while current online gearshift point decision method still encounters problems of imperfect power request definition,insufficient accuracy prediction model,and poor real-time performance.Therefore,this thesis conducted the research on real-time optimization and decision of gearshift point for stepped automatic transmission,which was supported by National Natural Science Foundation of China and Ph.D.Programs Foundation of Ministry of Education of China.The main contents are as follows:1)Construction of vehicle longitudinal dynamics simulation platform.The longitudinal dynamics model of the vehicle was built by AVL Cruise,and the gearshift decision control model was established by Matlab/Simulink/Stateflow,the vehicle simulation platform was built by co-simulation.At the same time,the simulation platform was revised and verified through the vehicle rotary drum fuel economy test and road dynamic test.2)Analysis of objective evaluation index of gearshift decision.Construction of effective objective evaluation index system is a prerequisite for quantitative analysis based on simulation or testing technologies to achieve coordinated optimization of multi-objective performance.On the basis of related evaluation indices of traditional full throttle dynamic and fuel economy,the objective evaluation index system of gearshift decision was constructed from three aspects of dynamic,economy and drivability.3)Optimal gearshift strategy design based on Dynamic Programming(DP).The gearshift decision process of the stepped automatic transmission is regarded as a multi-stage decision process.The DP algorithm was adopted to solve the optimal shift control sequence of the standard driving cycle.To improve the practicality of the model based offline optimal shift strategy,the numerical model of the traditional drivelines torque loss was established based on bench test data.Considering the shift time and the influence of the clutch friction loss on the gearshift decision,the control variable constraints were added to modify the DP algorithm.The optimized gearshift sequences on the condition of ten standard driving cycles were selected as the database,and cluster analysis method was adopted to extract the sub-optimal shift strategy which can be directly used in vehicle controller.4)Online optimization of gearshift point based on power demand.The Markov chain model is used to establish the longitudinal dynamic demand prediction model for vehicles.The initial Transfer Probability Matrix(TPM)was estimated using Bayesian estimation method.And the online update arithmetic for TPM was proposed,which can reflect the changes of the current vehicle driving conditions in real time.The on-line prediction of future driving conditions was realized using the latest TPM.The gearshift point predictive control strategy based on double-layer MPC was introduced,and the optimal control problem in predictive control time domain was solved by DP algorithm,and the computational efficiency in different predictive time domain was further analyzed.5)Real-time gearshift decision based on the pre obtaining road information.To solve the problem of unexpected and back-to-back shifting under common road conditions such as ramps,bends and traffic congestions,etc.,using the road information provided by the vehicle navigation system as the driver’s visual function supplement,they were introduced into the automatic transmission gearshift decision system.By analyzing the characteristics of the gearshift decision under typical working conditions and combining with the road condition information in the limited time domain in front of the vehicle,the real-time gearshift decision-making strategies under ramp,bend and traffic congestion conditions were designed.6)Verification through vehicle rotary drum test bench and road test.A domestic AMT mini-vehicle was selected as test platform,the software and hardware system of automatic transmission controller developed independently were adopted.In order to verify the validity of the gearshift strategy based on dynamic programming,the vehicle rotary drum test bench was used to test the fuel economy and the vehicle road dynamic performance test was carried out.Finally,under the condition that the vehicle navigation system was not yet available,the validity of the real-time gearshift strategies based on the pre obtaining road information were verified by pre-setting the test road conditions.The research results have important socioeconomic significance and broad engineering application prospects for perfecting automatic transmission theory,shortening the development cycle of automatic transmission products,improving product performance,and promoting the industrialization process of self-owned brand automatic transmission.