Fuel consumption abnormity identification method and device, equipment, storage medium and product

By extracting the vehicle status data and inputting it into the pre-training model, the fuel consumption difference value is calculated to determine the abnormal fuel consumption, the problem of poor fuel consumption recognition accuracy in the prior art is solved, and higher recognition accuracy and robustness are achieved .

CN119089185APending Publication Date: 2024-12-06SINO TRUK JINAN POWER CO LTD
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202411107160.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-13
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

In the prior art, the accuracy and robustness of fuel consumption abnormality recognition are poor, and it is impossible to effectively adapt to complex and changeable driving scenarios and driving modes.

Method used

By collecting vehicle status data and actual fuel consumption values, feature extraction and input into the pre-trained fuel consumption prediction model, calculate the fuel consumption difference value to determine the abnormal fuel consumption, and set the abnormal fuel consumption threshold using the fault data set to improve the accuracy and reliability of identification Great.

Benefits of technology

It improves the accuracy and robustness of fuel consumption abnormality recognition, can more effectively adapt to different driving scenarios and modes, and timely output alarm signals to prompt the driver.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119089185A_ABST
    Figure CN119089185A_ABST
Patent Text Reader

Abstract

The invention provides a fuel consumption abnormity identification method and device, equipment, a storage medium and a product. The method comprises the following steps: acquiring current vehicle state data and a current actual fuel consumption value; wherein the vehicle state data comprises road condition data, driving behavior data and engine operation data; performing feature extraction on the current vehicle state data to obtain current fuel consumption feature data; inputting the current fuel consumption characteristic data into a fuel consumption prediction model to obtain a current predicted fuel consumption value output by the fuel consumption prediction model; calculating a difference value result between the current actual fuel consumption value and the current fuel consumption predicted value to obtain a current fuel consumption difference value; if the oil consumption difference value is larger than the abnormal oil consumption threshold value, it is judged that the current oil consumption of the vehicle is abnormal. According to the scheme of the invention, the accuracy and robustness of fuel consumption abnormity identification are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of intelligent transportation technology, and in particular to a method, device, equipment, storage medium and product for identifying abnormal fuel consumption. Background Art

[0002] As cars become more popular and their usage increases, the monitoring and management of vehicle fuel consumption becomes increasingly important. Abnormal fuel consumption not only affects the economy and driving performance of the vehicle, but may also indicate potential faults or problems.

[0003] In actual applications, in order to meet the driving needs of users, the driving scenarios and driving modes of cars are diversified. At present, the recognition of abnormal fuel consumption relies on static thresholds or manually set rules, which cannot adapt well to complex and changeable driving scenarios and driving modes.

[0004] Therefore, in the above-mentioned prior art, the accuracy and robustness of abnormal fuel consumption identification are relatively poor. Summary of the invention

[0005] The present invention provides a fuel consumption abnormality identification method, device, equipment, storage medium and product, which are used to solve the defects of poor accuracy and robustness of fuel consumption abnormality identification in the prior art, and improve the accuracy and robustness of fuel consumption identification.

[0006] The present invention provides a method for identifying abnormal fuel consumption, comprising: collecting and obtaining current vehicle status data and current actual fuel consumption value; wherein the vehicle status data includes road condition data, driving behavior data and engine operation data; performing feature extraction on the current vehicle status data to obtain current fuel consumption feature data; inputting the current fuel consumption feature data into a fuel consumption prediction model to obtain a current predicted fuel consumption value output by the fuel consumption prediction model; wherein the fuel consumption prediction model is a machine learning model obtained by pre-training a training set based on a normal fuel consumption data set and testing a test set; the training set includes a plurality of historical fuel consumption feature data and an actual normal fuel consumption value corresponding to each historical fuel consumption feature data; the test set includes a plurality of historical fuel consumption feature data and an actual normal fuel consumption value corresponding to each historical fuel consumption feature data, and is used to test the current fuel consumption prediction model, and when the accuracy of the fuel consumption prediction model on the test set reaches a satisfactory result, the fuel consumption prediction model is obtained; calculating the difference result between the current actual fuel consumption value and the current fuel consumption prediction value to obtain a current fuel consumption difference value; if the fuel consumption difference value is greater than an abnormal fuel consumption threshold, it is determined that the current fuel consumption of the vehicle is abnormal.

[0007] According to a method for identifying fuel consumption anomalies provided by the present invention, before collecting and obtaining the current vehicle status data and the current actual fuel consumption value, the method also includes: obtaining a fault data set, the fault data set including a plurality of historical abnormal fuel consumption characteristic data and an actual fuel consumption value corresponding to each historical fuel consumption characteristic data; inputting each historical abnormal fuel consumption characteristic data into the fuel consumption prediction model to obtain a predicted fuel consumption value corresponding to each historical abnormal fuel consumption data output by the fuel consumption prediction model; calculating a difference result between the actual fuel consumption value corresponding to each historical fuel consumption characteristic data and the predicted fuel consumption value corresponding to each historical abnormal fuel consumption data to obtain a historical fuel consumption difference value corresponding to each historical fuel consumption characteristic data; and setting the abnormal fuel consumption threshold value by a fractional method based on all historical fuel consumption difference values.

[0008] According to a method for identifying fuel consumption anomalies provided by the present invention, before collecting and obtaining the current vehicle status data and the current actual fuel consumption value, the method also includes: constructing a normal fuel consumption data set, and dividing the normal fuel consumption data set to obtain the training set and the test set; wherein the normal fuel consumption data set includes a plurality of historical feature data within a historical period, and each of the historical feature data includes historical fuel consumption feature data and the actual normal fuel consumption value corresponding to the historical fuel consumption feature data; constructing an initial fuel consumption prediction model based on a machine learning model; training the initial fuel consumption prediction model based on the training set to obtain a current fuel consumption prediction model; and testing the current fuel consumption prediction model based on the test set until the current fuel consumption prediction model meets the requirements to obtain the fuel consumption prediction model.

[0009] According to a method for identifying abnormal fuel consumption provided by the present invention, the construction of a normal fuel consumption data set includes: acquiring multiple historical data within a historical period; wherein each historical data includes historical vehicle status data and an actual fuel consumption value corresponding to the historical vehicle status data; the historical vehicle status data includes historical road condition data, historical driving behavior data, and historical engine operation data; for each historical data, performing feature extraction processing to obtain the normal fuel consumption data set; wherein the feature extraction processing includes: performing feature extraction on the historical vehicle status data to obtain historical fuel consumption feature data; and using the actual fuel consumption value corresponding to the historical vehicle status data as the actual normal fuel consumption value corresponding to the historical fuel consumption feature data.

[0010] According to a method for identifying abnormal fuel consumption provided by the present invention, before performing feature extraction processing on each piece of historical data to obtain the normal fuel consumption data set, the method also includes: preprocessing multiple pieces of historical data within the historical period; the preprocessing includes at least one of the following: data cleaning, outlier processing, missing value filling and normalization processing.

[0011] According to a method for identifying abnormal fuel consumption provided by the present invention, if the fuel consumption difference value is greater than the abnormal fuel consumption threshold, after determining that the current fuel consumption of the vehicle is abnormal, the method also includes: outputting an alarm signal, and recording the current vehicle status data and the current actual fuel consumption value.

[0012] The present invention also provides a fuel consumption anomaly identification device, comprising: an acquisition module, used to acquire current vehicle status data and current actual fuel consumption value; wherein the vehicle status data includes road condition data, driving behavior data and engine operation data; a feature extraction module, used to extract features from the current vehicle status data to acquire current fuel consumption feature data; a processing module, used to input the current fuel consumption feature data into a fuel consumption prediction model to acquire the current predicted fuel consumption value output by the fuel consumption prediction model; wherein the fuel consumption prediction model is pre-trained based on a training set of a normal fuel consumption data set and tested with a test set, and the obtained machine learning The model comprises a training set including a plurality of historical fuel consumption characteristic data and an actual normal fuel consumption value corresponding to each historical fuel consumption characteristic data; the test set comprises a plurality of historical fuel consumption characteristic data and an actual normal fuel consumption value corresponding to each historical fuel consumption characteristic data, and is used to test the current fuel consumption prediction model. When the accuracy of the fuel consumption prediction model on the test set reaches a satisfactory result, the fuel consumption prediction model is obtained; a calculation module is used to calculate the difference between the current actual fuel consumption value and the current fuel consumption prediction value to obtain the current fuel consumption difference value; and a judgment module is used to judge that the current fuel consumption of the vehicle is abnormal if the fuel consumption difference value is greater than an abnormal fuel consumption threshold.

[0013] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, any of the above-mentioned methods for identifying abnormal fuel consumption is implemented.

[0014] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the fuel consumption abnormality identification method as described in any one of the above is implemented.

[0015] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any of the above-mentioned methods for identifying abnormal fuel consumption.

[0016] In the fuel consumption abnormality identification method, device, equipment, storage medium and product provided by the present invention, feature extraction is performed on the current vehicle status data collected to obtain the current fuel consumption feature data; the current fuel consumption feature data is input into the fuel consumption prediction model to obtain the current predicted fuel consumption value output by the fuel consumption prediction model; wherein, the vehicle status data includes road condition data, driving behavior data and engine operation data, and the vehicle status data can characterize the current driving behavior, road condition and driving scene, so the current predicted fuel consumption value has high accuracy and robustness. Further, it is determined whether the difference between the current actual fuel consumption value and the current predicted fuel consumption value is greater than the abnormal fuel consumption threshold; it is determined whether the difference between the current actual fuel consumption value and the current predicted fuel consumption value is large; if the fuel consumption difference value is greater than the abnormal fuel consumption threshold, it is determined that the current fuel consumption of the vehicle is abnormal, thereby improving the accuracy and robustness of abnormal fuel consumption identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 This is one of the flow charts of the fuel consumption abnormality identification method provided by the present invention.

[0019] Figure 2 This is the second flow chart of the fuel consumption abnormality identification method provided by the present invention.

[0020] Figure 3 It is a structural schematic diagram of the fuel consumption abnormality identification device provided by the present invention.

[0021] Figure 4 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] The following specific embodiments are used to describe in detail the technical solution of the present application and how the technical solution of the present application solves the above technical problems. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The fuel consumption prediction method of the present invention is described below in conjunction with the accompanying drawings.

[0024] In practical applications, the execution subject of the fuel consumption abnormality identification method can be a fuel consumption abnormality identification device, and there are many ways to implement the fuel consumption abnormality identification device, for example, it can be implemented through a computer program, such as application software, etc.; or, for example, a chip, etc. It can also be implemented as a medium storing relevant computer programs, such as a USB disk, a cloud disk, etc.; or, it can also be implemented through a physical device integrated or installed with relevant computer programs, such as a server, etc.

[0025] Figure 1 is one of the flow charts of the method for identifying abnormal fuel consumption provided by the present invention, such as Figure 1 As shown, the method includes steps 101 to 105.

[0026] Step 101: Collect and obtain current vehicle status data and current actual fuel consumption value; wherein the vehicle status data includes road condition data, driving behavior data and engine operation data.

[0027] As an example, the fuel consumption abnormality identification device can be integrated into the vehicle, for example, the fuel consumption abnormality identification device can be integrated into the vehicle controller to ensure that the fuel consumption abnormality identification device can collect the current vehicle status data and the current actual fuel consumption value in real time.

[0028] The road condition data represents the road condition of the vehicle. As an example, the road condition data includes at least one of the following: slope, road flatness, road section type (e.g., urban highway, expressway, suburban road section, etc.). The driving behavior data includes at least one of the following: vehicle speed, acceleration, braking force, gear position, throttle opening. The engine operation data includes at least one of the following: speed, torque, fuel injection amount.

[0029] It should be noted that this embodiment does not specifically limit the data types included in the vehicle status data. The above-mentioned vehicle status data includes road condition data, driving behavior data, and engine operation data, which are only used as examples. In actual applications, the data types in the vehicle status data can be increased or decreased according to the vehicle driving mode and vehicle driving scene.

[0030] It can be understood that the vehicle status data includes road condition data, driving behavior data and engine operation data. The vehicle status data can characterize the current driving behavior, road conditions and driving scenarios. Therefore, in this implementation, driving behavior, road conditions and driving scenarios are comprehensively considered to identify fuel consumption anomalies.

[0031] Step 102: extract features from the current vehicle status data to obtain current fuel consumption feature data.

[0032] In practical applications, there are many ways to implement feature extraction, for example, based on principal components analysis (PCA), attention mechanism, multilayer perceptron (MLP) neural network, etc., a feature extraction model is constructed, and the current vehicle state data is subjected to feature extraction through the feature extraction model to obtain the current fuel consumption feature data. In this embodiment, the implementation method of feature extraction is not specifically limited.

[0033] Combined with the above description, the vehicle status data includes road condition data, driving behavior data and engine operation data, that is, the vehicle status data includes multi-dimensional data that characterizes the vehicle driving mode and vehicle driving scene. Taking into account the differences between different data types, feature extraction can be performed on the current road condition data, the current driving behavior data and the current engine operation data to obtain the current road condition feature data, the current driving behavior feature data and the current engine feature data. Furthermore, the current road condition feature data, the current driving behavior feature data and the current engine feature data are feature fused to obtain the current fuel consumption feature data.

[0034] In practical applications, in order to obtain accurate fuel consumption characteristic data and thus improve the accuracy of fuel consumption anomaly identification, the current vehicle status data may be preprocessed. As an example, in a possible implementation, before step 102, the fuel consumption anomaly identification method further includes: preprocessing the current vehicle status data; the preprocessing includes at least one of the following: data cleaning, outlier processing, missing value filling, and normalization processing.

[0035] In one example, the current vehicle status data may be subject to noise interference, and the current vehicle status data may be cleaned. For example, a low-pass filter may be used to remove data noise. In another example, the current vehicle status data may contain outliers, which may cause serious deviations in the abnormal fuel consumption identification results. Deletion, mean correction, regression interpolation, multiple interpolation and other methods may be used to process the outliers. In another example, the current vehicle status data may have data missing problems, and front filling, back filling, mean filling, mode filling, median filling and other methods may be used to fill the missing values. In another example, different types of data in the current vehicle fuel consumption data are in different ranges, and normalization processing may be applied to limit the current vehicle fuel consumption data to a certain range. It should be noted that the above-mentioned preprocessing methods may be implemented separately or in combination.

[0036] Step 103: input the current fuel consumption characteristic data into the fuel consumption prediction model to obtain the current predicted fuel consumption value output by the fuel consumption prediction model.

[0037] The fuel consumption prediction model is a machine learning model obtained by pre-training a training set based on a normal fuel consumption data set and testing a test set. The training set includes multiple historical fuel consumption feature data and the actual normal fuel consumption value corresponding to each historical fuel consumption feature data. The test set includes multiple historical fuel consumption feature data and the actual normal fuel consumption value corresponding to each historical fuel consumption feature data, which is used to test the current fuel consumption prediction model. When the accuracy of the fuel consumption prediction model on the test set reaches a satisfactory result, the fuel consumption prediction model is obtained.

[0038] In this embodiment, the fuel consumption prediction model is a machine learning model pre-trained based on a training set. For example, the fuel consumption prediction model can be a BP (full name Back Propagation) neural network model pre-trained based on a training set. For another example, the fuel consumption prediction model can also be a gated recurrent neural network (Gate Recurrent Unit, GUR) model pre-trained based on a training set.

[0039] Specifically, the input of the fuel consumption prediction model is the current fuel consumption feature data, and the output of the fuel consumption prediction model is the current predicted fuel consumption value. As an example, the fuel consumption prediction model can be a BP neural network model pre-trained based on a training set. The activation functions of the hidden layer and the output layer of the BP neural network model are set to ReLU (full name Rectified Linear Unit) function and Softplus function respectively.

[0040] It can be understood that feature extraction is performed on the current vehicle status data collected to obtain current fuel consumption feature data; the current fuel consumption feature data is input into the fuel consumption prediction model to obtain the current predicted fuel consumption value output by the fuel consumption prediction model; wherein, the vehicle status data includes road condition data, driving behavior data and engine operation data, and the vehicle status data can characterize the current driving behavior, road conditions and driving scenarios, and therefore, the current predicted fuel consumption value has high accuracy and robustness.

[0041] Step 104: Calculate the difference between the current actual fuel consumption value and the current fuel consumption prediction value to obtain the current fuel consumption difference value.

[0042] Step 105: If the fuel consumption difference value is greater than the abnormal fuel consumption threshold, it is determined that the current fuel consumption of the vehicle is abnormal.

[0043] Among them, the abnormal fuel consumption threshold is used to determine whether the vehicle's fuel consumption is abnormal. If the current fuel consumption difference value is greater than the abnormal fuel consumption threshold, it means that the current actual fuel consumption value is significantly different from the current predicted fuel consumption value, and the vehicle's current fuel consumption is determined to be abnormal.

[0044] In this embodiment, the difference between the current actual fuel consumption value and the current predicted fuel consumption value is calculated to obtain the current fuel consumption difference value. Further, by comparing the magnitude relationship between the current fuel consumption difference value and the abnormal fuel consumption threshold, it is determined whether the fuel consumption is abnormal; if the current fuel consumption difference value is greater than the abnormal fuel consumption threshold, it means that the current actual fuel consumption value is significantly different from the current predicted fuel consumption value, and the current fuel consumption of the vehicle is determined to be abnormal, thereby improving the accuracy and robustness of abnormal fuel consumption identification.

[0045] In practical applications, before determining whether the current fuel consumption of the vehicle is abnormal based on the abnormal fuel consumption threshold, the abnormal fuel consumption threshold is predetermined. In one example, the abnormal fuel consumption threshold can be set according to the actual abnormal fuel consumption identification requirements. For example, for vehicles with higher accuracy in identifying abnormal fuel consumption, the abnormal fuel consumption threshold is set to a smaller value. In another example, the abnormal fuel consumption threshold can be obtained based on statistical calculations of the fault data set.

[0046] As for the method for determining the abnormal fuel consumption threshold, as an example, in a possible implementation, before step 101, the above-mentioned fuel consumption abnormality identification method also includes: obtaining a fault data set, the fault data set includes multiple historical abnormal fuel consumption characteristic data and the actual fuel consumption value corresponding to each historical fuel consumption characteristic data; inputting each historical abnormal fuel consumption characteristic data into the fuel consumption prediction model, and obtaining the predicted fuel consumption value corresponding to each historical abnormal fuel consumption data output by the fuel consumption prediction model; calculating the difference between the actual fuel consumption value corresponding to each historical fuel consumption characteristic data and the predicted fuel consumption value corresponding to each historical abnormal fuel consumption data, and obtaining the historical fuel consumption difference value corresponding to each historical fuel consumption characteristic data; and setting the abnormal fuel consumption threshold by a fractional method based on all historical fuel consumption difference values.

[0047] In practical applications, the fault data set is constructed based on multiple abnormal historical data collected in the historical period under abnormal fuel consumption conditions, and each abnormal historical data includes historical vehicle status data and actual fuel consumption values ​​corresponding to the historical vehicle status data. Specifically, the historical period can be set according to actual needs. For example, the historical period is 3 months before the current month. Further, feature extraction is performed on the historical vehicle status data in each abnormal historical data to obtain abnormal fuel consumption feature data corresponding to each historical vehicle status data. It can be understood that both the abnormal fuel consumption feature data and the actual fuel consumption value correspond to the historical vehicle status data. Therefore, based on the historical vehicle status data, a corresponding relationship between the historical fuel consumption feature data and the actual fuel consumption value is established to construct a fault data set. The fault data set includes multiple historical abnormal fuel consumption feature data and the actual fuel consumption value corresponding to each historical fuel consumption feature data.

[0048] Furthermore, each historical abnormal fuel consumption characteristic data in the fault data set is input into the fuel consumption prediction model, and the fuel consumption prediction model outputs the predicted fuel consumption value corresponding to each historical abnormal fuel consumption characteristic data. The difference between the actual fuel consumption value corresponding to each historical fuel consumption characteristic data and the predicted fuel consumption value corresponding to each historical abnormal fuel consumption data is calculated to obtain the historical fuel consumption difference value corresponding to each historical fuel consumption characteristic data.

[0049] Furthermore, according to the requirements for abnormal fuel consumption identification, the abnormal fuel consumption threshold can be set by calculating the percentile of the fuel consumption difference value. For example, if it is expected to identify 95% of the abnormal fuel consumption data, all the historical fuel consumption difference values ​​are arranged in order of size, and the abnormal fuel consumption threshold is set by calculating the percentile of the fuel consumption difference value, and the historical fuel consumption difference value corresponding to the 95th percentile is used as the abnormal fuel consumption threshold.

[0050] In this implementation, an abnormal fuel consumption threshold is obtained based on statistical calculations of the fault data set. On this basis, by comparing the current fuel consumption difference value with the abnormal fuel consumption threshold, it is determined whether the fuel consumption is abnormal, thereby improving the accuracy and robustness of abnormal fuel consumption identification.

[0051] In combination with the above description, the fuel consumption prediction model is a machine learning model pre-trained based on the training set. For the training process of the fuel consumption prediction model, as an example, in a possible implementation, Figure 2 This is a second flow chart of the method for identifying abnormal fuel consumption provided by the present invention, such as Figure 2 As shown, before the above step 101, the above abnormal fuel consumption identification method also includes steps 201 to 204.

[0052] Step 201: construct a normal fuel consumption data set, and divide the normal fuel consumption data set into a training set and a test set.

[0053] The normal fuel consumption data set includes a plurality of historical characteristic data in a historical period, and each piece of historical characteristic data includes historical fuel consumption characteristic data and an actual normal fuel consumption value corresponding to the historical fuel consumption characteristic data.

[0054] Specifically, for the process of constructing a normal fuel consumption data set, in one example, the construction of the normal fuel consumption data set in the above step 201 includes: obtaining multiple historical data within a historical period; wherein each historical data includes historical vehicle status data and actual fuel consumption values ​​corresponding to the historical vehicle status data; the historical vehicle status data includes historical road condition data, historical driving behavior data, and historical engine operation data. For each historical data, feature extraction processing is performed to obtain a normal fuel consumption data set; wherein the feature extraction processing includes: extracting features from the historical vehicle status data to obtain historical fuel consumption feature data; and using the actual fuel consumption values ​​corresponding to the historical vehicle status data as the actual normal fuel consumption values ​​corresponding to the historical fuel consumption feature data.

[0055] Step 202: construct an initial fuel consumption prediction model based on the machine learning model.

[0056] Step 203: Based on the training set, the initial fuel consumption prediction model is trained to obtain the current fuel consumption prediction model.

[0057] Step 204: Based on the test set, the current fuel consumption prediction model is tested until the current fuel consumption prediction model meets the requirements, thereby obtaining a fuel consumption prediction model.

[0058] In practical applications, the training set is used to train the current fuel consumption prediction model, and the test set is used to verify whether the current fuel consumption prediction model meets the requirements. If it meets the requirements, the training is stopped to obtain the fuel consumption prediction model; if it does not meet the requirements, the current fuel consumption prediction model is adjusted. For example, the gradient descent method can be used to adjust the parameters of the current fuel consumption prediction model, and after the adjustment, the training is continued based on the training set until the current fuel consumption prediction model meets the requirements and the fuel consumption prediction model is obtained.

[0059] In addition, in order to obtain a fuel consumption prediction model that meets the requirements and thus improve the accuracy of abnormal fuel consumption identification, multiple historical data in the historical period can be preprocessed before constructing the normal fuel consumption data set. As an example, in a possible implementation, the above-mentioned feature extraction processing is performed on each historical data, and before obtaining the normal fuel consumption data set, the above-mentioned fuel consumption abnormality identification method also includes: preprocessing multiple historical data in the historical period; the preprocessing includes at least one of the following: data cleaning, outlier processing, missing value filling, and normalization processing.

[0060] In one example, multiple historical data within a historical period may have noise interference, and each historical data can be cleaned. For example, a low-pass filter can be used to remove data noise. In another example, multiple historical data within a historical period may have outliers, which will cause serious deviations in the abnormal fuel consumption identification results. Deletion, mean correction, regression interpolation, multiple interpolation methods, etc. can be used to process the outliers. In another example, multiple historical data within a historical period may have data missing problems. Front filling, back filling, mean filling, mode filling, median filling and other methods can be used to fill the missing values. In another example, different types of data in the multiple historical data within a historical period are in different ranges. Normalization processing can be applied to limit the multiple historical data within the historical period to a certain range. It should be noted that the above-mentioned preprocessing methods can be implemented separately or in combination.

[0061] In this embodiment, before performing feature extraction processing on each piece of historical data to obtain a normal fuel consumption data set, multiple pieces of historical data within the historical period are preprocessed, which improves the accuracy of the normal fuel consumption data set and can obtain a fuel consumption prediction model that meets the requirements, thereby improving the accuracy of abnormal fuel consumption identification.

[0062] In addition, in order to further ensure the safety and reliability of the vehicle, after determining that the current fuel consumption of the vehicle is abnormal, an alarm mode can be triggered to remind the driver to handle it in time. As an example, in a possible implementation, after the above step 105, the fuel consumption abnormality identification method also includes: outputting an alarm signal, and recording the current vehicle status data and the current actual fuel consumption value.

[0063] In this embodiment, after determining that the vehicle's current fuel consumption is abnormal, an alarm signal is output, and the current vehicle status data and the current actual fuel consumption value are recorded, which can prompt the driver to deal with it in time. Moreover, the driver can determine the cause of the vehicle's abnormal fuel consumption in time based on the recorded current vehicle status data and the current actual fuel consumption value, thereby improving the safety and reliability of the vehicle.

[0064] In this embodiment, feature extraction is performed on the current vehicle status data collected to obtain the current fuel consumption feature data; the current fuel consumption feature data is input into the fuel consumption prediction model to obtain the current predicted fuel consumption value output by the fuel consumption prediction model; wherein, the vehicle status data includes road condition data, driving behavior data and engine operation data, and the vehicle status data can characterize the current driving behavior, road condition and driving scene, therefore, the current predicted fuel consumption value has high accuracy and robustness. Further, it is determined whether the difference between the current actual fuel consumption value and the current predicted fuel consumption value is greater than the abnormal fuel consumption threshold; it is determined whether the difference between the current actual fuel consumption value and the current predicted fuel consumption value is large; if the fuel consumption difference value is greater than the abnormal fuel consumption threshold, it is determined that the current fuel consumption of the vehicle is abnormal. The solution of this embodiment improves the accuracy and robustness of abnormal fuel consumption identification.

[0065] The fuel consumption abnormality identification device provided by the present invention is described below. The fuel consumption abnormality identification device described below and the fuel consumption abnormality identification method described above can be referenced to each other.

[0066] Figure 3 is a schematic diagram of the structure of the fuel consumption abnormality identification device provided by the present invention, such as Figure 3 As shown, the fuel consumption abnormality identification device includes: a collection module 31, a feature extraction module 32, a processing module 33, a calculation module 34 and a determination module 35.

[0067] The acquisition module 31 is used to acquire the current vehicle status data and the current actual fuel consumption value; wherein the vehicle status data includes road condition data, driving behavior data and engine operation data.

[0068] As an example, the fuel consumption abnormality identification device can be integrated into the vehicle, for example, the fuel consumption abnormality identification device can be integrated into the vehicle controller to ensure that the fuel consumption abnormality identification device can collect the current vehicle status data and the current actual fuel consumption value in real time.

[0069] The road condition data represents the road condition of the vehicle. As an example, the road condition data includes at least one of the following: slope, road flatness, road section type (e.g., urban highway, expressway, suburban road section, etc.). The driving behavior data includes at least one of the following: vehicle speed, acceleration, braking force, gear position, throttle opening. The engine operation data includes at least one of the following: speed, torque, fuel injection amount.

[0070] It should be noted that this embodiment does not specifically limit the data types included in the vehicle status data. The above-mentioned vehicle status data includes road condition data, driving behavior data, and engine operation data, which are only used as examples. In actual applications, the data types in the vehicle status data can be increased or decreased according to the vehicle driving mode and vehicle driving scene.

[0071] It can be understood that the vehicle status data includes road condition data, driving behavior data and engine operation data. The vehicle status data can characterize the current driving behavior, road conditions and driving scenarios. Therefore, in this implementation, driving behavior, road conditions and driving scenarios are comprehensively considered to identify fuel consumption anomalies.

[0072] The feature extraction module 32 is used to extract features from the current vehicle status data to obtain current fuel consumption feature data.

[0073] In practical applications, there are many ways to implement feature extraction, for example, based on principal components analysis (PCA), attention mechanism, multilayer perceptron (MLP) neural network, etc., a feature extraction model is constructed, and the current vehicle state data is subjected to feature extraction through the feature extraction model to obtain the current fuel consumption feature data. In this embodiment, the implementation method of feature extraction is not specifically limited.

[0074] Combined with the above description, the vehicle status data includes road condition data, driving behavior data and engine operation data, that is, the vehicle status data includes multi-dimensional data that characterizes the vehicle driving mode and the vehicle driving scene. Taking into account the differences between different data types, the feature extraction module 32 can perform feature extraction on the current road condition data, the current driving behavior data and the current engine operation data, respectively, to obtain the current road condition feature data, the current driving behavior feature data and the current engine feature data. Furthermore, the feature extraction module 32 performs feature fusion on the current road condition feature data, the current driving behavior feature data and the current engine feature data to obtain the current fuel consumption feature data.

[0075] In practical applications, in order to obtain accurate fuel consumption characteristic data and thus improve the accuracy of fuel consumption anomaly identification, the current vehicle status data can be preprocessed. As an example, in a possible implementation, the above fuel consumption anomaly identification device further includes: a first preprocessing module, which is used to preprocess the current vehicle status data; the preprocessing includes at least one of the following: data cleaning, outlier processing, missing value filling, and normalization processing.

[0076] In one example, the current vehicle status data may be subject to noise interference, and the first preprocessing module may perform data cleaning on the current vehicle status data. For example, a low-pass filter may be used to remove data noise. In another example, the current vehicle status data may contain outliers, which may cause serious deviations in the abnormal fuel consumption identification results. The first preprocessing module may apply deletion, mean correction, regression interpolation, multiple interpolation and other methods to process the outliers. In another example, the current vehicle status data may have data missing problems, and the first preprocessing module may apply front filling, back filling, mean filling, mode filling, median filling and other methods to fill in the missing values. In another example, different types of data in the current vehicle fuel consumption data are in different ranges, and the first preprocessing module may apply normalization processing to limit the current vehicle fuel consumption data to a certain range. It should be noted that the above-mentioned preprocessing methods can be implemented separately or in combination.

[0077] The processing module 33 is used to input the current fuel consumption characteristic data into the fuel consumption prediction model to obtain the current predicted fuel consumption value output by the fuel consumption prediction model; wherein the fuel consumption prediction model is a machine learning model obtained by pre-training a training set based on a normal fuel consumption data set and testing a test set. The training set includes a plurality of historical fuel consumption characteristic data and an actual normal fuel consumption value corresponding to each historical fuel consumption characteristic data. The test set includes a plurality of historical fuel consumption characteristic data and an actual normal fuel consumption value corresponding to each historical fuel consumption characteristic data, and is used to test the current fuel consumption prediction model. When the accuracy of the fuel consumption prediction model on the test set reaches a satisfactory result, the fuel consumption prediction model is obtained.

[0078] In this embodiment, the fuel consumption prediction model is a machine learning model pre-trained based on a training set. For example, the fuel consumption prediction model can be a BP (full name Back Propagation) neural network model pre-trained based on a training set. For another example, the fuel consumption prediction model can also be a gated recurrent neural network (Gate Recurrent Unit, GUR) model pre-trained based on a training set.

[0079] Specifically, the input of the fuel consumption prediction model is the current fuel consumption feature data, and the output of the fuel consumption prediction model is the current predicted fuel consumption value. As an example, the fuel consumption prediction model can be a BP neural network model pre-trained based on a training set. The activation functions of the hidden layer and the output layer of the BP neural network model are set to ReLU (full name Rectified Linear Unit) function and Softplus function respectively.

[0080] It can be understood that the feature extraction module 32 performs feature extraction on the collected current vehicle status data to obtain the current fuel consumption feature data; the processing module 33 obtains the current predicted fuel consumption value output by the fuel consumption prediction model by inputting the current fuel consumption feature data into the fuel consumption prediction model; wherein, the vehicle status data includes road condition data, driving behavior data and engine operation data, and the vehicle status data can characterize the current driving behavior, road conditions and driving scenarios, therefore, the current predicted fuel consumption value has high accuracy and robustness.

[0081] The calculation module 34 is used to calculate the difference between the current actual fuel consumption value and the current fuel consumption prediction value to obtain the current fuel consumption difference value.

[0082] The determination module 35 is configured to determine that the current fuel consumption of the vehicle is abnormal if the fuel consumption difference value is greater than the abnormal fuel consumption threshold value.

[0083] Among them, the abnormal fuel consumption threshold is used to determine whether the vehicle's fuel consumption is abnormal. If the current fuel consumption difference value is greater than the abnormal fuel consumption threshold, it means that the current actual fuel consumption value is significantly different from the current predicted fuel consumption value, then the determination module 35 determines that the vehicle's current fuel consumption is abnormal.

[0084] In this embodiment, the calculation module 34 calculates the difference between the current actual fuel consumption value and the current predicted fuel consumption value to obtain the current fuel consumption difference value. Further, the determination module 35 determines whether the fuel consumption is abnormal by comparing the magnitude relationship between the current fuel consumption difference value and the abnormal fuel consumption threshold value; if the current fuel consumption difference value is greater than the abnormal fuel consumption threshold value, it means that the current actual fuel consumption value is significantly different from the current predicted fuel consumption value, and the current fuel consumption of the vehicle is determined to be abnormal, thereby improving the accuracy and robustness of abnormal fuel consumption identification.

[0085] In practical applications, before determining whether the current fuel consumption of the vehicle is abnormal based on the abnormal fuel consumption threshold, the abnormal fuel consumption threshold is predetermined. In one example, the abnormal fuel consumption threshold can be set according to the actual abnormal fuel consumption identification requirements. For example, for vehicles with higher accuracy in identifying abnormal fuel consumption, the abnormal fuel consumption threshold is set to a smaller value. In another example, the abnormal fuel consumption threshold can be obtained based on statistical calculations of the fault data set.

[0086] As for the method of determining the abnormal fuel consumption threshold, as an example, in a possible implementation, the above-mentioned fuel consumption abnormality identification device also includes: a determination module; the determination module is used to: obtain a fault data set, the fault data set includes multiple historical abnormal fuel consumption characteristic data and an actual fuel consumption value corresponding to each historical fuel consumption characteristic data; input each historical abnormal fuel consumption characteristic data into the fuel consumption prediction model, and obtain the predicted fuel consumption value corresponding to each historical abnormal fuel consumption data output by the fuel consumption prediction model; calculate the difference between the actual fuel consumption value corresponding to each historical fuel consumption characteristic data and the predicted fuel consumption value corresponding to each historical abnormal fuel consumption data, and obtain the historical fuel consumption difference value corresponding to each historical fuel consumption characteristic data; based on all historical fuel consumption difference values, set the abnormal fuel consumption threshold by a fractional method.

[0087] In actual applications, the fault data set is constructed based on multiple abnormal historical data collected in the historical period under abnormal fuel consumption conditions, and each abnormal historical data includes historical vehicle status data and actual fuel consumption values ​​corresponding to the historical vehicle status data. Specifically, the historical period can be set according to actual needs. Furthermore, the determination module performs feature extraction on the historical vehicle status data in each abnormal historical data to obtain abnormal fuel consumption feature data corresponding to each historical vehicle status data. It can be understood that both the abnormal fuel consumption feature data and the actual fuel consumption value correspond to the historical vehicle status data. Therefore, the determination module establishes the correspondence between the historical fuel consumption feature data and the actual fuel consumption value based on the historical vehicle status data, and constructs a fault data set. The fault data set includes multiple historical abnormal fuel consumption feature data and the actual fuel consumption value corresponding to each historical fuel consumption feature data.

[0088] Furthermore, the determination module inputs each historical abnormal fuel consumption characteristic data in the fault data set into the fuel consumption prediction model, and obtains the predicted fuel consumption value corresponding to each historical abnormal fuel consumption characteristic data output by the fuel consumption prediction model. The determination module calculates the difference between the actual fuel consumption value corresponding to each historical fuel consumption characteristic data and the predicted fuel consumption value corresponding to each historical abnormal fuel consumption data, and obtains the historical fuel consumption difference value corresponding to each historical fuel consumption characteristic data.

[0089] Furthermore, the determination module can set the abnormal fuel consumption threshold by calculating the percentile of the fuel consumption difference value according to the fuel consumption abnormality identification requirements. For example, if it is expected to identify 95% of the abnormal fuel consumption data, all the historical fuel consumption difference values ​​are arranged in order of size, and the abnormal fuel consumption threshold is set by calculating the percentile of the fuel consumption difference value, and the historical fuel consumption difference value corresponding to the 95th percentile is used as the abnormal fuel consumption threshold.

[0090] In this embodiment, the determination module obtains the abnormal fuel consumption threshold value based on the statistical calculation of the fault data set. On this basis, by comparing the current fuel consumption difference value with the abnormal fuel consumption threshold value, it is judged whether the fuel consumption is abnormal, thereby improving the accuracy and robustness of abnormal fuel consumption identification.

[0091] In combination with the above description, the fuel consumption prediction model is a machine learning model pre-trained based on a training set. As an example, in a possible implementation, the fuel consumption abnormality identification device further includes a first construction module, a second construction module, a training module, and a test module.

[0092] The above-mentioned first construction module is used to construct a normal fuel consumption data set, and divide the normal fuel consumption data set into a training set and a test set; wherein the normal fuel consumption data set includes multiple historical feature data within a historical period, and each historical feature data includes historical fuel consumption feature data and the actual normal fuel consumption value corresponding to the historical fuel consumption feature data.

[0093] Specifically, for the construction process of the normal fuel consumption data set, in one example, when the above-mentioned first construction module is used to construct the normal fuel consumption data set, it is specifically used to: obtain multiple historical data within a historical period; wherein each historical data includes historical vehicle status data and actual fuel consumption values ​​corresponding to the historical vehicle status data; the historical vehicle status data includes historical road condition data, historical driving behavior data, and historical engine operation data. For each historical data, perform feature extraction processing to obtain a normal fuel consumption data set; wherein the feature extraction processing includes: extracting features from the historical vehicle status data to obtain historical fuel consumption feature data; and using the actual fuel consumption values ​​corresponding to the historical vehicle status data as the actual normal fuel consumption values ​​corresponding to the historical fuel consumption feature data.

[0094] The second building module is used to build an initial fuel consumption prediction model based on a machine learning model.

[0095] The training module is used to train the initial fuel consumption prediction model based on the training set to obtain the current fuel consumption prediction model.

[0096] The above-mentioned test module is used to test the current fuel consumption prediction model based on the test set until the current fuel consumption prediction model meets the requirements and obtains the fuel consumption prediction model.

[0097] In practical applications, the training set is used to train the current fuel consumption prediction model, and the test set is used to verify whether the current fuel consumption prediction model meets the requirements. If it meets the requirements, the training is stopped to obtain the fuel consumption prediction model; if it does not meet the requirements, the current fuel consumption prediction model is adjusted. For example, the gradient descent method can be used to adjust the parameters of the current fuel consumption prediction model, and after the adjustment, the training is continued based on the training set until the current fuel consumption prediction model meets the requirements and the fuel consumption prediction model is obtained.

[0098] In addition, in order to obtain a fuel consumption prediction model that meets the requirements and thus improve the accuracy of fuel consumption anomaly identification, multiple historical data in the historical period can be preprocessed before constructing the normal fuel consumption data set. As an example, in a possible implementation, the above-mentioned fuel consumption anomaly identification device also includes: a second preprocessing module, which is used to preprocess multiple historical data in the historical period; the preprocessing includes at least one of the following: data cleaning, outlier processing, missing value filling, and normalization processing.

[0099] In one example, multiple historical data within a historical period may be subject to noise interference, and the second preprocessing module may perform data cleaning on each piece of historical data. For example, a low-pass filter may be used to remove data noise. In another example, multiple historical data within a historical period may have outliers, which may cause serious deviations in the abnormal fuel consumption identification results. The second preprocessing module may use deletion, mean correction, regression interpolation, multiple interpolation and other methods to process the outliers. In another example, multiple historical data within a historical period may have data missing problems, and the second preprocessing module may use front filling, back filling, mean filling, mode filling, median filling and other methods to fill in the missing values. In another example, different types of data in the multiple historical data within a historical period are within different ranges, and the second preprocessing module may use normalization processing to limit the multiple historical data within the historical period to a certain range. It should be noted that the above preprocessing methods can be implemented separately or in combination.

[0100] In this embodiment, before the first construction module performs feature extraction processing on each piece of historical data to obtain a normal fuel consumption data set, the second preprocessing module preprocesses multiple pieces of historical data within the historical period, thereby improving the accuracy of the normal fuel consumption data set and being able to obtain a fuel consumption prediction model that meets the requirements, thereby improving the accuracy of fuel consumption anomaly identification.

[0101] In addition, in order to further ensure the safety and reliability of the vehicle, after determining that the current fuel consumption of the vehicle is abnormal, an alarm mode can be triggered to remind the driver to deal with it in time. As an example, in a possible implementation, the above-mentioned fuel consumption abnormality identification device also includes: an alarm module, which is used to output an alarm signal and record the current vehicle status data and the current actual fuel consumption value.

[0102] In this embodiment, after determining that the vehicle's current fuel consumption is abnormal, the alarm module outputs an alarm signal and records the current vehicle status data and the current actual fuel consumption value, which can prompt the driver to deal with it in time. Moreover, the driver can determine the cause of the vehicle's abnormal fuel consumption in time based on the recorded current vehicle status data and the current actual fuel consumption value, thereby improving the safety and reliability of the vehicle.

[0103] In this embodiment, the feature extraction module extracts features from the current vehicle state data collected by the collection module to obtain the current fuel consumption feature data; the processing module obtains the current predicted fuel consumption value output by the fuel consumption prediction model by inputting the current fuel consumption feature data into the fuel consumption prediction model; wherein, the vehicle state data includes road condition data, driving behavior data and engine operation data, and the vehicle state data can characterize the current driving behavior, road condition and driving scene, therefore, the current predicted fuel consumption value has high accuracy and robustness. Further, the determination module determines whether the difference between the current actual fuel consumption value and the current predicted fuel consumption value is greater than the abnormal fuel consumption threshold; determines whether the difference between the current actual fuel consumption value and the current predicted fuel consumption value is large; if the fuel consumption difference value is greater than the abnormal fuel consumption threshold, the current fuel consumption of the vehicle is determined to be abnormal. The solution of this embodiment improves the accuracy and robustness of abnormal fuel consumption identification.

[0104] Figure 4 is a schematic diagram of the structure of the electronic device provided by the present invention, such as Figure 4As shown, the electronic device may include: a processor (processor) 410, a communication interface (Communications Interface) 420, a memory (memory) 430 and a communication bus 440, wherein the processor 410, the communication interface 420, and the memory 430 communicate with each other through the communication bus 440. The processor 410 can call the logic instructions in the memory 430 to execute the fuel consumption abnormality identification method, which includes: collecting and obtaining the current vehicle status data and the current actual fuel consumption value; wherein the vehicle status data includes road condition data, driving behavior data and engine operation data; performing feature extraction on the current vehicle status data to obtain the current fuel consumption feature data; inputting the current fuel consumption feature data into the fuel consumption prediction model to obtain the current predicted fuel consumption value output by the fuel consumption prediction model; wherein the fuel consumption prediction model is pre-trained based on a training set of a normal fuel consumption data set, and is tested with a test set, The obtained machine learning model, the training set includes multiple historical fuel consumption characteristic data and the actual normal fuel consumption value corresponding to each historical fuel consumption characteristic data; the test set includes multiple historical fuel consumption characteristic data and the actual normal fuel consumption value corresponding to each historical fuel consumption characteristic data, which are used to test the current fuel consumption prediction model. When the accuracy of the fuel consumption prediction model on the test set reaches a satisfactory result, the fuel consumption prediction model is obtained; the difference between the current actual fuel consumption value and the current fuel consumption prediction value is calculated to obtain the current fuel consumption difference value; if the fuel consumption difference value is greater than the abnormal fuel consumption threshold, it is determined that the current fuel consumption of the vehicle is abnormal.

[0105] In addition, the logic instructions in the above-mentioned memory 430 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0106] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the fuel consumption abnormality identification method provided by the above methods, which includes: collecting and obtaining current vehicle status data and current actual fuel consumption value; wherein the vehicle status data includes road condition data, driving behavior data and engine operation data; performing feature extraction on the current vehicle status data to obtain current fuel consumption feature data; inputting the current fuel consumption feature data into a fuel consumption prediction model to obtain the current predicted fuel consumption value output by the fuel consumption prediction model; wherein the fuel The fuel consumption prediction model is a machine learning model obtained by pre-training a training set based on a normal fuel consumption data set and testing a test set. The training set includes multiple historical fuel consumption feature data and the actual normal fuel consumption value corresponding to each historical fuel consumption feature data; the test set includes multiple historical fuel consumption feature data and the actual normal fuel consumption value corresponding to each historical fuel consumption feature data, which are used to test the current fuel consumption prediction model. When the accuracy of the fuel consumption prediction model on the test set reaches a satisfactory result, the fuel consumption prediction model is obtained; the difference between the current actual fuel consumption value and the current fuel consumption prediction value is calculated to obtain the current fuel consumption difference value; if the fuel consumption difference value is greater than the abnormal fuel consumption threshold, it is determined that the current fuel consumption of the vehicle is abnormal.

[0107] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the fuel consumption anomaly identification method provided by the above-mentioned methods, the method comprising: collecting and obtaining current vehicle status data and current actual fuel consumption value; wherein the vehicle status data includes road condition data, driving behavior data and engine operation data; performing feature extraction on the current vehicle status data to obtain current fuel consumption feature data; inputting the current fuel consumption feature data into a fuel consumption prediction model to obtain the current predicted fuel consumption value output by the fuel consumption prediction model; wherein the fuel consumption prediction model is based on a normal fuel consumption data set The machine learning model is pre-trained with a training set and tested with a test set, wherein the training set includes multiple historical fuel consumption feature data and the actual normal fuel consumption value corresponding to each historical fuel consumption feature data; the test set includes multiple historical fuel consumption feature data and the actual normal fuel consumption value corresponding to each historical fuel consumption feature data, which are used to test the current fuel consumption prediction model. When the accuracy of the fuel consumption prediction model on the test set reaches a satisfactory result, the fuel consumption prediction model is obtained; the difference between the current actual fuel consumption value and the current fuel consumption prediction value is calculated to obtain the current fuel consumption difference value; if the fuel consumption difference value is greater than the abnormal fuel consumption threshold, it is determined that the current fuel consumption of the vehicle is abnormal.

[0108] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative effort.

[0109] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0110] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying abnormal fuel consumption, characterized in that: include: Collect and obtain current vehicle status data and current actual fuel consumption value; wherein the vehicle status data includes road condition data, driving behavior data and engine operation data; Extracting features from the current vehicle status data to obtain current fuel consumption feature data; The current fuel consumption characteristic data is input into the fuel consumption prediction model to obtain the current predicted fuel consumption value output by the fuel consumption prediction model; wherein the fuel consumption prediction model is a machine learning model obtained by pre-training a training set based on a normal fuel consumption data set and testing a test set; the training set includes a plurality of historical fuel consumption characteristic data and an actual normal fuel consumption value corresponding to each historical fuel consumption characteristic data; the test set includes a plurality of historical fuel consumption characteristic data and an actual normal fuel consumption value corresponding to each historical fuel consumption characteristic data, and is used to test the current fuel consumption prediction model, and when the accuracy of the fuel consumption prediction model on the test set reaches a satisfactory result, the fuel consumption prediction model is obtained; Calculating the difference between the current actual fuel consumption value and the current fuel consumption prediction value to obtain a current fuel consumption difference value; If the fuel consumption difference value is greater than the abnormal fuel consumption threshold, it is determined that the current fuel consumption of the vehicle is abnormal.

2. The method for identifying abnormal fuel consumption according to claim 1, characterized in that: Before acquiring the current vehicle status data and the current actual fuel consumption value, the method further includes: Acquire a fault data set, wherein the fault data set includes a plurality of historical abnormal fuel consumption characteristic data and an actual fuel consumption value corresponding to each historical fuel consumption characteristic data; Inputting each historical abnormal fuel consumption feature data into the fuel consumption prediction model to obtain a predicted fuel consumption value corresponding to each historical abnormal fuel consumption data output by the fuel consumption prediction model; Calculating the difference between the actual fuel consumption value corresponding to each historical fuel consumption characteristic data and the predicted fuel consumption value corresponding to each historical abnormal fuel consumption data to obtain the historical fuel consumption difference value corresponding to each historical fuel consumption characteristic data; The abnormal fuel consumption threshold is set by a fractional method based on all historical fuel consumption difference values.

3. The method for identifying abnormal fuel consumption according to claim 1, characterized in that: Before acquiring the current vehicle status data and the current actual fuel consumption value, the method further includes: Constructing a normal fuel consumption data set, and dividing the normal fuel consumption data set into the training set and the test set; wherein the normal fuel consumption data set includes a plurality of historical feature data in a historical period, and each of the historical feature data includes historical fuel consumption feature data and an actual normal fuel consumption value corresponding to the historical fuel consumption feature data; Build an initial fuel consumption prediction model based on the machine learning model; Based on the training set, the initial fuel consumption prediction model is trained to obtain a current fuel consumption prediction model; Based on the test set, the current fuel consumption prediction model is tested until the current fuel consumption prediction model meets the requirements, thereby obtaining the fuel consumption prediction model.

4. The method for identifying abnormal fuel consumption according to claim 3, characterized in that: The constructing of the normal fuel consumption data set includes: Acquire multiple pieces of historical data within a historical period; wherein each piece of historical data includes historical vehicle status data and actual fuel consumption values ​​corresponding to the historical vehicle status data; the historical vehicle status data includes historical road condition data, historical driving behavior data, and historical engine operation data; For each piece of historical data, feature extraction processing is performed to obtain the normal fuel consumption data set; wherein the feature extraction processing includes: extracting features from historical vehicle status data to obtain historical fuel consumption feature data; and using the actual fuel consumption value corresponding to the historical vehicle status data as the actual normal fuel consumption value corresponding to the historical fuel consumption feature data.

5. The method for identifying abnormal fuel consumption according to claim 4, characterized in that: Before performing feature extraction processing on each piece of historical data to obtain the normal fuel consumption data set, the method further includes: Preprocessing is performed on multiple historical data within the historical period; the preprocessing includes at least one of the following: data cleaning, outlier processing, missing value filling and normalization processing.

6. The method for identifying abnormal fuel consumption according to any one of claims 1 to 5, characterized in that: If the fuel consumption difference value is greater than the abnormal fuel consumption threshold, after determining that the current fuel consumption of the vehicle is abnormal, the method further includes: Output an alarm signal, and record the current vehicle status data and the current actual fuel consumption value.

7. A fuel consumption abnormality identification device, characterized in that: include: A collection module, used to collect and obtain current vehicle status data and current actual fuel consumption value; wherein the vehicle status data includes road condition data, driving behavior data and engine operation data; A feature extraction module, used to extract features from the current vehicle state data to obtain current fuel consumption feature data; A processing module, used for inputting the current fuel consumption characteristic data into a fuel consumption prediction model, and obtaining a current predicted fuel consumption value output by the fuel consumption prediction model; wherein the fuel consumption prediction model is a machine learning model obtained by pre-training a training set based on a normal fuel consumption data set and testing a test set, wherein the training set includes a plurality of historical fuel consumption characteristic data and an actual normal fuel consumption value corresponding to each historical fuel consumption characteristic data; the test set includes a plurality of historical fuel consumption characteristic data and an actual normal fuel consumption value corresponding to each historical fuel consumption characteristic data, and is used for testing the current fuel consumption prediction model, and when the accuracy of the fuel consumption prediction model on the test set reaches a satisfactory result, the fuel consumption prediction model is obtained; A calculation module, used for calculating the difference between the current actual fuel consumption value and the current fuel consumption prediction value to obtain a current fuel consumption difference value; The determination module is configured to determine that the current fuel consumption of the vehicle is abnormal if the fuel consumption difference value is greater than an abnormal fuel consumption threshold.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the fuel consumption abnormality identification method as described in any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for identifying abnormal fuel consumption as claimed in any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for identifying abnormal fuel consumption as claimed in any one of claims 1 to 6 is implemented.

Citation Information

Patent Citations

  • Vehicle operation condition evaluation method and device, computer equipment and storage medium

    CN111523701A

  • Fuel consumption monitoring method, fuel consumption monitoring device and engineering vehicle

    CN113624291A

  • BP neural network fuel consumption prediction method based on fused driving behavior characteristics

    CN113642227A

  • Early warning processing method and device for mechanical equipment and computer readable storage medium

    CN116665421A