CN-116394913-B - Plug-in hybrid electric vehicle energy management system and method based on road section information electric quantity distribution

CN116394913BCN 116394913 BCN116394913 BCN 116394913BCN-116394913-B

Abstract

The invention discloses a plug-in hybrid electric vehicle energy management system and method based on road section information electric quantity distribution. The data is trained to support a vector machine to identify working conditions, the combined new working conditions are processed by a DP algorithm, the result of the DP algorithm is processed, the other part of the electric quantity distribution relation of each working condition obtained by training the BP network is the off-line optimization of an ECMS algorithm, and the off-line optimization of equivalent factors is carried out on each class of working conditions. And the on-line part is used for obtaining an electric quantity distribution relation through the BP neural network according to the road section information and the identification information obtained by the traffic information obtaining module, planning the SOC, and then selecting an equivalent factor according to the working condition and the SOC planning track and carrying out operation. The invention ensures that the electric quantity distribution and the selection of the equivalent factors are more reasonable, and can better improve the fuel economy of the plug-in hybrid electric vehicle.

Inventors

WANG SHAOHUA
ZHENG YUNXIANG
SHI DEHUA
YIN CHUNFANG
LI CHUN

Assignees

江苏大学

Dates

Publication Date: 20260512
Application Date: 20230421

Claims (7)

1. The plug-in hybrid electric vehicle energy management system based on road section information electric quantity distribution is characterized by comprising an offline part and an online part, wherein the offline part comprises a DP module, a data analysis module and a neural network module which are sequentially connected, and the online part comprises a traffic information acquisition module, a working condition identification module, a vehicle module, an ECMS algorithm module and an SOC planning module, wherein the traffic information acquisition module is connected with the working condition identification module, and the vehicle module, the ECMS algorithm module and the SOC planning module are sequentially connected; The system comprises a neural network module, a working condition identification module and an ECMS algorithm module, wherein the neural network module, the working condition identification module and the ECMS algorithm module are provided with an online part and an offline part, the offline part is that a driver acquires data and processes the acquired processed data, the neural network module trains the working condition identification module to identify the working condition of a road section, the processed data is input into a DP module for calculation, the result data is input into a data analysis module, the neural network module is trained to obtain the ratio of the power consumption of each working condition unit mileage, the ECMS equivalent factor offline optimization is carried out according to the working condition and the power consumption of each working condition unit mileage, the online part is that a traffic information acquisition module acquires the mileage of the road section, the average speed of the road section and the average parking frequency of the road section, the working condition identification module carries out identification of working conditions of different road sections by using a trained support vector machine, the working condition classification of each road section is recorded, the current available SOC of the vehicle is input into the neural network module, the ratio of the power consumption of each working condition unit mileage is obtained by using the ratio to be input into an SOC module for reference track generation, the working condition classification of the road section and the SOC reference track is selected in the ECMS algorithm module according to carry out the ECMS equivalent factor EF, and finally the ECMS algorithm is controlled by calculation according to the information such as the vehicle demand torque.
2. The plug-in hybrid electric vehicle energy management method based on road section information electric quantity distribution is characterized by comprising the following steps of: (1) Firstly, data acquisition of a driver is carried out, and the data are divided into four types of congestion, smoother, unobstructed and high-speed working conditions; (2) Combining each working condition of the classified segments into working conditions with different distances, calculating the average speed and the average stopping times of each working condition combined, wherein the data are used for training a support vector machine, and the support vector machine is used for identifying the road section driving working conditions; (3) The relation research of the mutual influence of four working conditions is carried out by utilizing DP planning, the distance of 60Km is fixed, the driving mileage of each working condition occupies the driving mileage difference of the total working condition, a plurality of groups of operations are carried out, the result data of DP dynamic planning is processed, and the numerical ratio of SOC/Km of electric quantity consumption per kilometer of each working condition is obtained and recorded as : : : And the ratio of the values of each working condition, namely the mutual influence relation among the working conditions, is carried out to obtain the duty ratio rule of the electricity consumption of each working condition, and the calculation formula is as follows: ; Wherein the method comprises the steps of : : : The ratio is the ratio of the unit distance electric quantity consumption representing four working conditions; is the value of SOC/Km of electric consumption per kilometer of the working condition type 1, The value of SOC/Km for each kilometer of power consumption for operating condition class 2, The value of SOC/Km for power consumption per kilometer for operating condition class 3, For the value of SOC/Km of electric consumption per kilometer of the working condition type 4, mileage of each working condition is calculated And the available SOC: As an input there is provided, 、、、 The BP neural network training is carried out as output, and the neural network comprises three layers, namely an input layer, an hidden layer and an output layer, wherein the input layer, the hidden layer and the output layer are used for carrying out electric quantity distribution of road section working conditions for information transmitted by a subsequent map; Wherein the method comprises the steps of The road mileage is the road section mileage of the working condition category I, the working condition category 2, the working condition category 3 and the working condition category 4, The current available SOC value of the automobile; (4) Road section information obtained by the traffic information obtaining module, namely average speed of each road section, average parking times and road section mileage identified by the support vector machine Vehicle available electricity amount information: Data is processed 、 Input into BP neural network to obtain output, i.e. the ratio of consumed electric quantity under unit distance of each working condition : : : ; (5) The BP neural network can obtain the working condition information output of each section of the trip through the traffic information acquisition module to predict the electric quantity distribution of the trip, and the output obtained through the BP neural network is the ratio of the electric quantity consumption under each working condition unit distance : : : Obtaining a relation to calculate and plan the SOC, and relating to a calculation formula: ; ; For the current amount of power available to the vehicle, x is a set variable (used as an intermediate variable for planning), Is an SOC reference track; for the distance currently traveled over operating condition category 1, For the distance currently traveled over operating condition category 2, For the distance currently traveled over operating condition category 3, Distance of the current driving working condition class 4; (6) In order to reduce the influence of the error of the condition identification judgment on the reference track, the condition of each road section is judged by using the average speed and the average parking times of different road sections obtained by the traffic information acquisition module, the current road section condition module is identified online for enhancing the accuracy of identification, the identification time is the latest running data of 300s in the road section, if the inconsistent identification time exceeds 60s, the SOC is replaced and planned again according to the category of the online identified condition; (7) The classified working conditions are combined and spliced into four kinds of working conditions, namely four kinds of working conditions of congestion, smoother, unobstructed and high-speed, wherein for each kind of working conditions, equivalent factor optimizing is carried out, only SOC variation optimizing is carried out, the distance is not used as a variable, only SOC is used as a variable, model optimizing of equivalent fuel consumption is carried out, the minimum SOC is set to be 0.3, optimization is carried out through a GA genetic algorithm, EF equivalent factor is used as a variable, fuel consumption is used as an objective function, the distance of each working condition is set to be 60Km, the SOC is gradually reduced to be 0.3 from 1, the interval is 0.05, the optimizing equivalent factor optimizing is carried out, and the optimized EF is recorded to record corresponding SOC/Km values at the same time; (8) The obtained planned SOC corresponds to the working condition of each road section, the power consumption per unit mileage is calculated, and the EF corresponding to the SOC/Km under the corresponding working condition is found from the Map to perform equivalent factors; ; ; ; ; Wherein the method comprises the steps of For minimum fuel consumption, t is the operating mode running time, N is the final value of the operating mode running time, For the time interval to be set to 1S, In order to achieve a fuel consumption of the engine, For engine torque For the engine speed, For the equivalent fuel consumption of the motor, EF is an equivalent factor, The motor power, eta is the battery efficiency, Q is the fuel heating value, For the torque of the motor, For the rotational speed of the motor, For optimal fuel consumption, For the engine torque, For the torque of the motor, Is the total required torque; (9) The planned SOC curve and the actual SOC curve inevitably have fluctuation in the running process, when the situation of seriously deviating from the planned SOC occurs, EF correction is carried out to enable the planned curve to be close to the planned curve, the adopted method is to segment PI, small adjustment is carried out on the planned curve within a reasonable range deviating from the planned curve, adjustment is carried out when the planned curve exceeds a threshold value, and the segmented PI functions are as follows: ; Wherein the method comprises the steps of For the planned SOC track, the SOC is the SOC of the vehicle at the current moment, and EF is the selected equivalent factor = , 、 Is a coefficient.
3. The method of claim 2, wherein the offline optimization DP dynamic programming data is optimized data with different duty ratios under classified working conditions, the different duty ratios gradually increase the duty ratio of a certain working condition, and the final optimization result of dynamic programming is processed to obtain the ratio of the power consumption per unit distance of each working condition under different duty ratios under different available power.
4. The method of claim 2, wherein the training data of the BP neural network is data obtained by DP dynamic programming, the input is available electric quantity of the vehicle, mileage of four different working condition categories are sequentially discharged, and the output is a ratio of unit electric quantity consumption of the four different working condition categories.
5. The method of claim 2, wherein the SOC trajectory planning performs SOC reference trajectory planning according to a ratio of unit power consumption of four output working conditions of the BP neural network and available power of the vehicle, and specifically performs power distribution by establishing a functional relationship to obtain an SOC planned trajectory.
6. The method of claim 2, wherein the data of the condition identification is derived from a network map, the identification parameter is average speed and average number of stops, and the result is the overall condition category of a certain road.
7. The method of claim 2, wherein after the condition is identified in the condition category of the road section, the current condition part is identified online, the identification time is the latest running data of 300s in the road section, for example, the identification inconsistency replaces the re-planning of the SOC according to the condition category identified online, and the condition for replacing the condition category is that the identification inconsistency time exceeds 60s.

Description

Plug-in hybrid electric vehicle energy management system and method based on road section information electric quantity distribution Technical Field The invention belongs to the field of automobile energy management methods, and particularly relates to a plug-in hybrid electric vehicle energy management system and method based on working condition type electric quantity distribution. Background The traditional fuel automobile has been developed for many years, the defects of the traditional fuel automobile are also obvious, the energy utilization rate is low, the environmental pollution is serious, the hybrid electric automobile and the pure electric automobile are rapidly developed under the background of nonrenewable petroleum resources and shortage of resources, the defects of the battery of the pure electric automobile are not widely accepted, and the hybrid electric automobile at present shows the advantages. The contradiction between the economic development and the environmental pollution and the energy shortage is increasingly sharp, the plug-in hybrid electric vehicle has the combination of the advantages of the traditional fuel oil vehicle and the transition type vehicle between the pure electric vehicle, and the plug-in hybrid electric vehicle has the advantages that the power source is provided with an engine and a motor, and the energy is also fuel oil and electric energy, so that the worry of the pure electric vehicle on the driving mileage can be solved, the consumption of the fuel oil can be reduced, and the advantages are obvious. The plug-in hybrid electric vehicle belongs to one of hybrid electric vehicles, has the advantages of reducing fuel consumption and reducing emission, and the quantity of the reduced fuel consumption mainly depends on an energy management strategy, and different energy management strategies have great influence on fuel economy, so that the plug-in hybrid electric vehicle has great research significance on energy management of the plug-in hybrid electric vehicle. The method is characterized in that an equivalent fuel consumption strategy is used as an instantaneous optimal algorithm and can be applied to a plug-in hybrid electric vehicle, the working principle is that electric energy consumption is equivalent to fuel consumption, the core is that an equivalent factor is selected, the selection of the equivalent factor directly influences fuel economy, factors influencing the equivalent factor mainly comprise working conditions, driving distance, available SOC and the like, the reasonable selection of the equivalent factor can greatly improve the fuel economy, real-time road section information can be obtained from a map along with the development of networking information, the information can be used as the influence quantity of the equivalent factor selection, the road section information is fused into the selection of the equivalent factor, the reference SOC track can be planned through the road section information, the indirect influence EF selection is carried out, each section of road section information also comprises the influence of the working condition category, and the establishment of the energy management strategy based on the distribution of the working condition category electric quantity has important significance for improving the fuel economy Disclosure of Invention The invention aims to carry out SOC road section planning on different road sections under a total journey so as to enable electric quantity to be distributed more reasonably, thereby improving fuel economy and providing an energy management method of a plug-in hybrid electric vehicle based on road section information electric quantity distribution. The invention adopts the following technical proposal to achieve the aim. The plug-in hybrid electric vehicle energy management system based on road section information electric quantity distribution comprises an offline part and an online part, wherein the offline part comprises a DP module, a data analysis module and a neural network module which are sequentially connected, and the online part comprises a traffic information acquisition module, a working condition identification module, a vehicle module, an ECMS algorithm module and an SOC planning module, wherein the traffic information acquisition module is connected with the working condition identification module, and the vehicle module, the ECMS algorithm module and the SOC planning module are sequentially connected; The system comprises a neural network module, a working condition identification module and an ECMS algorithm module, wherein the neural network module, the working condition identification module and the ECMS algorithm module are provided with an online part and an offline part, the offline part is that a driver acquires data and processes the acquired processed data, the neural network module trains the working condition identification module to identify the working co