Vibration data anomaly detection method, device and equipment and storage medium

Abstract

The application provides a vibration data anomaly detection method and device, equipment and a storage medium, comprising: obtaining target vibration data to be detected, collection time of the target vibration data, and a historical vibration data sequence corresponding to the target vibration data; generating a probability distribution function corresponding to the collection time according to each historical vibration data in the historical vibration data sequence; obtaining a confidence interval based on the probability distribution function and statistical characteristics corresponding to the probability distribution function; and performing anomaly detection on the target vibration data based on the confidence interval. By determining the confidence interval for judging the target vibration data to be detected based on the historical data, and performing anomaly judgment on the target vibration data to be detected based on the confidence interval, the accuracy of the judgment result is ensured.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of vibration data processing, and particularly relates to a vibration data anomaly detection method and device, equipment and a storage medium. BACKGROUND

[0002] With the gradual mechanization of industrial production, the mechanical equipment used for production in the industry and manufacturing industry is also developing in the direction of large-scale, precision and automation. In order to ensure the safe and normal operation of these machines and avoid serious economic and property losses caused by equipment deterioration, damage and other faults, fault diagnosis of mechanical equipment is increasingly valued by people.

[0003] Since the vibration signal of the mechanical equipment contains a large amount of information about the working process of the equipment, the vibration signal anomaly detection is generally used for mechanical equipment fault diagnosis. In the related art, the vibration data is mainly judged by the experience accumulation of the workers, which has certain subjectivity and is limited by experience, resulting in low accuracy of abnormal vibration data judgment. SUMMARY

[0004] The present application provides a vibration data anomaly detection method, device, equipment and storage medium, which determines a confidence interval for judging target vibration data to be detected by historical data, and performs abnormal judgment on the target vibration data to be detected by the confidence interval, thereby ensuring the accuracy of the judgment result.

[0005] In a first aspect, the present application provides a vibration data anomaly detection method, comprising:

[0006] obtaining target vibration data to be detected, a collection time of the target vibration data, and a historical vibration data sequence corresponding to the target vibration data;

[0007] generating a probability distribution function corresponding to the collection time according to each historical vibration data in the historical vibration data sequence;

[0008] obtaining a confidence interval based on the probability distribution function and a statistical feature corresponding to the probability distribution function;

[0009] performing anomaly detection on the target vibration data based on the confidence interval.

[0010] In a possible implementation manner of the present application, the confidence interval is obtained based on the probability distribution function and the statistical feature corresponding to the probability distribution function, comprising:

[0011] obtaining a statistical feature of the random vibration data corresponding to the collection time according to the probability value of the random vibration data of the probability distribution function, the statistical feature comprising at least one of mean value and variance;

[0012] Based on the preset vibration data interval generation rules, the statistical characteristics are processed by standard deviation to obtain the confidence interval.

[0013] In one possible implementation of this application, after performing anomaly detection on the target vibration data based on the confidence interval, the method further includes:

[0014] If the target vibration data is abnormal vibration data, then the standard vibration data is determined according to the statistical characteristics corresponding to the probability distribution function;

[0015] The standard vibration data is added to the historical vibration data sequence, and the historical vibration data sequence is updated.

[0016] In one possible implementation of this application, generating the probability distribution function corresponding to the acquisition time based on each historical vibration data in the historical vibration data sequence includes:

[0017] The historical vibration data in the historical vibration data sequence is input into a preset Gaussian distribution model to obtain the core function of the Gaussian model. The Gaussian model is a vibration data statistical model formed by inputting Gaussian white noise under the Bayesian framework.

[0018] Obtain the function type of the core function of the Gaussian model; if the function type is exponential, obtain the log-likelihood function corresponding to the exponential type.

[0019] Based on the historical vibration data sequence, the corresponding vibration calculation parameters of the log-likelihood function are calculated, and the probability distribution function corresponding to the acquisition time is generated based on the vibration calculation parameters.

[0020] In one possible implementation of this application, the step of solving for the vibration calculation parameters corresponding to the log-likelihood function based on the historical vibration data sequence, and generating the probability distribution function corresponding to the acquisition time based on the vibration calculation parameters, includes:

[0021] The kernel function matrix corresponding to the historical vibration data sequence is determined based on the historical vibration data sequence and the Gaussian model kernel function.

[0022] Using the historical vibration data sequence and the kernel function matrix, the vibration calculation parameters corresponding to the log-likelihood function are solved;

[0023] The probability distribution function corresponding to the acquisition time is generated based on the preset posterior distribution function and the vibration calculation parameters.

[0024] In one possible implementation of this application, after performing anomaly detection on the target vibration data based on the confidence interval, the process includes:

[0025] When the target vibration data is detected as abnormal vibration data, the target vibration data is accumulated into the abnormal vibration dataset;

[0026] If the number of abnormal vibration data in the abnormal vibration dataset exceeds a preset threshold, then the target data change information of each abnormal vibration data in the abnormal vibration dataset is statistically analyzed.

[0027] Based on the preset mapping relationship between anomaly handling information and data change information, the target anomaly handling information corresponding to the target data change information is determined;

[0028] Vibration data anomaly information is generated and fed back based on the target anomaly processing information and the target data change information.

[0029] In one possible implementation of this application, the anomaly detection of the target vibration data based on the confidence interval includes:

[0030] If the target vibration data is within the confidence interval, then the target vibration data is determined to be normal vibration data;

[0031] If the target vibration data is outside the confidence interval, then the target vibration data is determined to be abnormal vibration data.

[0032] Secondly, this application provides a vibration data anomaly detection device, wherein the vibration data anomaly detection device...

[0033] Acquisition module: used to acquire the vibration data of the target to be detected, the acquisition time of the target vibration data, and the historical vibration data sequence corresponding to the target vibration data;

[0034] Determination module: used to generate a probability distribution function corresponding to the acquisition time based on each historical vibration data in the historical vibration data sequence;

[0035] Interval determination module: used to obtain confidence intervals based on the probability distribution function and the statistical characteristics corresponding to the probability distribution function;

[0036] Detection module: used to perform anomaly detection on the target vibration data based on the confidence interval.

[0037] Thirdly, this application provides a vibration data anomaly detection device, the vibration data anomaly detection device comprising:

[0038] One or more processors;

[0039] Memory; and

[0040] One or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the processor to implement any of the vibration data anomaly detection methods.

[0041] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, the computer program being loaded by a processor to perform the steps in any of the vibration data anomaly detection methods described above.

[0042] This application provides a vibration data anomaly detection method, apparatus, device, and storage medium. It acquires target vibration data to be detected, the acquisition time of the target vibration data, and a historical vibration data sequence corresponding to the target vibration data. Then, based on each historical vibration data in the historical vibration data sequence, it generates a probability distribution function corresponding to the acquisition time. Next, based on the probability distribution function and its corresponding statistical characteristics, it obtains a confidence interval. Finally, it performs anomaly detection on the target vibration data based on the confidence interval. By using the historical vibration data sequence and the probability distribution function corresponding to the acquisition time, i.e., by determining the probability distribution function of possible vibration data corresponding to the acquisition time based on the historical vibration data sequence, it ensures data diversity and temporal correlation between corresponding data. Then, based on the probability distribution function and its corresponding statistical characteristics, it obtains a confidence interval for judging whether the target vibration data to be detected is abnormal, avoiding the limitations of directly predicting using preset values ​​and ensuring data accuracy. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1 This is a schematic diagram of a scenario for the data anomaly detection method provided in an embodiment of this application;

[0045] Figure 2 This is a schematic flowchart of an embodiment of the data anomaly detection method provided in this application.

[0046] Figure 3 A schematic diagram of an implementation scheme for determining the probability distribution function in the vibration data anomaly detection method provided in this application;

[0047] Figure 4A schematic flowchart of an implementation scheme for determining the confidence interval in the vibration data anomaly detection method provided in this application;

[0048] Figure 5 This is a schematic flowchart of another embodiment of the data anomaly detection method provided in this application;

[0049] Figure 6 This is a schematic flowchart of another embodiment of the data anomaly detection method provided in this application.

[0050] Figure 7 This is a schematic diagram of an embodiment of the data anomaly detection device provided in this application.

[0051] Figure 8 This is a schematic diagram of an embodiment of the data anomaly detection device provided in this application. Detailed Implementation

[0052] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0054] In this application, the term "exemplary" is used to mean "serving as an example, illustration, or description." Any embodiment described as "exemplary" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed in this application.

[0055] This application provides a method, apparatus, device, and computer-readable storage medium for detecting abnormal vibration data, which will be described in detail below.

[0056] The vibration data anomaly detection method in this embodiment of the invention is applied to a vibration data anomaly detection device. The vibration data anomaly detection device is set in a vibration data anomaly detection equipment. The vibration data anomaly detection equipment is provided with one or more processors, a memory, and one or more application programs. The one or more application programs are stored in the memory and configured to be executed by the processor to implement the vibration data anomaly detection method. The vibration data anomaly detection equipment can be a terminal, such as a mobile phone or a tablet computer. The vibration data anomaly detection equipment can also be a server or a service cluster composed of multiple servers.

[0057] like Figure 1 As shown, Figure 1 This is a schematic diagram of a scenario for the vibration data anomaly detection method according to an embodiment of this application. The vibration data anomaly detection scenario in this embodiment includes a vibration data anomaly detection device 100 (the vibration data anomaly detection device 100 integrates a vibration data anomaly detection apparatus). The vibration data anomaly detection device 100 runs a computer-readable storage medium corresponding to vibration data anomaly detection to perform the vibration data anomaly detection steps.

[0058] Understandable, Figure 1 The vibration data anomaly detection device in the scenario of the vibration data anomaly detection method shown, or the device included in the vibration data anomaly detection device, does not constitute a limitation on the embodiments of the present invention. That is, the number or type of device included in the scenario of the vibration data anomaly detection method, or the number or type of device included in each device, does not affect the overall implementation of the technical solution in the embodiments of the present invention, and can all be considered as equivalent substitutions or derivatives of the technical solutions claimed in the embodiments of the present invention.

[0059] In this embodiment of the invention, the vibration data anomaly detection device 100 is mainly used for: acquiring target vibration data to be detected, the acquisition time of the target vibration data, and the historical vibration data sequence corresponding to the target vibration data; generating a probability distribution function corresponding to the acquisition time based on each historical vibration data in the historical vibration data sequence; obtaining a confidence interval based on the probability distribution function and the statistical characteristics corresponding to the probability distribution function; and performing anomaly detection on the target vibration data based on the confidence interval.

[0060] In this embodiment of the invention, the vibration data anomaly detection device 100 can be an independent vibration data anomaly detection device, or it can be a network or cluster of vibration data anomaly detection devices. For example, the vibration data anomaly detection device 100 described in this embodiment includes, but is not limited to, a computer, a network host, a single network vibration data anomaly detection device, a set of multiple network vibration data anomaly detection devices, or a cloud vibration data anomaly detection device composed of multiple vibration data anomaly detection devices. The cloud vibration data anomaly detection device is composed of a large number of computer or network vibration data anomaly detection devices based on cloud computing.

[0061] Those skilled in the art will understand that Figure 1 The application environment shown is merely one application scenario of the solution in this application and does not constitute a limitation on the application scenario of the solution in this application. Other application environments may include those that are more specific to this application. Figure 1 The number of vibration data anomaly detection devices shown, or the network connection relationship of vibration data anomaly detection devices, for example... Figure 1 Only one vibration data anomaly detection device is shown in the diagram. It is understood that the scenario of this vibration data anomaly detection method may also include one or more other vibration data anomaly detection devices, which are not specifically limited here. The vibration data anomaly detection device 100 may also include a memory, which can be used to store vibration data.

[0062] Furthermore, in the scenario of the vibration data anomaly detection method of this application, the vibration data anomaly detection device 100 can be equipped with a display device, or the vibration data anomaly detection device 100 can be connected to an external display device 200 without a display device. The display device 200 is used to output the results of the vibration data anomaly detection method executed in the vibration data anomaly detection device. The vibration data anomaly detection device 100 can access the background vibration database 300 (the background vibration database can be in the local storage of the vibration data anomaly detection device, or it can be located in the cloud). The background vibration database 300 stores information related to vibration data anomaly detection, such as historical vibration data sequences, or pre-set calculation functions, calculation models, etc.

[0063] It should be noted that, Figure 1 The schematic diagram of the vibration data anomaly detection method shown is merely an example. The scenario of the vibration data anomaly detection method described in this embodiment of the invention is for the purpose of more clearly illustrating the technical solution of this embodiment of the invention, and does not constitute a limitation on the technical solution provided by this embodiment of the invention.

[0064] Based on the scenarios described above for vibration data anomaly detection methods, an embodiment of the vibration data anomaly detection method is proposed.

[0065] like Figure 2 The diagram shown is a flowchart of an embodiment of the vibration data anomaly detection method in this application. The vibration data anomaly detection method includes steps S201-S204:

[0066] S201. Obtain the target vibration data to be detected, the acquisition time of the target vibration data, and the historical vibration data sequence corresponding to the target vibration data.

[0067] The vibration data refers to the vibration data to be detected collected at any given time. This means the vibration data can be the vibration data collected at the current time or the vibration data to be detected collected at the previous time. It is understood that the vibration data can be collected by a vibration data acquisition device installed on the vibration generating equipment, such as a vibration sensor.

[0068] It is understood that the acquisition time refers to the acquisition time of the target vibration data. The target vibration data can be any vibration data point from continuously acquired vibration data, or it can be the target data acquired by the vibration data acquisition device according to a preset vibration data acquisition frequency.

[0069] The target vibration data corresponds to the historical vibration data sequence, which is the historical vibration data within a preset time period prior to the acquisition time of the target vibration data. For example, if the target vibration data is y... t The acquisition time is t, that is, the historical vibration data sequence includes Y train =[y t-k y t-k+1 , ..., y t-1 ]; where k is a preset duration, it can be understood that each historical vibration data corresponds to a second acquisition time, it can be understood that the second time has no practical meaning, it is only used to distinguish acquisition times, it can be understood that y t With y t-1 The interval between collection cycles can be one second, one hour, or other similar periods.

[0070] It is understood that the data in the historical vibration sequence is arranged in a time series, that is, the historical vibration sequence is a time series matrix corresponding to a historical vibration data, and the historical vibration data sequence changes in response to the changes in the target vibration data.

[0071] S202. Generate a probability distribution function corresponding to the acquisition time based on each historical vibration data in the historical vibration data sequence.

[0072] The probability distribution function of the acquisition time is a prediction function of the probability of occurrence of all possible vibration data at the acquisition time. It can be understood that the possible vibration data at this time is the predicted vibration data, which can be predicted based on historical vibration data sequences and preset calculation models or formulas.

[0073] For details, see Figure 3 , Figure 3 One implementation scheme for determining the probability distribution function in the vibration data anomaly detection method provided in this application includes steps S301-S303:

[0074] S301. Input the historical vibration data in the historical vibration data sequence into the preset Gaussian distribution model to obtain the core function of the Gaussian model. The Gaussian distribution model is a vibration data statistical model formed by inputting Gaussian white noise under the Bayesian framework.

[0075] S302. Obtain the function type of the core function of the Gaussian model. If the function type is an exponential type, obtain the log-likelihood function corresponding to the exponential type.

[0076] S303. Solve for the vibration calculation parameters corresponding to the log-likelihood function based on the historical vibration data sequence, and generate the probability distribution function corresponding to the acquisition time based on the vibration calculation parameters.

[0077] Specifically, the core function of the Gaussian model can be found here:

[0078]

[0079] The core function of the Gaussian model includes the parameter σ to be optimized. c , l has an initial value, wherein, in one embodiment of this application, the (t) i -t j ) 2 The acquisition time t can be any two historical vibration data points in the corresponding historical vibration data sequence. i and t j The variance between them.

[0080] Furthermore, the log-likelihood function is described in the following document:

[0081]

[0082] Wherein, L is a preset log-likelihood function, and p(Y) train |X train , σ z Y is the likelihood function calculated based on the historical training data sequence, and the likelihood function conforms to a Gaussian white noise distribution. train The sequence is a historical vibration data (a matrix containing multiple historical vibration data points distributed according to acquisition time), where t is the acquisition time and σ is the vibration data. z Here, C represents the parameters to be optimized, and C is the kernel function matrix. It can be understood that the kernel function matrix can be calculated based on the kernel matrix function.

[0083] In the embodiments of this application, the x train This refers to the historical vibration data sequence obtained by correcting the vibration data of a historical vibration data sequence according to a preset system error. The calculation method for correcting the vibration data of the historical vibration data sequence based on the system error is not specifically limited in this application. In the embodiment of this application, the formula for correcting the vibration data of the historical vibration data sequence based on the system error is as follows:

[0084] Y = X + Z

[0085] Where Y is any historical vibration data in the historical vibration sequence data, Z is the preset system error, and X is any historical vibration data in the historical vibration sequence data after correction corresponding to Y.

[0086] Specifically, the vibration calculation parameters corresponding to the log-likelihood function obtained from the historical vibration data sequence include:

[0087] (1) Determine the kernel function matrix corresponding to the historical vibration data sequence based on the historical vibration data sequence and the Gaussian model kernel function;

[0088] (2) Using the historical vibration data sequence and the kernel function matrix, solve for the vibration calculation parameters corresponding to the log-likelihood function;

[0089] (3) Generate the probability distribution function corresponding to the acquisition time based on the preset posterior distribution function and the vibration calculation parameters.

[0090] Specifically, based on the historical vibration data obtained by correcting the historical vibration data in the historical vibration data sequence, the data in the corrected historical vibration data sequence are substituted into the core function of the Gaussian model to obtain the kernel function matrix. Then, the historical vibration data sequence and the kernel function matrix are substituted into the log-likelihood function, and the partial derivatives of the log-likelihood function are solved using the conjugate gradient method to obtain the σ that satisfies local optima. z σ c The solution of l yields the vibration calculation parameter θ, wherein the vibration calculation parameter θ = {σ} c ,l,σ z Based on the vibration calculation parameters, the core function of the Gaussian model includes the parameter σ to be optimized. c The core function of the Gaussian model is updated by updating l.

[0091] Then, based on the preset posterior distribution function and the vibration calculation parameters, a probability distribution function corresponding to the acquisition time is generated. Specifically, the preset posterior distribution function can be found here:

[0092]

[0093] Among them, Y * |Y train Let C be the posterior probability distribution, i.e., the probability distribution function of the acquisition time, with the sign ~ indicating proportionality to the sign; * C is the cross kernel function matrix for the training and test sets. ** This is the kernel function matrix for the test set. Specifically, the kernel function matrix is ​​calculated based on the updated Gaussian model kernel function (kernel matrix function) described above.

[0094] Specifically, the updated Gaussian model core function can obtain the C values ​​for future and past time steps. * Matrix, similarly C **C is a matrix representing all future moments, while C is a matrix that only depends on past times (training). In other words, it can be understood that C can be calculated by substituting the acquisition time from the historical vibration data sequence into the updated Gaussian model kernel function, and C can be calculated by substituting the acquisition time from the historical vibration data sequence into the acquisition time of the target vibration data into the updated Gaussian model kernel function. * Substituting the acquisition time of the target vibration data into the updated Gaussian model kernel function yields C. ** .

[0095] S203. Based on the probability distribution function and the statistical characteristics corresponding to the probability distribution function, the confidence interval is obtained.

[0096] The statistical characteristics mentioned above refer to data such as the variance and mean corresponding to the probability distribution function. For details, please refer to [link to relevant documentation]. Figure 4 , Figure 4 An implementation scheme for determining the confidence interval in the vibration data anomaly detection method provided in this application includes steps S401-S402:

[0097] S401. Based on the probability value of the random vibration data corresponding to the acquisition time according to the probability distribution function, the statistical characteristics of the random vibration data are obtained, and the statistical characteristics include at least one of mean and variance.

[0098] S402. Based on the preset vibration data interval generation rules, the statistical characteristics are processed by standard deviation to obtain the confidence interval.

[0099] Specifically, in the embodiments of this application, the statistical characteristics include the mean and standard deviation, wherein the mean and standard deviation are the probability distribution function. mean and standard deviation That is, the mean and standard deviation of the vibration data that may occur during the acquisition time. For example, we can calculate the cumulative probability density function through the obtained probability distribution function. When the cumulative probability density function takes the value of 0.5, the corresponding point is the mean. The standard deviation is calculated based on the mean and the vibration data that may occur during the acquisition time.

[0100] Specifically, the preset vibration data interval generation rule is the three sigma principle. In the embodiment of this application, the confidence interval I ∈ [I low I up See the formula for generating ]:

[0101]

[0102]

[0103] Specifically, the confidence interval of vibration data for the corresponding acquisition time is determined by calculation formulas, as well as the mean and variance.

[0104] S204. Anomaly detection is performed on the target vibration data based on the confidence interval.

[0105] Specifically, after obtaining the confidence interval, the target vibration data is compared with the confidence interval. If the target vibration data is within the confidence interval, the target vibration data is determined to be normal vibration data; if the target vibration data is outside the confidence interval, the target vibration data is determined to be abnormal vibration data.

[0106] Furthermore, based on any of the above implementation schemes, see [link to relevant documentation]. Figure 5 After performing anomaly detection on the target vibration data, steps S501-S502 are also included:

[0107] S501. If the target vibration data is abnormal vibration data, then the standard vibration data corresponding to the target vibration data is determined according to the statistical characteristics corresponding to the probability distribution function.

[0108] S502. Add the standard vibration data to the historical vibration data sequence and update the historical vibration data sequence.

[0109] Specifically, after anomaly detection of the target vibration data, if the target vibration data is abnormal, standard vibration data is determined based on the statistical characteristics corresponding to the probability distribution function. Specifically, the standard vibration data replaces the abnormal data to form a historical dataset. Specifically, the standard vibration data is the mean data among the statistical characteristics. It is understood that in some other embodiments of this application, the standard vibration data can also be a preset value. This application does not make specific limitations. That is, the standard vibration data is added to the historical vibration data sequence, the historical vibration data sequence is updated, and anomaly detection is performed on the target vibration data corresponding to the next acquisition time based on the updated historical vibration data sequence.

[0110] Furthermore, based on any of the above implementation schemes, see [link to relevant documentation]. Figure 6 After performing anomaly detection on the target vibration data, steps S601-S605 are also included:

[0111] S601. When the target vibration data is detected as abnormal vibration data, the target vibration data is accumulated into the abnormal vibration dataset.

[0112] In this context, it can be understood that the abnormal vibration dataset includes abnormal data that is at least one vibration data acquisition cycle apart. It can also be understood that the preset duration can be preset according to different vibration conditions. For example, if the vibration data acquisition cycle is relatively short, the preset duration can be set to be shorter; if the vibration data acquisition cycle is relatively long, the preset duration can be set to be longer. By accumulating the abnormal vibration data within the preset duration, the limitations of single-time abnormal judgment are avoided.

[0113] S602. If the number of abnormal vibration data in the abnormal vibration dataset exceeds a preset threshold, then the target data change information of each abnormal vibration data in the abnormal vibration dataset is statistically analyzed.

[0114] Specifically, if the number of abnormal vibration data accumulated in the abnormal vibration dataset reaches a preset number of times, it indicates that the vibration equipment corresponding to the vibration data may actually have malfunctioned, reducing the possibility of misjudgment of vibration data. Then, based on the relationship between the vibration data magnitudes of the abnormal vibration data in the abnormal vibration dataset, such as the mean value of vibration data error and the standard deviation of error between the abnormal vibration data in the abnormal vibration dataset, the target data change information between the abnormal vibration data is determined.

[0115] S603. Based on the preset mapping relationship between the anomaly handling information and the data change information, determine the target anomaly handling information corresponding to the target data change information.

[0116] S604. Generate vibration data anomaly information based on the target anomaly processing information and the target data change information, and then feed it back.

[0117] Specifically, the target anomaly handling information can be historical operation information or historical fault causes corresponding to target data change information. The vibration data anomaly information is generated based on the target anomaly handling information and the target data change information. That is, while giving the result of the anomaly data judgment, the processing operation and / or fault cause of the anomaly data corresponding to the anomaly data is fed back based on the historical data.

[0118] It is understandable that the vibration data anomaly detection method is applied to the vibration data anomaly detection equipment. The vibration data anomaly detection equipment can feed back the feedback results to the display end that communicates with the vibration data anomaly detection equipment, such as the display screen of the vibration device corresponding to the vibration data, or generate voice broadcast commands for corresponding feedback.

[0119] This implementation provides a vibration data anomaly detection method. It acquires the target vibration data to be detected, the acquisition time of the target vibration data, and the corresponding historical vibration data sequence. Then, based on each historical vibration data point in the historical vibration data sequence, it generates a probability distribution function corresponding to the acquisition time. Next, based on the probability distribution function and its corresponding statistical characteristics, it obtains a confidence interval. Finally, it performs anomaly detection on the target vibration data based on the confidence interval. By using the historical vibration data sequence and the probability distribution function corresponding to the acquisition time—that is, by determining the probability distribution function of possible vibration data corresponding to the acquisition time based on the historical vibration data sequence—it ensures data diversity and temporal correlation between corresponding data points. Then, based on the probability distribution function and its corresponding statistical characteristics, it obtains a confidence interval for judging whether the target vibration data to be detected is abnormal, avoiding the limitations of directly predicting using preset values ​​and ensuring data accuracy.

[0120] To better implement the vibration data anomaly detection method in the embodiments of this application, based on the vibration data anomaly detection method, the embodiments of this application also provide a vibration data anomaly detection device, such as... Figure 7 As shown, the vibration data anomaly detection device includes 701-704:

[0121] Acquisition module 701: used to acquire the vibration data of the target to be detected, the acquisition time of the target vibration data, and the historical vibration data sequence corresponding to the target vibration data;

[0122] Determining module 702: used to generate a probability distribution function corresponding to the acquisition time based on each historical vibration data in the historical vibration data sequence;

[0123] Interval determination module 703: used to obtain a confidence interval based on the probability distribution function and the statistical characteristics corresponding to the probability distribution function;

[0124] Detection module 704: used to perform anomaly detection on the target vibration data based on the confidence interval.

[0125] In some embodiments of this application, the interval determination module 703 is used to obtain a confidence interval based on the probability distribution function and the statistical characteristics corresponding to the probability distribution function, specifically including:

[0126] Based on the probability values ​​of random vibration data corresponding to the acquisition time according to the probability distribution function, the statistical characteristics of the random vibration data are obtained, and the statistical characteristics include at least one of mean and variance.

[0127] Based on the preset vibration data interval generation rules, the statistical characteristics are processed by standard deviation to obtain the confidence interval.

[0128] In some embodiments of this application, the vibration data anomaly detection device further includes a data sequence update module, used to perform anomaly detection on the target vibration data based on the confidence interval:

[0129] If the target vibration data is abnormal vibration data, then the standard vibration data corresponding to the target vibration data is determined according to the statistical characteristics corresponding to the probability distribution function;

[0130] The standard vibration data is added to the historical vibration data sequence to update the historical vibration data sequence.

[0131] In some embodiments of this application, the determining module 702 is configured to generate a probability distribution function corresponding to the acquisition time based on each historical vibration data in the historical vibration data sequence, specifically including:

[0132] The historical vibration data in the historical vibration data sequence is input into a preset Gaussian distribution model to obtain the core function of the Gaussian model. The Gaussian model is a vibration data statistical model formed by inputting Gaussian white noise under the Bayesian framework.

[0133] Obtain the function type of the core function of the Gaussian model; if the function type is exponential, obtain the log-likelihood function corresponding to the exponential type.

[0134] Based on the historical vibration data sequence, the corresponding vibration calculation parameters of the log-likelihood function are calculated, and the probability distribution function corresponding to the acquisition time is generated based on the vibration calculation parameters.

[0135] In some embodiments of this application, the determining module 702 further includes: calculating the vibration calculation parameters corresponding to the log-likelihood function based on the historical vibration data sequence, and generating the probability distribution function corresponding to the acquisition time based on the vibration calculation parameters, specifically including:

[0136] The kernel function matrix corresponding to the historical vibration data sequence is determined based on the historical vibration data sequence and the Gaussian model kernel function.

[0137] Using the historical vibration data sequence and the kernel function matrix, the vibration calculation parameters corresponding to the log-likelihood function are solved;

[0138] The probability distribution function corresponding to the acquisition time is generated based on the preset posterior distribution function and the vibration calculation parameters.

[0139] In some embodiments of this application, the vibration data anomaly detection device further includes a feedback module, used for:

[0140] After performing anomaly detection on the target vibration data based on the confidence interval,

[0141] When the target vibration data is detected as abnormal vibration data, the target vibration data is accumulated into the abnormal vibration dataset;

[0142] If the number of abnormal vibration data in the abnormal vibration dataset exceeds a preset threshold, then the target data change information of each abnormal vibration data in the abnormal vibration dataset is statistically analyzed.

[0143] Based on the preset mapping relationship between anomaly handling information and data change information, the target anomaly handling information corresponding to the target data change information is determined;

[0144] Vibration data anomaly information is generated and fed back based on the target anomaly processing information and the target data change information.

[0145] In some embodiments of this application, the detection module 704 is used to perform anomaly detection on the target vibration data based on the confidence interval, specifically including:

[0146] If the target vibration data is within the confidence interval, then the target vibration data is determined to be normal vibration data;

[0147] If the target vibration data is outside the confidence interval, then the target vibration data is determined to be abnormal vibration data.

[0148] This implementation provides a vibration data anomaly detection device. It acquires the target vibration data to be detected, the acquisition time of the target vibration data, and the corresponding historical vibration data sequence. Then, based on each historical vibration data point in the historical vibration data sequence, it generates a probability distribution function corresponding to the acquisition time. Next, based on the probability distribution function and its corresponding statistical characteristics, it obtains a confidence interval. Finally, it performs anomaly detection on the target vibration data based on the confidence interval. By using the historical vibration data sequence and the probability distribution function corresponding to the acquisition time—that is, by determining the probability distribution function of possible vibration data corresponding to the acquisition time based on the historical vibration data sequence—it ensures data diversity and temporal correlation between corresponding data points. Then, based on the probability distribution function and its corresponding statistical characteristics, it obtains a confidence interval for judging whether the target vibration data to be detected is abnormal, avoiding the limitations of directly predicting using preset values ​​and ensuring data accuracy.

[0149] This invention also provides a vibration data anomaly detection device, such as... Figure 8 As shown, Figure 8 This is a schematic diagram of an embodiment of the vibration data anomaly detection device provided in this application.

[0150] The vibration data anomaly detection device integrates any of the vibration data anomaly detection devices provided in the embodiments of the present invention, and the vibration data anomaly detection device includes:

[0151] One or more processors;

[0152] Memory; and

[0153] One or more applications, wherein the one or more applications are stored in the memory and configured by the processor to execute the steps of the vibration data anomaly detection method described in any of the embodiments of the above-described vibration data anomaly detection method.

[0154] Specifically, a vibration data anomaly detection device may include components such as a processor 801 with one or more processing cores, a memory 802 with one or more computer-readable storage media, a power supply 803, and an input unit 804. Those skilled in the art will understand that... Figure 8 The structure of the vibration data anomaly detection device shown does not constitute a limitation on the vibration data anomaly detection device. It may include more or fewer components than shown, or combine certain components, or have different component arrangements. Wherein:

[0155] The processor 801 is the control center of the vibration data anomaly detection device. It connects to various parts of the device via various interfaces and lines. By running or executing software programs and / or modules stored in the memory 802, and by calling data stored in the memory 802, it performs various functions and processes data, thereby providing overall monitoring of the vibration data anomaly detection device. Optionally, the processor 801 may include one or more processing cores; preferably, the processor 801 may integrate an application processor and a modem processor, wherein the application processor mainly handles the operating system, user interface, and applications, while the modem processor mainly handles wireless communication. It is understood that the modem processor may not be integrated into the processor 801.

[0156] The memory 802 can be used to store software programs and modules. The processor 801 executes various functional applications and data processing by running the software programs and modules stored in the memory 802. The memory 802 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, application programs required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created based on the use of the vibration data anomaly detection device, etc. In addition, the memory 802 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device. Accordingly, the memory 802 may also include a memory controller to provide the processor 801 with access to the memory 802.

[0157] The vibration data anomaly detection device also includes a power supply 803 that supplies power to various components. Preferably, the power supply 803 can be logically connected to the processor 801 through a power management system, thereby enabling functions such as charging, discharging, and power consumption management through the power management system. The power supply 803 may also include one or more DC or AC power supplies, recharging systems, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components.

[0158] The vibration data anomaly detection device may also include an input unit 804, which can be used to receive input digital or character information, and generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function control.

[0159] Although not shown, the vibration data anomaly detection device may also include a display unit, etc., which will not be described in detail here. Specifically, in this embodiment, the processor 801 in the vibration data anomaly detection device loads the executable files corresponding to the processes of one or more application programs into the memory 802 according to the following instructions, and the processor 801 runs the application programs stored in the memory 802 to realize various functions, as follows:

[0160] Acquire the vibration data of the target to be detected, the acquisition time of the target vibration data, and the historical vibration data sequence corresponding to the target vibration data;

[0161] Based on each historical vibration data point in the historical vibration data sequence, a probability distribution function corresponding to the acquisition time is generated;

[0162] Based on the probability distribution function and the statistical characteristics corresponding to the probability distribution function, the confidence interval is obtained;

[0163] Anomaly detection is performed on the target vibration data based on the confidence interval.

[0164] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be performed by instructions, or by instructions controlling related hardware. These instructions can be stored in a computer-readable storage medium and loaded and executed by a processor.

[0165] Therefore, embodiments of the present invention provide a computer-readable storage medium, which may include: read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk, etc. A computer program is stored thereon, which is loaded by a processor to execute the steps in any of the vibration data anomaly detection methods provided in the embodiments of the present invention. For example, the computer program loaded by the processor can execute the following steps:

[0166] Acquire the vibration data of the target to be detected, the acquisition time of the target vibration data, and the historical vibration data sequence corresponding to the target vibration data;

[0167] Based on each historical vibration data point in the historical vibration data sequence, a probability distribution function corresponding to the acquisition time is generated;

[0168] Based on the probability distribution function and the statistical characteristics corresponding to the probability distribution function, the confidence interval is obtained;

[0169] Anomaly detection is performed on the target vibration data based on the confidence interval.

[0170] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the detailed descriptions of other embodiments above, which will not be repeated here.

[0171] In practice, each of the above units or structures can be implemented as an independent entity or can be arbitrarily combined to be implemented as the same or several entities. For the specific implementation of each of the above units or structures, please refer to the previous method embodiments, which will not be repeated here.

[0172] For details on the implementation of each of the above operations, please refer to the previous examples, which will not be repeated here.

[0173] The above provides a detailed description of a vibration data anomaly detection method, apparatus, device, and storage medium provided in the embodiments of this application. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for detecting vibration data anomalies, characterized in that, include: Acquire the vibration data of the target to be detected, the acquisition time of the target vibration data, and the historical vibration data sequence corresponding to the target vibration data; Based on each historical vibration data in the historical vibration data sequence, a probability distribution function corresponding to the acquisition time is generated; the probability distribution function of the acquisition time is a prediction function for the probability of occurrence of all possible vibration data at the acquisition time. The step of generating the probability distribution function corresponding to the acquisition time based on each historical vibration data in the historical vibration data sequence includes: The historical vibration data in the historical vibration data sequence is input into a preset Gaussian distribution model to obtain the core function of the Gaussian model. The Gaussian model is a vibration data statistical model formed by inputting Gaussian white noise under the Bayesian framework. Obtain the function type of the core function of the Gaussian model; if the function type is exponential, obtain the log-likelihood function corresponding to the exponential type. Based on the historical vibration data sequence, the vibration calculation parameters corresponding to the log-likelihood function are calculated, and the probability distribution function corresponding to the acquisition time is generated based on the vibration calculation parameters. Based on the probability distribution function and the statistical characteristics corresponding to the probability distribution function, the confidence interval is obtained; Anomaly detection is performed on the target vibration data based on the confidence interval.

2. The vibration data anomaly detection method according to claim 1, characterized in that, The process of obtaining the confidence interval based on the probability distribution function and the corresponding statistical characteristics includes: Based on the probability values ​​of the random vibration data corresponding to the acquisition time according to the probability distribution function, the statistical characteristics of the random vibration data are obtained, and the statistical characteristics include at least one of mean and variance. Based on the preset vibration data interval generation rules, the statistical characteristics are processed by standard deviation to obtain the confidence interval.

3. The vibration data anomaly detection method according to claim 1, characterized in that, After performing anomaly detection on the target vibration data based on the confidence interval, the method further includes: If the target vibration data is abnormal vibration data, then the standard vibration data corresponding to the target vibration data is determined according to the statistical characteristics corresponding to the probability distribution function; The standard vibration data is added to the historical vibration data sequence to update the historical vibration data sequence.

4. The vibration data anomaly detection method according to claim 1, characterized in that, The step of solving the vibration calculation parameters corresponding to the log-likelihood function based on the historical vibration data sequence, and generating the probability distribution function corresponding to the acquisition time based on the vibration calculation parameters, includes: The kernel function matrix corresponding to the historical vibration data sequence is determined based on the historical vibration data sequence and the Gaussian model kernel function. Using the historical vibration data sequence and the kernel function matrix, the vibration calculation parameters corresponding to the log-likelihood function are solved; The probability distribution function corresponding to the acquisition time is generated based on the preset posterior distribution function and the vibration calculation parameters.

5. The vibration data anomaly detection method according to claim 1, characterized in that, After performing anomaly detection on the target vibration data based on the confidence interval, the process includes: When the target vibration data is detected as abnormal vibration data, the target vibration data is accumulated into the abnormal vibration dataset; If the number of abnormal vibration data in the abnormal vibration dataset exceeds a preset threshold, then the target data change information of each abnormal vibration data in the abnormal vibration dataset is statistically analyzed. Based on the preset mapping relationship between anomaly handling information and data change information, the target anomaly handling information corresponding to the target data change information is determined; Vibration data anomaly information is generated and fed back based on the target anomaly processing information and the target data change information.

6. The vibration data anomaly detection method according to claim 1, characterized in that, The anomaly detection of the target vibration data based on the confidence interval includes: If the target vibration data is within the confidence interval, then the target vibration data is determined to be normal vibration data; If the target vibration data is outside the confidence interval, then the target vibration data is determined to be abnormal vibration data.

7. A vibration data anomaly detection device, characterized in that, The vibration data anomaly detection device includes: Acquisition module: used to acquire the vibration data of the target to be detected, the acquisition time of the target vibration data, and the historical vibration data sequence corresponding to the target vibration data; Determining module: used to generate a probability distribution function corresponding to the acquisition time based on each historical vibration data in the historical vibration data sequence; the probability distribution function of the acquisition time is a prediction function of the probability of occurrence of all possible vibration data at the acquisition time; The determining module specifically includes functions for: The historical vibration data in the historical vibration data sequence is input into a preset Gaussian distribution model to obtain the core function of the Gaussian model. The Gaussian model is a vibration data statistical model formed by inputting Gaussian white noise under the Bayesian framework. Obtain the function type of the core function of the Gaussian model; if the function type is exponential, obtain the log-likelihood function corresponding to the exponential type. Based on the historical vibration data sequence, the vibration calculation parameters corresponding to the log-likelihood function are calculated, and the probability distribution function corresponding to the acquisition time is generated based on the vibration calculation parameters. Interval determination module: used to obtain confidence intervals based on the probability distribution function and the statistical characteristics corresponding to the probability distribution function; Detection module: used to perform anomaly detection on the target vibration data based on the confidence interval.

8. A vibration data anomaly detection device, characterized in that, The vibration data anomaly detection device includes: One or more processors; Memory; and One or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the processor to implement the vibration data anomaly detection method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, It stores a computer program, which is loaded by a processor to execute the steps in the vibration data anomaly detection method according to any one of claims 1 to 6.

Images