Research On DCT Ramp Shift Strategy Based On Data Mining

Gear decision is the core problem of automatic transmission research and the basis of improving the shift quality of automatic transmission.The decision-making of gear position solves the problem of choosing gear shift time or gear shifting point reasonably under complex driving conditions,and the decision-making of gear shifting directly affects the power,economy and driving comfort of the vehicle.At present,scholars at home and abroad generally study the gear decision-making based on the theory design method of two parameters and three parameters.Through the identification of complex driving environment and driving intention,more control parameters are introduced to optimize the shift law.After introducing more parameters in the shift law,although the shift efficiency is more intelligent than the traditional shift law,the shift logic based on map still has a big gap with the skilled driver.In order to achieve the optimization goal of automatic variable speed vehicle design,the intelligent gear decision-making design method of vehicle power transmission system is established by using the massive driving data obtained by skilled drivers to operate the manual stop vehicles of the same type of vehicles,and through the method of data mining,the intelligent gear decision-making design method of vehicle power transmission system is established.Due to the limitation of test conditions,the test data of manual vehicle retaining cannot be provided.Therefore,this paper takes the SUV vehicle with 7-speed DCT carried by a vehicle manufacturer as the research object temporarily.The method of setting up the shift law of the vehicle under special working conditions is carried out based on the vehicle driving data.The paper also elaborates the method in detail with the ramp condition as an example,and the main research contents include:(1)Complete the collection and preprocessing of driving data,and obtain the pure driving data of ramp condition.First of all,the driving data acquisition scheme and route are formulated,and the appropriate skilled drivers are selected to drive the test vehicle to complete the driving data acquisition.Then,the collected driving data is preprocessed,and the cubic spline interpolation algorithm with sliding time window is used to synchronize and resample the data.Then,the missing values in the original data are removed,and the Butterworth filter is used to filter the original data.Finally,the characteristic signal is used to remove the driving data outside the ramp condition,The pure driving data of ramp condition is obtained.(2)Based on the driving data,the shift law of ramp condition is constructed.Firstly,considering the influence of driving intention,the ramp condition is further subdivided into acceleration uphill condition and uniform uphill condition according to acceleration.Then,the driving data of the two conditions are extracted according to the shift parameters.In the acceleration uphill condition,the data points are extracted which are composed of four shift parameters: speed,acceleration,throttle opening and road resistance coefficient,Under the condition of uniform uphill speed,the data points composed of speed,throttle opening and road resistance coefficient are extracted.After obtaining the driving data of the two conditions,one class SVM algorithm is used to detect and process the outliers in the data,and Chi shape algorithm is used to extract the boundary points of each gear.Finally,the moving least square method is used to fit the shift boundary points of each gear,so as to construct the shift law of the ramp condition.(3)The construction of engine actual operation model and the formulation of power and economy shift rules.The engine torque model is established based on the engine bench test data,and the real vehicle data is used as input for simulation verification.The analysis results show that the engine torque model is quite different from the real vehicle driving data.Therefore,it is necessary to reconstruct the actual engine model based on the real vehicle driving data,use the engine bench data to compensate the incomplete model,and then according to the reconstructed real vehicle engine model,formulate the best power performance and the best economic shift rules.(4)Compensation for the missing part of shift schedule.Firstly,the change trend of shift curve(surface)based on data mining is analyzed.Then,aiming at the missing part of shift law,two kinds of shift law compensation rules are proposed: in the acceleration uphill condition,the existing shift surface within the range of throttle opening is compensated by the adjacent acceleration shift line,The missing throttle opening range is compensated by power or economic shift line;Under the condition of constant speed uphill,the shift surface uses the shift line under the resistance coefficient of the adjacent road to compensate the difference within the range of the existing throttle opening,and uses the dynamic and economic shift line to compensate in the range of the missing throttle opening.