Abnormal data extraction method and device based on stability detection, electronic equipment and storage medium

Abstract

The invention provides an abnormal data extraction method and device based on stationarity detection, electronic equipment and a storage medium, and the method comprises the steps: marking a starting and ending point search range of a historical abnormal fluctuation data set, and determining a starting and ending point search range matrix; performing mean value processing on the starting and ending point searching range parameter matrix, and determining the size of a target window for stability detection of the historical abnormal fluctuation set, the number of target hops and the number of continuous and stable target windows; performing abnormal starting and ending point optimization on the historical abnormal fluctuation data set based on a time sequence stability detection method, and determining an effective abnormal fluctuation data set in the historical abnormal fluctuation data set; and performing merging processing on the plurality of effective abnormal fluctuation data sets, and determining a merged effective abnormal fluctuation data set. Accurate extraction of an abnormal fluctuation set on actual battery data is facilitated, redundant data is avoided, context analysis of data with similar time is facilitated, and the method has very important significance for postmortem analysis.

Description

Technical Field

[0001] The present application relates to the technical field of big data processing, and in particular to a method, device, electronic device and storage medium for extracting abnormal data based on stationarity detection. Background Art

[0002] The abnormal fluctuation set is extremely important for the analysis of battery monitoring data. If the range of the abnormal fluctuation segment cannot be accurately extracted, it will increase the difficulty of subsequent analysis and even interfere with the accuracy of subsequent analysis. Existing methods usually have the following two problems when extracting time series of abnormal time periods: (1) simply assume that the time series is stable, that is, the abnormal fluctuation segment has ended, and ignore the context of the time series itself, or the time series is not stable, that is, the abnormal fluctuation segment has begun (2) focus on the impact of the context on the subsequent time series, but do not remove the impact of adjacent data. This may cause the abnormal fluctuation of the time series itself to not stop, or divide one abnormal fluctuation segment into several parts. This will cause the abnormal fluctuation set to occupy more index overhead and increase data fragmentation. Therefore, how to accurately extract abnormal data has become a technical issue that cannot be underestimated. Summary of the invention

[0003] In view of this, the purpose of the present application is to provide an abnormal data extraction method, device, electronic device and storage medium based on stationarity detection, which processes the known abnormal fluctuation set through time series stationarity detection to accurately locate the abnormal start and end points, obtain an effective abnormal fluctuation data set, merge some abnormal fluctuation sets that are close in time, avoid the generation of too many abnormal fluctuation fragments, and be beneficial to the accurate extraction of abnormal fluctuation sets on actual battery data, avoid redundant data, and facilitate contextual analysis of data that are close in time, which is of great significance for post-analysis.

[0004] The embodiment of the present application provides a method for extracting abnormal data based on stationarity detection, and the method for extracting abnormal data includes:

[0005] Marking the start and end point search ranges of the historical abnormal fluctuation data set of the vehicle during driving, and determining the start and end point search range matrix of the historical abnormal fluctuation data set; wherein the start and end point search range matrix includes the starting search position, the ending search position, the number of hops, and the number of windows where continuous stability occurs;

[0006] Performing mean processing on the start and end point search range parameter matrix to determine the size of the target window, the target number of jumps, and the number of continuous and stable target windows for the stability detection of the historical abnormal fluctuation set;

[0007] The time series stationarity detection method optimizes the abnormal start and end points of the historical abnormal fluctuation data set based on the size of the target window, the target number of jumps and the target number of windows, and determines the effective abnormal fluctuation data set in the historical abnormal fluctuation data set;

[0008] A plurality of valid abnormal fluctuation data sets are merged to determine a merged valid abnormal fluctuation data set, so that data analysis is performed based on the merged valid abnormal fluctuation data set and the valid abnormal fluctuation data set.

[0009] In a possible implementation manner, after the base performs a merging process on a plurality of valid abnormal fluctuation data sets to determine a merged valid abnormal fluctuation data set, the abnormal data extraction method further includes:

[0010] Determine a new start and end point search range for detecting a new abnormal fluctuation data set, and detect whether an error value between the new start and end point search range and the start and end point search range is greater than a preset threshold;

[0011] If yes, then the start and end point search range matrix is ​​updated by half to determine a new start and end point search range matrix;

[0012] If not, a stationarity test is performed on the new abnormal fluctuation data set based on the start and end point search range matrix.

[0013] In a possible implementation manner, the marking of the start and end point search range parameters of the historical abnormal fluctuation data set during the driving process of the vehicle to determine the start and end point search range parameter matrix of the historical abnormal fluctuation data set includes:

[0014] Marking the starting search position, the ending search position and the stable window position of each abnormal fluctuation data set in the historical abnormal fluctuation data set;

[0015] For any of the abnormal fluctuation data sets, a starting sliding window and an ending sliding window are set, and the number of hops, the length of the starting sliding window, and the length of the ending sliding window are gradually reduced during the sliding window process until the window position of a stable abnormal fluctuation data set is detected to be the stable window position, and the length of the sliding window and the number of hops of the abnormal fluctuation data set are determined.

[0016] In a possible implementation manner, the time series stationarity detection method optimizes the abnormal start and end points of the historical abnormal fluctuation data set based on the size of the target window, the target number of hops, and the target number of windows, and determines a valid abnormal fluctuation data set in the historical abnormal fluctuation data set, including:

[0017] When the target window slides according to the target number of jumps, the first stable window of the abnormal fluctuation data set is detected based on the starting sliding window, and when the number of consecutive stable windows that appear after the first stable window reaches the target number of windows, it is determined that the starting search position of the abnormal fluctuation data set is valid;

[0018] After determining that the starting search position is valid, a second stable window appears based on the end sliding window detection. When the number of consecutive stable windows that appear after the second stable window reaches the target window number, it is determined that the end search position of the abnormal fluctuation data set is valid.

[0019] In a possible implementation manner, the merging of multiple valid abnormal fluctuation data sets to determine a merged valid abnormal fluctuation data set includes:

[0020] The valid abnormal fluctuation data sets are merged based on the starting search position and the ending search position of at least two adjacent valid abnormal fluctuation data sets to determine a merged valid abnormal fluctuation data set.

[0021] In a possible implementation manner, the merging of the valid abnormal fluctuation data sets based on the starting search position and the ending search position of at least two adjacent valid abnormal fluctuation data sets to determine the merged valid abnormal fluctuation data set includes:

[0022] Detecting whether the end time corresponding to the end search position of the first valid abnormal fluctuation data set is greater than or equal to the start time corresponding to the start search position of the adjacent valid abnormal fluctuation data set;

[0023] If so, detecting whether the first valid abnormal fluctuation data set has reached the merging number, and if not, merging the first valid abnormal fluctuation data set and the adjacent valid abnormal fluctuation data set to generate a merged valid abnormal fluctuation data set;

[0024] If not, determine whether the time difference between the end time of the first valid abnormal fluctuation data set and the start time of the adjacent valid abnormal fluctuation data set is less than or equal to the preset time difference. If it is less than or equal to the preset time difference and the number of merging times has not been reached, merge the first valid abnormal fluctuation data set and the adjacent valid abnormal fluctuation data set to generate a merged valid abnormal fluctuation data set.

[0025] In a possible implementation manner, after detecting whether the first valid abnormal fluctuation data set reaches the merging times, the abnormal data extraction method further includes:

[0026] If the merging times are reached, the first valid abnormal fluctuation data set is not merged with the adjacent valid abnormal fluctuation data set.

[0027] The embodiment of the present application also provides an abnormal data extraction device based on stationarity detection, and the abnormal data extraction device includes:

[0028] A marking module is used to mark the start and end point search range of the historical abnormal fluctuation data set of the vehicle during driving, and determine the start and end point search range matrix of the historical abnormal fluctuation data set; wherein the start and end point search range matrix includes the starting search position, the ending search position, the number of hops, and the number of windows with continuous stability;

[0029] A parameter processing module, used for performing mean processing on the start and end point search range parameter matrix to determine the size of the target window, the target number of jumps and the number of continuous stable target windows for the stability detection of the historical abnormal fluctuation set;

[0030] An ADF detection module is used for a time series stationarity detection method to optimize the abnormal start and end points of the historical abnormal fluctuation data set based on the size of the target window, the target number of jumps and the target number of windows, and determine a valid abnormal fluctuation data set in the historical abnormal fluctuation data set;

[0031] The merging module is used to merge multiple valid abnormal fluctuation data sets to determine a merged valid abnormal fluctuation data set, so as to perform data analysis based on the merged valid abnormal fluctuation data set and the valid abnormal fluctuation data set.

[0032] An embodiment of the present application also provides an electronic device, comprising: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate through the bus, and when the machine-readable instructions are executed by the processor, the steps of the abnormal data extraction method based on stationarity detection as described above are performed.

[0033] An embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the abnormal data extraction method based on stationarity detection as described above are executed.

[0034] An abnormal data extraction method, device, electronic device and storage medium based on stationarity detection are provided in an embodiment of the present application. The abnormal data extraction method includes: marking the start and end point search ranges of the historical abnormal fluctuation data set of the vehicle during driving, and determining the start and end point search range matrix of the historical abnormal fluctuation data set; wherein the start and end point search range matrix includes the starting search position, the ending search position, the number of hops, and the number of windows with continuous stationarity; performing mean processing on the start and end point search range parameter matrix to determine the size of the target window, the target number of hops and the number of target windows with continuous stationarity for the stationarity detection of the historical abnormal fluctuation set; a time series stationarity detection method optimizes the abnormal start and end points of the historical abnormal fluctuation data set based on the size of the target window, the target number of hops and the target number of windows to determine the effective abnormal fluctuation data set in the historical abnormal fluctuation data set; and merging multiple effective abnormal fluctuation data sets to determine the merged effective abnormal fluctuation data set, so as to perform data analysis based on the merged effective abnormal fluctuation data set and the effective abnormal fluctuation data set. Through time series stationarity detection, the known abnormal fluctuation sets are processed to accurately locate the abnormal start and end points, obtain effective abnormal fluctuation data sets, merge some abnormal fluctuation sets with similar time distances, avoid generating too many abnormal fluctuation fragments, which is conducive to the accurate extraction of abnormal fluctuation sets on actual battery data, avoid redundant data, and facilitate contextual analysis of data with similar time, which is of great significance for post-analysis.

[0035] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0037] Figure 1 A flowchart of an abnormal data extraction method based on stationarity detection provided in an embodiment of the present application;

[0038] Figure 2 One of the structural schematic diagrams of an abnormal data extraction device based on stationarity detection provided in an embodiment of the present application;

[0039] Figure 3 A second structural diagram of an abnormal data extraction device based on stationarity detection provided in an embodiment of the present application;

[0040] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0041] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application usually described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, each other embodiment obtained by those skilled in the art without making creative work belongs to the scope of protection of the present application.

[0042] First, the application scenarios to which this application is applicable are introduced. This application can be applied in the field of big data processing technology.

[0043] Based on this, an embodiment of the present application provides an abnormal data extraction method based on stationarity detection, which processes the known abnormal fluctuation set through time series stationarity detection to accurately locate the abnormal start and end points, obtain an effective abnormal fluctuation data set, and merge some abnormal fluctuation sets that are close in time to avoid generating too many abnormal fluctuation fragments, which is beneficial to the accurate extraction of abnormal fluctuation sets on actual battery data, avoids redundant data, and facilitates contextual analysis of data that are close in time, which is of great significance for post-analysis.

[0044] See also Figure 1 , Figure 1 The flowchart of an abnormal data extraction method based on stationarity detection provided by an embodiment of the present application is as follows. Figure 1 As shown in , the abnormal data extraction method provided by the embodiment of the present application includes:

[0045] S101: Mark the start and end point search ranges of the historical abnormal fluctuation data set of the vehicle during driving, and determine the start and end point search range matrix of the historical abnormal fluctuation data set; wherein the start and end point search range matrix includes the starting search position, the ending search position, the number of jumps, and the number of windows where continuous stability occurs.

[0046] In this step, the start and end point search ranges of the historical abnormal fluctuation data set of the vehicle during driving are marked to determine the start and end point search range matrix of the historical abnormal fluctuation data set.

[0047] The start and end point search range matrix includes a start search position, an end search position, a hop count, and a number of windows where continuous stability occurs.

[0048] In a possible implementation manner, the marking of the start and end point search range parameters of the historical abnormal fluctuation data set during the driving process of the vehicle to determine the start and end point search range parameter matrix of the historical abnormal fluctuation data set includes:

[0049] A: Mark the starting search position, ending search position and stable window position of each abnormal fluctuation data set in the historical abnormal fluctuation data set.

[0050] Here, the starting search position, the ending search position and the stable window position of each abnormal fluctuation data set in the historical abnormal fluctuation data set are marked.

[0051] In a specific embodiment, according to the historical abnormal fluctuation set χ={χ1,χ2,…χ i ,,…χ n}, n = |χ|, manually mark multiple groups or customize several accurate abnormal fluctuation set ranges, and divide them into Y = {c1, c2, ..., c k} class, and give the relevant starting search position Ε s =t s (x 10 ) and end search position E n =t e (x 10 ), denoted as χ 10 ={χ1,χ2,…χ 10}, m indicates the category it belongs to. The duration of each abnormal fluctuation set is

[0052] B: For any of the abnormal fluctuation data sets, set the starting sliding window and the ending sliding window, and gradually reduce the number of hops, the length of the starting sliding window, and the length of the ending sliding window during the sliding window process until it is detected that the window position of a stable abnormal fluctuation data set is the stable window position, and then determine the length of the sliding window and the number of hops of the abnormal fluctuation data set.

[0053] Here, for each abnormal fluctuation data set, a starting sliding window and an ending sliding window are set. During the sliding window process, the number of hops, the length of the starting sliding window, and the length of the ending sliding window are gradually reduced until the window position of a stable abnormal fluctuation data set is detected to be a stable window position, and the length of the sliding window and the number of hops of the abnormal fluctuation data set are determined.

[0054] In a specific embodiment, for a certain abnormal fluctuation set Set a starting sliding window win rs , an end sliding window win re , the length of the initial window is the length of the time series, slide these windows, let the number of hops s equal to the window size, and keep reducing s and win r (including win rs and win re ) length until the stable (non-stationary) window position is detected as the correct position of the manual mark, and the size value of the sliding window win is saved r , number of hops s.

[0055] S102: Perform mean processing on the start and end point search range parameter matrix to determine the size of the target window, the target number of jumps, and the number of continuous and stable target windows for the stability detection of the historical abnormal fluctuation set.

[0056] In this step, the start and end point search range parameter matrix is ​​processed by mean value to determine the size of the target window, the target number of jumps, and the number of continuous and stable target windows for the stability detection of the historical abnormal fluctuation set.

[0057] Here, for the start and end point search range parameter matrix Φ, Φ={win r· |s · |σ ·}={(win r1· ,s 1· ,σ 1· ),(win r2· ,s 2· ,σ 2· ),…,(win rn· ,s n· ,σ n· ) Calculate the average value of each column as the window size for the final abnormal fluctuation set stationarity detection and hop count as well as That is Φ f ={win f ,s f ,σ f}.

[0058] S103: The time series stationarity detection method optimizes the abnormal start and end points of the historical abnormal fluctuation data set based on the size of the target window, the target number of jumps and the target number of windows, and determines a valid abnormal fluctuation data set in the historical abnormal fluctuation data set.

[0059] In this step, the time series stationarity detection method optimizes the abnormal start and end points of the historical abnormal fluctuation data set according to the target window size, target jump number and target window number, and determines the effective abnormal fluctuation data set in the historical abnormal fluctuation data set.

[0060] In a possible implementation manner, the time series stationarity detection method optimizes the abnormal start and end points of the historical abnormal fluctuation data set based on the size of the target window, the target number of hops, and the target number of windows, and determines a valid abnormal fluctuation data set in the historical abnormal fluctuation data set, including:

[0061] a: When the target window slides according to the target number of jumps, the first stable window of the abnormal fluctuation data set is detected based on the starting sliding window. When the number of consecutive stable windows that appear after the first stable window reaches the target number of windows, it is determined that the starting search position of the abnormal fluctuation data set is valid.

[0062] Here, when the target window slides according to the target number of hops, the first stable window of the abnormal fluctuation data set is detected according to the starting sliding window. When the number of consecutive stable windows after the first stable window reaches the target window number, it is determined that the starting search position of the abnormal fluctuation data set is valid.

[0063] b: After determining that the starting search position is valid, a second stable window appears based on the end sliding window detection. When the number of consecutive stable windows that appear after the second stable window reaches the target number of windows, it is determined that the end search position of the abnormal fluctuation data set is valid.

[0064] Here, after determining that the starting search position is valid, a second stable window appears according to the end sliding window detection. When the number of consecutive stable windows after the second stable window reaches the target window number, it is determined that the end search position of the abnormal fluctuation data set is valid.

[0065] Among them, the abnormal fluctuation data set between the effective search positions is the effective abnormal fluctuation data set.

[0066] In the specific embodiment, the same abnormal fluctuation set When the window win′ r When sliding, use win rs Detect the first stable window, and the number of subsequent stable windows reaches σ s (initial value is 0, and increases by 1 each time) to identify the starting point of the current abnormal fluctuation set Valid. When a valid s After that, follow the win reDetect the second stable window, and the number of subsequent consecutive stable windows reaches σ e (initial value is 0, and increases by 1 each time) to identify the end point of the current abnormal fluctuation set. efficient.

[0067] S104: merging multiple valid abnormal fluctuation data sets to determine a merged valid abnormal fluctuation data set, so as to perform data analysis based on the merged valid abnormal fluctuation data set and the valid abnormal fluctuation data set.

[0068] In this step, a plurality of valid abnormal fluctuation data sets are merged to determine a merged valid abnormal fluctuation data set, so that data analysis is performed based on the merged valid abnormal fluctuation data set and the valid abnormal fluctuation data set.

[0069] In a possible implementation manner, the merging of multiple valid abnormal fluctuation data sets to determine a merged valid abnormal fluctuation data set includes:

[0070] The valid abnormal fluctuation data sets are merged based on the starting search position and the ending search position of at least two adjacent valid abnormal fluctuation data sets to determine a merged valid abnormal fluctuation data set.

[0071] Here, the valid abnormal fluctuation data set is merged according to the starting search position and the ending search position of at least two adjacent valid abnormal fluctuation data sets to determine the merged valid abnormal fluctuation data set. Some abnormal fluctuation sets that are close in time are merged to avoid generating too many abnormal fluctuation fragments.

[0072] In a possible implementation manner, the merging of the valid abnormal fluctuation data sets based on the starting search position and the ending search position of at least two adjacent valid abnormal fluctuation data sets to determine the merged valid abnormal fluctuation data set includes:

[0073] (1): Check whether the end time corresponding to the end search position of the first valid abnormal fluctuation data set is greater than or equal to the start time corresponding to the start search position of the adjacent valid abnormal fluctuation data set.

[0074] Here, it is detected whether the end time corresponding to the end search position of the first valid abnormal fluctuation data set is greater than or equal to the start time corresponding to the start search position of the adjacent valid abnormal fluctuation data set.

[0075] (2): If so, check whether the first valid abnormal fluctuation data set has reached the merging number. If not, merge the first valid abnormal fluctuation data set and the adjacent valid abnormal fluctuation data set to generate a merged valid abnormal fluctuation data set.

[0076] Here, if yes, it is detected whether the first valid abnormal fluctuation data set reaches the merging number, if not, the first valid abnormal fluctuation data set and the adjacent valid abnormal fluctuation data set are merged to generate a merged valid abnormal fluctuation data set.

[0077] Among them, when End time When for The starting search position of the merged valid abnormal fluctuation data set End search location

[0078] (3): If not, determine whether the time difference between the end time of the first valid abnormal fluctuation data set and the start time of the adjacent valid abnormal fluctuation data set is less than or equal to the preset time difference. If it is less than or equal to the preset time difference and the number of merging times has not been reached, merge the first valid abnormal fluctuation data set and the adjacent valid abnormal fluctuation data set to generate a merged valid abnormal fluctuation data set.

[0079] Here, if not, determine whether the time difference between the end time of the first valid abnormal fluctuation data set and the start time of the adjacent valid abnormal fluctuation data set is less than or equal to the preset time difference. If it is less than or equal to the preset time difference and the number of merges has not been reached, the first valid abnormal fluctuation data set and the adjacent valid abnormal fluctuation data set are merged to generate a merged valid abnormal fluctuation data set.

[0080] Among them, when the effective abnormal fluctuation data set End time When when When is χ′ i Yes, its

[0081] In a possible implementation manner, after detecting whether the first valid abnormal fluctuation data set reaches the merging times, the abnormal data extraction method further includes:

[0082] If the merging times are reached, the first valid abnormal fluctuation data set is not merged with the adjacent valid abnormal fluctuation data set.

[0083] Here, after multiple merging, since multiple abnormal fluctuation sets with similar time distances may be merged multiple times, the length of some abnormal fluctuation sets after the final merger becomes longer and longer, resulting in and Time interval To avoid this, set each The maximum number of merging times is ω, which is determined according to the actual situation. After merging once, the number of merges is recorded separately. Included in Then the current Can no longer be merged.

[0084]

[0085] In a possible implementation manner, after the base performs a merging process on a plurality of valid abnormal fluctuation data sets to determine a merged valid abnormal fluctuation data set, the abnormal data extraction method further includes:

[0086] i: Determine a new start and end point search range for detecting a new abnormal fluctuation data set, and detect whether the error value between the new start and end point search range and the start and end point search range is greater than a preset threshold.

[0087] Here, the new start and end point search range of the new abnormal fluctuation data set is determined through the above-mentioned step of determining the start and end point search range.

[0088] ii: If yes, then the parameters of the start-end point search range matrix are updated by half to determine a new start-end point search range matrix; if no, then the new abnormal fluctuation data set is tested for stability based on the start-end point search range matrix.

[0089] Here, each time we add a new abnormal fluctuation set χ new The new start and end point search range uses the existing start and end point search range matrix Φ f Mark and search the new start and end points (E s ,Ε n ) and the search range of the start and end points marked manually Do the difference and get the error Set the correction error threshold ε, when ε>ε, the matrix Φ={win′ r |s|σ} adopts the half-update method, the number of hops s and the start and end windows win′ r , you can quickly locate the required appropriate size interval through binary search. Each time, the parameter size will be modified to half of the original size. When the parameter is not suitable, increase the existing parameter size by half until the appropriate parameter is found. This way, you can quickly find the appropriate parameter.

[0090] This application uses time series stationarity detection to determine the known abnormal fluctuation set, determine the precise abnormal start and end points of the abnormal fluctuation set, use time series stationarity detection to optimize the range of the abnormal fluctuation set, and merge some abnormal fluctuation sets that are close in time to avoid generating too many abnormal fluctuation fragments. Through the manual correction of the start and end points of the abnormal fluctuation set, the parameter update method is used to automatically update the search range parameter matrix of the start and end points of the abnormal fluctuation set based on expert experience to avoid inaccurate start and end points due to parameter errors. The half-parameter update method is used to quickly locate the parameter interval and realize fast search of parameter tuples.

[0091] An embodiment of the present application provides an abnormal data extraction method based on stationarity detection, the abnormal data extraction method comprising: marking the start and end point search ranges of a historical abnormal fluctuation data set of a vehicle during driving, and determining a start and end point search range matrix of the historical abnormal fluctuation data set; wherein the start and end point search range matrix includes a starting search position, an ending search position, a hop number, and a number of windows with continuous stationarity; performing mean processing on the start and end point search range parameter matrix, and determining the size of a target window, a target hop number, and a number of target windows with continuous stationarity for stationarity detection of the historical abnormal fluctuation set; a time series stationarity detection method performs abnormal start and end point optimization on the historical abnormal fluctuation data set based on the target window size, the target hop number, and the target window number, and determines a valid abnormal fluctuation data set in the historical abnormal fluctuation data set; and merging multiple valid abnormal fluctuation data sets to determine a merged valid abnormal fluctuation data set, so as to perform data analysis based on the merged valid abnormal fluctuation data set and the valid abnormal fluctuation data set. Through time series stationarity detection, the known abnormal fluctuation sets are processed to accurately locate the abnormal start and end points, obtain effective abnormal fluctuation data sets, merge some abnormal fluctuation sets with similar time distances, avoid generating too many abnormal fluctuation fragments, which is conducive to the accurate extraction of abnormal fluctuation sets on actual battery data, avoid redundant data, and facilitate contextual analysis of data with similar time, which is of great significance for post-analysis.

[0092] See also Figure 2 , Figure 3 , Figure 2 One of the structural schematic diagrams of an abnormal data extraction device based on stationarity detection provided in an embodiment of the present application; Figure 3 The second structural diagram of an abnormal data extraction device based on stationarity detection provided in an embodiment of the present application. Figure 2 As shown in , the abnormal data extraction device 200 includes:

[0093] The marking module 210 is used to mark the start and end point search range of the historical abnormal fluctuation data set of the vehicle during driving, and determine the start and end point search range matrix of the historical abnormal fluctuation data set; wherein the start and end point search range matrix includes the starting search position, the ending search position, the number of hops, and the number of windows with continuous stability;

[0094] The parameter processing module 220 is used to perform mean processing on the start and end point search range parameter matrix to determine the size of the target window, the target number of jumps, and the number of continuous stable target windows for the stability detection of the historical abnormal fluctuation set;

[0095] ADF detection module 230, used for time series stationarity detection method, optimizes the abnormal start and end points of the historical abnormal fluctuation data set based on the size of the target window, the target number of jumps and the target number of windows, and determines the effective abnormal fluctuation data set in the historical abnormal fluctuation data set;

[0096] The merging module 240 is used to merge multiple valid abnormal fluctuation data sets to determine a merged valid abnormal fluctuation data set, so as to perform data analysis based on the merged valid abnormal fluctuation data set and the valid abnormal fluctuation data set.

[0097] Further, such as Figure 3 As shown, the abnormal data extraction device 200 further includes a parameter updating module 250, and the parameter updating module 250 is used to:

[0098] Determine a new start and end point search range for detecting a new abnormal fluctuation data set, and detect whether an error value between the new start and end point search range and the start and end point search range is greater than a preset threshold;

[0099] If yes, then the start and end point search range matrix is ​​updated by half to determine a new start and end point search range matrix;

[0100] If not, a stationarity test is performed on the new abnormal fluctuation data set based on the start and end point search range matrix.

[0101] Further, when the marking module 210 is used to mark the start and end point search range parameters of the historical abnormal fluctuation data set of the vehicle during driving and determine the start and end point search range parameter matrix of the historical abnormal fluctuation data set, the marking module 210 is specifically used to:

[0102] Marking the starting search position, the ending search position and the stable window position of each abnormal fluctuation data set in the historical abnormal fluctuation data set;

[0103] For any of the abnormal fluctuation data sets, a starting sliding window and an ending sliding window are set, and the number of hops, the length of the starting sliding window, and the length of the ending sliding window are gradually reduced during the sliding window process until the window position of a stable abnormal fluctuation data set is detected to be the stable window position, and the length of the sliding window and the number of hops of the abnormal fluctuation data set are determined.

[0104] Further, when the ADF detection module 230 optimizes the abnormal start and end points of the historical abnormal fluctuation data set based on the size of the target window, the target number of hops, and the target number of windows in the time series stationarity detection method, and determines a valid abnormal fluctuation data set in the historical abnormal fluctuation data set, the ADF detection module 230 is specifically used to:

[0105] When the target window slides according to the target number of jumps, the first stable window of the abnormal fluctuation data set is detected based on the starting sliding window, and when the number of consecutive stable windows that appear after the first stable window reaches the target number of windows, it is determined that the starting search position of the abnormal fluctuation data set is valid;

[0106] After determining that the starting search position is valid, a second stable window appears based on the end sliding window detection. When the number of consecutive stable windows that appear after the second stable window reaches the target window number, it is determined that the end search position of the abnormal fluctuation data set is valid.

[0107] Further, when the merging module 240 is used to merge the multiple valid abnormal fluctuation data sets to determine the merged valid abnormal fluctuation data set, the merging module 240 is specifically used to:

[0108] The valid abnormal fluctuation data sets are merged based on the starting search position and the ending search position of at least two adjacent valid abnormal fluctuation data sets to determine a merged valid abnormal fluctuation data set.

[0109] Further, when the merging module 240 is used to merge the valid abnormal fluctuation data set based on the starting search position and the ending search position of at least two adjacent valid abnormal fluctuation data sets to determine the merged valid abnormal fluctuation data set, the merging module 240 is specifically used to:

[0110] Detecting whether the end time corresponding to the end search position of the first valid abnormal fluctuation data set is greater than or equal to the start time corresponding to the start search position of the adjacent valid abnormal fluctuation data set;

[0111] If so, detecting whether the first valid abnormal fluctuation data set has reached the merging number, and if not, merging the first valid abnormal fluctuation data set and the adjacent valid abnormal fluctuation data set to generate a merged valid abnormal fluctuation data set;

[0112] If not, determine whether the time difference between the end time of the first valid abnormal fluctuation data set and the start time of the adjacent valid abnormal fluctuation data set is less than or equal to the preset time difference. If it is less than or equal to the preset time difference and the number of merging times has not been reached, merge the first valid abnormal fluctuation data set and the adjacent valid abnormal fluctuation data set to generate a merged valid abnormal fluctuation data set.

[0113] Furthermore, after the merging module 240 is used to detect whether the first valid abnormal fluctuation data set reaches the merging times, the merging module 240 is also used to:

[0114] If the merging times are reached, the first valid abnormal fluctuation data set is not merged with the adjacent valid abnormal fluctuation data set.

[0115] An abnormal data extraction device based on stationarity detection is provided in an embodiment of the present application. The abnormal data extraction device includes: a labeling module, which is used to label the start and end point search range of the historical abnormal fluctuation data set of the vehicle during driving, and determine the start and end point search range matrix of the historical abnormal fluctuation data set; wherein the start and end point search range matrix includes a starting search position, an end search position, a hop number, and a number of windows with continuous stationarity; a parameter processing module, which is used to perform mean processing on the start and end point search range parameter matrix, and determine the size of the target window, the target hop number, and the number of target windows with continuous stationarity for the stationarity detection of the historical abnormal fluctuation set; an ADF detection module, which is used for optimizing the abnormal start and end points of the historical abnormal fluctuation data set based on the size of the target window, the target hop number, and the target window number in the time series stationarity detection method, and determining the effective abnormal fluctuation data set in the historical abnormal fluctuation data set; a merging module, which is used to merge multiple effective abnormal fluctuation data sets, and determine the merged effective abnormal fluctuation data set, so as to perform data analysis based on the merged effective abnormal fluctuation data set and the effective abnormal fluctuation data set. Through time series stationarity detection, the known abnormal fluctuation sets are processed to accurately locate the abnormal start and end points, obtain effective abnormal fluctuation data sets, merge some abnormal fluctuation sets with similar time distances, avoid generating too many abnormal fluctuation fragments, which is conducive to the accurate extraction of abnormal fluctuation sets on actual battery data, avoid redundant data, and facilitate contextual analysis of data with similar time, which is of great significance for post-analysis.

[0116] See also Figure 4 , Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 4 As shown in , the electronic device 400 includes a processor 410 , a memory 420 and a bus 430 .

[0117] The memory 420 stores machine-readable instructions executable by the processor 410. When the electronic device 400 is running, the processor 410 communicates with the memory 420 via the bus 430. When the machine-readable instructions are executed by the processor 410, the above-mentioned Figure 1 The specific implementation of the steps of the abnormal data extraction method based on stationarity detection in the method embodiment shown can be found in the method embodiment, which will not be repeated here.

[0118] The present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the computer program can execute the above-mentioned Figure 1 The steps of the abnormal data extraction method based on stationarity detection in the method embodiment are shown in the figure. The specific implementation method can be found in the method embodiment, which will not be repeated here.

[0119] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0120] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of devices or units can be electrical, mechanical or other forms.

[0121] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0122] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0123] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application can essentially be embodied in the form of a software product, or in other words, the part that contributes to the prior art or the part of the technical solution. The computer software product is stored in a storage medium, including several 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 application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0124] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the above-mentioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-mentioned embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A method for extracting abnormal data based on stationarity detection, characterized in that: The abnormal data extraction method comprises: Marking the start and end point search ranges of the historical abnormal fluctuation data set of the vehicle during driving, and determining the start and end point search range matrix of the historical abnormal fluctuation data set; wherein the start and end point search range matrix includes the starting search position, the ending search position, the number of hops, and the number of windows where continuous stability occurs; Performing mean processing on the start and end point search range parameter matrix to determine the size of the target window, the target number of jumps, and the number of continuous and stable target windows for the stability detection of the historical abnormal fluctuation set; The time series stationarity detection method optimizes the abnormal start and end points of the historical abnormal fluctuation data set based on the size of the target window, the target number of jumps and the target number of windows, and determines the effective abnormal fluctuation data set in the historical abnormal fluctuation data set; A plurality of valid abnormal fluctuation data sets are merged to determine a merged valid abnormal fluctuation data set, so that data analysis is performed based on the merged valid abnormal fluctuation data set and the valid abnormal fluctuation data set.

2. The abnormal data extraction method according to claim 1, characterized in that: After the base performs a merging process on a plurality of valid abnormal fluctuation data sets to determine a merged valid abnormal fluctuation data set, the abnormal data extraction method further comprises: Determine a new start and end point search range for detecting a new abnormal fluctuation data set, and detect whether an error value between the new start and end point search range and the start and end point search range is greater than a preset threshold; If yes, then the start and end point search range matrix is ​​updated by half to determine a new start and end point search range matrix; If not, a stationarity test is performed on the new abnormal fluctuation data set based on the start and end point search range matrix.

3. The abnormal data extraction method according to claim 1, characterized in that: The step of marking the start and end point search range parameters of the historical abnormal fluctuation data set of the vehicle during driving, and determining the start and end point search range parameter matrix of the historical abnormal fluctuation data set, includes: Marking the starting search position, the ending search position and the stable window position of each abnormal fluctuation data set in the historical abnormal fluctuation data set; For any of the abnormal fluctuation data sets, a starting sliding window and an ending sliding window are set, and the number of hops, the length of the starting sliding window, and the length of the ending sliding window are gradually reduced during the sliding window process until the window position of a stable abnormal fluctuation data set is detected to be the stable window position, and the length of the sliding window and the number of hops of the abnormal fluctuation data set are determined.

4. The abnormal data extraction method according to claim 1, characterized in that: The time series stationarity detection method optimizes the abnormal start and end points of the historical abnormal fluctuation data set based on the size of the target window, the target number of jumps, and the target number of windows, and determines a valid abnormal fluctuation data set in the historical abnormal fluctuation data set, including: When the target window slides according to the target number of jumps, the first stable window of the abnormal fluctuation data set is detected based on the starting sliding window, and when the number of consecutive stable windows that appear after the first stable window reaches the target number of windows, it is determined that the starting search position of the abnormal fluctuation data set is valid; After determining that the starting search position is valid, a second stable window appears based on the end sliding window detection. When the number of consecutive stable windows that appear after the second stable window reaches the target window number, it is determined that the end search position of the abnormal fluctuation data set is valid.

5. The abnormal data extraction method according to claim 1, characterized in that: The merging of the multiple valid abnormal fluctuation data sets to determine a merged valid abnormal fluctuation data set includes: The valid abnormal fluctuation data sets are merged based on the starting search position and the ending search position of at least two adjacent valid abnormal fluctuation data sets to determine a merged valid abnormal fluctuation data set.

6. The abnormal data extraction method according to claim 5, characterized in that: The merging of the valid abnormal fluctuation data sets based on the starting search position and the ending search position of at least two adjacent valid abnormal fluctuation data sets to determine the merged valid abnormal fluctuation data set includes: Detecting whether the end time corresponding to the end search position of the first valid abnormal fluctuation data set is greater than or equal to the start time corresponding to the start search position of the adjacent valid abnormal fluctuation data set; If so, detecting whether the first valid abnormal fluctuation data set has reached the merging number, and if not, merging the first valid abnormal fluctuation data set and the adjacent valid abnormal fluctuation data set to generate a merged valid abnormal fluctuation data set; If not, determine whether the time difference between the end time of the first valid abnormal fluctuation data set and the start time of the adjacent valid abnormal fluctuation data set is less than or equal to the preset time difference. If it is less than or equal to the preset time difference and the number of merging times has not been reached, merge the first valid abnormal fluctuation data set and the adjacent valid abnormal fluctuation data set to generate a merged valid abnormal fluctuation data set.

7. The abnormal data extraction method according to claim 6, characterized in that: After detecting whether the first valid abnormal fluctuation data set reaches the merging number, the abnormal data extraction method further includes: If the merging times are reached, the first valid abnormal fluctuation data set is not merged with the adjacent valid abnormal fluctuation data set.

8. An abnormal data extraction device based on stationarity detection, characterized in that: The abnormal data extraction device comprises: A marking module is used to mark the start and end point search range of the historical abnormal fluctuation data set of the vehicle during driving, and determine the start and end point search range matrix of the historical abnormal fluctuation data set; wherein the start and end point search range matrix includes the starting search position, the ending search position, the number of hops, and the number of windows with continuous stability; A parameter processing module, used for performing mean processing on the start and end point search range parameter matrix to determine the size of the target window, the target number of jumps and the number of continuous stable target windows for the stability detection of the historical abnormal fluctuation set; An ADF detection module is used for a time series stationarity detection method to optimize the abnormal start and end points of the historical abnormal fluctuation data set based on the size of the target window, the target number of jumps and the target number of windows, and determine a valid abnormal fluctuation data set in the historical abnormal fluctuation data set; The merging module is used to merge multiple valid abnormal fluctuation data sets to determine a merged valid abnormal fluctuation data set, so as to perform data analysis based on the merged valid abnormal fluctuation data set and the valid abnormal fluctuation data set.

9. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate through the bus, and the machine-readable instructions are executed by the processor to execute the steps of the abnormal data extraction method based on stationarity detection as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the abnormal data extraction method based on stationarity detection according to any one of claims 1 to 7 are executed.

Images