A data analysis method for fault prediction and health management of automotive electronic control systems

Abstract

The present invention relates to the field of data processing technology, and specifically to a data analysis method for automobile electronic control system fault prediction and health management. The present invention obtains the fluctuation amplitude by analyzing the speed difference between a data point in a speed time series and its adjacent data points before and after it; obtains the change trend by analyzing the data point and the reference data points selected before and after it; obtains the abnormality degree according to the fluctuation amplitude and the change trend; screens out suspected abnormal data points according to the abnormality degree; analyzes the relationship between the abnormality degree of the suspected abnormal data point and the data point after it, and obtains the engine stutter probability of the suspected abnormal data point in combination with the number of data points in the suspected abnormal time series connected domain; screens out the engine stutter data points in the suspected abnormal data points according to the engine stutter probability, and obtains the abnormal data points in the speed time series; and predicts the automobile electronic control system fault according to the abnormal data points. The method of the present invention improves the accuracy of screening out abnormal data generated by engine stutter.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a data analysis method for fault prediction and health management of an automobile electronic control system. Background Art

[0002] The automotive electronic control system is a very important part of modern cars. It covers many aspects, including the engine control unit (ECU), sensors, actuators, and communication with various vehicle systems. The air flow sensor is responsible for monitoring the engine intake volume. If the air flow sensor fails, the engine control system will not be able to accurately adjust the air-fuel ratio, causing unstable engine speed. Frequent fluctuations in engine speed will increase the wear of transmission systems, engines and other parts, shorten their service life, and increase maintenance and replacement costs. Therefore, for equipment safety, it is necessary to collect engine speed data for real-time monitoring and analysis, timely and accurate identification of abnormal conditions and early warning.

[0003] In the prior art, the fluctuation of engine speed data is generally analyzed to obtain abnormal data, and then the automobile electronic control system fault is predicted based on the abnormal data. However, when the vehicle accelerates, decelerates, or shifts gears, the engine speed will also fluctuate greatly. Therefore, a single analysis of the fluctuation of engine speed data will produce a large error in the prediction of automobile electronic control system faults. Summary of the invention

[0004] In order to solve the above technical problems, the purpose of the present invention is to provide a data analysis method for fault prediction and health management of an automotive electronic control system.

[0005] The present application provides a data analysis method for fault prediction and health management of an automotive electronic control system, and the technical solution adopted is as follows:

[0006] Collecting the engine's running speed time series data to obtain a speed time series sequence;

[0007] Analyze the speed difference between a data point in the speed time series and its adjacent data points before and after it to obtain the fluctuation amplitude of the data point;

[0008] Analyze the data points in the speed time series and the reference data points selected before and after them to obtain the change trend of the data points;

[0009] Obtaining the abnormality degree of the data point according to the fluctuation amplitude and the change trend;

[0010] According to the abnormality degree, suspected abnormal data points are screened out;

[0011] Analyze the relationship between the suspected abnormal data point and the abnormality degree of the subsequent data points, and obtain the engine stall probability of the suspected abnormal data point in combination with the number of data points in the suspected abnormal time series connected domain;

[0012] According to the engine jerking probability, filter out the engine jerking data points in the suspected abnormal data points to obtain the abnormal data points in the speed time series;

[0013] Based on the abnormal data points, the failure of the automobile electronic control system is predicted.

[0014] In some embodiments of the present invention, analyzing the speed difference between a data point in the speed time series and its adjacent data points before and after to obtain the fluctuation amplitude of the data point includes:

[0015] Calculate the first Data points and The first speed difference between the data points , and Data points and The second speed difference between the data points , and then calculate the absolute value of the difference between the first speed difference and the second speed difference , normalize the absolute value data of the difference to obtain the first The fluctuation range of a data point.

[0016] In some embodiments of the present invention, analyzing the data points in the speed time series and the reference data points selected before and after them to obtain the change trend of the data points includes:

[0017] The least square method is used to perform straight-line fitting on the data points in the speed time series and the reference data points selected before and after them to obtain a trend line of the data points; based on the trend line, the slope value of the trend line is obtained; the slope value is the changing trend of the data points.

[0018] In some embodiments of the present invention, the method for acquiring the reference data point is to select a reference data point of the data point according to the fluctuation amplitude.

[0019] In some embodiments of the present invention, obtaining the abnormality degree of the data point according to the fluctuation amplitude and the change trend includes:

[0020] Averaging the fluctuation amplitudes of all the reference data points to obtain an average fluctuation amplitude of the reference data points;

[0021] The abnormality degree of the data point is obtained according to the fluctuation amplitude of the data point, the average fluctuation amplitude of the reference data points, and the slope value.

[0022] In some embodiments of the present invention, the calculation formula of the abnormality degree is:

[0023]

[0024] In the formula: Indicates the first The degree of abnormality of the data point, Indicates the first The fluctuation range of the data points, Indicates the first Around the data point The fluctuation range of the selected reference data points, Indicates the first The number of reference data points selected around the data point, Indicates the first The slope value of the trend line corresponding to the data point, represents the linear normalization function.

[0025] In some embodiments of the present invention, the engine stall probability is: calculating the sum of the differences between the abnormality degree of the suspected abnormal data point and the abnormality degrees of multiple data points thereafter, multiplying the sum of the differences by the inverse of the number of data points in the suspected abnormal time series connectivity domain, and then linearly normalizing the product value.

[0026] In some embodiments of the present invention, the suspected abnormal time series connected domain is a suspected abnormal time series connected domain composed of continuous suspected abnormal data points.

[0027] In some embodiments of the present invention, according to the engine jerking probability, filtering out the engine jerking data points in the suspected abnormal data points to obtain the abnormal data points in the speed time series includes:

[0028] comparing the engine stall probability with a preset engine stall threshold;

[0029] The engine jerking probability is greater than a preset engine jerking threshold, the suspected abnormal data point is an engine jerking data point, and the engine jerking data point is screened out;

[0030] The engine jerking probability is less than or equal to a preset engine jerking threshold, and the suspected abnormal data point is an abnormal data point.

[0031] In some embodiments of the present invention, predicting an electric control system failure according to the abnormal data point includes:

[0032] Comparing the number of abnormal data points with a preset threshold value of the number of abnormal vehicle data;

[0033] When the number of abnormal data points is greater than or equal to a preset threshold value of the number of abnormal data of the automobile, it is determined that there is a fault in the automobile electronic control system and a fault light is turned on.

[0034] Compared with the prior art, the data analysis method for automobile electronic control system fault prediction and health management provided by the present invention has the following beneficial effects:

[0035] In order to identify abnormal data points of engine speed, the present invention obtains a speed time series sequence by collecting the engine's operating speed time series data, obtains the fluctuation amplitude of the data point by quantifying the speed difference between the data point in the speed time series and its adjacent data points before and after it; obtains the degree of abnormality of the data point by analyzing the fluctuation amplitude of the data point within the reference range corresponding to each data point and the overall change trend of the speed data within the current reference range of each data point; obtains abnormal data points by analyzing the degree of abnormality of each data point and screening out special engine stuttering conditions, providing good data support for subsequent fault prediction of the automobile's electronic control system. In order to screen out points with a large degree of abnormality caused by engine stuttering, the present method quantifies and compares the speed at which the degree of abnormality is reduced by utilizing the characteristics of engine stuttering that the degree of abnormality is large but the time is short, thereby improving the accuracy of screening. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only 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.

[0037] Figure 1 A schematic diagram of the basic flow of a data analysis method for fault prediction and health management of an automotive electronic control system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0038] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the data analysis method for fault prediction and health management of an automotive electronic control system proposed by the present invention, its specific implementation method, structure, characteristics and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0039] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0040] When the automobile engine is running, the abnormality of the automobile electronic control system causes the abnormal engine speed. Frequent engine speed fluctuations will increase the wear of the transmission system, engine and other parts. Therefore, the purpose of the present invention is to find the abnormal engine speed data points caused by the abnormal automobile electronic control system by analyzing the automobile engine speed time series data.

[0041] In order to achieve the above purpose, the present invention provides a data analysis method for fault prediction and health management of automobile electronic control systems. The overall logic of the data analysis method is: obtain the fluctuation amplitude according to the difference between the previous and next data points in the speed time series; obtain the abnormality degree through the fluctuation amplitude of the data point and the surrounding data points and the change trend of the data point; analyze the data points by analyzing the abnormality degree and the abnormal characteristics of the speed during gear shifting, and filter out the abnormal data points of the speed.

[0042] The specific scheme of the data analysis method for fault prediction and health management of an automobile electronic control system provided by the present invention is described in detail below with reference to the accompanying drawings.

[0043] See also Figure 1 , which shows a basic flow chart of a data analysis method for automobile electronic control system fault prediction and health management provided by an embodiment of the present invention.

[0044] like Figure 1 As shown, a data analysis method for automobile electronic control system fault prediction and health management provided by an embodiment of the present invention specifically includes the following steps:

[0045] S1: Collect the engine's running speed time series data to obtain a speed time series sequence.

[0046] The data collector is connected to the on-board driving computer of the car, the engine is started, the engine start time is recorded as the starting time, the engine speed data is collected (the collection interval in this embodiment is set to 1 / 1000s), and the engine speed timing data when the car is running is collected in real time when the engine is working, and the speed timing sequence is obtained. The speed timing sequence is recorded as .

[0047] At this point, the engine speed timing sequence is obtained.

[0048] Regarding the engine speed, the engine speed is relatively stable during normal operation, but due to the abnormality of the engine electronic control system, the engine speed may fluctuate abnormally. In order to find the abnormal data points, it is necessary to obtain the fluctuation amplitude of each data point through the data points before and after each data point; however, due to the intense driving of the car driver and the engine gear shifting frustration, the speed fluctuations will also be large, so the abnormal points are screened according to the fluctuation characteristics of the abnormal speed points, where the speed fluctuations of the abnormal speed part are large and the duration is long; and during intense driving, the speed changes greatly but the speed fluctuations are small; and for the engine frustration caused by the engine gear shifting, the speed fluctuations are large during the frustration, but the duration is short; thus, the data points with relatively abnormal changes are screened. Therefore, this embodiment processes the speed timing sequence when the engine is running, and the steps of obtaining abnormal data points are steps S2 to S7.

[0049] S2: Analyze the speed difference between a data point in the speed time series and its adjacent data points before and after to obtain the fluctuation amplitude of the data point.

[0050] For the speed time series of the car engine when it is running, if the speed changes smoothly, then the speed of the data point at one moment in the speed time series is smaller than the speed of the data point at the previous moment and the speed change value of the data point at the next moment. If the speed fluctuates greatly, the opposite is true. Therefore, the speed difference between the data point in the speed time series and its adjacent data points before and after is analyzed. The speed time series data of the first and last endpoints and their neighborhood data are obtained through data extension (existing algorithm) to obtain the fluctuation amplitude of the data point. Specifically, the speed time series data of the first and last endpoints is calculated. Data points and The first speed difference between the data points , and Data points and The second speed difference between the data points , and then calculate the absolute value of the difference between the first speed difference and the second speed difference , normalize the absolute value of the difference data to obtain the first The fluctuation range of the data points. The quantitative formula for the fluctuation range of a data point is:

[0051]

[0052] In the formula: Indicates the number of times in the speed sequence when the car engine is running. The fluctuation range of the data, Indicates the first The speed of the data point, Indicates the first The speed of the data point, Indicates the first The speed of the data point, represents the linear normalization function.

[0053] When the engine is operating normally, its speed changes relatively smoothly, so the difference between the data points before and after the engine speed time series is small. Therefore, the speed of each data point is compared with the speed of the previous data point to obtain , and compare the speed of the data point with the speed of the next data point to obtain , subtract the two differences and find the absolute value, then we can get the difference before and after the data point The larger the value, the greater the difference between the data before and after, and the greater the fluctuation of the data. There is a positive correlation.

[0054] At this point, the fluctuation amplitude of each data point is obtained by comparing the speed difference before and after the engine speed time series.

[0055] As for the speed fluctuation when the engine is running, when the engine is operating normally, its speed will also produce relatively small fluctuations; therefore, for the abnormal data points of the speed, the speed is unstable due to the abnormality of the automobile electronic control system. The data characteristics of the unstable speed are characterized by large fluctuations and long duration, and its fluctuations are relatively sudden. Therefore, the fluctuation amplitude of each data point and its surrounding data points in the speed time series is analyzed. If the fluctuation amplitude of the data points around the data point is higher, the degree of abnormality is greater; and for each data point, if the current automobile engine is idling at a constant speed, or when accelerating and decelerating slowly, the speed changes relatively slowly, then when the speed data of the data point has a large fluctuation abnormality, the degree of abnormality is higher; and when driving fiercely, although the speed changes greatly, the fluctuation of the speed change is small, so when the engine speed data fluctuates, it is more likely to be a fluctuation caused by abnormal engine operation.

[0056] Based on the above analysis, in order to obtain the abnormality degree of the engine speed data point, steps S3 and S4 are included.

[0057] S3: Analyze the data points in the speed time series and the reference data points selected before and after them to obtain the change trend of the data points.

[0058] First, the method of selecting reference data points before and after the data points in the speed time series is to select the reference data points of the data points according to the fluctuation amplitude, and select a certain number of data points on the left and right sides of the data points as reference data points. Among them, the neighborhood data of the speed time series data of the first and last endpoints are obtained by data extension (existing algorithm). Furthermore, data points with larger fluctuation amplitudes require more reference data points, so the quantitative formula for the number of reference data points is:

[0059]

[0060] In the formula, Indicates the first The number of reference data points to be selected on the left and right sides of each data point, Indicates the number of times in the speed sequence when the car engine is running. The fluctuation range of the data.

[0061] Furthermore, the data points in the speed time series and the reference data points selected before and after them are analyzed to obtain the changing trend of the data points, including using the least squares method to perform straight line fitting on the data points in the speed time series and the reference data points selected before and after them to obtain the trend line of the data points; based on the trend line, the slope value of the trend line is obtained; the slope value is the changing trend of the data points.

[0062] S4: Obtain the abnormality degree of the data point based on the fluctuation amplitude and change trend.

[0063] According to the fluctuation range and change trend, the abnormality of the data point is obtained, including: averaging the fluctuation ranges of all reference data points to obtain the average fluctuation range of the reference data points; according to the fluctuation range of the data point and the average fluctuation range of the reference data points, as well as the slope value, the abnormality of the data point is obtained. The calculation formula for the abnormality of the data point is:

[0064]

[0065] In the formula: Indicates the first The degree of abnormality of the data point, Indicates the first The fluctuation range of the data points, Indicates the first Around the data point The fluctuation range of the selected reference data points, Indicates the first The number of reference data points selected around the data point, Indicates the first The slope value of the trend line corresponding to the data point, represents the linear normalization function.

[0066] By obtaining the speed timing sequence The number of data points to be selected on the left and right sides of each data point Adaptively perform anomaly analysis on each data point.

[0067] For the data points in the engine speed time series, the abnormal data will fluctuate greatly and last for a certain period of time. Therefore, it is necessary to analyze the abnormality of the data around each data point, that is, to adaptively select the data points around each data point in the speed time series. The average fluctuation range of the selected data points around each data point is calculated The fluctuation amplitude of the data point is weighted. The higher the average fluctuation amplitude of the reference data points selected around it, the higher the abnormality of the data point. The trend line of the data point is obtained by fitting the reference data points around the data point with the data point using the least square method, which reflects the change trend of the speed near the data point. When the car is driven vigorously, its speed trend is larger. Therefore, the larger the trend, the smaller the weight of the abnormality. Therefore, the slope of the trend line of each data point is and There is a negative correlation.

[0068] At this point, the degree of abnormality of each data point is obtained by adaptively analyzing the fluctuation amplitude of the data points around the data point and the change trend of the data point.

[0069] S5: Filter out suspected abnormal data points based on the degree of abnormality.

[0070] The abnormality degree of each data point in the speed time series The preset abnormal threshold For comparison, in this embodiment, the value of the abnormal threshold can be ,Will The data points are screened as suspected abnormal data points, and the suspected abnormal data points.

[0071] S6: Analyze the relationship between the abnormal degree of the suspected abnormal data point and the data points after it, and obtain the engine stall probability of the suspected abnormal data point based on the number of data points in the suspected abnormal time series connected domain.

[0072] The abnormality degree of the data points in the engine speed time series obtained through the above steps , where there may be engine jerking due to engine shifting, which causes the engine to form a large fluctuation in a short period of time, which may make the speed data abnormal at the time of gear shifting too high; therefore, it is necessary to screen out possible engine jerking; where the speed fluctuation caused by engine jerking occurs in a short period of time, the abnormal data at the time of gear shifting is manifested as a large speed value but immediately becomes smaller. Therefore, in order to screen out the anomalies caused by gear shifting engine jerking, the latter of each suspected abnormal data point in the speed time series is selected The abnormal degree of each data point is compared, and the data point with the largest decrease in abnormal degree is the engine stall data point. And for each suspected abnormal data point, if two suspected abnormal data points are adjacent in the time series, they are said to belong to the same suspected abnormal time series connected domain, that is, the suspected abnormal time series connected domain is, and continuous suspected abnormal data points constitute a suspected abnormal time series connected domain. For each suspected abnormal data point, if the number of data points in the suspected abnormal time series connected domain is greater, the abnormal speed lasts longer, and therefore the probability of it belonging to engine stall is smaller.

[0073] Based on the above analysis, the relationship between the abnormality degree of the suspected abnormal data point and the data points after it is analyzed, and the engine stall probability of the suspected abnormal data point is obtained by combining the number of data points in the suspected abnormal time series connected domain. The engine stall probability is: calculate the sum of the differences between the abnormality degree of the suspected abnormal data point and the abnormality degrees of multiple data points after it, then multiply the sum of the differences by the inverse of the number of data points in the suspected abnormal time series connected domain, and then perform linear normalization on the product value. That is, the calculation formula for the engine stall probability is:

[0074]

[0075] In the formula: Indicates The probability that the suspected abnormal data point belongs to engine stalling, Indicates The abnormality degree of the suspected abnormal data point, Indicates The suspected abnormal data point is the second one in the speed time series. The degree of abnormality of the data point, Indicates The number of data points in the suspected abnormal time series connected domain where the suspected abnormal data point is located, represents the linear normalization function.

[0076] For the phenomenon of excessive abnormality caused by engine stuttering, since the speed fluctuation caused by engine stuttering occurs in a short period of time, the abnormality of the stuttering data point is large but the speed of reduction is also fast. Therefore, the later abnormality of each suspected abnormal data point is compared in the time series. The difference between the abnormality degree of the data point and the abnormality degree of the suspected abnormal data point and , the greater the difference, the more abnormal the degree of abnormality decreases, so the higher the probability of engine setback. positively correlated; and for each suspected abnormal data point, the number of data points in the suspected temporal connected domain where it is located The more the abnormal speed lasts, the longer it lasts, while the abnormal speed caused by engine stalling lasts for a shorter time. Therefore, the probability that it belongs to engine stalling is smaller. There is a negative correlation.

[0077] S7: According to the engine stall probability, filter out the engine stall data points from the suspected abnormal data points to obtain the abnormal data points in the speed time series.

[0078] According to the probability of engine stall, filter out the engine stall data points in the suspected abnormal data points, and obtain the abnormal data points in the speed time series, including: preset engine stall threshold In this embodiment, ; Compare the engine stall probability with the preset engine stall threshold The engine stall probability is greater than the preset engine stall threshold, and the suspected abnormal data point is the engine stall data point, that is, , filter out engine stall data points; the engine stall probability is less than or equal to the preset engine stall threshold, that is , the suspected abnormal data point is the abnormal data point, and we get Abnormal data points in the speed time series of the engine when it is running.

[0079] So far, by analyzing the abnormality of each data point, we can obtain Abnormal data points in the speed time series of the engine when it is running.

[0080] S8: Predict automotive electronic control system failures based on abnormal data points.

[0081] Based on the abnormal data points, predict the failure of the automobile electronic control system, including: comparing the number of abnormal data points with a preset threshold number of abnormal automobile data points; when the number of abnormal data points is greater than or equal to the preset threshold number of abnormal automobile data, it is determined that there is a failure in the automobile electronic control system, and a fault light is turned on to remind the driver to maintain the automobile electronic control system.

[0082] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0083] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A data analysis method for fault prediction and health management of an automobile electronic control system, characterized in that: The method comprises: Collecting the engine's running speed time series data to obtain a speed time series sequence; Analyze the speed difference between a data point in the speed time series and its adjacent data points before and after it to obtain the fluctuation amplitude of the data point; Analyze the data points in the speed time series and the reference data points selected before and after them to obtain the change trend of the data points; Obtaining the abnormality degree of the data point according to the fluctuation amplitude and the change trend; According to the abnormality degree, suspected abnormal data points are screened out; Analyze the relationship between the suspected abnormal data point and the abnormality degree of the subsequent data points, and obtain the engine stall probability of the suspected abnormal data point in combination with the number of data points in the suspected abnormal time series connected domain; According to the engine jerking probability, filter out the engine jerking data points in the suspected abnormal data points to obtain the abnormal data points in the speed time series; Predicting automotive electronic control system failures based on the abnormal data points; Analyzing the data points in the speed time series and the reference data points selected before and after them to obtain the change trend of the data points includes: Using the least squares method, a linear fit is performed between the data points in the speed time series and the reference data points selected before and after the data points to obtain a trend line of the data points; based on the trend line, a slope value of the trend line is obtained; the slope value is a change trend of the data points; The method for obtaining the reference data point is to select the reference data point of the data point according to the fluctuation amplitude; Obtaining the abnormality degree of the data point according to the fluctuation amplitude and the change trend includes: Averaging the fluctuation amplitudes of all the reference data points to obtain an average fluctuation amplitude of the reference data points; Obtaining the degree of abnormality of the data point according to the fluctuation amplitude of the data point, the average fluctuation amplitude of the reference data point, and the slope value; The calculation formula of the abnormality degree is: In the formula: Indicates the first The degree of abnormality of the data point, Indicates the first The fluctuation range of the data points, Indicates the first Around the data point The fluctuation range of the selected reference data points, Indicates the first The number of reference data points selected around the data point, Indicates the first The slope value of the trend line corresponding to the data point, represents the linear normalization function; The engine stall probability is: calculate the sum of the differences between the abnormality degree of the suspected abnormal data point and the abnormality degrees of multiple data points after it, then multiply the sum of the differences by the inverse of the number of data points in the suspected abnormal time series connectivity domain, and then perform linear normalization on the product value.

2. The data analysis method for automobile electronic control system fault prediction and health management according to claim 1 is characterized in that: Analyzing the speed difference between a data point in the speed time series and its adjacent data points before and after it to obtain the fluctuation amplitude of the data point includes: Calculate the first Data points and The first speed difference between the data points , and Data points and The second speed difference between the data points , and then calculate the absolute value of the difference between the first speed difference and the second speed difference , normalize the absolute value data of the difference to obtain the first The fluctuation range of a data point.

3. The data analysis method for automobile electronic control system fault prediction and health management according to claim 1 is characterized in that: The suspected abnormal time series connected domain is that the continuous suspected abnormal data points form a suspected abnormal time series connected domain.

4. The data analysis method for automobile electronic control system fault prediction and health management according to claim 1, characterized in that: According to the engine jerking probability, filter out the engine jerking data points in the suspected abnormal data points to obtain the abnormal data points in the speed time series, including: comparing the engine stall probability with a preset engine stall threshold; The engine jerking probability is greater than a preset engine jerking threshold, the suspected abnormal data point is an engine jerking data point, and the engine jerking data point is screened out; The engine jerking probability is less than or equal to a preset engine jerking threshold, and the suspected abnormal data point is an abnormal data point.

5. The data analysis method for automobile electronic control system fault prediction and health management according to claim 3 is characterized in that: Predicting electronic control system failures based on the abnormal data points includes: Comparing the number of abnormal data points with a preset threshold value of the number of abnormal vehicle data; When the number of abnormal data points is greater than or equal to a preset threshold value of the number of abnormal data of the automobile, it is determined that there is a fault in the automobile electronic control system and a fault light is turned on.

Images