A high-speed driving-based electric vehicle energy consumption prediction method and system
By extracting high-speed operating conditions and acceleration segment data of electric vehicles, and combining K-means clustering and regression algorithms, an energy consumption prediction model was constructed. This model solved the problem of the accuracy of energy consumption prediction under high-speed driving conditions of electric vehicles, and improved the accuracy of prediction and the driver's travel planning ability.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHONGQING INNOVATION CENTER OF BEIJING INSTITUTE OF TECHNOLOGY
- Filing Date
- 2023-03-07
- Publication Date
- 2026-07-21
AI Technical Summary
Existing technology makes it difficult to accurately predict the energy consumption of electric vehicles at high speeds, leading to misjudgments by owners about battery level and remaining range, thus increasing range anxiety.
By acquiring vehicle operation data of electric vehicles, extracting high-speed driving segments and acceleration segments, calculating energy consumption data, and using K-means clustering and regression algorithms to train a machine learning model, an energy consumption prediction model is obtained to predict energy consumption in the future.
It improves the accuracy and reliability of electric vehicle energy consumption prediction, helping drivers to rationally allocate battery power and plan routes, and alleviate range anxiety.
Smart Images
Figure CN116238337B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of electric vehicle energy consumption prediction technology, and in particular to a method and system for predicting the energy consumption of electric vehicles based on high-speed driving. Background Technology
[0002] In recent years, new energy electric vehicles have gradually matured, and charging infrastructure has become more complete. However, because the standard driving range displayed by the vehicle often differs significantly from the actual driving range, under high-speed conditions, the limited data capacity of charging stations may lead drivers to misjudge the battery level and remaining range, reducing public confidence in using electric vehicles. Therefore, predicting the energy demand of electric vehicles under high-speed conditions and providing accurate energy consumption forecasts can help drivers better manage their driving range and plan reasonable travel routes.
[0003] Energy consumption research of electric vehicles has always been a hot topic in electric vehicle research. Under different environmental and usage conditions, the energy consumption of electric vehicles varies, which makes it extremely difficult to predict the energy consumption of electric vehicles. It is often more difficult to find a perfect algorithm to predict the energy consumption and remaining range of electric vehicles under various conditions.
[0004] Therefore, there is an urgent need for a method to predict the energy consumption of electric vehicles at high speeds, improve the accuracy of battery remaining power prediction, and alleviate range anxiety. Summary of the Invention
[0005] Therefore, it is necessary to provide a method and system for predicting the energy consumption of electric vehicles based on high-speed driving to address the aforementioned technical problems.
[0006] A method for predicting the energy consumption of electric vehicles based on high-speed driving includes the following steps: acquiring vehicle operation data of the electric vehicle within a target time period; extracting high-speed driving segments from the vehicle operation data; and extracting several acceleration segments from the high-speed driving segments; acquiring energy consumption data from the several acceleration segments; and calculating the energy consumption of the electric vehicle during acceleration based on the energy consumption data; calculating the vehicle slope based on the vehicle driving level collected at the vehicle end; and acquiring historical data of the electric vehicle driving under various conditions, including energy consumption; processing the historical data using the K-means clustering algorithm to obtain collected features; performing principal component analysis on the collected features, selecting several principal components, and combining them with dynamic and static parameters from the vehicle end to input them into a regression algorithm machine learning model to train an electric vehicle energy consumption prediction model; acquiring the target principal components of the electric vehicle to be driven and inputting them into the electric vehicle energy consumption prediction model to obtain the predicted energy consumption of the electric vehicle during acceleration in the future.
[0007] In one embodiment, the vehicle operation data includes a unique vehicle identifier, time, vehicle status, vehicle speed, vehicle level, engine operating status, brake pedal, control lever information, accelerator pedal position, and drive motor power.
[0008] In one embodiment, acquiring vehicle operation data of an electric vehicle within a target time period and extracting high-speed driving segments from the vehicle operation data includes: identifying time points in the vehicle operation data where the electric vehicle is in power mode, not charging, and has a vehicle speed greater than 0, and marking these as start times; identifying time points after the start times where the electric vehicle is not in power mode for multiple consecutive periods, or is in charging mode for multiple consecutive periods, or where the vehicle speed is 0 for a duration exceeding a preset time threshold, and marking these as end times; obtaining vehicle driving segments based on the start and end times; acquiring a preset high-speed driving speed threshold and a segment time threshold, and extracting high-speed driving segments from the vehicle driving segments based on the high-speed driving speed threshold and the segment time threshold.
[0009] In one embodiment, the step of extracting several acceleration segments from the high-speed driving segment includes: in the high-speed driving segment, detecting when the electric vehicle is in power mode, not charging, the brake pedal is not engaged, the control lever is engaged, and the current accelerator pedal position minus the previous accelerator pedal position being greater than a multiple threshold of the maximum throttle opening slope, marking this as the acceleration start time; after the acceleration start time, detecting when the accelerator pedal position remains in a preset duration threshold and continues to decrease, marking this as the acceleration end time; and extracting several acceleration segments from the high-speed driving segment based on the acceleration start time and the corresponding acceleration end time.
[0010] In one embodiment, acquiring energy consumption data from the plurality of acceleration segments and calculating the energy consumed during the acceleration process of the electric vehicle based on the energy consumption data includes: extracting energy consumption data from the plurality of acceleration segments based on a preset sampling period, wherein the energy consumption data includes battery voltage and battery current; and calculating the energy consumed by the electric vehicle during the sampling period by combining the sampling period, battery voltage, and battery current, using the following formula:
[0011] E = U * I * T
[0012] In the formula, U represents the battery voltage, I represents the battery current, and T represents the sampling period; the start time and end time of acceleration are taken as a sampling period to calculate the energy consumption of the electric vehicle during the acceleration segment.
[0013] In one embodiment, the collected features include vehicle speed, vehicle steering wheel angle, vehicle lateral acceleration, vehicle longitudinal acceleration, vehicle total current, and the combined distribution of vehicle speed-longitudinal acceleration, vehicle speed-braking deceleration, and vehicle speed-steering wheel rotation speed.
[0014] In one embodiment, the step of performing principal component analysis on the collected features and selecting several feature principal components includes: using the collected features as initial analysis variables, and calculating principal components based on the data characteristics of the initial analysis variables using a coefficient correlation matrix or covariance matrix, wherein the data characteristics include standardized or unstandardized features; calculating the eigenvalues and corresponding standard eigenvectors of the coefficient correlation matrix or covariance matrix; determining whether the eigenvalues and the standard eigenvectors exhibit multicollinearity; if so, reselecting the initial analysis variables; otherwise, performing a goodness test of principal component analysis based on the eigenvalues and the standard eigenvectors; obtaining the expressions of the principal components and determining the number of principal components, and selecting several feature principal components from all principal components.
[0015] A high-speed driving-based electric vehicle energy consumption prediction system is provided to implement the high-speed driving-based electric vehicle energy consumption prediction method described above. The system includes: an acceleration segment extraction module for acquiring vehicle operation data of the electric vehicle within a target time period, extracting high-speed driving segments from the vehicle operation data, and extracting several acceleration segments from the high-speed driving segments; an energy consumption calculation module for acquiring energy consumption data from the several acceleration segments and calculating the energy consumption of the electric vehicle during acceleration based on the energy consumption data; a historical data acquisition module for calculating the vehicle slope based on the vehicle driving level collected from the vehicle end, and acquiring historical data of the electric vehicle under various conditions, including energy consumption; a historical data processing module for processing the historical data using a K-means clustering algorithm to acquire collected features; a prediction model training module for performing principal component analysis on the collected features, selecting several feature principal components, and combining them with dynamic and static parameters from the vehicle end to input them into a regression algorithm machine learning model to train an electric vehicle energy consumption prediction model; and a predicted energy consumption acquisition module for acquiring the target feature principal components of the electric vehicle to be driven and inputting them into the electric vehicle energy consumption prediction model to obtain the predicted energy consumption of the electric vehicle during acceleration over a future period.
[0016] In one embodiment, the system further includes: a start-of-driving detection module, used to detect, in the vehicle operation data, time points when the electric vehicle is in power mode, not in charging state, and the vehicle speed is greater than 0, and mark them as start-of-driving times; an end-of-driving detection module, used to detect, after the start-of-driving times, time points when the electric vehicle is not in power mode for multiple consecutive times, or is in charging state for multiple consecutive times, or the duration of the vehicle speed being 0 exceeds a preset time threshold, and mark them as end-of-driving times; a driving segment acquisition module, used to obtain vehicle driving segments based on the start-of-driving times and end-of-driving times; and a high-speed driving segment extraction module, used to acquire a preset high-speed driving speed threshold and a segment time threshold, and extract high-speed driving segments from the vehicle driving segments according to the high-speed driving speed threshold and the segment time threshold.
[0017] In one embodiment, the system further includes: a start acceleration detection module, used to detect, during the high-speed driving segment, the time point in which the electric vehicle is in power mode, not charging, the brake pedal is not engaged, the control lever is engaged, and the current accelerator pedal position minus the previous accelerator pedal position is greater than a multiple of the maximum throttle opening slope threshold, and mark it as the acceleration start time; an end acceleration detection module, used to detect, after the acceleration start time, the time point in which the accelerator pedal position remains in a preset duration threshold and continues to decrease, and mark it as the acceleration end time; and an acceleration segment extraction module, used to extract several acceleration segments from the high-speed driving segment based on the acceleration start time and the corresponding acceleration end time.
[0018] Compared to existing technologies, the advantages and beneficial effects of this invention are as follows: By acquiring vehicle operation data of an electric vehicle within a target time period, high-speed driving segments are extracted, and several acceleration segments are extracted from these high-speed driving segments. Energy consumption data in these acceleration segments is obtained, and the energy consumption of the electric vehicle during acceleration is calculated. Based on the vehicle's driving level collected from the vehicle end, the vehicle slope is calculated, and historical data of the electric vehicle under various conditions is acquired. This historical data includes energy consumption. The historical data is processed using the K-means clustering algorithm to obtain collected features. Principal component analysis is performed on these features, and several principal components are selected. These features are then combined with the vehicle's dynamic and static parameters and input into a regression algorithm machine learning model to train an electric vehicle energy consumption prediction model. When it is necessary to predict the energy consumption of an electric vehicle, the target principal components of the electric vehicle can be obtained. Through the electric vehicle energy consumption prediction model, the predicted energy consumption of the electric vehicle during acceleration in the future period can be obtained, improving the accuracy and reliability of energy consumption prediction. This allows drivers to rationally allocate vehicle battery power and plan routes, alleviating range anxiety. Attached Figure Description
[0019] Figure 1 This is a flowchart illustrating an electric vehicle energy consumption prediction method based on high-speed driving in one embodiment.
[0020] Figure 2 This is a schematic diagram of the structure of an electric vehicle energy consumption prediction system based on high-speed driving in one embodiment. Detailed Implementation
[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.
[0022] In one embodiment, such as Figure 1 As shown, a method for predicting the energy consumption of electric vehicles based on high-speed driving is provided, including the following steps:
[0023] Step S101: Obtain vehicle operation data of electric vehicle within the target time period, extract high-speed driving segments from the vehicle operation data, and extract several acceleration segments from the high-speed driving segments.
[0024] Specifically, the electric vehicle driving behavior prediction algorithm proposed in this embodiment is based on real-vehicle operation data from the vehicle network platform. The electric vehicle is a pure electric vehicle. To predict the energy consumption of the electric vehicle during continuous acceleration from the perspective of the electric vehicle, dynamically updated vehicle operation data is required. Therefore, vehicle operation data of the electric vehicle over a continuous period of time can be extracted from the vehicle network monitoring data platform. High-speed driving segments of the electric vehicle can then be queried from this data, and several acceleration segments can be extracted from these segments. Based on these acceleration segments, the energy consumption of the electric vehicle over a future period can be predicted, which aligns with the energy consumption of the electric vehicle during acceleration.
[0025] The vehicle operation data includes the vehicle's unique identifier, time, vehicle status, vehicle speed, vehicle levelness, engine operating status, brake pedal, control lever information, accelerator pedal position, and drive motor power.
[0026] The steps for obtaining high-speed driving segments are as follows: In the vehicle operation data, the time points when the electric vehicle is in power mode, not in charging state, and the vehicle speed is greater than 0 are marked as the start time of driving; after the start time of driving, the time points when the electric vehicle is not in power mode for several consecutive times, or is in charging state for several consecutive times, or the duration of the vehicle speed being 0 exceeds a preset time threshold are marked as the end time of driving; based on the start time of driving and the end time of driving, the vehicle driving segment is obtained; a preset high-speed driving speed threshold and segment time threshold are obtained, and the high-speed driving segment in the vehicle driving segment is extracted according to the high-speed driving speed threshold and segment time threshold.
[0027] Specifically, to extract vehicle driving segments from the acquired vehicle operation data, it is necessary to determine whether the vehicle is in a driving state. Therefore, the time node when the electric vehicle is in power mode, charging status is 0, and vehicle speed is greater than 0 can be marked as the start time of driving. After the start time of driving, if the vehicle starts to be out of power mode multiple times in a row, or is in charging status multiple times in a row, or the duration of vehicle speed being 0 exceeds a preset time threshold, such as 5 minutes, this time node is marked as the end time of driving. Therefore, the corresponding vehicle driving segments are obtained based on the start time and the end time of driving.
[0028] By pre-setting a high-speed driving speed threshold and a segment time threshold, for example, setting the high-speed driving speed threshold to 80 km / h and the segment time threshold to 10 minutes, segments with a vehicle speed greater than 80 km / h and a duration of more than 10 minutes can be extracted from the vehicle driving segments as high-speed driving segments, so as to extract acceleration segments from the high-speed driving segments.
[0029] The steps for extracting acceleration segments are as follows: After acquiring the high-speed driving segment, in the high-speed driving segment, the time node where the electric vehicle is in power mode, not charging, the brake pedal is not engaged, the control lever is engaged, and the current accelerator pedal position minus the previous accelerator pedal position is greater than a multiple threshold of the maximum throttle opening slope is detected and marked as the acceleration start time; after the acceleration start time, the time node where the accelerator pedal position continues for a preset duration threshold and remains in a decreasing state is detected and marked as the acceleration end time; based on the acceleration start time and the corresponding acceleration end time, several acceleration segments are extracted from the high-speed driving segment.
[0030] Specifically, after acquiring a high-speed driving segment, the system detects changes in the vehicle's driving status, charging status, brake pedal status, control lever status, and accelerator pedal position within the high-speed driving segment. In the high-speed driving segment, when the electric vehicle is in power mode, not charging, the brake pedal is not engaged, the control lever is engaged, and the current accelerator pedal position minus the previous accelerator pedal position is greater than a multiple threshold of the maximum throttle opening slope, the corresponding time point is marked as the acceleration start time.
[0031] After the acceleration begins, the duration of the accelerator pedal position is detected. If the duration exceeds a preset threshold, such as 1 second, and the accelerator pedal position continues to decrease, the corresponding time point is marked as the acceleration end time.
[0032] Finally, based on the marked acceleration start time and corresponding acceleration end time, several acceleration segments of the electric vehicle are extracted from the high-speed driving segment, so as to predict the vehicle's energy consumption based on the acceleration segments.
[0033] Step S102: Obtain energy consumption data from several acceleration segments, and calculate the energy consumed during the acceleration process of the electric vehicle based on the energy consumption data.
[0034] Specifically, after acquiring the acceleration segment, the energy consumption data of the electric vehicle during acceleration is extracted from the acceleration segment, and the energy consumed during the acceleration process of the electric vehicle is calculated based on the acquired energy consumption data.
[0035] Step S102 includes: extracting energy consumption data from several acceleration segments based on a preset sampling period, the energy consumption data including battery voltage and battery current; and calculating the energy consumed by the electric vehicle during the sampling period by combining the sampling period, battery voltage, and battery current, using the following formula:
[0036] E = U * I * T
[0037] In the formula, U represents the battery voltage, I represents the battery current, and T represents the sampling period; the acceleration start time and acceleration end time are taken as a sampling period to calculate the energy consumption of the electric vehicle in the acceleration segment.
[0038] Specifically, based on a preset sampling period, the battery voltage and battery current in several acceleration segments are extracted. By combining the sampling period, battery voltage, and battery current, the energy consumed by the electric vehicle in the sampling period can be calculated. Thus, the acceleration start time and acceleration end time in the acceleration segment can be used as a sampling period to calculate the energy consumed by the electric vehicle in the corresponding acceleration segment, so as to train the prediction model based on the calculated energy consumption.
[0039] Step S103: Calculate the vehicle slope based on the vehicle driving level collected by the vehicle terminal, and obtain historical data of the electric vehicle driving under various conditions, including energy consumption.
[0040] Specifically, since vehicles consume different amounts of energy under different slope conditions, the slope value during vehicle travel is an important factor affecting vehicle energy consumption. The vehicle's driving level can be collected at the vehicle end to calculate the vehicle slope, and historical data of the vehicle under different vehicle type conditions, different temperature conditions, and different driving speed conditions can be uploaded to record the historical data, which includes the energy consumption of electric vehicles.
[0041] Step S104: Use the Kmeans clustering algorithm to process historical data and obtain collected features.
[0042] Specifically, since driving conditions are mainly affected by road information and driving style, it is necessary to use the Kmeans clustering algorithm to process the acquired historical data to obtain collected features, so as to identify driving style and road information based on the collected features and better predict the energy consumption of electric vehicles.
[0043] The collected features include vehicle speed, vehicle steering wheel angle, vehicle lateral acceleration, vehicle longitudinal acceleration, vehicle total current, and the combined distribution of vehicle speed-longitudinal acceleration, vehicle speed-braking deceleration, and vehicle speed-steering wheel rotation speed.
[0044] Step S105: Perform principal component analysis on the collected features, select several principal components, and combine them with the vehicle's dynamic and static parameters to input them into the regression algorithm machine learning model to train and obtain the electric vehicle energy consumption prediction model.
[0045] Specifically, in order to save computational load and improve computational efficiency, it is necessary to perform principal component analysis on the collected features. Based on the contribution rate and cumulative contribution rate of all principal components, the top few feature principal components are selected. For example, if the vehicle speed, total vehicle current, and vehicle-braking deceleration rank in the top three among the calculated contribution rates and cumulative contribution rates of the collected features, then vehicle speed, total vehicle current, and vehicle-braking deceleration are selected as feature principal components.
[0046] The obtained principal components are combined with dynamic and static parameters from the vehicle, such as vehicle weight, remaining battery power, maximum speed, driving force, and acceleration time—all parameters related to vehicle energy consumption. These parameters are then fed into a support vector machine regression algorithm machine learning model for training, resulting in an electric vehicle energy consumption prediction model. This model can accurately predict the energy consumption of electric vehicles over a future period, facilitating route planning. Furthermore, this electric vehicle energy consumption prediction model has good versatility and can be applied to different types of electric vehicles, expanding its applicability.
[0047] The steps for performing principal component analysis (PCA) on the collected features and selecting the principal components are as follows: Using the collected features as initial analysis variables, calculate the principal components using the coefficient correlation matrix or covariance matrix based on the data characteristics of the initial analysis variables. The data characteristics may be standardized or unstandardized. Calculate the eigenvalues and corresponding standard eigenvectors of the coefficient correlation matrix or covariance matrix. Determine if multicollinearity exists between the eigenvalues and the standard eigenvectors. If it does, reselect the initial analysis variables. Otherwise, perform a goodness-of-fit test for PCA based on the eigenvalues and standard eigenvectors. Obtain the expressions for the principal components and determine the number of principal components. Select several principal components from all principal components.
[0048] Specifically, in principal component analysis, the acquired features are used as initial analysis variables. Based on the data characteristics of the initial analysis variables, the coefficient correlation matrix or covariance matrix is used to find the principal components. The data characteristics include both standardized and unstandardized data. Specifically, when the acquired features are standardized, the coefficient correlation matrix is used to find the principal components; when the acquired features are unstandardized, the covariance matrix is used.
[0049] Calculate the eigenvalues and corresponding standard eigenvectors of the correlation matrix or covariance matrix, and determine whether multicollinearity exists between the eigenvalues and standard eigenvectors. If multicollinearity exists, reselect the initial analysis variables and perform principal component analysis. If multicollinearity does not exist, perform a goodness test of principal component analysis based on the eigenvalues and standard eigenvectors. The goodness test can use the sphericity test or the KMO statistic test. If the goodness test passes, obtain the expression of the principal components and determine the number of principal components. Select several eigenvalues from all the obtained principal components. Eigenvalues have a higher correlation with vehicle energy consumption, thereby improving the accuracy and reliability of the electric vehicle energy consumption prediction model.
[0050] Step S106: Obtain the target feature principal components of the electric vehicle to be used and input them into the electric vehicle energy consumption prediction model to obtain the predicted energy consumption of the electric vehicle during the acceleration process in the future.
[0051] Specifically, when it is necessary to predict the energy consumption of an electric vehicle under high-speed conditions, the principal components of the target features of the electric vehicle are obtained and input into the electric vehicle energy consumption prediction model to obtain the predicted energy consumption of the electric vehicle during acceleration over a period of time in the future. This allows the driver to allocate the battery power reasonably, plan travel routes, and alleviate range anxiety.
[0052] In this embodiment, by acquiring vehicle operation data of an electric vehicle within a target time period, high-speed driving segments are extracted, and several acceleration segments are extracted from these high-speed driving segments. Energy consumption data in these acceleration segments is obtained, and the energy consumed by the electric vehicle during acceleration is calculated. Based on the vehicle's driving level collected from the vehicle end, the vehicle slope is calculated, and historical data of the electric vehicle under various conditions is acquired. This historical data includes energy consumption. The historical data is processed using the K-means clustering algorithm to obtain collected features. Principal component analysis is performed on the collected features, and several principal components are selected. These features are then combined with the vehicle's dynamic and static parameters and input into a regression algorithm machine learning model to train an electric vehicle energy consumption prediction model. When it is necessary to predict the energy consumption of an electric vehicle, the target principal components of the electric vehicle can be obtained. Through the electric vehicle energy consumption prediction model, the predicted energy consumption of the electric vehicle during acceleration in the future period can be obtained, improving the accuracy and reliability of energy consumption prediction. This allows drivers to rationally allocate vehicle battery power and plan routes, alleviating range anxiety.
[0053] This invention first segments the continuous acceleration phase of a vehicle after it enters high-speed operation from normal operating conditions. Then, it combines vehicle operation data to construct an energy consumption model of the vehicle under continuous acceleration during its most recent high-speed operation. Finally, it predicts the vehicle's future driving conditions by considering external driving factors and the driver's driving style. By focusing on vehicle-side data and incorporating external and spatiotemporal factors, this method offers greater accuracy and timeliness compared to previous data sources, and can accurately predict individual vehicles, significantly improving the accuracy of vehicle energy consumption prediction.
[0054] like Figure 2 As shown, an electric vehicle energy consumption prediction system 20 based on high-speed driving is provided to implement the electric vehicle energy consumption prediction method based on high-speed driving as described above. It includes: an acceleration segment extraction module 21, an energy consumption calculation module 22, a historical data acquisition module 23, a historical data processing module 24, a prediction model training module 25, and a predicted energy consumption acquisition module 26, wherein:
[0055] The acceleration segment extraction module 21 is used to acquire vehicle operation data of electric vehicles within a target time period, extract high-speed driving segments from the vehicle operation data, and extract several acceleration segments from the high-speed driving segments.
[0056] The energy consumption calculation module 22 is used to acquire energy consumption data in several acceleration segments and calculate the energy consumption of the electric vehicle during acceleration based on the energy consumption data.
[0057] The historical data acquisition module 23 is used to calculate the vehicle slope based on the vehicle driving level collected by the vehicle terminal, and to acquire historical data of electric vehicles driving under various conditions, including energy consumption.
[0058] The historical data processing module 24 is used to process historical data using the Kmeans clustering algorithm to obtain the collected features;
[0059] The prediction model training module 25 is used to perform principal component analysis on the collected features, select several feature principal components, and combine them with the vehicle-side dynamic and static parameters to input them into the regression algorithm machine learning model to train the electric vehicle energy consumption prediction model.
[0060] The predicted energy consumption acquisition module 26 is used to acquire the target feature principal components of the electric vehicle to be used and input them into the electric vehicle energy consumption prediction model to obtain the predicted energy consumption of the electric vehicle during the acceleration process in the future.
[0061] In one embodiment, the system further includes: a start-of-driving detection module, used to detect, in the vehicle operation data, the time node when the electric vehicle is in power mode, not in charging state, and the vehicle speed is greater than 0, and mark it as the start-of-driving time; an end-of-driving detection module, used to detect, after the start-of-driving time, the time node when the electric vehicle is not in power mode for multiple consecutive times, or is in charging state for multiple consecutive times, or the duration of the vehicle speed being 0 exceeds a preset time threshold, and mark it as the end-of-driving time; a driving segment acquisition module, used to obtain a vehicle driving segment based on the start-of-driving time and the end-of-driving time; and a high-speed driving segment extraction module, used to obtain a preset high-speed driving speed threshold and a segment time threshold, and extract the high-speed driving segment from the vehicle driving segment according to the high-speed driving speed threshold and the segment time threshold.
[0062] In one embodiment, the system further includes: a start acceleration detection module, used to detect, during a high-speed driving segment, the time point at which the electric vehicle is in power mode, not charging, the brake pedal is not engaged, the control lever is engaged, and the current accelerator pedal position minus the previous accelerator pedal position is greater than a multiple of the maximum throttle opening slope threshold, and mark it as the acceleration start time; an end acceleration detection module, used to detect, after the acceleration start time, the time point at which the accelerator pedal position remains in a preset duration threshold and continues to decrease, and mark it as the acceleration end time; and an acceleration segment extraction module, used to extract several acceleration segments from the high-speed driving segment based on the acceleration start time and the corresponding acceleration end time.
[0063] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.
[0064] It is obvious to those skilled in the art that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby storing them in a computer storage medium (ROM / RAM, magnetic disk, optical disk) for execution by the computing device. In some cases, the steps shown or described can be performed in a different order than those described herein, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Therefore, the present invention is not limited to any particular hardware and software combination.
[0065] The above description, in conjunction with specific embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such deductions or substitutions should be considered within the scope of protection of the present invention.
Claims
1. A method for predicting the energy consumption of electric vehicles based on high-speed driving, characterized in that, include: Acquire vehicle operation data of electric vehicles within a target time period, extract high-speed driving segments from the vehicle operation data, and extract several acceleration segments from the high-speed driving segments. The energy consumption data of the acceleration segments are obtained, and the energy consumption of the electric vehicle during acceleration is calculated based on the energy consumption data. The vehicle slope is calculated based on the vehicle driving level collected from the vehicle terminal, and historical data of electric vehicles driving under various conditions is obtained, including energy consumption. The historical data was processed using the K-means clustering algorithm to obtain the collected features; Principal component analysis is performed on the collected features to select several principal components, and these components are combined with the vehicle's dynamic and static parameters and input into the regression algorithm machine learning model to train and obtain an electric vehicle energy consumption prediction model. Obtain the principal components of the target features of the electric vehicle to be used and input them into the electric vehicle energy consumption prediction model to obtain the predicted energy consumption of the electric vehicle during the acceleration process in the future.
2. The method for predicting the energy consumption of electric vehicles based on high-speed driving according to claim 1, characterized in that, The vehicle operation data includes the vehicle's unique identifier, time, vehicle status, vehicle speed, vehicle levelness, engine operating status, brake pedal, control lever information, accelerator pedal position, and drive motor power.
3. The method for predicting the energy consumption of electric vehicles based on high-speed driving according to claim 2, characterized in that, The step of acquiring vehicle operation data of electric vehicles within a target time period and extracting high-speed driving segments from the vehicle operation data includes: In the vehicle operation data, the time point when the electric vehicle is detected to be in power mode, not in charging state, and the vehicle speed is greater than 0 is marked as the start time of driving. After the start of driving, the time point when the electric vehicle is not in power mode for multiple consecutive times, or is in charging state for multiple consecutive times, or the duration of the vehicle speed being 0 exceeds a preset time threshold is marked as the end of driving. Based on the start and end times of driving, a vehicle driving segment is obtained; Obtain a preset high-speed driving speed threshold and a segment time threshold, and extract high-speed driving segments from the vehicle driving segments based on the high-speed driving speed threshold and the segment time threshold.
4. The method for predicting the energy consumption of electric vehicles based on high-speed driving according to claim 3, characterized in that, The extraction of several acceleration segments from the high-speed driving segment includes: In the high-speed driving segment, the acceleration start time is marked as the time when the electric vehicle is in power mode, not charging, the brake pedal is not engaged, the control lever is engaged, and the current accelerator pedal position minus the previous accelerator pedal position is greater than a multiple threshold of the maximum throttle opening slope. After the acceleration start time, the time point in which the accelerator pedal position remains in a preset duration threshold and continues to decrease after the acceleration start time is marked as the acceleration end time. Based on the acceleration start time and the corresponding acceleration end time, several acceleration segments are extracted from the high-speed driving segment.
5. The method for predicting the energy consumption of electric vehicles based on high-speed driving according to claim 4, characterized in that, The step of acquiring energy consumption data in the plurality of acceleration segments and calculating the energy consumed during the acceleration process of the electric vehicle based on the energy consumption data includes: Based on a preset sampling period, energy consumption data is extracted from the several acceleration segments, including battery voltage and battery current. Combining the sampling period, battery voltage, and battery current, the energy consumed by the electric vehicle during the sampling period is calculated using the following formula: E = U * I * T In the formula, U represents the battery voltage, I represents the battery current, and T represents the sampling period; By taking the start and end times of acceleration as a sampling period, the energy consumption of the electric vehicle during the acceleration segment can be calculated.
6. The method for predicting the energy consumption of electric vehicles based on high-speed driving according to claim 1, characterized in that, The collected features include vehicle speed, vehicle steering wheel angle, vehicle lateral acceleration, vehicle longitudinal acceleration, vehicle total current, and the combined distribution of vehicle speed-longitudinal acceleration, vehicle speed-braking deceleration, and vehicle speed-steering wheel rotation speed.
7. The method for predicting the energy consumption of electric vehicles based on high-speed driving according to claim 1, characterized in that, The process involves performing principal component analysis on the collected features, selecting several principal components, including: The collected features are used as initial analysis variables. Principal components are obtained by using a coefficient correlation matrix or a covariance matrix based on the data features of the initial analysis variables. The data features include standardized or unstandardized features. Calculate the eigenvalues and corresponding standard eigenvectors of the coefficient correlation matrix or covariance matrix; Determine whether the eigenvalues and the standard eigenvectors are multicollinear; if so, reselect the initial analysis variables. Conversely, a goodness test of principal component analysis is performed based on the eigenvalues and the standard eigenvectors. Obtain the expressions of the principal components and determine the number of principal components. Select several characteristic principal components from all principal components.
8. An energy consumption prediction system for electric vehicles based on high-speed driving, characterized in that, A method for predicting the energy consumption of an electric vehicle based on high-speed driving, as described in any one of claims 1-7, comprises: An acceleration segment extraction module is used to acquire vehicle operation data of electric vehicles within a target time period, extract high-speed driving segments from the vehicle operation data, and extract several acceleration segments from the high-speed driving segments. An energy consumption calculation module is used to acquire energy consumption data in the several acceleration segments and calculate the energy consumption of the electric vehicle during acceleration based on the energy consumption data. The historical data acquisition module is used to calculate the vehicle slope based on the vehicle driving level collected by the vehicle terminal, and to acquire historical data of electric vehicles driving under various conditions, including energy consumption. The historical data processing module is used to process the historical data using the K-means clustering algorithm to obtain the collected features; The prediction model training module is used to perform principal component analysis on the collected features, select several feature principal components, and combine them with the vehicle's dynamic and static parameters to input them into the regression algorithm machine learning model to train and obtain the electric vehicle energy consumption prediction model. The predicted energy consumption acquisition module is used to acquire the target feature principal components of the electric vehicle to be used and input them into the electric vehicle energy consumption prediction model to obtain the predicted energy consumption of the electric vehicle during the acceleration process in the future.
9. The electric vehicle energy consumption prediction system based on high-speed driving according to claim 8, characterized in that, The system also includes: The start-of-driving detection module is used to detect, in the vehicle operation data, the time points when the electric vehicle is in power mode, not in charging state, and the vehicle speed is greater than 0, and mark them as the start-of-driving time. The end-of-driving detection module is used to detect, after the start-of-driving time, the time point when the electric vehicle is not in power mode for multiple consecutive times, or is in charging state for multiple consecutive times, or the duration of the vehicle speed being 0 exceeds a preset time threshold, and mark it as the end-of-driving time. The driving segment acquisition module is used to obtain vehicle driving segments based on the start time and end time of driving. The high-speed driving segment extraction module is used to obtain a preset high-speed driving speed threshold and a segment time threshold, and extract the high-speed driving segment from the vehicle driving segment based on the high-speed driving speed threshold and the segment time threshold.
10. The electric vehicle energy consumption prediction system based on high-speed driving according to claim 9, characterized in that, The system also includes: The acceleration detection module is used to detect, during the high-speed driving segment, the electric vehicle is in power mode, not charging, brake pedal is not engaged, control lever is engaged, and the time point when the current accelerator pedal position minus the previous accelerator pedal position is greater than a multiple threshold of the maximum throttle opening slope is marked as the acceleration start time. The acceleration detection module is used to detect the time point after the acceleration start time when the accelerator pedal position remains in a preset duration threshold and continues to decrease, and mark it as the acceleration end time. The acceleration segment extraction module is used to extract several acceleration segments from the high-speed driving segment based on the acceleration start time and the corresponding acceleration end time.