Thermal management energy consumption abnormity identification method and equipment based on working condition clustering and medium

Abstract

The invention discloses a thermal management energy consumption abnormity identification method and device based on working condition clustering and a medium, and the method comprises the steps: obtaining historical travel slice data, employing a clustering algorithm to carry out the classification of vehicle driving working conditions, and generating a working condition classification model; acquiring historical thermal management energy consumption data, wherein the historical thermal management energy consumption data comprises a plurality of energy consumption points; classifying the working condition of each energy consumption point based on the working condition classification model; and performing cross comparison on each energy consumption point and the corresponding working condition, and screening energy consumption abnormal points of the vehicle. According to the method, the normal fluctuation range of heat management energy consumption under different working conditions is fully considered, misjudgment and missed judgment are effectively avoided, and the efficiency of screening the energy consumption abnormal points is high.

Description

TECHNICAL FIELD

[0001] The present application relates to automobile thermal management, in particular to a thermal management energy consumption anomaly identification method and device based on working condition clustering and a medium. BACKGROUND

[0002] How to accurately identify thermal management high energy consumption anomaly is a problem that thermal management research focuses on. The traditional technical solution adopts a fixed threshold judgment method, which specifically is to preset a fixed energy consumption threshold, and it is considered that high energy consumption of thermal management occurs when the energy consumption exceeds the energy consumption threshold. However, since when the vehicle is in a high-speed, climbing and other working conditions that have high requirements on vehicle driving performance, even if the energy consumption exceeds the energy consumption threshold, it is not necessarily an abnormal high energy consumption, and in the low-speed or other working conditions without high requirements on driving performance, even if the energy consumption does not exceed the energy consumption threshold, it is not necessarily normal high energy consumption. The fixed threshold judgment method cannot effectively integrate vehicle multi-condition data to accurately identify thermal management abnormal energy consumption, and has problems such as insufficient adaptability to complex conditions and inability to comprehensively compare different driver behaviors. SUMMARY

[0003] Therefore, the purpose of the present application is to provide a thermal management energy consumption anomaly identification method and device based on working condition clustering, which fully considers the normal fluctuation range of thermal management energy consumption under different working conditions, effectively avoids misjudgment and omission, and has high efficiency in screening energy consumption anomaly points.

[0004] The thermal management energy consumption anomaly identification method based on working condition clustering in the present application comprises: Obtaining historical trip slice data, using a clustering algorithm to classify vehicle driving working conditions, and generating a working condition classification model; Obtaining historical thermal management energy consumption data, wherein the historical thermal management energy consumption data comprises a plurality of energy consumption points; Classifying the working condition of each energy consumption point based on the working condition classification model; Cross-comparing each energy consumption point and its corresponding working condition to screen the energy consumption anomaly points of the vehicle.

[0005] Further, the obtaining of the historical trip slice data, the use of the clustering algorithm to classify the vehicle driving working conditions, and the generation of the working condition classification model comprise: Obtaining historical trip slice data, wherein the historical trip slice data comprises driving behavior data and external environment data; Segmenting the indicators of the driving behavior data and classifying the indicators of the external environment data; Performing dummy variable processing and normalization processing on the historical trip slice data to obtain data points; Randomly selecting k initial cluster centers; The distance of each data point of the historical travel slice data to each cluster center is calculated, each data point is assigned to the nearest cluster center, and a new cluster center is calculated according to the data points in the cluster; if the cluster center result changes, the step is repeated, and if the cluster center does not change, the working condition classification model is output.

[0006] Further, the driving behavior data includes vehicle speed and braking intensity, and the external environment data includes road type.

[0007] Further, the classification of the working condition in which each energy consumption point is located based on the working condition classification model includes: Obtaining travel slice data to be classified corresponding to the collection time of each energy consumption point; Performing dummy variable processing and normalization processing on the travel slice data to be classified to obtain data points; Inputting the data points of each travel slice data to be classified into the working condition classification model to obtain the working condition classification to which the travel slice data belongs.

[0008] Further, the cross comparison of each energy consumption point and its corresponding working condition to screen the energy consumption abnormal point of the vehicle includes: Performing set operation on energy consumption points with the same working condition classification to form a plurality of energy consumption point sets corresponding to the working condition classification; Determining the abnormal threshold value corresponding to each working condition classification based on the energy consumption point values under the same set; Comparing each energy consumption point with the abnormal threshold value corresponding to its working condition classification, if the energy consumption point value is not less than the corresponding abnormal threshold value, the energy consumption point is determined as the energy consumption abnormal point of the vehicle; if the energy consumption point value is less than the corresponding abnormal threshold value, the energy consumption point is determined as not being the energy consumption abnormal point of the vehicle.

[0009] Further, the determination of the abnormal threshold value corresponding to each working condition classification based on the energy consumption point values under the same set includes: Based on the energy consumption point values under the same set, the energy consumption average value and the energy consumption standard deviation of the energy consumption points in the set are calculated respectively; Based on the energy consumption average value and the energy consumption standard deviation of the corresponding set of each working condition classification, the abnormal threshold value corresponding to each working condition classification is determined.

[0010] Further, the energy consumption average value is μ, , the energy consumption standard deviation is σ, , and the abnormal threshold value is a, ; Wherein, n is the number of energy consumption points in the set, X1, X2,..., Xn are respectively the energy consumption point values in the set.

[0011] Further, the method further comprises verifying the trip data and optimizing the working condition classification model, specifically comprising: obtaining trip detail data of a certain time period before and after the energy consumption abnormal point of the vehicle; analyzing whether the judgment of the energy consumption abnormal point is correct based on the driving behavior data and external environment data of the trip detail data; if not, re-generate the working condition classification model, and if correct, do not re-generate the working condition classification model.

[0012] An apparatus in the present application, the apparatus comprises one or more processors; a storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, the device realizes the above-mentioned working condition clustering-based thermal management energy consumption anomaly identification method.

[0013] A computer readable storage medium in the present application, characterized in that a computer program is stored thereon, when the computer program is executed by the processor of the computer, the computer executes the above-mentioned working condition clustering-based thermal management energy consumption anomaly identification method.

[0014] The beneficial effects of the present application are: the present application optimizes the traditional fixed threshold judgment method to the cross comparison of the energy consumption point and its corresponding working condition in the present embodiment, fully considers the normal fluctuation range of the thermal management energy consumption under different working conditions, effectively avoids misjudgment and omission, and has high efficiency in screening energy consumption abnormal points, can adapt to various driving scenes and environmental conditions, has strong universality and adaptability. And, by verifying the trip data and optimizing the working condition classification model, the accuracy of the working condition classification model can be optimized and improved. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to make the purpose, technical scheme and beneficial effects of the present application clearer, the present application provides the following drawings for illustration: Figure 1 The flowchart of the working condition clustering-based thermal management energy consumption anomaly identification method of the present application; Figure 2 The detailed flowchart of step S1 of the working condition clustering-based thermal management energy consumption anomaly identification method of the present application; Figure 3 The detailed flowchart of step S4 of the working condition clustering-based thermal management energy consumption anomaly identification method of the present application; Figure 4 The working condition classification diagram of the present application; Figure 5 The cross judgment diagram of the thermal management energy consumption data and the working condition of the present application. DETAILED DESCRIPTION

[0016] The technical solutions of the present application will be described in detail below with reference to the drawings and embodiments.

[0017] As shown in the embodiment, the heat management energy consumption anomaly identification method based on working condition clustering includes the following steps: Figures 1-3 S1, obtain historical trip slice data, and classify vehicle driving conditions by using a clustering algorithm to generate a working condition classification model.

[0018] The clustering algorithm can be a Kmeans algorithm, and step S1 specifically includes: S101, obtain historical trip slice data, which includes driving behavior data and external environment data.

[0019] S102, segment the indicators of the driving behavior data and classify the indicators of the external environment data.

[0020] The classification and segmentation refer to historical driving data statistics and CLTC domestic working condition standards. The driving behavior data includes vehicle speed and braking intensity. The vehicle speed can be obtained from the data collected by the vehicle speed sensor or from the navigation information. The braking intensity is the average deceleration of the vehicle, which can be obtained from the data collected by the vehicle speed sensor. The external environment data includes road type, which can be obtained from the data collected by the external camera or from the navigation information. The historical trip slice data can be data uploaded by multiple vehicles stored in the cloud database, or data uploaded by a certain vehicle stored in the cloud database. The working condition classification diagram is shown in Figure 4 The indicators of the driving behavior data are segmented, i.e., the vehicle speed and braking intensity are divided into multiple segments, such as <20km / h, 20~60km / h, 60~90km / h, and 90~150km / h for vehicle speed; and <-15m / s 2 , -15~-30m / s 2 for braking intensity. The road type can be urban road, township road, highway, and mountain. At this time, k ≤32.

[0021] S103, perform dummy variable processing and normalization processing on the historical trip slice data to obtain data points.

[0022] S104, randomly select k initial cluster centers.

[0023] The number of working condition types is k, and k < speed segmentation number x braking intensity segmentation number x road classification number. For example, if in step S102, the vehicle speed is divided into <20km / h, 20~60km / h, 60~90km / h, and 90~150km / h; and the braking intensity is divided into <-15m / s 2 , -15~-30m / s​2 The road type can be: urban road, town road, highway, mountain. At this time, k < 32. Some working conditions do not exist in actual driving process, such as urban road high-speed driving, town road high-speed driving, etc., so usually k is less than the number of speed segments x the number of braking intensity segments x the number of road classification.

[0024] The vehicle is determined to be low-speed, medium-speed, medium-high-speed, high-speed driving by the vehicle speed, and the vehicle is determined to be smooth driving or intense driving by the braking intensity. After being combined with the road type, it is the working condition classification, such as urban road low-speed smooth driving, highway medium-speed intense driving, town road low-speed smooth driving, and the like.

[0025] S105, calculate the distance from each data point of the historical trip slice data to each cluster center, assign each data point to the nearest cluster center of the cluster, and calculate a new cluster center according to the data points in the cluster. If the cluster center result changes, repeat this step S105. If the cluster center does not change, output the working condition classification model.

[0026] The cluster center result is shown in the following table. Qualitative working condition classification is determined according to the statistical results of each index: S2, obtain historical thermal management energy consumption data, the historical thermal management energy consumption data including a plurality of energy consumption points. The thermal management energy consumption of the vehicle needing to be judged can be read within a certain period (for example, the past 7 days) of the vehicle.

[0027] S3, classify the working condition of each energy consumption point based on the working condition classification model.

[0028] Step S3 specifically includes: S301, obtain the to-be-classified trip slice data corresponding to the collection time of each energy consumption point.

[0029] S302, perform dummy variable processing and normalization processing on the to-be-classified trip slice data to obtain data points.

[0030] S303, input the data points of each to-be-classified trip slice data into the working condition classification model to obtain the working condition classification to which the trip slice data belongs.

[0031] By inputting the data points of the to-be-classified trip slice data corresponding to the collection time of the energy consumption point into the working condition classification model, the working condition classification corresponding to the data point can be obtained according to the result of the cluster center assigned thereto.

[0032] S4, cross-comparison of each energy consumption point and its corresponding working condition, screening of vehicle energy consumption abnormal points.

[0033] Step S4 specifically comprises: S401, performing set operation on the energy consumption points of the same working condition classification to form a set of energy consumption points corresponding to each working condition classification.

[0034] S402, determining the abnormal threshold corresponding to each working condition classification based on the energy consumption point values in the same set.

[0035] Step S402 specifically comprises: S4021, calculating the energy consumption average and energy consumption standard deviation of the energy consumption points in the set based on the energy consumption point values in the same set; S4022, determining the abnormal threshold corresponding to each working condition classification based on the energy consumption average and energy consumption standard deviation of the set corresponding to each working condition classification. Specifically, the energy consumption average is μ, , the energy consumption standard deviation is σ, , and the abnormal threshold is a, ; wherein n is the number of energy consumption points in the set, X1, X2,..., Xn are respectively the values of each energy consumption point in the set.

[0036] S403, comparing each energy consumption point with the abnormal threshold corresponding to its working condition classification, if the energy consumption point value is not less than the corresponding abnormal threshold, the energy consumption point is determined as an energy consumption abnormal point of the vehicle; if the energy consumption point value is less than the corresponding abnormal threshold, the energy consumption point is determined as not an energy consumption abnormal point of the vehicle. The heat management energy consumption data and working condition cross judgment schematic diagram is shown in Figure 5 , the heat energy consumption management values of the working conditions such as high-speed highway high-intensity driving and mountainous medium-speed intense driving are all relatively high, but only a part of the abnormally high points are taken as energy consumption abnormal points.

[0037] S5, verifying the trip data and optimizing the working condition classification model.

[0038] Specifically comprising: S501, obtaining the trip detail data of a certain time period (for example, half an hour before and after the energy consumption abnormal point) before and after the energy consumption abnormal point of the vehicle.

[0039] S502, analyzing whether the judgment of the energy consumption abnormal point is correct based on the driving behavior data and external environment data of the trip detail data; if not, re-generate the working condition classification model, if correct, do not re-generate the working condition classification model.

[0040] By manually analyzing the driving behavior and external environment of the driver near the energy consumption abnormal point within this period of time, it is analyzed whether the judgment of these energy consumption abnormal points is correct. If not, return to step S1 to re-generate the working condition classification model, and when re-generating, the segmentation of vehicle speed and braking intensity can be further subdivided in step S102, for example, the braking intensity is subdivided from the original two segments: <-15m / s 2-15~-30 m / s 2 -10~-20 m / s 2 -15~-30 m / s 2 -15~-30 m / s 2 Then, the operating condition classification model is regenerated to achieve optimized operating condition clustering results. If correct, it indicates that the operating condition classification model is qualified, and the steps S2-S4 can be repeated to screen the energy consumption abnormal points of different vehicles, achieving accurate and rapid screening of energy consumption abnormal points.

[0041] The screening of energy consumption abnormal points can be applied to vehicle performance analysis, driving habit analysis and other research scenarios. By optimizing the traditional fixed threshold judgment method to the cross comparison of energy consumption points and their corresponding operating conditions in this embodiment, the normal fluctuation range of thermal management energy consumption under different operating conditions is fully considered, effectively avoiding misjudgment and omission, and the efficiency of screening energy consumption abnormal points is high, which can adapt to various driving scenes and environmental conditions, has strong universality and adaptability. Moreover, by verifying the trip data and optimizing the operating condition classification model, the accuracy of the operating condition classification model can be optimized and improved.

[0042] A device in this embodiment includes one or more processors; A storage device is configured to store one or more programs, which, when executed by the one or more processors, cause the device to implement the above-mentioned thermal management energy consumption abnormality identification method based on operating condition clustering.

[0043] A computer-readable storage medium in this embodiment is characterized in that a computer program is stored thereon, which, when executed by a processor of a computer, causes the computer to execute the above-mentioned thermal management energy consumption abnormality identification method based on operating condition clustering.

[0044] Finally, it should be pointed out that the above embodiments are only used to illustrate the technical solutions of the present application and are not limiting. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the purpose and scope of the present application, which should be covered by the claims of the present application.

Claims

1. A method for identifying anomalies in thermal management energy consumption based on operating condition clustering, characterized in that, include: Historical trip slice data is obtained, and a clustering algorithm is used to classify vehicle driving conditions to generate a condition classification model; Acquire historical thermal management energy consumption data, which includes multiple energy consumption points; The operating conditions of each energy consumption point are classified based on the operating condition classification model. Cross-compare each energy consumption point and its corresponding operating condition to screen out abnormal energy consumption points in the vehicle.

2. The thermal management energy consumption anomaly identification method based on operating condition clustering according to claim 1, characterized in that, The process of acquiring historical trip slice data and using clustering algorithms to classify vehicle driving conditions to generate a driving condition classification model includes: Acquire historical trip slice data, which includes driving behavior data and external environment data; The indicators for the driving behavior data are segmented, and the indicators for the external environment data are categorized. The historical travel slice data is processed by dummy variable processing and normalization to obtain data points; Randomly select k initial cluster centers; For each data point in the historical travel slice data, calculate the distance to each cluster center, assign each data point to the nearest cluster center, and calculate a new cluster center based on the data points within the cluster. If the cluster center result changes, repeat this step; if the cluster center does not change, output the working condition classification model.

3. The thermal management energy consumption anomaly identification method based on operating condition clustering according to claim 2, characterized in that: The driving behavior data includes vehicle speed and braking intensity, and the external environment data includes road type.

4. The thermal management energy consumption anomaly identification method based on operating condition clustering according to claim 1, characterized in that, The classification of the operating conditions of each energy consumption point based on the operating condition classification model includes: Obtain the travel slice data to be classified corresponding to the collection time of each of the energy consumption points; The dummy variable processing and normalization processing are performed on the trip slice data to be classified to obtain data points; Input each data point of the travel slice data to be classified into the working condition classification model to obtain the working condition classification to which the travel slice data belongs.

5. The thermal management energy consumption anomaly identification method based on operating condition clustering according to claim 1, characterized in that, The process of cross-comparing each energy consumption point and its corresponding operating condition to screen for abnormal energy consumption points in the vehicle includes: Perform a set operation on energy consumption points with the same operating condition classification to form multiple sets of energy consumption points corresponding to different operating condition classifications; Based on the energy consumption point values ​​under the same set, determine the abnormal threshold corresponding to each operating condition category; Each energy consumption point is compared with the abnormal threshold corresponding to its respective operating condition category. If the energy consumption point value is not less than the corresponding abnormal threshold, the energy consumption point is determined to be an abnormal energy consumption point of the vehicle; if the energy consumption point value is less than the corresponding abnormal threshold, the energy consumption point is determined not to be an abnormal energy consumption point of the vehicle.

6. The thermal management energy consumption anomaly identification method based on operating condition clustering according to claim 5, characterized in that, The determination of the abnormal threshold corresponding to each operating condition category based on energy consumption point values ​​within the same set includes: Based on the energy consumption point values ​​within the same set, calculate the average energy consumption and standard deviation of energy consumption for each energy consumption point in the set. Based on the average energy consumption and standard deviation of energy consumption of the corresponding sets for each operating condition category, the abnormal threshold corresponding to each operating condition category is determined.

7. The thermal management energy consumption anomaly identification method based on operating condition clustering according to claim 6, characterized in that: The average energy consumption is μ. The standard deviation of the energy consumption is σ. The abnormal threshold is a. ; Where n is the number of energy consumption points in the set, and X1, X2, ..., Xn are the values ​​of each energy consumption point in the set.

8. The thermal management energy consumption anomaly identification method based on operating condition clustering according to claim 1, characterized in that, The method also includes verifying travel data and optimizing the operating condition classification model, specifically including: Obtain detailed travel data for a certain time period before and after the vehicle's energy consumption anomaly point; The analysis is based on the driving behavior data and external environment data of the trip details to determine whether the judgment of energy consumption anomalies is correct; if incorrect, the operating condition classification model is regenerated; if correct, the operating condition classification model is not regenerated.

9. A device, characterized in that: The device includes one or more processors; A storage device for storing one or more programs, which, when executed by one or more processors, cause the device to implement the thermal management energy consumption anomaly identification method based on operating condition clustering as described in any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed by the computer's processor, causes the computer to perform the thermal management energy consumption anomaly identification method based on operating condition clustering as described in any one of claims 1 to 8.

Images