A method and device for predicting the driving range of a pure electric vehicle

Abstract

The present invention provides a method and a prediction device for predicting the cruising range of a pure electric vehicle, comprising the following steps: measuring different initial battery powers of the vehicle and a battery capacity correction coefficient; defining travel segments to form a travel condition library; calculating various parameter characteristics of the vehicle; constructing a cruising range prediction model; obtaining vehicle state parameters; obtaining parameter information between the driver and the destination; selecting travel segments with corresponding feature errors within 5%, and calculating the average driving characteristics of the travel segments; obtaining an electric energy consumption prediction value EC and a discharge rate C; and calculating the cruising range of the vehicle. The beneficial effects of the present invention are as follows: fully considering the influence of environmental temperature and discharge rate on the actual capacity of the battery under different battery power states; through the method of machine learning, associating the prediction of the cruising range with the driving behavior of the driver, and obtaining a cruising range prediction result that is more consistent with the user's driving habits.

Description

Technical Field

[0001] The present invention belongs to the field of electric vehicle control, and particularly relates to a method and a device for predicting the cruising range of a pure electric vehicle. Background Art

[0002] Range anxiety refers to the concern of vehicle owners or drivers that the vehicle does not have enough range to reach their destination and the fear of breaking down on the way. Range anxiety mainly occurs in pure electric vehicles, which is also considered a major obstacle to the large-scale promotion of electric vehicles. Especially at the present stage, without breakthroughs in battery technology and the lack of popularization of charging piles, the problem of range anxiety greatly affects the user experience and also affects the sales of electric vehicles and other electric driving tools.

[0003] The operating conditions of electric vehicles are complex, and the real-time energy consumption of the vehicles varies greatly, which pose challenges to accurately predicting the feasible driving range under the current remaining energy state of the vehicle. Currently, the main way to predict the remaining range of the vehicle is based on the method of energy consumption prediction, that is, estimating the possible future energy consumption rate of the vehicle, and then calculating the remaining range based on the current remaining energy. Among them, some methods calculate the average energy consumption value per unit mileage (energy consumption per 100 kilometers) recently, and divide the current remaining energy by the average energy consumption to obtain the remaining mileage; some methods perform system identification on the energy consumption process, and predict the future energy consumption change by establishing a dynamic model of energy consumption to calculate the remaining mileage; some methods comprehensively consider driving data and environmental traffic data to model the energy consumption to calculate the remaining mileage. The above methods based on the energy consumption model are the mainstream methods currently in application and cannot accurately predict the remaining mileage. The characteristics of the present invention are to use chassis dynamometer test data, actual vehicle driving condition data, navigation data, and comprehensively consider various influencing factors of the automotive air conditioner to establish a method and a device for predicting the cruising range of a pure electric vehicle. Summary of the Invention

[0004] In view of this, the present invention aims to propose a method and a device for predicting the cruising range of a pure electric vehicle, so as to effectively improve the prediction accuracy of the remaining range of the electric vehicle and greatly reduce the range anxiety of the driver.

[0005] To achieve the above object, the technical solution of the present invention is realized as follows:

[0006] A method for predicting the cruising range of a pure electric vehicle includes the following steps:

[0007] S1. For a selected vehicle model, measure the battery capacity correction coefficients of the vehicle at different initial battery levels and with different discharge rates in high and low temperature environments in the laboratory respectively;

[0008] S2. Obtain the historical driving data of the vehicle. Divide the driving data from point A to the destination point B of the vehicle into one trip segment. According to the definition of the trip segment, divide the historical data into multiple trip segments to form a trip driving condition library. Screen the trip segments in the database to eliminate abnormal trips.

[0009] S3. Calculate the speed - related features, acceleration - related features, pedal operation - related features, and the energy consumption of each trip segment in the trip driving condition library, and calculate the mutual information number between each feature parameter and the energy consumption. Select the feature parameters with a mutual information value greater than 0.3 with the energy consumption as the input during the training of the remaining driving range prediction model.

[0010] S4. Obtain the state of charge (SOC) of the vehicle's battery and the usage status of the vehicle air - conditioner through the CAN line.

[0011] S5. Obtain the average vehicle speed and the maximum speed limit of the road through the in - vehicle navigation system.

[0012] S6. Input the driving features in steps S4 and S5 into the remaining driving range prediction model obtained in step S3 to obtain the predicted value of the electricity consumption EC and the discharge rate C.

[0013] S7. Obtain the current ambient temperature of the vehicle. Correct the dischargeable capacity of the battery through the discharge rate to obtain the actual capacity of the battery SOC_real under the current vehicle environment. Divide the actual capacity of the battery by the predicted electricity consumption rate to calculate the remaining driving range of the vehicle under the current vehicle environment state.

[0014] Furthermore, the historical trip segment feature data in step S2 includes:

[0015] Speed - related features: average vehicle speed, maximum vehicle speed, standard deviation of vehicle speed;

[0016] Acceleration - related features: maximum acceleration, average acceleration, maximum deceleration, average deceleration, relative positive acceleration;

[0017] Pedal operation - related features: average acceleration pedal opening, average deceleration pedal opening;

[0018] Air - conditioner - related features: cooling on, heating on;

[0019] Furthermore, the definition of the trip segment in step S2: The trip segment represents the driving process of the driver from the starting point A to the destination point B, and is composed of multiple idle segments and adjacent motion segments.

[0020] An electric vehicle driving range prediction device includes a power supply module, a storage module, a travel segment division module, a feature analysis module, a machine learning module, a driving range prediction module, a display module, a navigation module, and a temperature module. The storage module, the travel segment division module, the feature analysis module, the machine learning module, the driving range prediction module, and the display module are sequentially connected by signals. The output end of the feature analysis module is also connected to the storage module. The navigation module and the temperature module are both connected to the output end of the driving range prediction module. The storage module, the travel segment division module, the feature analysis module, the machine learning module, the driving range prediction module, the display module, the navigation module, and the temperature module are all electrically connected to the power supply module;

[0021] The travel segment division module is used to obtain the historical driving data of the vehicle in step S2, divide the driving data between the vehicle starting from point A and reaching the destination point B into one travel segment, and according to the definition of the travel segment, divide the historical data into multiple travel segments to form a travel condition database. By screening the travel segments in the database, abnormal trips are eliminated;

[0022] The feature analysis module is used to calculate the speed - type features, acceleration - type features, pedal operation - type features, air - conditioning features, and energy consumption of each travel segment in the travel condition database in step S3, and calculate the mutual information number between each feature parameter and the energy consumption, and select the feature parameters with a mutual information value greater than 0.3 with the energy consumption as the input during model training;

[0023] The machine learning module and the driving range prediction module are used to construct a driving range prediction model using machine learning algorithms in step S3. The input of the driving range prediction model is the travel segment features, and the output of the driving range prediction model is the energy consumption prediction value EC and the battery discharge rate C;

[0024] The navigation module is used to obtain the average vehicle speed and the maximum road speed limit of the driver reaching the destination through the in - vehicle navigation system in step S5;

[0025] The temperature module is used to obtain the current ambient temperature of the vehicle in step S7, correct the available discharge capacity of the battery through the discharge rate to obtain the actual battery capacity SOC_real under the current vehicle environment, and calculate the driving range of the vehicle under the current vehicle environment state by dividing the actual battery capacity by the predicted power consumption rate;

[0026] The display module is used to display the driving range calculated in step S7.

[0027] Compared with the prior art, the electric vehicle driving range prediction method and prediction device of the present invention have the following advantages:

[0028] A method and device for predicting the cruising range of a pure electric vehicle according to the present invention fully consider the influence of environmental temperature and discharge rate on the actual capacity of the battery under different power states; through machine learning, the prediction of the cruising range is associated with the driving behavior of the driver, and a prediction result of the cruising range that is more consistent with the user's driving habits can be obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The drawings forming a part of the present invention are used to provide a further understanding of the present invention. The schematic embodiments and descriptions thereof of the present invention are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0030] Figure 1 It is a schematic diagram of the relationship between the battery capacity, environmental temperature and discharge rate according to the embodiment of the present invention;

[0031] Figure 2 It is a schematic diagram of the division of travel segments according to the embodiment of the present invention;

[0032] Figure 3 It is a schematic flowchart of a method for predicting the cruising range of a pure electric vehicle provided by the embodiment of the present invention;

[0033] Figure 4 It is a schematic structural diagram of a device for predicting the cruising range of a pure electric vehicle provided by the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0034] It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.

[0035] The present invention will be described in detail below with reference to the drawings and in conjunction with the embodiments.

[0036] Glossary:

[0037] Maximal Information Coefficient method: MIC (Maximal Information Coefficient). It is used to measure the degree of association (linear or non-linear relationship) between two feature variables and has higher accuracy compared to Mutual Information (MI).

[0038] Machine learning algorithms: Used for the training and prediction of the energy consumption prediction model. Common algorithms include support vector machines, principal component analysis, and Q-learning.

[0039] As Figures 1 to 4 shown, a method for predicting the cruising range of a pure electric vehicle mainly includes the following steps:

[0040] 1. For a certain vehicle model, the battery capacity correction coefficients are measured separately in the laboratory under high and low temperature environments with different initial battery powers and different discharge rates.

[0041] 2. Obtain the historical driving data of the vehicle. Divide the driving data from point A to the destination point B of the vehicle into one travel segment. According to the definition of the travel segment, divide the historical data into multiple travel segments to form a travel condition database. Screen the travel segments in the database to eliminate abnormal trips.

[0042] 3. Calculate the speed - related features, acceleration - related features, pedal operation - related features, air - conditioner - related features, and energy consumption of each travel segment in the travel database, and calculate the mutual information number between each feature parameter and the energy consumption. Select the feature parameters with a mutual information value greater than 0.3 with the energy consumption as the input during model training.

[0043] 4. Obtain the current SOC of the vehicle and the usage status of the vehicle air - conditioner through the CAN line.

[0044] 5. Through the in - vehicle navigation system, obtain the average vehicle speed and the maximum speed limit of the road between the driver and the destination.

[0045] 6. Input the data obtained in steps 4 and 5 into the model obtained in step 3 to obtain the predicted power consumption value EC and the discharge rate C.

[0046] 7. Obtain the current ambient temperature, correct the dischargeable capacity of the battery in combination with the discharge rate to obtain the actual battery capacity SOC_real in this environment. Finally, divide the actual battery capacity by the predicted power consumption rate to calculate the cruising range of the vehicle in this state.

[0047] The historical travel segment feature data described in step 2 includes but is not limited to:

[0048] 1) Speed - related features: average vehicle speed, maximum vehicle speed;

[0049] 2) Acceleration - related features: maximum acceleration, average acceleration, maximum deceleration, relative positive acceleration (RPA);

[0050] 3) Pedal operation - related features: average opening of the acceleration pedal, average opening of the brake pedal;

[0051] 4) Air - conditioner - related features: refrigeration on, heating on;

[0052] The travel definition in step 2 is as follows: A travel segment represents the driving process of the driver from the starting point A to the destination point B, and is composed of multiple idle segments and adjacent moving segments. The schematic diagram of the travel segment is as Figure 1 shown.

[0053] The advantages of the present invention are:

[0054] Fully consider the influence of ambient temperature and discharge rate on the actual capacity of the battery under different power states; through machine learning methods, the prediction of the cruising range is correlated with the driving behavior of the driver, and a more accurate prediction result of the cruising range that is more in line with the user's driving habits can be obtained.

[0055] Embodiment 1

[0056] The technical solution of the present application will be described below with reference to the accompanying drawings.

[0057] 1. Conduct battery capacity tests in a laboratory under high and low temperature environments with different initial battery powers and different discharge rates. The high temperature test temperatures are 20°C, 25°C, 30°C, 35°C, 40°C, and 45°C respectively, and the low temperature test temperatures are -20°C, -15°C, -10°C, -5°C, 0°C, and 5°C respectively. The discharge rates of the battery are 1 / 3C, 1 / 2C, 1C, 2C, 3C, 4C, 5C, and 6C respectively. Through multiple linear regression, the battery capacity correction coefficients under different temperatures, different discharge rates, and different initial SOCs are fitted. Taking the ambient temperature of 0°C and the initial SOC equal to 80% as an example, the correction coefficients fitted according to the test results are as Figure 1 shown.

[0058] 2. Use the method of dividing travel segments to cut historical data as Figure 2 shown to obtain a travel segment library, and screen the segments in the travel library according to the following 5 rules to eliminate abnormal trips:

[0059] 1) Missing rate rule: If the missing rate of the motion segment data exceeds 10%, the segment will be deleted. If the missing rate is lower than 10%, the missing data will be supplemented by interpolation.

[0060] 2) Travel time rule: The duration of a single trip is not less than 5 minutes.

[0061] 3) Acceleration rule: The instantaneous acceleration of the motion segment is in the range of [-6m / s 2 , 6m / s 2 . The instantaneous acceleration is calculated using the method of calculating at every other point. Using the method of calculating at every other point can reduce accidental errors and make the calculation results more accurate. The specific calculation method is shown in formula (1).

[0062]

[0063] 4) Speed rule: The maximum speed of the motion segment does not exceed 130 km / h, and the minimum speed is not less than 5 km / h.

[0064] 5) Idle time rule: The duration of the idle segment does not exceed 30%.

[0065] Fragments that do not conform to the above five rules are deleted, and the remaining fragments are used for subsequent model construction. In the present invention, the energy consumption of travel fragments is calculated by formula (2), and the unit is kW·h / 100km. Wherein, v is the driving speed of the vehicle, and U and I are the DC current and DC voltage of the battery respectively, which are read from the OBD module.

[0066]

[0067] 3. Calculate the mutual information value between the characteristic parameters and the energy consumption in each trip in the trip library, as shown in Table 1:

[0068] Table 1 Fragment Feature Table

[0069]

[0070] Taking the average speed, maximum speed, average acceleration, average deceleration, RPA, air conditioner switch, average accelerator pedal opening, and average brake pedal opening of the fragment as inputs, using machine learning algorithms to train the energy consumption and cruising range prediction models, and the outputs are energy consumption and discharge rate.

[0071] 4. Obtain the current ambient temperature, remaining battery charge SOC, and the usage status acon / acoff of the vehicle air conditioner through the CAN line.

[0072] 5. Through the in-vehicle navigation, obtain the average speed v1 to the destination and the maximum speed limit v2 of the route.

[0073] 6. Input the parameters obtained in steps 5 and 6 into the model constructed in step 4 to obtain the energy consumption prediction value EC and the discharge rate C.

[0074] 7. According to the ambient temperature and the current SOC of the battery in step 5, combined with the predicted discharge rate in step 7, use Figure 1 to correct the battery capacity. The corrected battery capacity is calculated by the following formula:

[0075] SOC_real = SOC * λ1

[0076] Where: λ1 is the battery capacity correction coefficient for different starting SOCs at the current ambient temperature T_out. Finally, according to the corrected battery capacity, combined with step 8, the cruising range prediction value is obtained:

[0077] s = SOC_real * Cr / EC

[0078] Wherein, Cr is the capacity of the battery, in kW·h.

[0079] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for predicting the cruising range of a pure electric vehicle, characterized in that: It includes the following steps: S1. For the selected vehicle model, measure the battery capacity correction coefficients of the vehicle at different initial battery levels and different discharge rates in high and low temperature environments in the laboratory; S2. Obtain the historical driving data of the vehicle, divide the driving data from point A to destination point B of the vehicle into a single trip segment. According to the definition of the trip segment, divide the historical data into multiple trip segments to form a driving condition library. Screen the trip segments in the database to eliminate abnormal trips; The definition of the trip segment in step S2: The trip segment represents the driving process of the driver from the starting point A to the destination point B, and is composed of multiple idle segments and adjacent motion segments; S3. Calculate the speed-related features, acceleration-related features, pedal operation-related features, and trip energy consumption of each trip segment in the driving condition library, and calculate the mutual information value between each characteristic parameter and the energy consumption. Select the characteristic parameters with a mutual information value greater than 0.3 with the energy consumption as the input during the training of the cruising range prediction model; S4. Obtain the state of charge (SOC) of the vehicle's battery and the usage status of the vehicle air conditioner through the CAN line; S5. Obtain the average vehicle speed and the maximum speed limit of the road through the in-vehicle navigation system; S6. Input the driving characteristics in steps S4 and S5 into the cruising range prediction model obtained in step S3 to obtain the predicted value of the electricity consumption EC and the discharge rate C; S7. Obtain the current ambient temperature of the vehicle, correct the dischargeable capacity of the battery through the discharge rate to obtain the actual capacity SOC_real of the battery under the current vehicle environment, and calculate the cruising range of the vehicle under the current vehicle environment state by dividing the actual battery capacity by the predicted electricity consumption rate.

2. A method for predicting the cruising range of a pure electric vehicle according to claim 1, characterized in that: In step S2, the trip segment characteristic data includes: Speed-related features: average vehicle speed, maximum vehicle speed; Acceleration-related features: average acceleration, average deceleration, relative positive acceleration; Pedal operation-related features: average opening of the accelerator pedal, average opening of the brake pedal; Air conditioner-related features: refrigeration on, heating on.

3. A device for predicting the cruising range of a pure electric vehicle, applying the method for predicting the cruising range of a pure electric vehicle according to any one of claims 1-2, characterized in that: It includes a power supply module, a storage module, a trip segment division module, a feature analysis module, a machine learning module, a cruising range prediction module, a display module, a navigation module, and a temperature module. The storage module, the trip segment division module, the feature analysis module, the machine learning module, the cruising range prediction module, and the display module are sequentially connected by signals; the output end of the feature analysis module is also connected to the storage module. The navigation module and the temperature module are both connected to the output end of the cruising range prediction module. The storage module, the trip segment division module, the feature analysis module, the machine learning module, the cruising range prediction module, the display module, the navigation module, and the temperature module are all electrically connected to the power supply module; The travel segment division module is used to obtain the historical driving data of the vehicle in step S2, divide the driving data of the vehicle from point A to the destination point B into one travel segment, and according to the definition of the travel segment, divide the historical data into multiple travel segments to form a travel driving cycle database. By screening the travel segments in the database, abnormal trips are eliminated; The feature analysis module is used to calculate the speed-related features, acceleration-related features, pedal operation-related features, air-conditioning features, and travel energy consumption of each travel segment in the travel driving cycle database in step S3, and calculate the mutual information number between each feature parameter and the energy consumption. The feature parameters with a mutual information value greater than 0.3 with the energy consumption are selected as the inputs during model training; The machine learning module and the remaining driving range prediction module are used to construct a remaining driving range prediction model using machine learning algorithms in step S3. The input of the remaining driving range prediction model is the travel segment features, and the outputs of the remaining driving range prediction model are the energy consumption prediction value EC and the battery discharge rate C; The navigation module is used to obtain the average vehicle speed and the maximum road speed limit of the driver to reach the destination through the in-vehicle navigation system in step S5; The temperature module is used to obtain the current ambient temperature of the vehicle in step S7, correct the available discharge capacity of the battery through the discharge rate to obtain the actual battery capacity SOC_real under the current vehicle environment, and calculate the remaining driving range of the vehicle under the current vehicle environment state by dividing the actual battery capacity by the predicted power consumption rate; The display module is used to display the remaining driving range calculated in step S7.

4. An electronic device, including a processor and a memory communicatively connected to the processor and used to store instructions executable by the processor, wherein: The processor is used to execute a method for predicting the remaining driving range of a pure electric vehicle according to any one of claims 1-2 above.

5. A server, wherein: It includes at least one processor and a memory communicatively connected to the processor. The memory stores instructions executable by the at least one processor. The instructions are executed by the processor so that the at least one processor executes a method for predicting the remaining driving range of a pure electric vehicle according to any one of claims 1-2.

6. A computer-readable storage medium storing a computer program, wherein: When the computer program is executed by a processor, it implements a method for predicting the remaining driving range of a pure electric vehicle according to any one of claims 1-2.

Images