A trend early warning level determination method, device, equipment and medium

Abstract

The application discloses a trend early warning level determination method and device, equipment and medium, relates to the field of mechanical equipment state monitoring and fault diagnosis, and comprises the following steps: performing change point detection on motor trend data by using a Bayesian change point detection algorithm to obtain one or more change point position information in the motor trend data; determining each change point type corresponding to each change point position information in the motor trend data; performing data division on the motor trend data according to each change point position information and by using a mean change point to obtain one or more motor trend data segments; screening out a motor trend data segment in a running state as motor running trend data; and determining a trend early warning level of the motor trend data based on the change point type in the motor running trend data.

Description

Technical Field

[0001] This invention relates to the field of mechanical equipment condition monitoring and fault diagnosis, and in particular to a method, apparatus, equipment and medium for determining trend early warning levels. Background Technology

[0002] Electric motors are widely used in industrial production. By monitoring the magnitude and trend changes of vibration characteristic quantities closely related to the motor's operating status, early warnings of abnormal motor conditions can be provided. Commonly used motor condition early warning methods include threshold warnings and trend warnings. Threshold warnings are based on absolute vibration indicators and are typically triggered only when vibration energy is high and the motor fault is in its middle or late stages. They cannot effectively monitor changes in vibration characteristics below the threshold. Trend warnings are based on relative vibration indicators and are generally divided into short-term and long-term trend warnings. Current trend warning methods typically focus on monitoring relatively obvious and observable significant changes in vibration characteristics. For weak trend information with small vibration characteristic fluctuations, such as the peak acceleration trend of motor rolling bearings, since early rolling bearing faults usually present impact components with randomly distributed amplitudes, the peak acceleration trend shows little energy change but a larger distribution range compared to normal operating conditions. Analysts and existing trend warning methods find it difficult to observe changes in characteristic values ​​from the trend data, making timely and accurate anomaly monitoring impossible and failing to detect weak early motor faults in a timely manner.

[0003] As can be seen from the above, how to overcome the missed reports caused by the need to manually set step size and iteratively search for trend change points in traditional trend early warning, and how to effectively detect subtle changes in trend data while completing the reliable detection of obvious changes in trend data, so as to realize early warning of subtle changes in the motor's early state and improve the reliability and accuracy of early warning, is a problem to be solved in this field. Summary of the Invention

[0004] In view of this, the purpose of this invention is to provide a method, apparatus, device, and medium for determining trend warning levels. This method overcomes the missed detections caused by the need for manually setting step sizes and iteratively searching for trend change points in traditional trend warning systems. While reliably detecting significant changes in trend data, it also effectively detects subtle changes, enabling early warning of minor changes in the motor's early state, thus improving the reliability and accuracy of the warning. The specific solution is as follows:

[0005] Firstly, this application discloses a method for determining trend early warning levels, including:

[0006] Acquire motor trend data, and use the Bayesian change point detection algorithm to detect change points in the motor trend data to obtain one or more change point location information in the motor trend data;

[0007] Identify the types of change points corresponding to the location information of each change point in the motor trend data; wherein, the change point types include mean change points;

[0008] Based on the location information of each variable point and using the mean variable point, the motor trend data is divided to obtain one or more motor trend data segments. The motor trend data segments in the running state are selected from each motor trend data segment as motor running trend data.

[0009] The trend warning level of the motor trend data is determined based on the type of change point in the motor operation trend data.

[0010] Optionally, acquiring motor trend data includes:

[0011] Acquire motor trend data uploaded by edge data acquisition devices; wherein, the motor trend data includes trend data and the acquisition time corresponding to the trend data; the data type of the motor operation trend data includes the peak trend of vibration acceleration signal.

[0012] Optionally, the data type of the motor operating trend may also include at least one of the following: vibration acceleration signal kurtosis trend, vibration acceleration signal envelope value trend, and vibration acceleration signal RMS value trend.

[0013] Optionally, the step of using a Bayesian change point detection algorithm to detect change points in the motor trend data to obtain one or more change point location information in the motor trend data includes:

[0014] The statistical distribution model of the motor trend data is determined, and the change point detection parameters are set. The change point detection parameters are input into the statistical distribution model to obtain a probability matrix that characterizes the current trend point as a change point.

[0015] The probability matrix is ​​summed column by column to obtain the actual probability that the current trend point is a change point, and it is determined whether the actual probability is greater than a preset threshold.

[0016] If the actual probability is greater than a preset threshold, the change point location information is determined, and the statistical distribution model is initialized with parameters. The change point detection parameters are updated, and then the process jumps to the step of inputting the change point detection parameters into the statistical distribution model to obtain one or more change point location information in the motor trend data. If the actual probability is not greater than the preset threshold, the current motor trend data is obtained, and the change point detection parameters of the statistical distribution model are updated based on the current motor trend data and the motor trend data.

[0017] Optionally, determining the type of each variable point corresponding to the location information of each variable point in the motor trend data includes:

[0018] The change point location information in the motor trend data is input into the statistical distribution model to determine the change point type, so as to obtain the change point type corresponding to the change point location information in the motor trend data.

[0019] The variable point type also includes variance variable points.

[0020] Optionally, determining the trend warning level of the motor trend data based on the change point type in the motor operation trend data includes:

[0021] The actual probability is used as the confidence level at the change point;

[0022] The trend warning level of the motor trend data is determined based on the change point type and the change point confidence level in the motor operation trend data.

[0023] Optionally, the step of inputting the location information of each variable point in the motor trend data into the statistical distribution model for variable point type determination includes:

[0024] Using the statistical distribution model, the F-test method is applied to the variable points corresponding to the variable point location information in the motor trend data to perform the homogeneity of variance assumption on the variable point.

[0025] If the homogeneity of variance assumption is not met, the point of change is determined to be a point of change of mean. If the homogeneity of variance assumption is met, the Aspin-Welch approximate T-test is used to assume the equality of mean for the point of change. If the equality of mean assumption is not met, the point of change is determined to be a point of change of variance.

[0026] Optionally, the step of dividing the motor trend data according to the location information of each of the changing points and using the mean changing point to obtain one or more motor trend data segments includes:

[0027] From the motor trend data, data of the type of vibration acceleration signal peak trend are determined. The time segment of each data segment is determined according to the position information of each change point in the data. The motor trend data is divided using the mean change point and according to the time segment of each data segment to obtain each motor trend data segment.

[0028] The state of each motor trend data segment is determined;

[0029] The states of each motor trend data segment include running state and shutdown state.

[0030] Secondly, this application discloses a trend early warning level determination device, comprising:

[0031] The change point detection module is used to acquire motor trend data and use the Bayesian change point detection algorithm to detect change points in the motor trend data in order to obtain one or more change point position information in the motor trend data.

[0032] The variable point type determination module is used to determine the variable point type corresponding to the variable point position information in the motor trend data; wherein, the variable point type includes the mean variable point;

[0033] The data segmentation and filtering module is used to segment the motor trend data according to the location information of each variable point and using the mean variable point to obtain one or more motor trend data segments, and to filter out the motor trend data segments in the running state from each motor trend data segment as motor running trend data.

[0034] The trend warning level determination module is used to determine the trend warning level of the motor trend data based on the change point type in the motor operation trend data.

[0035] Thirdly, this application discloses an electronic device, including:

[0036] Memory, used to store computer programs;

[0037] A processor is used to execute the computer program to implement the aforementioned trend warning level determination method.

[0038] Fourthly, this application discloses a non-volatile computer storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the steps of the aforementioned disclosed trend warning level determination method.

[0039] As can be seen, this application provides a method for determining the trend warning level, including acquiring motor trend data; using a Bayesian change point detection algorithm to perform change point detection on the motor trend data to obtain one or more change point location information in the motor trend data; determining each change point type corresponding to each change point location information in the motor trend data; wherein, the change point type includes mean change points; dividing the motor trend data according to each change point location information and using the mean change points to obtain one or more motor trend data segments; selecting the motor trend data segments in the operating state from each motor trend data segment as motor operating trend data; and determining the trend warning level of the motor trend data based on the change point type in the motor operating trend data. This application utilizes motor trend data to accurately identify change points within the trend data, overcoming the missed detections caused by the need for manually setting step sizes and iteratively searching for trend change points in traditional trend early warning systems. While reliably detecting obvious changes in trend data, it can effectively detect subtle changes hidden within the trend data, enabling early warning of subtle changes in the motor's state. Furthermore, the employed Bayesian change point detection algorithm fully utilizes motor trend data, eliminating false alarms caused by traditional trend early warning systems that rely solely on trend change information. This improves the reliability and accuracy of early warnings, increases equipment availability, reduces maintenance costs, and enhances operational efficiency. It effectively detects subtle changes hidden within trend data, providing timely warnings of early motor faults. Attached Figure Description

[0040] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0041] Figure 1 This is a flowchart of a trend early warning level determination method disclosed in this application;

[0042] Figure 2 A flowchart of another trend warning level determination method disclosed in this application;

[0043] Figure 3 Here is a flowchart of a method for determining a trend early warning level as disclosed in this application;

[0044] Figure 4 This is an example diagram showing the results after implementing a trend early warning system as disclosed in this application;

[0045] Figure 5 This is an example diagram showing the results after another implementation of the trend warning disclosed in this application;

[0046] Figure 6 This is a schematic diagram of a trend early warning level determination device disclosed in this application;

[0047] Figure 7 This application provides a structural diagram of an electronic device. Detailed Implementation

[0048] The technical solutions of the embodiments of the present invention 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.

[0049] Electric motors are widely used in industrial production. By monitoring the magnitude and trend changes of vibration characteristic quantities closely related to the motor's operating status, early warnings of abnormal motor conditions can be provided. Commonly used motor condition early warning methods include threshold warnings and trend warnings. Threshold warnings are based on absolute vibration indicators and are typically triggered only when vibration energy is high and the motor fault is in its middle or late stages. They cannot effectively monitor changes in vibration characteristics below the threshold. Trend warnings are based on relative vibration indicators and are generally divided into short-term and long-term trend warnings. Current trend warning methods typically focus on monitoring relatively obvious and observable significant changes in vibration characteristics. For weak trend information with small vibration characteristic fluctuations, such as the peak acceleration trend of motor rolling bearings, since early rolling bearing faults usually present impact components with randomly distributed amplitudes, the peak acceleration trend shows little energy change but a larger distribution range compared to normal operating conditions. Analysts and existing trend warning methods find it difficult to observe changes in characteristic values ​​from the trend data, making timely and accurate anomaly monitoring impossible and failing to detect weak early motor faults in a timely manner. As can be seen from the above, how to overcome the missed reports caused by the need to manually set step size and iteratively search for trend change points in traditional trend early warning, and how to effectively detect subtle changes in trend data while completing the reliable detection of obvious changes in trend data, so as to realize early warning of subtle changes in the motor's early state and improve the reliability and accuracy of early warning, is a problem to be solved in this field.

[0050] See Figure 1 As shown in the figure, an embodiment of the present invention discloses a method for determining a trend warning level, which may specifically include:

[0051] Step S11: Obtain motor trend data, and use the Bayesian change point detection algorithm to detect change points in the motor trend data to obtain one or more change point location information in the motor trend data.

[0052] In this embodiment, motor trend data uploaded by a vibration acceleration sensor and an edge data acquisition device is acquired. The motor trend data includes trend feature data and the acquisition time corresponding to the trend data. The data type of the motor operation trend data includes the peak trend of the vibration acceleration signal. Then, a Bayesian change point detection algorithm is used to detect change points in the motor trend data to obtain the location information of each change point in the motor trend data. Specifically, the motor trend data x uploaded by the vibration acceleration sensor and the edge data acquisition device is acquired. t ,t=1,2,...,T, where x t For trend data, t = 1, 2, ..., T represents the collection time corresponding to the trend data; a segment of the trend data can be represented as x. t1：t2 Where t1 and t2 represent the start and end times of the trend data, respectively, and t1 ≤ t2. Since Bayesian change point detection is a predictive trend discrimination method, it needs to utilize the statistical distribution information of existing trend data. Therefore, the duration of the acquired motor operation trend data needs to be no less than one week. The data type of the acquired motor operation trend data should be able to characterize the early subtle changes in the motor's state, and should include at least one of the following: peak trend of vibration acceleration signal, kurtosis trend of vibration acceleration signal, and envelope value trend of vibration acceleration signal.

[0053] In this embodiment, the change point detection process is as follows: A statistical distribution model of the motor trend data is determined, and change point detection parameters are set. The change point detection parameters are input into the statistical distribution model to obtain a probability matrix representing the current trend point as a change point. The probability matrix is ​​summed column-wise to obtain the actual probability representing the current trend point as a change point. It is determined whether the actual probability is greater than a preset threshold. If the actual probability is greater than the preset threshold, the change point location information is determined, and the statistical distribution model is initialized. The change point detection parameters are updated, and then the process jumps to the step of inputting the change point detection parameters into the statistical distribution model to obtain one or more change point location information in the motor trend data. Specifically, a statistical model is selected based on the prior information of the motor trend data distribution, and initial conditions are set. Where p(r0) represents the probability that the first time point of the motor trend data is a turning point, v prior x prior The hyperparameters of the statistical distribution model are automatically updated according to the rules in each iteration. The probability matrix that the current trend point is a change point is calculated according to the following formula:

[0054]

[0055] Where, r t p(r) represents the length of the current trend segment.t-1 ,x 1:t-1 ) is the recursive travel length matrix. For prior probability prediction, (L) represents any possible travel length of the current trend segment, and H(r) represents the travel length of the current trend segment. t-1 ) is the risk function. f(r t-1 ) indicates that the current trend segment has a travel length of r. t-1 The probability, S(r) t-1 Let be the survival function at time t-1;

[0056] The actual probability of the current trend point being a change point can be obtained by summing the columns of the probability matrix of the change points in the motor trend data. When the actual probability of the change point is greater than a preset threshold, the change point detection parameters of the statistical distribution model are reinitialized and the trend data change point detection process is restarted. This avoids changes in the statistical distribution model causing changes in the accuracy of change point detection. When the actual probability of the change point is not greater than the preset threshold, the current motor trend data is obtained, and the change point detection parameters of the statistical distribution model are updated based on the current motor trend data and the motor trend data.

[0057]

[0058] Step S12: Determine the type of each change point corresponding to the change point location information in the motor trend data; wherein, the change point type includes the mean change point.

[0059] Step S13: Divide the motor trend data according to the location information of each variable point and the mean variable point to obtain one or more motor trend data segments, and select the motor trend data segment in the running state from each motor trend data segment as the motor running trend data.

[0060] Step S14: Determine the trend warning level of the motor trend data based on the change point type in the motor operation trend data.

[0061] In this embodiment, the actual probability is used as the change point confidence level; the trend warning level of the motor trend data is determined based on the change point type and the change point confidence level in the motor operation trend data.

[0062] Specifically, the trend warning level is determined based on the type and confidence level of the change points in the motor trend data. Variance change points are mainly used to detect trend data containing vibration impact information, while mean change points are used to detect trend data containing vibration energy changes. The type of change point refers to which parameter of the trend distribution model changes, such as mean change point and variance change point, which are two parameters of the Gaussian model. The confidence level of the change point is the actual probability calculated above.

[0063] In this embodiment, motor trend data is acquired, and a Bayesian change point detection algorithm is used to detect change points in the motor trend data to obtain one or more change point location information in the motor trend data; each change point type corresponding to each change point location information in the motor trend data is determined; wherein, the change point type includes mean change points; the motor trend data is divided according to each change point location information and using the mean change points to obtain one or more motor trend data segments; the motor trend data segments in the running state are selected from each of the motor trend data segments as motor running trend data; the trend warning level of the motor trend data is determined based on the change point type in the motor running trend data. This application utilizes motor trend data to accurately identify change points within the trend data, overcoming the missed detections caused by the need for manually setting step sizes and iteratively searching for trend change points in traditional trend early warning systems. While reliably detecting obvious changes in trend data, it can effectively detect subtle changes hidden within the trend data, enabling early warning of subtle changes in the motor's state. Furthermore, the employed Bayesian change point detection algorithm fully utilizes motor trend data, eliminating false alarms caused by traditional trend early warning systems that rely solely on trend change information. This improves the reliability and accuracy of early warnings, increases equipment availability, reduces maintenance costs, and enhances operational efficiency. It effectively detects subtle changes hidden within trend data, providing timely warnings of early motor faults.

[0064] See Figure 2 As shown in the figure, an embodiment of the present invention discloses a method for determining a trend warning level, which may specifically include:

[0065] Step S21: Obtain motor trend data, and use the Bayesian change point detection algorithm to detect change points in the motor trend data to obtain one or more change point location information in the motor trend data.

[0066] Step S22: Input the location information of each variable point in the motor trend data into the statistical distribution model to determine the variable point type, so as to obtain the variable point type corresponding to the location information of each variable point in the motor trend data; wherein, the variable point type also includes variance variable points.

[0067] In this embodiment, the statistical distribution model is used to apply the F-test to the variable point corresponding to the variable point location information in the motor trend data to assume homogeneity of variance. If the homogeneity of variance assumption is not met, the variable point is determined to be a mean variable point. If the homogeneity of variance assumption is met, the Aspin-Welch approximate T-test is used to assume equal mean for the variable point. If the equal mean assumption is not met, the variable point is determined to be a variance variable point.

[0068] In this embodiment, the trend data of the motor undergoes structural changes before and after a change point, thereby determining the type of change point. Change point types include mean change points and variance change points, etc. The method moves from directly identifying significant changes in the trend using a statistical distribution model to determining trend changes based on the statistical distribution model of the trend data. The hyperparameters of the statistical distribution model are used as the basis for judging whether the parameter data has changed. The statistical distribution model is selected according to the type of motor trend, which can effectively detect subtle changes hidden in the trend.

[0069] Step S23: Determine the data type as peak trend of vibration acceleration signal from the motor trend data; determine the time segment of each data segment based on the position information of each change point in the data; divide the motor trend data using the mean change point and according to the time segment of each data segment to obtain each motor trend data segment; determine the state of each motor trend data segment; and select the motor trend data segment in the running state from each motor trend data segment as the motor running trend data; wherein, the state of each motor trend data segment includes the running state and the shutdown state.

[0070] In this embodiment, the adaptive operating condition division of motor trend data is achieved by utilizing the mean change point. The duration of each segment is obtained based on the trend data change point, and the motor trend data is divided according to the duration to obtain each motor trend data segment. The state of each motor trend data segment is determined, thereby extracting the motor trend data segment in the running state as the motor running trend data. Here, the adaptive operating condition division refers to removing the motor trend data segment in the shutdown state based on the trend change and extracting the motor trend data segment in the running state for further analysis later.

[0071] The adaptive working condition division rule for trend data using mean change points is as follows: The duration of each segment is obtained based on the change points of the peak vibration acceleration trend data, and the trend data is divided according to the duration. The trend average value of each trend segment is calculated, and the threshold for judging the shutdown state based on the peak acceleration characteristic is set to 0.5 m / s². 2 Based on the discrimination threshold, it is determined whether each motor trend data segment is in a stopped state or a running state, thereby extracting the motor running trend data. Since the magnitude of the acceleration kurtosis trend and envelope value trend amplitude is not related to the motor vibration energy, it is necessary to use the stopping time identified by the acceleration peak trend to realize the working condition division of acceleration kurtosis trend and envelope trend.

[0072] Step S24: Determine the trend warning level of the motor trend data based on the change point type in the motor operation trend data.

[0073] The trend warning level is determined based on the type and confidence level of the change points in the motor operation trend data. Variance change points are mainly used to detect trend data containing vibration impact information in the vibration trend. For example, early and weak state changes in motor operation are uniformly set as Level 1 warning. Mean change points are used to detect trend data containing changes in vibration energy in the vibration trend. For significant state changes in motor operation, a Level 2 warning threshold of 0.5 is set. If the probability of a mean change point is greater than the Level 2 warning threshold, it is a Level 2 warning. A Level 3 trend warning threshold of 0.8 is set. If the probability of a mean change point is greater than or equal to the Level 3 warning threshold, it is a Level 3 trend alarm warning. Other situations are considered normal trends.

[0074] The specific process for this application is as follows: Figure 3 As shown, (1) obtain motor trend data; (2) use the Bayesian variable point detection algorithm to perform variable point detection on the motor trend data to obtain one or more variable point location information in the motor trend data; (3) determine each variable point type corresponding to each variable point location information in the motor trend data; the variable point type includes mean variable point and variance variable point; (4) use mean variable point to realize adaptive working condition division of trend data: according to each variable point location information and using mean variable point to divide the motor trend data to obtain one or more motor trend data segments; (5) select the motor trend data segment in the running state from each motor trend data segment as motor running trend data; (6) use the actual probability as the variable point confidence level, and determine the trend warning level of the motor trend data based on the variable point type and the variable point confidence level in the motor running trend data.

[0075] The results of this application after implementing trend prediction for the peak vibration acceleration of a certain fixed-frequency three-phase asynchronous motor are as follows: Figure 4 As shown in the figure, the peak motor acceleration fluctuates significantly during this period, indicating a rapid change in the equipment's operating status. In such cases, an immediate alarm is required to indicate an abnormal equipment condition. By applying the trend warning technology of this invention, multiple vibration energy fluctuations can be reliably and quickly detected. In this embodiment, the probability of the first and third abrupt change points is greater than the third trend warning threshold, and the probability of the second abrupt change point is greater than the second trend warning threshold.

[0076] The results of this application after implementing trend prediction for the peak vibration acceleration of a fixed-frequency three-phase asynchronous motor are as follows: Figure 5As shown in the figure, the peak motor acceleration distribution remained stable overall during this period, with slight distribution anomalies observed in local timeframes, corresponding to a slow change in the equipment's operating status. By applying the trend warning technology of this invention, multiple distribution anomalies can still be reliably and quickly detected. In this embodiment, the first, second, and third abrupt change points are all variance abrupt change types, corresponding to a first-level trend warning according to the trend warning rules.

[0077] In this embodiment, motor trend data is acquired, and a Bayesian change point detection algorithm is used to detect change points in the motor trend data to obtain one or more change point location information in the motor trend data; each change point type corresponding to each change point location information in the motor trend data is determined; wherein, the change point type includes mean change points; the motor trend data is divided according to each change point location information and using the mean change points to obtain one or more motor trend data segments; the motor trend data segments in the running state are selected from each of the motor trend data segments as motor running trend data; the trend warning level of the motor trend data is determined based on the change point type in the motor running trend data. This application utilizes motor trend data to accurately identify change points within the trend data, overcoming the missed detections caused by the need for manually setting step sizes and iteratively searching for trend change points in traditional trend early warning systems. While reliably detecting obvious changes in trend data, it can effectively detect subtle changes hidden within the trend data, enabling early warning of subtle changes in the motor's state. Furthermore, the employed Bayesian change point detection algorithm fully utilizes motor trend data, eliminating false alarms caused by traditional trend early warning systems that rely solely on trend change information. This improves the reliability and accuracy of early warnings, increases equipment availability, reduces maintenance costs, and enhances operational efficiency. It effectively detects subtle changes hidden within trend data, providing timely warnings of early motor faults.

[0078] See Figure 6 As shown in the figure, an embodiment of the present invention discloses a trend early warning level determination device, which may specifically include:

[0079] The change point detection module 11 is used to acquire motor trend data and use the Bayesian change point detection algorithm to detect change points in the motor trend data in order to obtain one or more change point position information in the motor trend data.

[0080] The variable point type determination module 12 is used to determine the variable point type corresponding to the variable point position information in the motor trend data; wherein, the variable point type includes the mean variable point;

[0081] The data segmentation and filtering module 13 is used to segment the motor trend data according to the location information of each variable point and using the mean variable point to obtain one or more motor trend data segments, and to filter out the motor trend data segments in the running state from each motor trend data segment as motor running trend data.

[0082] The trend warning level determination module 14 is used to determine the trend warning level of the motor trend data based on the change point type in the motor operation trend data.

[0083] In this embodiment, motor trend data is acquired, and a Bayesian change point detection algorithm is used to detect change points in the motor trend data to obtain one or more change point location information in the motor trend data; each change point type corresponding to each change point location information in the motor trend data is determined; wherein, the change point type includes mean change points; the motor trend data is divided according to each change point location information and using the mean change points to obtain one or more motor trend data segments; the motor trend data segments in the running state are selected from each of the motor trend data segments as motor running trend data; the trend warning level of the motor trend data is determined based on the change point type in the motor running trend data. This application utilizes motor trend data to accurately identify change points within the trend data, overcoming the missed detections caused by the need for manually setting step sizes and iteratively searching for trend change points in traditional trend early warning systems. While reliably detecting obvious changes in trend data, it can effectively detect subtle changes hidden within the trend data, enabling early warning of subtle changes in the motor's state. Furthermore, the employed Bayesian change point detection algorithm fully utilizes motor trend data, eliminating false alarms caused by traditional trend early warning systems that rely solely on trend change information. This improves the reliability and accuracy of early warnings, increases equipment availability, reduces maintenance costs, and enhances operational efficiency. It effectively detects subtle changes hidden within trend data, providing timely warnings of early motor faults.

[0084] In some specific embodiments, the change point detection module 11 may specifically include:

[0085] The data acquisition module is used to acquire motor trend data uploaded by the edge data acquisition device; wherein, the motor trend data includes trend data and the acquisition time corresponding to the trend data; the data type of the motor operation trend data includes the peak trend of vibration acceleration signal.

[0086] In some specific embodiments, the data type of the motor operating trend in the trend warning level determination device may also include at least one of the vibration acceleration signal kurtosis trend, vibration acceleration signal envelope value trend, and vibration acceleration signal effective value trend.

[0087] In some specific embodiments, the change point detection module 11 may specifically include:

[0088] The probability matrix determination module is used to determine the statistical distribution model of the motor trend data and set the change point detection parameters. The change point detection parameters are input into the statistical distribution model to obtain a probability matrix that characterizes the current trend point as a change point.

[0089] The judgment module is used to sum the probability matrix column by column to obtain the actual probability that the current trend point is a change point, and to determine whether the actual probability is greater than a preset threshold.

[0090] The variable point location information determination module is used to determine the variable point location information if the actual probability is greater than a preset threshold, and to perform parameter initialization operation on the statistical distribution model, update the variable point detection parameters, and then jump to the step of inputting the variable point detection parameters into the statistical distribution model to obtain one or more variable point location information in the motor trend data.

[0091] The change point detection parameter update module is used to obtain the current motor trend data if the actual probability is not greater than a preset threshold, and update the change point detection parameters of the statistical distribution model based on the current motor trend data and the motor trend data.

[0092] In some specific embodiments, the variable point type determination module 12 may specifically include:

[0093] The variable point type discrimination module is used to input the variable point location information in the motor trend data into the statistical distribution model to discriminate the variable point type, so as to obtain the variable point type corresponding to the variable point location information in the motor trend data.

[0094] The variable point type also includes variance variable points.

[0095] In some specific embodiments, the trend warning level determination module 14 may specifically include:

[0096] The change-point confidence determination module is used to use the actual probability as the change-point confidence.

[0097] The trend warning level determination module is used to determine the trend warning level of the motor trend data based on the change point type and the change point confidence level in the motor operation trend data.

[0098] In some specific embodiments, the variable point type determination module 12 may specifically include:

[0099] The homogeneity of variance assumption module is used to apply the F-test to assume homogeneity of variance for the variable points corresponding to the variable point location information in the motor trend data using the statistical distribution model.

[0100] The mean equality assumption module is used to determine if a point of change is a mean change point if the homogeneity of variance assumption is not met; if the homogeneity of variance assumption is met, the Aspin-Welch approximate T-test is used to assume the mean equality of the point of change; if the mean equality assumption is not met, the point of change is determined to be a variance change point.

[0101] In some specific embodiments, the data segmentation and filtering module 13 may specifically include:

[0102] The data segmentation module is used to determine the data with the data type of vibration acceleration signal peak trend from the motor trend data, determine the time segment of each data segment according to the position information of each change point in the data, and divide the motor trend data using the mean change point and according to the time segment of each data segment to obtain each motor trend data segment.

[0103] The data segmentation status determination module is used to determine the status of each motor trend data segment; wherein, the status of each motor trend data segment includes running status and shutdown status.

[0104] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device 20 may specifically include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the trend warning level determination method performed by the electronic device disclosed in any of the foregoing embodiments.

[0105] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0106] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored on it include operating system 221, computer program 222 and data 223, etc., and the storage method can be temporary storage or permanent storage.

[0107] The operating system 221 manages and controls the various hardware devices on the electronic device 20 and the computer program 222 to enable the processor 21 to perform calculations and processing on the data 223 in the memory 22. The operating system 221 can be Windows, Unix, Linux, etc. The computer program 222, in addition to including a computer program capable of performing the trend warning level determination method executed by the electronic device 20 as disclosed in any of the foregoing embodiments, may further include computer programs capable of performing other specific tasks. The data 223 may include data received by the trend warning level determination device from external devices, as well as data collected by its own input / output interface 25.

[0108] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0109] Furthermore, this application also discloses a non-volatile computer-readable storage medium storing a computer program. When the computer program is loaded and executed by a processor, it implements the trend warning level determination method steps disclosed in any of the foregoing embodiments.

[0110] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0111] The above provides a detailed description of the trend warning level determination method, apparatus, device, and storage medium provided by the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is 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 determining trend early warning levels, characterized in that, include: Acquire motor trend data, and use the Bayesian change point detection algorithm to detect change points in the motor trend data to obtain one or more change point location information in the motor trend data; Identify the types of change points corresponding to the location information of each change point in the motor trend data; wherein, the change point types include mean change points; Based on the location information of each variable point and using the mean variable point, the motor trend data is divided to obtain one or more motor trend data segments. The motor trend data segments in the running state are selected from each motor trend data segment as motor running trend data. The trend warning level of the motor trend data is determined based on the type of change point in the motor operation trend data. The acquisition of motor trend data includes: Acquire motor trend data uploaded by edge data acquisition devices; wherein, the motor trend data includes trend data and the acquisition time corresponding to the trend data; the data type of the motor operation trend data includes the peak trend of vibration acceleration signal; The data type of the motor operation trend also includes at least one of the following: vibration acceleration signal kurtosis trend, vibration acceleration signal envelope value trend, and vibration acceleration signal effective value trend; The step of dividing the motor trend data according to the location information of each of the changing points and using the mean changing points to obtain one or more motor trend data segments includes: From the motor trend data, data of the type of vibration acceleration signal peak trend are determined. The time segment of each data segment is determined according to the position information of each change point in the data. The motor trend data is divided using the mean change point and according to the time segment of each data segment to obtain each motor trend data segment. The state of each motor trend data segment is determined; The states of each motor trend data segment include running state and shutdown state.

2. The method for determining the trend early warning level according to claim 1, characterized in that, The step of using a Bayesian change point detection algorithm to detect change points in the motor trend data, in order to obtain one or more change point location information in the motor trend data, includes: The statistical distribution model of the motor trend data is determined, and the change point detection parameters are set. The change point detection parameters are input into the statistical distribution model to obtain a probability matrix that characterizes the current trend point as a change point. The probability matrix is ​​summed column by column to obtain the actual probability that the current trend point is a change point, and it is determined whether the actual probability is greater than a preset threshold. If the actual probability is greater than a preset threshold, the change point location information is determined, and the statistical distribution model is initialized with parameters. The change point detection parameters are updated, and then the process jumps to the step of inputting the change point detection parameters into the statistical distribution model to obtain one or more change point location information in the motor trend data. If the actual probability is not greater than the preset threshold, the current motor trend data is obtained, and the change point detection parameters of the statistical distribution model are updated based on the current motor trend data and the motor trend data.

3. The method for determining the trend early warning level according to claim 2, characterized in that, The determination of the change point type corresponding to the change point position information in the motor trend data includes: The change point location information in the motor trend data is input into the statistical distribution model to determine the change point type, so as to obtain the change point type corresponding to the change point location information in the motor trend data. The variable point type also includes variance variable points.

4. The method for determining the trend early warning level according to claim 2, characterized in that, The determination of the trend warning level of the motor trend data based on the change point type in the motor operation trend data includes: The actual probability is used as the confidence level at the change point; The trend warning level of the motor trend data is determined based on the change point type and the change point confidence level in the motor operation trend data.

5. The method for determining the trend early warning level according to claim 3, characterized in that, The step of inputting the location information of each change point in the motor trend data into the statistical distribution model for change point type determination includes: Using the statistical distribution model, the F-test method is applied to the variable points corresponding to the variable point location information in the motor trend data to perform the homogeneity of variance assumption on the variable point. If the homogeneity of variance assumption is not met, the point of change is determined to be a point of change of mean. If the homogeneity of variance assumption is met, the Aspin-Welch approximate T-test is used to assume the equality of mean for the point of change. If the equality of mean assumption is not met, the point of change is determined to be a point of change of variance.

6. A trend early warning level determination device, characterized in that, include: The change point detection module is used to acquire motor trend data and use the Bayesian change point detection algorithm to detect change points in the motor trend data in order to obtain one or more change point position information in the motor trend data. The variable point type determination module is used to determine the variable point type corresponding to the variable point position information in the motor trend data; wherein, the variable point type includes the mean variable point; The data segmentation and filtering module is used to segment the motor trend data according to the location information of each variable point and using the mean variable point to obtain one or more motor trend data segments, and to filter out the motor trend data segments in the running state from each motor trend data segment as motor running trend data. The trend warning level determination module is used to determine the trend warning level of the motor trend data based on the change point type in the motor operation trend data. The change point detection module includes: The data acquisition module is used to acquire motor trend data uploaded by the edge data acquisition device; wherein, the motor trend data includes trend data and the acquisition time corresponding to the trend data; the data type of the motor operation trend data includes the peak trend of vibration acceleration signal; The data type of the motor operation trend in the trend early warning level determination device also includes at least one of the following: vibration acceleration signal kurtosis trend, vibration acceleration signal envelope value trend, and vibration acceleration signal effective value trend. The data segmentation and filtering module is specifically used to determine data with vibration acceleration signal peak trend as the data type from the motor trend data, determine the time segment of each data segment based on the position information of each change point in the data, and divide the motor trend data using the mean change point and according to the time segment of each data segment to obtain each motor trend data segment; determine the state of each motor trend data segment; wherein, the state of each motor trend data segment includes running state and stopping state.

7. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the trend warning level determination method as described in any one of claims 1 to 5.

8. A non-volatile computer-readable storage medium, characterized in that: Used to store computer programs; When the computer program is executed by the processor, it implements the trend warning level determination method as described in any one of claims 1 to 5.

Images