CN-115640740-B - ECMS strategy optimization method based on working condition clustering result of navigation information

CN115640740BCN 115640740 BCN115640740 BCN 115640740BCN-115640740-B

Abstract

The invention provides an ECMS strategy optimization method based on a working condition clustering result of navigation information, which comprises the steps of performing off-line simulation according to real car running data to obtain running segment division of each section of running car data, training a working condition identification neural network model for the running segments, cleaning and repairing key road condition signals in the navigation information to obtain repairing data, performing mileage window division on the basis of the repairing data to obtain correction on congestion and gradient, calculating working condition characteristic parameters of each time window under the real navigation repairing data by utilizing time windows based on the actual navigation repairing data division in the third and fourth steps, performing working condition clustering division according to the working condition characteristic parameters and the neural network model, matching corresponding optimal equivalent factors according to a database of the first step, and correcting navigation clustering working condition output results according to the optimal equivalent factors obtained by historical vehicle speed clustering. The method of the invention greatly reduces the oil consumption.

Inventors

WANG ZHENGQIANG
FU CHANGBO
WANG JIAWEI
HU ZHICHENG

Assignees

上汽大众汽车有限公司

Dates

Publication Date: 20260505
Application Date: 20220920

Claims (6)

1. An ECMS strategy optimization method based on a working condition clustering result of navigation information is characterized by comprising the following steps: Step one, performing off-line simulation according to real car sports data, dividing the running segments of each piece of sports car data according to preset dividing rules, and performing simulation experiments through a cyclic targeting method based on the divided running segments to obtain a database of SOC and oil consumption corresponding to different effective factors under each running segment; Training a working condition recognition neural network model for the operation segment; Step three, cleaning and repairing key road condition signals in the navigation information to obtain repairing data; Dividing mileage windows based on the repair data of the third step, and correcting the division according to road congestion conditions and gradients; dividing a time window by the repair data, wherein the division is corrected according to road congestion conditions and gradients; step five, calculating working condition characteristic parameters of each time window under the actual navigation repair data based on the time windows divided by the actual navigation repair data in the step three and the step four; Step six, carrying out clustering division of working conditions according to the working condition characteristic parameters in the step five and the neural network model obtained through training in the step two, and matching corresponding optimal equivalent factors according to the database obtained in the step one; Step seven, correcting the navigation clustering working condition output result according to the optimal equivalent factor obtained by the historical vehicle speed clustering; The first step further comprises: Step 11, dividing the running segments of each piece of sports car data according to a certain dividing rule, wherein the dividing rule comprises that the time between the starting and stopping of a car is more than 60s, and the running segment is defined as a limit according to 60s, and the time between the starting and stopping of the car is less than 60s, and the running segment is defined as a running segment between the starting and stopping of the car; And 12, performing simulation experiments according to the data after the segmentation, and recording the final SOC and oil consumption of each operation segment under different discharging equivalent factors and charging equivalent factors by a cyclic targeting method to finally obtain a database of the SOC and oil consumption corresponding to the different equivalent factors under each operation segment.
2. The ECMS policy optimization method based on the working condition clustering result of the navigation information according to claim 1, wherein the working condition characteristic parameter calculation method of each time window adopted in the second operation segment and the fifth operation segment includes: Average vehicle speed: Acceleration average value: , Deceleration average value: , idle time ratio: , average running speed , Standard deviation of speed , Standard deviation of acceleration , Standard deviation of deceleration , Wherein, the For the vehicle speed at this operation segment i, The acceleration time of the run-time segment, The deceleration time of the run-out segment is, For the duration of the run-time segment, For the time when the running segment vehicle speed is not 0, For the time when the total speed change for the run segment is not 0, For times when the run segment acceleration is not 0, For the time when the run-segment deceleration is not 0, Is the forward acceleration value for the moment i of the operating segment, The running segment is the reverse acceleration value at the moment i, and the forward direction is the forward direction of the vehicle.
3. The ECMS policy optimization method based on the condition clustering result of the navigation information according to claim 1, wherein the third step further includes: Cleaning and repairing key road condition signals in the navigation information, wherein the traffic information of each road section comprises average speed, road speed limit value, gradient and congestion levels; Wherein the data predicted by the secondary exponential smoothing method is replaced by the cleaned data.
4. The ECMS policy optimization method based on the operation condition clustering result of the navigation information according to claim 3, wherein the cleaning rule in the third step includes: When the average vehicle speed is greater than the maximum vehicle speed limit value multiplied by 1.1, clearing the road section data; when the average speed is greater than the maximum speed limit value multiplied by 0.7, the congestion level is 4, and the road section data is cleared; and when the average vehicle speed is smaller than the maximum vehicle speed limit value multiplied by 0.2, the congestion level is 1, and the road section data is cleared.
5. The ECMS policy optimization method based on the condition clustering result of the navigation information according to claim 1, wherein the cyclic targeting method in step 12 includes: The initial condition is SOC=50%, the increment condition is discharge equivalent factor 2.0-4.0, charge equivalent factor 2.0-4.0, and increment step length is 0.5.
6. The ECMS policy optimization method based on the operation condition clustering result of the navigation information according to claim 5, wherein the step two further includes: training and verifying as neural network model input according to the obtained working condition characteristic parameters and the labels of each operation segment in the first step; And selecting 10 neurons and 4 neurons from an implicit layer and an output layer of the neural network model respectively, and performing 46 iterations to obtain the neural network model capable of performing working condition clustering.

Description

ECMS strategy optimization method based on working condition clustering result of navigation information Technical Field The invention mainly relates to an ECMS strategy optimization method, in particular to an ECMS strategy optimization method based on a working condition clustering result of navigation information. Background The hybrid electric vehicle is a whole vehicle driving system formed by a plurality of power sources, and the power and the energy among the power sources are reasonably distributed and coordinated, so that the fuel economy of the whole vehicle can be optimized on the premise that the power requirement is met. The existing equivalent fuel consumption minimum energy management strategy (Equivalent Fuel Consumption Minimization Strategy, ECMS for short) is based on the standard circulation working condition for energy consumption optimization, however, the actual driving working condition of the driver is greatly different from the standard circulation working condition, the existing ECMS strategy cannot be adaptively adjusted along with the actual driving working condition, and further energy consumption optimization is difficult to be carried out aiming at the actual driving working condition of the driver. Disclosure of Invention In view of the above, the invention introduces navigation information, and realizes a working condition clustering method based on the navigation information and ECMS policy optimization based on a clustering result. The method comprises the steps of dividing the real vehicle data into two parts, obtaining fuel consumption and charge State (SOC) sample libraries corresponding to different fuel consumption factors under different working conditions according to the real vehicle data offline simulation, carrying out sectional clustering on road conditions to be driven according to navigation information, and matching optimal fuel consumption conversion factors such as ECMS and the like according to clustering results. Under different working conditions, oil consumption and an SOC sample library corresponding to different oil consumption factors are obtained according to the offline simulation of the real vehicle data, firstly, the real vehicle data driving data are preprocessed, and then, in a simulation environment, an iterative targeting method (ITERATIVE SHOOTING ALGORITHM, ISA for short) is adopted based on the processed data to obtain the oil consumption and the SOC sample library corresponding to different oil consumption factors under different working conditions. And carrying out sectional clustering on road conditions to be driven according to navigation information, matching optimal ECMS (electronic fuel system) and other fuel consumption conversion factor parts according to clustering results, firstly, preprocessing navigation data, carrying out adjacent line segmentation, calculating operation characteristics for each section of operation window based on segmentation results, clustering to obtain corresponding sample working conditions, and selecting the corresponding optimal fuel consumption factor. It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory and are intended to provide further explanation of the disclosure. In order to solve the technical problems, the invention provides an ECMS strategy optimization method based on a working condition clustering result of navigation information, which is characterized by comprising the following steps: step one, performing off-line simulation according to real car sports data to obtain the division of running segments of each piece of sports car data; Training a working condition recognition neural network model for the operation segment; Step three, cleaning and repairing key road condition signals in the navigation information to obtain repairing data; dividing mileage windows on the basis of the repair data to obtain correction on congestion and gradient; Step five, calculating working condition characteristic parameters of each time window under the actual navigation repair data by utilizing the time windows divided based on the actual navigation repair data in the step three and the step four; step six, carrying out clustering division of working conditions according to the working condition characteristic parameters of the step five and the neural network model of the step two, and matching corresponding optimal equivalent factors according to the database of the step one; And step seven, correcting the navigation clustering working condition output result according to the optimal equivalent factor obtained by the historical vehicle speed clustering. Preferably, the present invention further provides an ECMS policy optimization method based on a working condition clustering result of navigation information, which is characterized in that the working condition characteristic parameter calculation method of each time