Single-peak distribution vehicle detection index anomaly detection method and terminal

Abstract

The invention discloses a unimodal distribution vehicle detection index anomaly detection method, and the method comprises the steps: obtaining the historical detection index data of a vehicle battery, removing the abnormal values, meeting a preset rule, of the historical detection index data, and obtaining reference detection index data; calculating a single-peak trend value of the reference detection index data, and obtaining a difference value abnormal interval according to a difference value between the single-peak trend value and the reference detection index data; obtaining current detection index data of the vehicle battery, and calculating a current trend value of the vehicle battery according to the current detection index data; and judging whether a difference value between the current trend value and the current detection index data falls into a difference value abnormal interval, if so, outputting that the vehicle battery is abnormal, and otherwise, outputting that the vehicle battery is normal. According to the invention, possible front and back changes of the detection index data of the vehicle battery are compared and analyzed, whether the vehicle battery detection index is normal is accurately judged, and the vehicle driving safety is guaranteed.

Description

Technical Field

[0001] The present invention relates to the field of electronic information technology, and in particular to a method and a terminal for detecting abnormality of vehicle detection indicators with unimodal distribution. Background Art

[0002] With the rapid development of new energy applications, the popularity of electric vehicles is increasing. As the core component of electric vehicles, the testing indicators of car batteries need to be strictly controlled during production and operation. Therefore, it is particularly important to accurately judge whether the vehicle battery testing indicators are normal. Summary of the invention

[0003] The technical problem to be solved by the present invention is to propose a method and terminal for detecting abnormal vehicle detection indicators with unimodal distribution, accurately determine whether the vehicle battery detection indicators are normal, and ensure vehicle driving safety.

[0004] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0005] A method for detecting abnormality of vehicle detection indicators with unimodal distribution comprises the following steps:

[0006] S1. Obtain historical detection index data of vehicle batteries, remove abnormal values ​​of the historical detection index data that meet preset rules, and obtain reference detection index data;

[0007] S2. Calculate the unimodal trend value of the reference detection index data, and obtain a difference abnormality interval according to the difference between the unimodal trend value and the reference detection index data;

[0008] S3, obtaining current detection index data of the vehicle battery, and calculating a current trend value of the vehicle battery according to the current detection index data;

[0009] S4. Determine whether the difference between the current trend value and the current detection index data falls within the abnormal difference interval. If so, output that the vehicle battery is abnormal; otherwise, output that the vehicle battery is normal.

[0010] In order to solve the above technical problems, another technical solution adopted by the present invention is:

[0011] A unimodal distribution vehicle detection index abnormality detection terminal includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the following steps are implemented:

[0012] S1. Obtain historical detection index data of vehicle batteries, remove abnormal values ​​of the historical detection index data that meet preset rules, and obtain reference detection index data;

[0013] S2. Calculate the unimodal trend value of the reference detection index data, and obtain a difference abnormality interval according to the difference between the unimodal trend value and the reference detection index data;

[0014] S3, obtaining current detection index data of the vehicle battery, and calculating a current trend value of the vehicle battery according to the current detection index data;

[0015] S4. Determine whether the difference between the current trend value and the current detection index data falls within the abnormal difference interval. If so, output that the vehicle battery is abnormal; otherwise, output that the vehicle battery is normal.

[0016] The beneficial effects of the present invention are: providing a vehicle detection index abnormality detection method and terminal with a unimodal distribution, obtaining a difference abnormal interval as comparison data based on the difference between the reference detection index data after the abnormal value of the vehicle battery is eliminated and the corresponding unimodal trend value, and then determining whether the vehicle battery is abnormal according to whether the difference between the current trend value and the current detection index data falls into the difference abnormal interval, comparing and analyzing the possible changes before and after the detection index data of the vehicle battery, accurately judging whether the vehicle battery detection index is normal, and ensuring the driving safety of the vehicle. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A schematic diagram of the steps of a method for detecting abnormality of vehicle detection indicators with unimodal distribution according to the present invention;

[0018] Figure 2 A flowchart of online anomaly detection for a method for detecting anomalies of vehicle detection indicators with unimodal distribution according to the present invention;

[0019] Figure 3 The present invention provides a system block diagram of a vehicle detection index abnormality detection terminal with unimodal distribution.

[0020] Description of labels:

[0021] 1. A vehicle detection index anomaly detection terminal with unimodal distribution; 2. A memory; 3. A processor. DETAILED DESCRIPTION

[0022] In order to explain the technical content, achieved objectives and effects of the present invention in detail, the following is an explanation in combination with the implementation modes and the accompanying drawings.

[0023] Please refer to Figure 1 and Figure 2 , a method for detecting abnormality of vehicle detection indicators with unimodal distribution, comprising the following steps:

[0024] S1. Obtain historical detection index data of vehicle batteries, remove abnormal values ​​of the historical detection index data that meet preset rules, and obtain reference detection index data;

[0025] S2. Calculate the unimodal trend value of the reference detection index data, and obtain a difference abnormality interval according to the difference between the unimodal trend value and the reference detection index data;

[0026] S3, obtaining current detection index data of the vehicle battery, and calculating a current trend value of the vehicle battery according to the current detection index data;

[0027] S4. Determine whether the difference between the current trend value and the current detection index data falls within the abnormal difference interval. If so, output that the vehicle battery is abnormal; otherwise, output that the vehicle battery is normal.

[0028] From the above description, it can be seen that the beneficial effects of the present invention are: based on the difference between the reference detection index data after the abnormal values ​​of the vehicle battery are eliminated and the corresponding single-peak trend value, the difference abnormal interval is obtained as the comparison data, and then according to whether the difference between the current trend value and the current detection index data falls into the difference abnormal interval, it is determined whether the vehicle battery is abnormal, and the possible changes before and after of the detection index data of the vehicle battery are compared and analyzed, so as to accurately judge whether the vehicle battery detection indicators are normal and ensure the driving safety of the vehicle.

[0029] Furthermore, the calculation of the single peak trend value of the reference detection index data is specifically as follows:

[0030] The trend item of the reference detection index data is extracted according to a time series decomposition algorithm to obtain the single peak trend value.

[0031] From the above description, it can be seen that by introducing time series analysis, the changes that occur in the vehicle battery during use can be accurately represented as trend items of the reference detection index data, thereby deriving a single-peak trend value for theoretical analysis.

[0032] Furthermore, the abnormal difference interval obtained according to the difference between the single peak trend value and the reference detection index data is specifically:

[0033] Calculate the difference between each item of the reference detection index data and the single peak trend value to obtain a difference sequence;

[0034] Based on the SH-ESD algorithm, the abnormal points in the difference sequence and the abnormal interval of the last point are identified to obtain the abnormal interval of the difference.

[0035] From the above description, we can know that we first calculate the difference between each data item in the reference detection index data and the single-peak trend value, and then identify the abnormal points in the difference sequence based on the high recognition ability of the SH-ESD algorithm, and select the abnormal interval of the last abnormal point as the difference abnormal interval.

[0036] Furthermore, the preset rules include:

[0037] Calculate the mean and standard deviation of the historical detection index data;

[0038] The difference between each item of the historical detection index data and the average value and the ratio of the difference to the standard deviation are calculated to determine and eliminate the abnormal values.

[0039] From the above description, it can be seen that by introducing the mean value and standard deviation, the degree of difference between each data and the mean value of the entire historical detection index data can be accurately evaluated, and then the outliers can be determined to ensure that the subsequent calculated difference anomaly interval is more accurate and reasonable.

[0040] Furthermore, the step S2 further includes:

[0041] According to the time series decomposition algorithm, a time series decomposition model of the detection index corresponding to the vehicle battery is established;

[0042] The current trend value of the vehicle battery currently predicted according to the current detection index data is specifically:

[0043] Substitute the current detection index data into the detection index time series decomposition model to calculate and obtain the current trend value.

[0044] From the above description, it can be seen that after using the time series decomposition algorithm to analyze the trend items of the reference detection index data, a detection index time series decomposition model based on the time series decomposition algorithm is established, so that when the current data is subsequently analyzed, the detection index time series decomposition model is used to calculate the current trend value, so as to improve the speed and accuracy of data processing.

[0045] Please refer to Figure 3 A vehicle detection index abnormality detection terminal 1 with unimodal distribution includes a memory 2, a processor 3, and a computer program stored in the memory 2 and executable on the processor 3. When the processor 3 executes the computer program, the following steps are implemented:

[0046] S1. Obtain historical detection index data of vehicle batteries, remove abnormal values ​​of the historical detection index data that meet preset rules, and obtain reference detection index data;

[0047] S2. Calculate the unimodal trend value of the reference detection index data, and obtain a difference abnormality interval according to the difference between the unimodal trend value and the reference detection index data;

[0048] S3, obtaining current detection index data of the vehicle battery, and calculating a current trend value of the vehicle battery according to the current detection index data;

[0049] S4. Determine whether the difference between the current trend value and the current detection index data falls within the abnormal difference interval. If so, output that the vehicle battery is abnormal; otherwise, output that the vehicle battery is normal.

[0050] From the above description, it can be seen that the beneficial effects of the present invention are: based on the difference between the reference detection index data after the abnormal values ​​of the vehicle battery are eliminated and the corresponding single-peak trend value, the difference abnormal interval is obtained as the comparison data, and then according to whether the difference between the current trend value and the current detection index data falls into the difference abnormal interval, it is determined whether the vehicle battery is abnormal, and the possible changes before and after of the detection index data of the vehicle battery are compared and analyzed, so as to accurately judge whether the vehicle battery detection indicators are normal and ensure the driving safety of the vehicle.

[0051] Furthermore, the calculation of the single peak trend value of the reference detection index data is specifically as follows:

[0052] The trend item of the reference detection index data is extracted according to a time series decomposition algorithm to obtain the single peak trend value.

[0053] From the above description, it can be seen that by introducing time series analysis, the changes that occur in the vehicle battery during use can be accurately represented as trend items of the reference detection index data, thereby deriving a single-peak trend value for theoretical analysis.

[0054] Furthermore, the abnormal difference interval obtained according to the difference between the single peak trend value and the reference detection index data is specifically:

[0055] Calculate the difference between each item of the reference detection index data and the single peak trend value to obtain a difference sequence;

[0056] Based on the SH-ESD algorithm, the abnormal points in the difference sequence and the abnormal interval of the last point are identified to obtain the abnormal interval of the difference.

[0057] From the above description, we can know that we first calculate the difference between each data item in the reference detection index data and the single-peak trend value, and then identify the abnormal points in the difference sequence based on the high recognition ability of the SH-ESD algorithm, and select the abnormal interval of the last abnormal point as the difference abnormal interval.

[0058] Furthermore, the preset rules include:

[0059] Calculate the mean and standard deviation of the historical detection index data;

[0060] The difference between each item of the historical detection index data and the average value and the ratio of the difference to the standard deviation are calculated to determine and eliminate the abnormal values.

[0061] From the above description, it can be seen that by introducing the mean value and standard deviation, the degree of difference between each data and the mean value of the entire historical detection index data can be accurately evaluated, and then the outliers can be determined to ensure that the subsequent calculated difference anomaly interval is more accurate and reasonable.

[0062] Furthermore, the step S2 further includes:

[0063] According to the time series decomposition algorithm, a time series decomposition model of the detection index corresponding to the vehicle battery is established;

[0064] The current trend value of the vehicle battery currently predicted according to the current detection index data is specifically:

[0065] Substitute the current detection index data into the detection index time series decomposition model to calculate and obtain the current trend value.

[0066] From the above description, it can be seen that after using the time series decomposition algorithm to analyze the trend items of the reference detection index data, a detection index time series decomposition model based on the time series decomposition algorithm is established, so that when the current data is subsequently analyzed, the detection index time series decomposition model is used to calculate the current trend value, so as to improve the speed and accuracy of data processing.

[0067] Please refer to Figure 1 and Figure 2 , Embodiment 1 of the present invention is:

[0068] A method for detecting abnormality of vehicle detection indicators with unimodal distribution comprises the following steps:

[0069] S1. Obtain historical test index data of vehicle batteries, remove abnormal values ​​of historical test index data that meet preset rules, and obtain reference test index data;

[0070] In this embodiment, the preset rules include: obtaining historical detection index data of the vehicle battery, calculating the average value and standard deviation of the historical detection index data; and eliminating abnormal values ​​of the historical detection index data that meet the preset rules based on the difference between each item of the historical detection index data and the average value and the standard deviation.

[0071] The above steps specifically use 3 times the standard deviation to eliminate outliers in historical data based on the z-score method, as follows:

[0072] Z-score is a commonly used concept in statistics. It is a standardized score used to represent the position of a data point in the data set to which it belongs. Z-score measures how different a data point is from the mean of the data set to which it belongs, and how it performs compared to the standard deviation of the data set. Generally, data points with a Z-score between -3 and +3 are considered to be within the normal range. The Z-score calculation formula is as follows:

[0073]

[0074] Where Z is the z-score, X is the value of the data point, u is the mean of the data points, and σ is the standard deviation.

[0075] In addition, it should be noted that in addition to the indicator data of automobile batteries with unimodal distribution, the present embodiment can also be applied to data objects such as capacity retention rate, maximum voltage of a single cell during charging, maximum temperature of a single cell during charging, maximum muzzle temperature during charging, charging SOC accuracy, maximum muzzle voltage during charging, battery DC internal resistance, temperature difference, temperature rise, etc.

[0076] S2. Calculate the unimodal trend value of the reference detection index data, and obtain the difference abnormal interval according to the difference between the unimodal trend value and the reference detection index data; wherein, extract the trend item of the reference detection index data according to the time series decomposition algorithm to obtain the unimodal trend value, and then establish the detection index time series decomposition model of the corresponding vehicle battery according to the time series decomposition algorithm;

[0077] In this embodiment, the time series decomposition algorithm is preferably STL (Seasonal-Trend Decomposition Procedure based on Loess): In time series analysis, the STL method is used to decompose time series data into three parts: seasonality, trend, and residual.

[0078] After obtaining the unimodal trend value according to the STL algorithm, the difference between each data in the reference detection index data and the unimodal trend value is calculated to obtain the difference sequence; based on the SH-ESD algorithm, the abnormal points in the difference sequence and the abnormal interval of the last point are identified to obtain the difference abnormal interval.

[0079] Among them, SH-ESD is an anomaly detection method, the full name is "Seasonal Hybrid Extreme Studentized Deviate". This method is used to detect outliers in time series data, especially for seasonal data sets. Combined with the foregoing, SH-ESD processes the time series data through STL (Seasonal-Temporal decomposition using Loess) decomposition, first removing the seasonal term, and then using the median as the trend term to perform an ESD test on the residual term. This processing method can not only detect global outliers, but also identify local outliers. Preferably, the abnormal interval of the last outlier identified is used as the difference abnormal interval.

[0080] S3, obtaining current detection index data of the vehicle battery, and calculating the current trend value of the vehicle battery according to the current detection index data;

[0081] Corresponding to the above content, the current detection index data is substituted into the detection index time series decomposition model to calculate the current trend value.

[0082] S4. Determine whether the difference between the current trend value and the current detection index data falls into the abnormal difference interval. If so, output that the vehicle battery is abnormal; otherwise, output that the vehicle battery is normal.

[0083] Please refer to Figure 3 , Embodiment 2 of the present invention is:

[0084] A unimodal distribution vehicle detection index anomaly detection terminal 1 includes a memory 2, a processor 3, and a computer program stored in the memory 2 and executable on the processor 3. When the processor 3 executes the computer program, a unimodal distribution vehicle detection index anomaly detection method of embodiment 1 is implemented.

[0085] To sum up, the present invention provides a method and terminal for detecting abnormalities in vehicle detection indicators with a unimodal distribution, which combine the use of Z-score, STL algorithm and SH-ESD algorithm to calculate the difference between the reference detection index data of the vehicle battery after removing abnormal values ​​and the corresponding unimodal trend value, and obtain the difference abnormal interval as comparison data, and then determine whether the vehicle battery is abnormal based on whether the difference between the current trend value and the current detection index data falls into the difference abnormal interval, compare and analyze the possible changes before and after the detection index data of the vehicle battery, accurately judge whether the vehicle battery detection indicators are normal, and ensure vehicle driving safety.

[0086] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent transformations made using the contents of the present invention's specification and drawings, or directly or indirectly applied in related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for detecting abnormality of vehicle detection indicators with unimodal distribution, characterized in that: The steps include: S1. Obtain historical detection index data of vehicle batteries, remove abnormal values ​​of the historical detection index data that meet preset rules, and obtain reference detection index data; S2. Calculate the unimodal trend value of the reference detection index data, and obtain a difference abnormality interval according to the difference between the unimodal trend value and the reference detection index data; S3, obtaining current detection index data of the vehicle battery, and calculating a current trend value of the vehicle battery according to the current detection index data; S4. Determine whether the difference between the current trend value and the current detection index data falls within the abnormal difference interval. If so, output that the vehicle battery is abnormal; otherwise, output that the vehicle battery is normal.

2. The method for detecting abnormality of vehicle detection indicators with unimodal distribution according to claim 1, characterized in that: The calculation of the single peak trend value of the reference detection index data is specifically as follows: The trend item of the reference detection index data is extracted according to a time series decomposition algorithm to obtain the single peak trend value.

3. The method for detecting abnormality of vehicle detection indicators with unimodal distribution according to claim 1, characterized in that: The abnormal difference interval obtained according to the difference between the single peak trend value and the reference detection index data is specifically: Calculate the difference between each item of the reference detection index data and the single peak trend value to obtain a difference sequence; Based on the SH-ESD algorithm, the abnormal points in the difference sequence and the abnormal interval of the last point are identified to obtain the abnormal interval of the difference.

4. The method for detecting abnormality of vehicle detection indicators with unimodal distribution according to claim 1, characterized in that: The preset rules include: Calculate the mean and standard deviation of the historical detection index data; The difference between each item of the historical detection index data and the average value and the ratio of the difference to the standard deviation are calculated to determine and eliminate the abnormal values.

5. The method for detecting abnormality of vehicle detection indicators with unimodal distribution according to claim 2, characterized in that: The step S2 further comprises: According to the time series decomposition algorithm, a time series decomposition model of the detection index corresponding to the vehicle battery is established; The current trend value of the vehicle battery currently predicted according to the current detection index data is specifically: Substitute the current detection index data into the detection index time series decomposition model to calculate and obtain the current trend value.

6. A vehicle detection index abnormality detection terminal with unimodal distribution, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the following steps are implemented: S1. Obtain historical test index data of the vehicle battery, remove abnormal values ​​of the historical test index data that meet preset rules, and obtain reference test index data; S2. Calculate the unimodal trend value of the reference detection index data, and obtain the difference abnormality interval according to the difference between the unimodal trend value and the reference detection index data; S3, obtaining current detection index data of the vehicle battery, and calculating a current trend value of the vehicle battery according to the current detection index data; S4. Determine whether the difference between the current trend value and the current detection index data falls within the abnormal difference interval. If so, output that the vehicle battery is abnormal; otherwise, output that the vehicle battery is normal.

7. The vehicle detection index abnormality detection terminal with unimodal distribution according to claim 6 is characterized in that: The calculation of the single peak trend value of the reference detection index data is specifically as follows: The trend item of the reference detection index data is extracted according to a time series decomposition algorithm to obtain the single peak trend value.

8. The vehicle detection index abnormality detection terminal with unimodal distribution according to claim 6, characterized in that: The difference between the single peak trend value and the reference detection index data is used to obtain the abnormal difference interval: Calculate the difference between each item of the reference detection index data and the single peak trend value to obtain a difference sequence; Based on the SH-ESD algorithm, the abnormal points in the difference sequence and the abnormal interval of the last point are identified to obtain the abnormal interval of the difference.

9. The vehicle detection index abnormality detection terminal with unimodal distribution according to claim 6, characterized in that: The preset rules include: Calculate the mean and standard deviation of the historical detection index data; The difference between each item of the historical detection index data and the average value and the ratio of the difference to the standard deviation are calculated to determine and eliminate the abnormal values.

10. The vehicle detection index abnormality detection terminal with unimodal distribution according to claim 7, characterized in that: The step S2 further comprises: According to the time series decomposition algorithm, a time series decomposition model of the detection index corresponding to the vehicle battery is established; The current trend value of the vehicle battery currently predicted according to the current detection index data is specifically: Substitute the current detection index data into the detection index time series decomposition model to calculate and obtain the current trend value.

Images