Research On Development Of Typical Driving Cycles Based On Markov Chain And Application Research Of Driving Cycles

As a common basic technology in the automotive industry,automotive driving cycles play an important role in fuel consumption testing,emission testing,the establishment of new energy vehicle technology evaluation system,and the upgrading of energy saving and emission reduction technology level.Driving cycles have strong regional characteristics and type differences.There are obvious differences in driving cycles between different countries,regions and different types of vehicles.At present,our country mainly refers to the standard of driving cycles in Europe,which is obviously different from the actual road characteristics of our country.Therefore,it is necessary to study the construction of automotive driving cycles with local characteristics,develop new methods of construction of driving cycles,and further explore the late application of driving cycles,so as to enhance the technical reserve and application of driving cycles research in China.In this thesis,the data acquisition and processing of a line hybrid bus is carried out firstly.Sampling lines and equipment are identified,and then the raw data is cleaned.Through data integration and extraction,long-distance interception,abnormal data processing and other processes,data for the construction of driving cycles are obtained,and then the amount of data is analyzed.Subsequently,the data are divided into fragments,classified states,statistics of state transition probability matrix and calculation of characteristic parameters,which provide data support for subsequent construction and processing of driving cycles.According to the speed correlation at different time scales,it is proved that the data of the vehicle's driving cycle has Markov property in a small times scale.Based on Markov chain method,the alternative driving cycles of Xi'an typical city bus line are constructed.With ten characteristic parameters such as average speed as the evaluation index,the driving cycle with the minimum deviation degree of the total data is selected as the typical driving cycle.Compared with the Velocity-Acceleration(V-A)method,the validity of the driving cycle based on Markov chain is verified.Driving cycles classification and identification are studied,and the process of driving cycles identification research is established.The characteristic parameters for driving cycles classification and recognition are selected by Spielman correlation coefficient.A standard condition library for driving cycles identification is established based on K-means clustering algorithm.Then a model for driving cycles identification is built and trained based on BP neural network.Finally,the correctness of the model is verified by the confusion matrix and ROC curve.Based on the typical economic driving cycle of the developed line bus,the parameters of the hybrid power system,including engine,motor,battery and transmission system,are matched.Then Cruise is used to build the vehicle model of HEV and Simulink is used to build the energy management strategy model based on logic gates.Typical economic driving cycle of bus routes developed in this thesis are used as input of driving cycles task.After designing dynamic calculation tasks,joint simulation analysis is carried out.