Vehicle driving cost evaluation method based on data driving scene

Abstract

The invention relates to a vehicle driving cost evaluation method based on a data driving scene, and belongs to the field of new energy vehicles. The method comprises the steps of obtaining driving historical data in a specific area, performing data preprocessing, and performing working condition fragment division; carrying out dimensionality reduction on multi-dimensional characteristic parameters of working condition segments based on principal component analysis; utilizing an IABC-Kmeans algorithm to establish a typical working condition feature set, reflecting a synthetic working condition of a specific driving style and a driving habit in a certain area according to the quantitative proportion of each clustering sample and correlation recombination of samples in the class and clustering center parameters, and carrying out comparison verification of statistical features on the synthetic working condition and an original data set; designing a unified quantification method for battery aging, fuel consumption and electric quantity maintenance, and establishing a DDPG multi-target energy management optimization model integrated with expert experience, so that the strategy has higher training efficiency on the premise of ensuring optimality. According to the invention, a reference can be provided for a more accurate vehicle driving cost evaluation method.

Description

Technical field [0001] The present invention belongs to the field of new energy vehicles, involving a method of evaluating the vehicle driving cost of the vehicle based on data drive scenarios. Background technique [0002] The vehicle test cycle is a curve describing the vehicle's driving state, and the total time is about 1500s, which is a centralized reflection of the dynamic driving behavior of the vehicle. It can be analyzed for vehicle driving costs, energy consumption / emissions, and limited value analysis and new models. The basis of technical development and assessment provides the basis. All countries in the world have attached importance to the development of typical driving conditions suitable for the country. At present, Europe and America Japan and other car developed countries have developed testing conditions that reflect the driving characteristics of their own driving characteristics. my country's model test conditions initially adopted NEDC (new standard Europ...

AI Technical Summary

Problems solved by technology

In addition, most evaluations of energy management strategies are based on current standard operating conditions, and the differences in driving styles and habits in various regions make it difficult to conduct more accurate vehicle cost evaluations based on specific driving habits in a certain region. There are many parameters in the working condition segment, and the parameters are coupled and interacted with each other. How to combine the actual road scene and include as much original data set information as possible to perform dimensionality reduction and cluster analysis on the multi-parameters of the working condition segment, It is the key problem that needs to be solved in the construction of working conditions

Moreover, existing RL-based control strategies focus one-sidedly on improving fuel economy while ignoring the impact of battery aging during vehicle operation.

The lifespan of a battery is closely related to its working condition. An unreasonable control strategy will significantly increase the cost of battery replacement, even accelerate battery inconsistency, and risk triggering dangerous thermal runaway

Faced with higher maintenance or replacement costs due to battery aging, how to extend battery life is a problem that must be carefully considered in the design of energy management strategies

Images