Device abnormal state monitoring method and computer device

Abstract

This application relates to the field of industrial Internet of Things (IIoT) technology, and discloses a method and computer device for monitoring abnormal equipment conditions. The method includes: acquiring real-time vibration acceleration data and real-time vibration velocity data of a target device; determining real-time acceleration trend data and real-time velocity trend data based on a pre-confirmed grouping capacity, the real-time vibration acceleration data, and the real-time vibration velocity data; determining whether the target device is in an abnormal state based on the real-time acceleration trend data, a pre-confirmed alarm threshold, and a preset alarm strategy; and if in an abnormal state, determining an alarm level for the abnormal state based on the real-time velocity trend data and a preset alarm level threshold. This application can improve the accuracy of monitoring abnormal equipment conditions.

Description

Technical Field

[0001] This application relates to the field of industrial Internet of Things (IoT) technology, specifically to a method for monitoring abnormal equipment conditions and a computer device. Background Technology

[0002] In the field of Industrial Internet of Things (IIoT) and equipment maintenance, with the widespread adoption of sensor technology, vibration monitoring of critical equipment such as motors and speed reducers has become an important means of predictive maintenance. However, existing equipment condition monitoring technologies still have significant limitations in practical applications. Traditional monitoring methods mainly rely on manually set fixed thresholds or simple statistical process control (SPC) upper and lower limits. This approach is usually based on general standards or static experience of experts, making it difficult to adapt to the dynamic characteristics of industrial equipment under different operating conditions (such as load changes and start-up / shutdown phases), resulting in alarm thresholds that cannot adaptively adjust with the natural aging of the equipment or changes in the environment.

[0003] Furthermore, in terms of data processing and anomaly detection mechanisms, existing technologies typically calculate statistical indicators directly based on unfiltered, full-volume historical data. This fails to effectively remove dirty data generated by equipment non-steady-state operation or sensor erroneous signals, causing the constructed monitoring benchmark to deviate from the true stable characteristics of the equipment. Regarding specific alarm strategies, a single instantaneous value exceeding the limit is commonly used as the criterion, lacking comprehensive consideration of multi-dimensional characteristics such as data continuity, trends, and frequency distribution. This crude monitoring model struggles to distinguish between occasional random noise interference and genuine fault symptoms. Furthermore, it is insensitive to early, minor faults that have not yet reached the alarm threshold but are already showing a deteriorating trend. This leads to a high likelihood of false alarms or missed alarms in equipment status monitoring, making it impossible to accurately assess the health status and severity of faults. Summary of the Invention

[0004] The purpose of this application is to provide a method and computer device for monitoring abnormal equipment conditions, so as to solve the problem of low accuracy in monitoring abnormal equipment conditions in the prior art.

[0005] To achieve the above objectives, the first aspect of this application provides a method for monitoring abnormal equipment conditions, the method comprising:

[0006] Acquire real-time vibration acceleration data and real-time vibration velocity data of the target device;

[0007] Based on the pre-confirmed grouping capacity, real-time vibration acceleration data, and real-time vibration velocity data, determine the real-time acceleration trend data and real-time velocity trend data;

[0008] Determine whether the target device is in an abnormal state based on real-time acceleration trend data, pre-confirmed alarm thresholds, and preset alarm strategies;

[0009] If an abnormal state is detected, the alarm level for the abnormal state will be determined based on real-time speed trend data and preset alarm level thresholds.

[0010] In this embodiment, the steps of confirming the grouping capacity and alarm threshold include: acquiring historical vibration acceleration time-series data and equipment operating status time-series data of the target device; determining robust vibration acceleration data based on the historical vibration acceleration time-series data, a preset acceleration benchmark value, a preset stable variation range, and the equipment operating status time-series data; determining the grouping capacity based on the robust vibration acceleration data and a preset grouping capacity upper limit; determining the overall mean and overall standard deviation of the robust vibration acceleration data based on the grouping capacity; and determining the alarm threshold based on the grouping capacity, the overall mean, the overall standard deviation, and the preset capacity threshold.

[0011] In this embodiment, the step of determining robust vibration acceleration data based on historical vibration acceleration time-series data, preset acceleration reference values, preset stable variation ranges, and equipment operating status time-series data includes: marking historical vibration acceleration time-series data with identifiers according to preset acceleration reference values ​​and preset stable variation ranges, and outputting first vibration acceleration time-series data with identifiers; determining at least one set of constant segment data from the first vibration acceleration time-series data through a sliding window; filtering each set of constant segment data according to equipment operating status time-series data, and outputting robust vibration acceleration data.

[0012] In this embodiment, the identifier includes constant points and fluctuation points; the timestamps of the first vibration acceleration time series data include multiple moments, each corresponding one-to-one with an identifier; the sliding window covers a maximum of two moments simultaneously during sliding; the step of determining at least one set of constant segment data from the first vibration acceleration time series data through the sliding window includes: controlling the sliding window to slide from the start point to the end point of the timestamps of the first vibration acceleration time series data to determine each set of constant segment data; the head of the sliding window points to the start point of the timestamps of the first vibration acceleration time series data, and the tail points to the end point of the timestamps of the first vibration acceleration time series data; in the sliding window During the sliding process of the opening, if the marker corresponding to the head of the sliding window is a fluctuation point and the marker corresponding to the tail is a constant point, then the head time is taken as the starting time of the current group of constant segment data; until the marker corresponding to the head of the sliding window is a constant point and the marker corresponding to the tail is a fluctuation point, then the head time is taken as the ending time of the current group of constant segment data; the time interval of the corresponding group of constant segment data is determined according to the starting time and ending time of each group of constant segment data; and each group of constant segment data is determined from the first vibration acceleration time series data according to the time interval of each group of constant segment data.

[0013] In this embodiment of the application, the step of marking historical vibration acceleration time-series data with identifiers according to a preset acceleration reference value and a preset stable variation range, and outputting the first vibration acceleration time-series data with the identifiers, includes: splitting the historical vibration acceleration time-series data into multiple vibration acceleration values ​​according to multiple time points; using the preset acceleration reference value as the current reference value; sequentially determining the absolute value of the difference between each vibration acceleration value and the current reference value; if the absolute value of the difference between the current vibration acceleration value and the current reference value is within the preset stable variation range, then marking the current vibration acceleration value as a constant point, and the current reference value remains unchanged; if the absolute value of the difference between the current vibration acceleration value and the current reference value is outside the preset stable variation range, then marking the current vibration acceleration value as a fluctuation point, and using the current vibration acceleration value as the new current reference value; and determining the first vibration acceleration time-series data based on each marked vibration acceleration value.

[0014] In this embodiment of the application, the steps of determining the overall mean and overall standard deviation of robust vibration acceleration data based on the grouping capacity include: dividing the robust vibration acceleration data into multiple continuous sub-vibration acceleration data according to the grouping capacity; generating global acceleration trend data based on the mean corresponding to each sub-vibration acceleration data; and determining the overall mean and overall standard deviation based on the global acceleration trend data.

[0015] In this embodiment of the application, determining the grouping capacity based on robust vibration acceleration data and a preset upper limit for grouping capacity includes: determining the grouping capacity according to the following formula:

[0016]

[0017] in, For group capacity; For robust vibration acceleration data, the data frequency is used. Set the maximum group size; This is the scaling adjustment factor.

[0018] In this embodiment, determining the alarm threshold based on the group capacity, overall mean, overall standard deviation, and preset capacity threshold includes: when the group capacity is greater than the preset capacity threshold, determining a control coefficient based on the group capacity, satisfying the following formula:

[0019]

[0020] The alarm threshold is determined using the following formula:

[0021]

[0022] in, For control coefficients; For group capacity; It is a gamma function; This is the alarm threshold; This is the population mean; The standard deviation is the population standard deviation.

[0023] In this embodiment, the real-time acceleration trend data includes multiple trend data points, each sorted according to its corresponding time; the preset alarm strategy includes a preset alarm threshold coefficient, a preset alarm time threshold, and a preset alarm point quantity threshold; the step of determining whether the target device is in an abnormal state based on the real-time acceleration trend data, the alarm threshold, and the preset alarm strategy includes: determining that the target device is in an abnormal state when the value of the latest trend data point in the real-time acceleration trend data is greater than the alarm threshold; determining that the target device is in an abnormal state when the values ​​of multiple consecutive trend data points in the real-time acceleration trend data are all greater than the alarm threshold; determining that the target device is in an abnormal state when the values ​​of multiple consecutive trend data points in the real-time acceleration trend data show an upward trend and each value is greater than the product of the alarm threshold and the preset alarm threshold coefficient; determining the number of trend data points in the real-time acceleration trend data whose values ​​are greater than the alarm threshold within the preset alarm time threshold; and determining that the target device is in an abnormal state when the number of exceeding the threshold is greater than or equal to the preset alarm point quantity threshold.

[0024] A second aspect of this application provides a computer device, comprising:

[0025] The memory is configured to store instructions; and

[0026] The processor is configured to retrieve instructions from memory and to implement the methods described above when executing instructions.

[0027] The above technical solution firstly aggregates and processes real-time collected vibration acceleration and vibration velocity data using pre-confirmed grouping capacity to generate trend data. This effectively filters out random noise in the original data and unifies the statistical granularity, providing a high-quality and time-aligned data foundation for subsequent analysis. Secondly, it fully utilizes the high sensitivity of vibration acceleration data to early equipment faults and impact signals, combined with a preset alarm strategy containing multiple verification logics to determine abnormal states. This allows for the keen detection of weak fault signs and effectively avoids false alarms caused by fluctuations in a single indicator, achieving early and accurate detection of abnormal states. Finally, after confirming that the equipment is in an abnormal state, real-time velocity trend data, which better characterizes vibration energy and equipment operational stability, is introduced as a grading basis. Quantitative evaluation is performed against preset alarm level thresholds. This step-by-step collaborative mechanism, which uses high-sensitivity acceleration to trigger alarms and high-stability velocity to determine the level, ensures that the monitoring system does not miss early faults and ensures the objectivity and compliance of fault severity assessment, thereby significantly improving the accuracy of monitoring abnormal equipment states.

[0028] Other features and advantages of the embodiments of this application will be described in detail in the following detailed description section. Attached Figure Description

[0029] The accompanying drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the following detailed description to explain the embodiments of this application, but do not constitute a limitation on the embodiments of this application. In the drawings:

[0030] Figure 1 A flowchart illustrating a method for monitoring abnormal equipment status according to an embodiment of this application is shown schematically.

[0031] Figure 2 A flowchart illustrating another device status monitoring method according to an embodiment of this application is shown schematically;

[0032] Figure 3 A flowchart illustrating a robust baseline period identification method according to an embodiment of this application is shown schematically.

[0033] Figure 4 An illustrative diagram illustrating an alarm strategy according to an embodiment of this application is shown schematically;

[0034] Figure 5 This schematic diagram illustrates a structural diagram of a computer device according to an embodiment of the present application;

[0035] Figure 6 The diagram illustrates the structure of a device abnormality monitoring system according to an embodiment of this application. Detailed Implementation

[0036] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for illustration and explanation of the embodiments of this application and are not intended to limit the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0037] It should be noted that if the embodiments of this application involve directional indicators (such as up, down, left, right, front, back, etc.), the directional indicators are only used to explain the relative positional relationship and movement of each component in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indicators will also change accordingly.

[0038] Furthermore, if the embodiments of this application involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, features defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions of various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed in this application.

[0039] The acquisition, transmission, storage, use, and processing of data in this application comply with relevant laws and regulations. Furthermore, it should be noted that certain software, components, models, and other existing industry solutions may be mentioned in the embodiments of this application. These should be considered exemplary, intended only to illustrate the feasibility of implementing the technical solution of this application, and do not imply that the applicant has already used or necessarily used such solutions.

[0040] It should be noted that the customer information (including but not limited to the initial conversation text, current conversation text, historical work order conversation text, various work order information, etc.) and various data (including but not limited to data used for analysis, stored data, and displayed data) involved in this application are all information and data authorized by the customer or fully authorized by all parties, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0041] Figure 1 A flowchart illustrating a method for monitoring abnormal device conditions according to an embodiment of this application is shown schematically. Figure 1As shown in the figure, this application provides a method for monitoring abnormal equipment status, which may include the following steps.

[0042] Step 101: Obtain real-time vibration acceleration data and real-time vibration velocity data of the target device.

[0043] Step 102: Determine the real-time acceleration trend data and real-time velocity trend data based on the pre-confirmed grouping capacity, real-time vibration acceleration data, and real-time vibration velocity data.

[0044] In this embodiment, the target equipment refers to various industrial machinery or its key transmission components, such as motors, reducers, gearboxes, or bearing systems, whose operating status is to be monitored. Real-time vibration acceleration data and real-time vibration velocity data refer to time-series numerical streams reflecting the physical vibration characteristics of the equipment, continuously collected by sensors installed on the target equipment during its current operation. Vibration acceleration data focuses on characterizing changes in internal impact forces and high-frequency vibration characteristics, and is more sensitive to early wear or minor faults. Vibration velocity data focuses on characterizing the energy magnitude and fatigue degree of the equipment's vibration, and can be used to measure the overall operational stability of the equipment.

[0045] The pre-confirmed grouping capacity refers to the granularity parameter used for discretizing and statistically processing continuously acquired raw time-series data. It is expressed as the number of data points or the length of the time window contained in each statistical unit. Determining real-time acceleration trend data and real-time velocity trend data based on the pre-confirmed grouping capacity, real-time vibration acceleration data, and real-time vibration velocity data involves dividing the high-frequency raw data stream into multiple consecutive subsets according to the grouping capacity, and then performing aggregation calculations on the data within each subset, such as calculating the arithmetic mean or eigenvalues, thereby generating a smooth data sequence that reflects the evolution of the equipment's operating status over time. The real-time acceleration trend data and real-time velocity trend data are the time-aligned statistical feature sequences obtained after the aforementioned noise reduction and feature extraction.

[0046] This step allows for the simultaneous acquisition of equipment operational data from both impact and energy characteristics, and the transformation of the raw high-frequency signal into more representative trend data using a uniform statistical granularity. This not only effectively filters out random noise and transient interference from the raw signal, ensuring strict temporal alignment of different physical quantity data, but also preserves key characteristic information of the equipment at different fault stages, laying a high-quality data foundation for subsequent accurate identification of abnormal symptoms.

[0047] Step 103: Determine whether the target device is in an abnormal state based on real-time acceleration trend data, pre-confirmed alarm thresholds, and preset alarm strategies.

[0048] In this embodiment, the pre-confirmed alarm threshold refers to a critical value calculated based on the statistical characteristics of historical data when the equipment is in a stable operating state, used to determine whether the equipment's vibration acceleration deviates from the normal range. It reflects the upper limit of the allowable vibration amplitude of the equipment in a healthy state. The preset alarm strategy refers to a set of composite logical judgment rules and supporting parameters pre-set to comprehensively capture different types of abnormal modes of the equipment. Specifically, it includes a preset alarm threshold coefficient, a preset alarm time threshold, and a preset alarm point number threshold, as well as various logical combinations based on these parameters, such as single-point value exceeding limit test, continuous multi-point exceeding limit test, trend increase magnitude test, and abnormal frequency test within a specific time window. Determining whether the target equipment is in an abnormal state involves inputting the real-time acquired trend data into the aforementioned preset alarm strategy. By comparing and analyzing the values ​​of data points, their continuous patterns, trends, and distribution density within a certain time period with the pre-confirmed alarm threshold and related strategy parameters in multiple dimensions, a judgment result is outputting whether the equipment currently exhibits fault symptoms.

[0049] This step moves beyond relying solely on a single hard indicator for judgment. Instead, it combines the high sensitivity of vibration acceleration data to early, subtle faults with the logical diversity of composite alarm strategies. Utilizing alarm thresholds derived from the equipment's inherent stability characteristics as a reference benchmark, and employing inspection rules targeting different anomaly modes such as sudden impacts, continuous degradation, slow drift, and intermittent faults, it can comprehensively capture abnormal signals. This multi-dimensional, integrated judgment mechanism effectively avoids false alarms caused by occasional random fluctuations, while enhancing the ability to identify highly concealed, gradual early faults.

[0050] Step 104: If the situation is abnormal, determine the alarm level for the abnormal situation based on the real-time speed trend data and the preset alarm level threshold.

[0051] In this embodiment, the preset alarm level threshold can be a graded critical value standard pre-set according to national standards, industry specifications, or expert experience databases to measure the magnitude of equipment vibration energy and the degree of damage. It typically includes multiple incremental numerical limits to distinguish different deterioration stages of equipment from normal operation to severe failure. Determining the alarm level for abnormal conditions based on real-time speed trend data and preset alarm level thresholds means that, after the equipment has been determined to have abnormal symptoms, further comparison is made between real-time speed trend data, which better reflects vibration energy and overall stability, and the aforementioned graded critical value standard. This quantifies the abnormal condition into different severity categories such as warning, general alarm, severe alarm, or danger alarm. The alarm level for abnormal conditions directly reflects the urgency and severity of equipment failure.

[0052] The above technical solution firstly aggregates and processes real-time collected vibration acceleration and vibration velocity data using pre-confirmed grouping capacity to generate trend data. This effectively filters out random noise in the original data and unifies the statistical granularity, providing a high-quality and time-aligned data foundation for subsequent analysis. Secondly, it fully utilizes the high sensitivity of vibration acceleration data to early equipment faults and impact signals, combined with a preset alarm strategy containing multiple verification logics to determine abnormal states. This allows for the keen detection of weak fault signs and effectively avoids false alarms caused by fluctuations in a single indicator, achieving early and accurate detection of abnormal states. Finally, after confirming that the equipment is in an abnormal state, real-time velocity trend data, which better characterizes vibration energy and equipment operational stability, is introduced as a grading basis. Quantitative evaluation is performed against preset alarm level thresholds. This step-by-step collaborative mechanism, which uses high-sensitivity acceleration to trigger alarms and high-stability velocity to determine the level, ensures that the monitoring system does not miss early faults and ensures the objectivity and compliance of fault severity assessment, thereby significantly improving the accuracy of monitoring abnormal equipment states.

[0053] In this embodiment, the steps of confirming the group capacity and alarm threshold may include: acquiring historical vibration acceleration time-series data and equipment operating status time-series data of the target device; determining robust vibration acceleration data based on the historical vibration acceleration time-series data, a preset acceleration benchmark value, a preset stable variation range, and the equipment operating status time-series data; determining the group capacity based on the robust vibration acceleration data and a preset group capacity upper limit; determining the overall mean and overall standard deviation of the robust vibration acceleration data based on the group capacity; and determining the alarm threshold based on the group capacity, the overall mean, the overall standard deviation, and the preset capacity threshold.

[0054] In this embodiment, historical vibration acceleration time-series data refers to a collection of vibration acceleration numerical records covering a relatively long operating period of the target equipment, preserving the complete trajectory of physical quantities changing over time. Equipment operating status time-series data can refer to log information recording the specific working modes of the equipment at different points in time, such as shutdown, operation, maintenance, or standby. The preset acceleration benchmark value can be an initial reference value set based on expert experience or the initial state of the equipment, used to initiate stability scan analysis. The preset stable variation range can refer to the maximum numerical range within which fluctuations between adjacent data points are allowed; changes falling within this range are considered normal stable fluctuations. Robust vibration acceleration data refers to a high-quality subset of data selected from the historical full data, representing data after the latest acceptance state and during a stable operating phase, eliminating interference data from non-steady-state periods and shutdown periods. The preset grouping capacity upper limit refers to the maximum constraint value set to ensure the sensitivity of statistical analysis for the number of subgroup data points. Grouping capacity refers to the optimal sample size calculated based on the robust data acquisition frequency and statistical requirements for dividing continuous data into discrete statistical units. The population mean and population standard deviation refer to the central tendency and dispersion measures of the trend data sequence obtained after aggregating robust vibration acceleration data based on group capacity. The preset capacity threshold is a critical value for group size used to distinguish different statistical coefficient calculation methods or lookup logic, such as the boundary used to determine which formula to use to calculate the control coefficient. This approach abandons the traditional method of setting fixed thresholds based on general empirical values, and instead establishes a dynamic benchmark based on the actual stable operating characteristics of the equipment. By rigorously screening robust data and using adaptive group capacity for statistical modeling, it ensures that the calculated alarm threshold accurately matches the unique normal fluctuation range of the equipment, effectively eliminating misjudgments caused by differences in operating conditions or impure benchmark data.

[0055] In this embodiment, the step of determining robust vibration acceleration data based on historical vibration acceleration time-series data, preset acceleration reference values, preset stable variation ranges, and equipment operating status time-series data may include: marking historical vibration acceleration time-series data with identifiers according to preset acceleration reference values ​​and preset stable variation ranges, and outputting first vibration acceleration time-series data with identifiers; determining at least one set of constant segment data from the first vibration acceleration time-series data through a sliding window; filtering each set of constant segment data according to equipment operating status time-series data, and outputting robust vibration acceleration data.

[0056] In this embodiment, the identifier refers to the state label attached to each discrete data point in historical data based on whether the numerical increment exceeds the allowable range. Specifically, it is distinguished into constant points representing a stable state and fluctuating points representing abrupt changes. The first vibration acceleration time series data with the identifier refers to the intermediate process data set after the aforementioned stability state label is added to the original physical quantity values. The sliding window refers to a data observation unit of fixed length that moves sequentially along the time axis, used to identify the continuity characteristics of time series data by analyzing the state combination changes of adjacent data points within the window. The constant segment data refers to the data subsequence identified by window scanning, consisting of continuous constant points, which can represent the stable operation characteristics of the equipment over a certain period of time. By generating first vibration acceleration time-series data with state attributes through point-by-point comparison and labeling, and using a continuous scanning mechanism with a sliding window to accurately lock the stable interval in the data stream, and then combining the actual operating status records of the equipment to perform secondary cleaning and filtering on these potential stable intervals, it is possible to extract robust vibration acceleration data that truly matches the current healthy and operating state of the equipment from massive historical messy data. This progressive screening mechanism effectively eliminates all non-stationary and non-operating interference data, ensuring the purity and representativeness of the subsequent threshold calculation benchmark, thereby significantly improving the accuracy of monitoring abnormal equipment conditions.

[0057] In this embodiment, the identifier includes constant points and fluctuation points; the timestamps of the first vibration acceleration time series data include multiple moments, each corresponding one-to-one with an identifier; the sliding window covers a maximum of two moments simultaneously during sliding; the step of determining at least one set of constant segment data from the first vibration acceleration time series data through the sliding window may include: controlling the sliding window to slide from the start point to the end point of the timestamps of the first vibration acceleration time series data to determine each set of constant segment data; the head of the sliding window points to the start point of the timestamps of the first vibration acceleration time series data, and the tail points to the end point of the timestamps of the first vibration acceleration time series data; during sliding... During the sliding of the window, if the marker corresponding to the head of the sliding window is a fluctuation point and the marker corresponding to the tail of the sliding window is a constant point, then the head time is taken as the starting time of the current group of constant segment data; until the marker corresponding to the head of the sliding window is a constant point and the marker corresponding to the tail of the sliding window is a fluctuation point, then the head time is taken as the ending time of the current group of constant segment data; the time interval of the corresponding group of constant segment data is determined according to the starting time and ending time of each group of constant segment data; and each group of constant segment data is determined from the first vibration acceleration time series data according to the time interval of each group of constant segment data.

[0058] In this embodiment, a constant point refers to a state value in the identifier that represents the current moment's vibration acceleration value change being within a preset allowable range, reflecting the stable operation characteristics of the equipment. A fluctuating point refers to a state value in the identifier that represents the current moment's vibration acceleration value change exceeding a preset allowable range, reflecting the non-stationary or abrupt operation characteristics of the equipment. A timestamp, including multiple moments, refers to a discrete record sequence of time-series data in the time dimension, providing a temporal positioning basis for retrieving state identifiers. A sliding window is a logical observation unit that can move along the time axis, constructing a comparative environment of preceding and following states by simultaneously covering two adjacent moments. The head and tail refer to the two ends of the sliding window in the temporal direction, corresponding to the earlier and later moments in the time series, respectively, used to distinguish the causes and consequences of state changes. The start and end times refer to the time boundaries that define the beginning and end of a continuous stable data segment, locked by a specific combination of jumps between adjacent identifiers within the window. A time interval refers to the time span containing continuous constant points, defined by the start and end times.

[0059] By controlling a sliding window to perform dual-point comparison scanning along the time axis, the system can accurately capture the critical states where the data stream transitions from fluctuation to constancy and from constancy to fluctuation. This allows for the precise identification and extraction of stable operating segments scattered throughout historical data into independent constant data segments. This extraction mechanism effectively eliminates non-steady-state data containing noise and transient impacts, ensuring that the benchmark data used to calculate alarm thresholds originates from the actual stable operating phase of the equipment and possesses extremely high purity.

[0060] In this embodiment of the application, the step of marking historical vibration acceleration time-series data with identifiers according to a preset acceleration reference value and a preset stable variation range, and outputting the first vibration acceleration time-series data with the identifiers, may include: splitting the historical vibration acceleration time-series data into multiple vibration acceleration values ​​according to multiple time points; using the preset acceleration reference value as the current reference value; sequentially determining the absolute value of the difference between each vibration acceleration value and the current reference value; if the absolute value of the difference between the current vibration acceleration value and the current reference value is within the preset stable variation range, then marking the current vibration acceleration value as a constant point, and the current reference value remains unchanged; if the absolute value of the difference between the current vibration acceleration value and the current reference value is outside the preset stable variation range, then marking the current vibration acceleration value as a fluctuation point, and using the current vibration acceleration value as the new current reference value; and determining the first vibration acceleration time-series data based on each marked vibration acceleration value.

[0061] In this embodiment, multiple vibration acceleration values ​​refer to a set of specific physical sampled values ​​obtained by discretizing continuous historical time-series data according to the time dimension. The current benchmark value refers to a dynamic reference standard used to evaluate data stability during iterative traversal; its value is adaptively adjusted according to data fluctuations to adapt to equipment state drift. The current vibration acceleration value refers to a specific data point currently undergoing judgment processing during time-series traversal. The absolute value of the difference is a numerical measure quantifying the deviation of the current data point from the dynamic reference standard. The new current benchmark value refers to the latest reference standard updated and used for subsequent comparisons when data fluctuations exceed the allowable range. Through this dynamic benchmark update mechanism, the judgment standard can be adaptively adjusted during point-by-point scanning, maintaining benchmark stability to filter noise when data is in a permissible small oscillation, and updating the benchmark in a timely manner to follow equipment state changes when substantial data mutations occur. This accurately marks each data point as constant or fluctuating, ensuring that the first vibration acceleration time-series data can truly reflect the stability details of the equipment's historical operation, providing a reliable basis for subsequently extracting clean and robust benchmark data, thereby significantly improving the accuracy of monitoring abnormal equipment states.

[0062] In this embodiment of the application, the step of determining the overall mean and overall standard deviation of robust vibration acceleration data based on the grouping capacity may include: dividing the robust vibration acceleration data into multiple continuous sub-vibration acceleration data according to the grouping capacity; generating global acceleration trend data based on the mean corresponding to each sub-vibration acceleration data; and determining the overall mean and overall standard deviation based on the global acceleration trend data.

[0063] In this embodiment, continuous multi-set sub-vibration acceleration data refers to multiple independent and continuous data segments formed by dividing a long-period robust data stream according to the grouping capacity. Each segment constitutes the smallest statistical unit for subsequent feature extraction. Global acceleration trend data refers to a smooth numerical sequence generated after averaging each sub-data segment, removing high-frequency random noise and retaining the main fluctuation pattern of the data. The overall mean and overall standard deviation are statistical indicators calculated based on the denoised trend sequence, representing the central horizontal position and discrete fluctuation range of the equipment's operating trend within the reference period, respectively.

[0064] In one embodiment, robust vibration acceleration data may include vibration acceleration time-series data and vibration velocity time-series data. Vibration acceleration time-series data and vibration velocity time-series data refer to two types of physical quantity numerical sequences that are included in the filtered robust data set, respectively characterizing the equipment's impact characteristics and energy characteristics, and are strictly aligned on the time axis. The former is used to construct a highly sensitive alarm benchmark, and the latter is used for subsequent level determination.

[0065] This secondary processing method, based on grouping and mean aggregation, transforms the direct statistics of the raw massive data into statistical analysis of trend data that better represents the operational patterns of the equipment. This approach effectively smooths out non-essential random interference in the original signal, ensuring that the final calculated population mean and standard deviation accurately reflect the core fluctuation patterns of the equipment under optimal stable conditions. Using these statistical parameters derived from pure trend data to construct alarm thresholds ensures that the monitoring system avoids false alarms due to momentary spikes and can also accurately detect any early, subtle anomalies deviating from the normal trend.

[0066] In this embodiment of the application, determining the grouping capacity based on robust vibration acceleration data and a preset upper limit for grouping capacity may include: determining the grouping capacity according to the following formula:

[0067]

[0068] in, For group capacity; For robust vibration acceleration data, the data frequency is used. Set the maximum group size; This is the scaling adjustment factor.

[0069] In this embodiment, determining the alarm threshold based on the group capacity, the overall mean, the overall standard deviation, and the preset capacity threshold may include: when the group capacity is greater than the preset capacity threshold, determining a control coefficient based on the group capacity, satisfying the following formula:

[0070]

[0071] The alarm threshold is determined using the following formula:

[0072]

[0073] in, For control coefficients; For group capacity; It is a gamma function; This is the alarm threshold; This is the population mean; The standard deviation is the population standard deviation.

[0074] In this embodiment, the real-time acceleration trend data includes multiple trend data points, each sorted according to its corresponding time; the preset alarm strategy includes a preset alarm threshold coefficient, a preset alarm time threshold, and a preset alarm point quantity threshold; the step of determining whether the target device is in an abnormal state based on the real-time acceleration trend data, the alarm threshold, and the preset alarm strategy may include: determining that the target device is in an abnormal state when the value of the latest trend data point in the real-time acceleration trend data is greater than the alarm threshold; determining that the target device is in an abnormal state when the values ​​of multiple consecutive trend data points in the real-time acceleration trend data are all greater than the alarm threshold; determining that the target device is in an abnormal state when the values ​​of multiple consecutive trend data points in the real-time acceleration trend data show an upward trend and each value is greater than the product of the alarm threshold and the preset alarm threshold coefficient; determining the number of trend data points in the real-time acceleration trend data whose values ​​are greater than the alarm threshold within the preset alarm time threshold; and determining that the target device is in an abnormal state when the number of exceeding the threshold is greater than or equal to the preset alarm point quantity threshold.

[0075] In this embodiment, a trend data point refers to each discrete statistical numerical unit constituting the real-time acceleration trend data sequence, representing the vibration acceleration characteristic value of the device after aggregation calculation within a specific subgroup time window. The preset alarm threshold coefficient is a proportional adjustment parameter used to construct specific trend judgment conditions; it, combined with the alarm threshold, forms a dynamic auxiliary benchmark for identifying anomalies with specific upward patterns. The preset alarm time threshold refers to the time window span set for forward backtracking or sliding observation during frequency distribution testing. The preset alarm point quantity threshold refers to the minimum abnormal point count standard required to determine equipment instability within the aforementioned time span. The number of exceedances refers to the total number of trend data points whose values ​​are actually higher than the alarm threshold within the currently set observation time window. By performing numerical over-limit checks for single-point instantaneous impacts, continuous multi-point checks for persistent faults, upward trend pattern checks for early gradual faults, and time window frequency checks for intermittent faults in parallel, this application constructs a multi-dimensional composite anomaly identification mechanism. This mechanism can comprehensively cover various complex anomaly modes of equipment, from sudden failures to early slow degradation. While effectively shielding occasional random noise interference, it greatly enhances the ability to capture weak fault signs.

[0076] The above technical solution achieves the following: First, based on the historical vibration acceleration time-series data of the target equipment, the equipment operating status time-series data, and the dynamically updated benchmark value and sliding window mechanism, robust vibration acceleration data representing the stable operating characteristics of the equipment are accurately selected. Based on this, the grouping capacity and alarm threshold adapted to the current operating condition of the equipment are calculated, thus establishing an adaptive monitoring benchmark. Second, the grouping capacity is used to process real-time vibration acceleration data and real-time vibration velocity data, determining real-time acceleration trend data and real-time velocity trend data respectively. This ensures that the processing granularity of real-time data is consistent with the calculation benchmark of the alarm threshold, thereby guaranteeing the effectiveness of subsequent anomaly comparisons. Furthermore, by fully utilizing the high sensitivity of real-time acceleration trend data to minor faults, combined with preset alarm strategies and alarm thresholds containing multiple logical dimensions, a comprehensive judgment can be made, enabling sensitive and accurate determination of whether the target equipment is in an abnormal state. Finally, after determining that it is in an abnormal state, the alarm level is determined based on the real-time velocity trend data and preset alarm level thresholds. The stability of the vibration energy characterization using velocity data achieves objective quantification of the fault severity, thereby significantly improving the accuracy of monitoring abnormal equipment conditions.

[0077] Figure 2 A flowchart illustrating another device status monitoring method according to an embodiment of this application is shown schematically. Figure 2 As shown, the overall process is divided into an alarm threshold calculation stage and a composite strategy alarm stage. In the alarm threshold calculation stage, the historical vibration acceleration time series data of the target equipment is first acquired and preprocessed. Simultaneously, a robust baseline extraction step is performed, which determines the robust baseline dataset (robust vibration acceleration data) based on historical data, preset acceleration baseline values, and a sliding window mechanism. Then, combined with standard parameter matching, dynamic threshold calculation is performed using the robust data and the determined grouping capacity to generate a pre-confirmed alarm threshold. In the composite strategy alarm stage, trend data calculation is performed on the real-time acquired equipment monitoring data based on the pre-confirmed grouping capacity to determine the real-time acceleration trend data. This trend data is then combined with the alarm threshold, and a composite verification triggering step is performed. This involves using a preset alarm strategy to determine whether the target equipment is in an abnormal state. If it is in an abnormal state, a graded alarm is finally triggered, which determines the alarm level for the abnormal state based on the real-time velocity trend data and the preset alarm level threshold.

[0078] The following is an example of this application:

[0079] This study focuses on a single motor in a typical transmission system (motor + reducer + gearbox) used in industrial production.

[0080] Step 1: Extract historical data from the time series database and perform preprocessing.

[0081] Vibration acceleration and vibration velocity data of the motor drive end over the past 6 months were collected. The data frequency was reduced to the second level using the maximum value aggregation method. Sensor erroneous signals were screened out and removed according to the sensor value range. The time series data of vibration acceleration and vibration velocity were aligned, and the time points of missing data were screened out and removed from the time series data.

[0082] Step 2: Establish a robust baseline dataset.

[0083] Collect motor acceptance information to determine the latest acceptance time. Starting from the acceptance time, use the constant segment identification method to identify multiple motor stabilization period time ranges.

[0084] First, the baseline value is initialized as vibration acceleration based on expert experience. Data is then scanned sequentially according to time sequence. If the numerical increment is within the range of stable change specified by expert knowledge, it is marked as a constant point, and the baseline value remains unchanged. Otherwise, it is marked as a fluctuating point, and the baseline value is updated to the current value. After the scan is completed, the labeled data is obtained.

[0085] Furthermore, the labeled data is scanned using a sliding window of size 2. If the first position in the window represents a fluctuating point and the second position represents a constant point, then the start time of the current window is marked as the beginning of a constant segment. If the first position in the window represents a constant point and the second position represents a fluctuating point or the last item, then the start time of the current window is marked as the end of a constant segment. If the first position in the window represents a fluctuating point and the second position represents a fluctuating point, then the start time of the current window is marked as a non-constant segment. The labeled data is obtained after the scan is complete.

[0086] Furthermore, the data is scanned sequentially according to the time series, recording the start and end points of each constant segment. After the scan is completed, multiple robust reference periods are obtained. Motor operation information is collected, and robust reference periods in operation are selected. If multiple periods are selected, the latest one is chosen. After selection, the robust reference period dataset is read from the time series database.

[0087] Figure 3 A flowchart illustrating a robust baseline period identification method according to an embodiment of this application is shown schematically. Figure 3As shown in the embodiment of this application, the process first initializes the reference value, that is, uses a preset acceleration reference value as the initial current reference value; then, it scans the historical vibration acceleration time series data one by one, and marks the data according to whether the absolute value of the difference between the current vibration acceleration value and the current reference value is within a preset stable change range. If it is within the range, it is marked as a constant point and the reference value remains unchanged; if it is outside the range, it is marked as a fluctuation point and the current vibration acceleration value is used as the new current reference value, thereby outputting the first vibration acceleration time series data with the mark; then, it uses a sliding window mechanism to scan the first vibration acceleration time series data with the mark one by one to determine the start and end points of the constant segment. The specific determination logic includes: when the mark corresponding to the head time of the sliding window is a fluctuation point and the mark corresponding to the tail time is a constant point, it is marked as the start point of the constant segment; when the mark corresponding to the head time is a constant point and the mark corresponding to the tail time is a fluctuation point or the last item, it is marked as the end point of the constant segment; if both are fluctuation points, it is marked as a non-constant segment; finally, by scanning these marked data item by item, the start and end points of multiple robust reference periods are obtained, thereby determining the robust vibration acceleration data.

[0088] Step 3: Divide the robust baseline dataset into multiple subgroups.

[0089] Considering the time span and data volume of the robust baseline dataset, the subgroup size is automatically matched. The matching references Shewhart control chart theory and extends its application. For datasets with a data frequency of less than 1 minute, the subgroup size is expanded using the following method.

[0090]

[0091] in, The data volume within 1 minute (data frequency of robust vibration acceleration data). This is the reference subgroup size limit (preset group capacity limit), which defaults to 25. Set the scaling factor for the 1-minute subgroup; the default value is 8. For the group capacity, if calculated as follows If the number is greater than 200, then the group capacity is 200.

[0092] The calculation results of the robust baseline dataset obtained in step 2:

[0093]

[0094] Furthermore, scan the data intervals within each subgroup; if an interval within a subgroup exceeds the data frequency... The current subgroup is divided into two subgroups by taking the number of adjacent points that are twice the size of the current subgroup. After the scan is complete, a list of the start and end points of the subgroup window is obtained.

[0095] Step 4: Calculate the alarm threshold.

[0096] Using the robust baseline dataset obtained in step 2, and taking the start and end times of the subgroup windows obtained in step 3 as the starting and ending points, we first calculate the arithmetic mean of each subgroup window as the trend data, and then calculate the overall mean of all vibration acceleration trend data. and the overall standard deviation of acceleration trend data The overall mean of all vibration velocity trend data and the overall standard deviation of speed trend data Simultaneously calculate the average subgroup duration span minute.

[0097] Furthermore, according to expert knowledge, vibration velocity is subject to the applicable vibration intensity standard parameters. The alarm threshold is calculated using the national standard for vibration intensity. The national standard for vibration intensity is GB / T 6075.3-2011.

[0098] Furthermore, the vibration acceleration is not subject to standard parameters; instead, the average value from the measurement control chart is used. The upper control limit is calculated from the control chart and used as the alarm threshold. The control coefficient is obtained by looking up the table. (when =25), Calculation result .

[0099] when When the value is greater than 25, the control factor is calculated using the following formula:

[0100]

[0101] in, It is a gamma function; This represents the group capacity.

[0102] Step 5: Calculate trend data and perform a composite test to determine the alarm.

[0103] Extract the dataset of historical vibration acceleration trend data from the latest time point to the current time point, and sort it by subgroup size. =25, calculate the subgroup mean and input it into the database as new trend data.

[0104] Furthermore, the trend data is searched forward and symptoms are identified according to the composite inspection strategy (preset alarm strategy).

[0105] R1 single-point test symptom: Search forward 20 minutes (i.e., 1 trend data point), the single point of vibration acceleration trend data is greater than 4.92;

[0106] R2 consecutive test signs: Searching forward 60 minutes (i.e., 3 trend data points), the vibration acceleration trend data is greater than 4.92 for 3 consecutive points;

[0107] R3 trend test indicators: Searching backwards for 160 minutes (i.e., 8 trend data points), the oscillation acceleration trend data shows a stable increase for 7 consecutive times, with the deviation between the last point and the first point greater than 2.22 (i.e., greater than 4.92). 45%)

[0108] R4 frequency test indicators: Searching forward for trend data over a duration of 300 minutes, there are 5 points in the vibration acceleration trend data that are greater than 4.92.

[0109] The trend data over the past 300 minutes is [3.684, 3.785, 4.594, 4.83, 3.696, 5.713, 4.605, 5.342, 5.500, 4.402, 5.449, 5.298, 5.392, 4.204, 5.344]. R1 single-point test indicators, R2 continuous test indicators, and R4 frequency test indicators are identified and are judged as alarms according to the selected medium sensitivity strategy.

[0110] Figure 4 An illustrative diagram illustrating an alarm strategy according to an embodiment of this application is shown. Figure 4 As shown in this embodiment, based on real-time acceleration trend data and a pre-confirmed alarm threshold, a combination of multiple logical dimensions is used to determine whether the target device is in an abnormal state. The real-time acceleration trend data includes multiple trend data points sorted by time. The preset alarm strategy specifically includes: determining whether the value of the latest trend data point in the real-time acceleration trend data is greater than the alarm threshold (corresponding to the single-point test in the figure); determining whether the values ​​of multiple consecutive trend data points are all greater than the alarm threshold (corresponding to the continuity test in the figure); determining whether the values ​​of multiple consecutive trend data points show an upward trend and each value is greater than the product of the alarm threshold and the preset alarm threshold coefficient (corresponding to the trend test in the figure); and determining whether the number of trend data points exceeding the alarm threshold within the preset alarm time threshold is greater than or equal to the preset alarm point number threshold (corresponding to the frequency test in the figure). If any of the above conditions are met, the target device is determined to be in an abnormal state.

[0111] Step 6: Calculate the level threshold and determine the alarm level.

[0112] Extract the dataset of historical vibration velocity trend data from the latest time point to the current time point, and sort it by subgroup size. =25, calculate the subgroup mean and input it into the database as new trend data.

[0113] The vibration intensity standard grade formula is used to calculate the grade threshold, where = 2.8, = 4.5.

[0114]

[0115]

[0116]

[0117] , and All are preset alarm level thresholds; This refers to the upper limit of the area boundary B in the national standard for vibration intensity. This refers to the upper limit of the area boundary C in the national standard for vibration intensity.

[0118] Furthermore, the alarm level is determined by comparing the symptoms identified in step 5.

[0119] R1 single-point test grading: Searching forward for 20 minutes (i.e., one trend data point), vibration velocity trend data within the range [2.8, 3.5) indicates alarm level one, and [3.5, 5.63) indicates alarm level two. The alarm level is level three; the rest are warning levels.

[0120] R2 Continuity Test Rating: Searching forward 60 minutes (i.e., 3 trend data points), if three consecutive vibration velocity trend data points exceed the limit, the last point within the range [2.8, 3.5) is considered alarm level 1, and [3.5, 5.63) is considered alarm level 2. The alarm level is level three; the rest are warning levels.

[0121] R4 frequency test grading: Searching forward for trend data over a 300-minute time span, among the 5 points where vibration acceleration trend data exceeds the limit, the last point within the range [2.8, 3.5) is considered alarm level 1, and [3.5, 5.63) is considered alarm level 2. The first level is alarm level 3; the rest are alert levels.

[0122] The trend data over the past 300 minutes is [1.218, 1.449, 1.593, 1.574, 1.336, 1.743, 1.678, 1.786, 1.766, 1.557, 1.735, 1.821, 1.859, 1.532, 1.786]. The single-point test rating is "prompt", the continuous test rating is "prompt", the frequency test rating is "prompt", and the overall rating is "prompt level".

[0123] Figure 5 A schematic diagram illustrating the structure of a computer device according to an embodiment of this application is provided. Figure 5 As shown in the illustration, this application provides a computer device that may include:

[0124] Memory 510 is configured to store instructions; and

[0125] Processor 520 is configured to retrieve instructions from memory 510 and to implement the methods described above when executing instructions.

[0126] Figure 6 A schematic diagram illustrating the structure of a device abnormality monitoring system according to an embodiment of this application is provided. Figure 6 As shown in the embodiments of this application, the system mainly consists of a data acquisition module, a data storage module, a data processing module, an application data management module, a composite inspection and alarm judgment module, and a human-computer interaction and alarm output module.

[0127] The data acquisition module may include a data protocol reading unit for interfacing with different data acquisition protocols, receiving data in real time from multiple data sources, assembling measurement point paths, processing numerical values, and aligning time formats to obtain the target data. It is responsible for acquiring real-time vibration acceleration and velocity data of the target device and storing them in the data storage module.

[0128] The storage module may include a raw data storage unit for storing time-series data received from the data source; and a trend data storage unit for storing trend data calculated and generated by the target data processing module.

[0129] The data processing module may include a data preprocessing unit for high-frequency data down-processing, sensor malfunction signal denoising, and time alignment with sensor data, and then sending it to the target point data storage module; a trend data generation unit for extracting raw data, calculating trend data, and sending it to the target point data storage module according to a trend data generation method; and a threshold calculation unit for mining robust baseline datasets, dividing them into subgroup datasets, calculating alarm thresholds and level thresholds, and sending them to the application data management module according to a dynamic threshold calculation method. The data processing module performs data down-processing, data alignment, and denoising, and is responsible for determining robust vibration acceleration data, pre-confirmed grouping capacity, and alarm thresholds based on historical data, while also generating real-time trend data.

[0130] The application data management module may include a parameter management unit for storing standard parameters, historical robust baseline periods, and the results of automatic subgroup parameter matching, as well as historical alarm thresholds and level thresholds by location; and an expert knowledge unit for storing Shewhart control chart parameter tables, national standard level values, level standard information derived from expert experience, and standard content and expert experience such as the stable data range for each measurement point type. It is used to store and manage standard parameters, subgroup parameters, robust baseline periods, and various thresholds.

[0131] The composite inspection and alarm judgment module may include a symptom inspection unit, used to trigger single inspection rules and discover early warning symptoms that meet the inspection rules from trend data; and a composite inspection and alarm classification unit, used to combine multiple inspection rules, generate alarms according to sensitivity settings, and read classification trend data to classify alarm levels. Based on real-time acceleration trend data, pre-confirmed alarm thresholds, and preset alarm strategies, it determines whether the target device is in an abnormal state, and determines the alarm level based on real-time speed trend data when an abnormality occurs.

[0132] The human-machine interaction and alarm output module may include: an equipment status monitoring unit, used to display the operating status and real-time alarm status of monitored equipment, sorted by alarm level; a threshold generation unit, used to provide operators with interpretable thresholds and rules throughout the entire threshold generation process, from data preprocessing, robust baseline dataset, subset partitioning, alarm threshold calculation, and level threshold calculation; and an alarm details unit, used to display trend data changes, symptom generation, and composite test results, while also displaying relevant information about the alarm equipment and location, helping operators understand the alarm triggering mechanism, alarm history, and diagnose equipment defects. It is used to display equipment status monitoring information, the threshold generation process, and specific alarm details.

[0133] This application also provides a machine-readable storage medium storing instructions that cause a machine to perform the above-described method.

[0134] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0135] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0136] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0137] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0138] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0139] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0140] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0141] It should also be noted that 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 process, method, article, or apparatus. Unless otherwise specified, 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 that element.

[0142] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for monitoring abnormal equipment conditions, characterized in that, The method includes: Acquire real-time vibration acceleration data and real-time vibration velocity data of the target device; Real-time acceleration trend data and real-time velocity trend data are determined based on the pre-confirmed grouping capacity, the real-time vibration acceleration data, and the real-time vibration velocity data. The system determines whether the target device is in an abnormal state based on the real-time acceleration trend data, the pre-confirmed alarm threshold, and the preset alarm strategy. If the abnormal state is in the aforementioned state, the alarm level for the abnormal state is determined based on the real-time speed trend data and the preset alarm level threshold. The steps for confirming the group capacity and the alarm threshold include: Acquire historical vibration acceleration time-series data and equipment operating status time-series data of the target device; The historical vibration acceleration time series data are marked with an identifier based on a preset acceleration reference value and a preset stable variation range, and the first vibration acceleration time series data with the identifier is output. At least one set of constant segment data is determined from the first vibration acceleration time series data by using a sliding window; Based on the time-series data of the device's operating status, each group of constant segment data is filtered to output robust vibration acceleration data; The grouping capacity is determined based on the robust vibration acceleration data and the preset grouping capacity limit. The population mean and population standard deviation of the robust vibration acceleration data are determined based on the grouping capacity. The alarm threshold is determined based on the group capacity, the overall mean, the overall standard deviation, and the preset capacity threshold.

2. The method of claim 1, wherein, The identifier includes constant points and fluctuation points; the timestamp of the first vibration acceleration time series data includes multiple moments, each of which corresponds one-to-one with the identifier; the sliding window can cover a maximum of two moments simultaneously when sliding.

3. The method of claim 2, wherein, The step of determining at least one set of constant segment data from the first vibration acceleration time series data through a sliding window includes: The sliding window is controlled to slide from the start point of the timestamp of the first vibration acceleration time series data to the end point to determine each group of constant segment data; the head of the sliding window points to the start point of the timestamp of the first vibration acceleration time series data, and the tail points to the end point of the timestamp of the first vibration acceleration time series data. During the sliding process of the sliding window, if the identifier corresponding to the time at the head of the sliding window is the fluctuation point and the identifier corresponding to the time at the tail is the constant point, then the time at the head is taken as the starting time of the current group of constant segment data. Until the identifier corresponding to the head time of the sliding window is the constant point and the identifier corresponding to the tail time is the fluctuation point, then the head time is taken as the end time of the current group of constant segment data. The time interval of the corresponding group of constant segment data is determined based on the start and end times of each group of constant segment data. The constant segment data is determined from the first vibration acceleration time series data based on the time interval of each constant segment data.

4. The method of claim 3, wherein, The step of marking the historical vibration acceleration time series data with an identifier based on a preset acceleration reference value and a preset stable variation range, and outputting the first vibration acceleration time series data with the identifier, includes: The historical vibration acceleration time series data is divided into multiple vibration acceleration values ​​according to the multiple time points; Use the preset acceleration reference value as the current reference value; Determine the absolute value of the difference between each vibration acceleration value and the current reference value in turn; If the absolute value of the difference between the current vibration acceleration value and the current reference value is within the preset stable change range, then the current vibration acceleration value is marked as the constant point, and the current reference value remains unchanged; If the absolute value of the difference between the current vibration acceleration value and the current reference value is outside the preset stable change range, then the current vibration acceleration value is marked as the fluctuation point, and the current vibration acceleration value is used as the new current reference value; The first vibration acceleration timing data is determined based on the vibration acceleration value after each mark.

5. The method according to any one of claims 2 to 4, characterized in that, The steps for determining the population mean and population standard deviation of the robust vibration acceleration data based on the grouping capacity include: The robust vibration acceleration data is divided into multiple continuous sub-groups of vibration acceleration data according to the grouping capacity. Global acceleration trend data is generated based on the mean value corresponding to each group of sub-vibration acceleration data. The population mean and population standard deviation are determined based on the global acceleration trend data.

6. The method according to any one of claims 2 to 4, characterized in that, Determining the grouping capacity based on the robust vibration acceleration data and the preset grouping capacity upper limit includes: The grouping capacity is determined according to the following formula: in, The group capacity; The data frequency of the robust vibration acceleration data; The preset group capacity limit; This is the scaling adjustment factor.

7. The method according to any one of claims 2 to 4, characterized in that, The step of determining the alarm threshold based on the group capacity, the overall mean, the overall standard deviation, and the preset capacity threshold includes: If the group capacity is greater than the preset capacity threshold, a control coefficient is determined based on the group capacity, and satisfies the following formula: The alarm threshold is determined according to the following formula: in, The control coefficient is mentioned above; The group capacity; It is a gamma function; The alarm threshold; This is the population mean; denoted as the overall standard deviation.

8. The method according to any one of claims 2 to 4, characterized in that, The real-time acceleration trend data includes multiple trend data points, each of which is sorted according to its corresponding time; the preset alarm strategy includes a preset alarm threshold coefficient, a preset alarm time threshold, and a preset alarm point quantity threshold. The step of determining whether the target device is in an abnormal state based on the real-time acceleration trend data, the alarm threshold, and the preset alarm strategy includes: If the value of the latest trend data point in the real-time acceleration trend data is greater than the alarm threshold, the target device is determined to be in an abnormal state. If the values ​​of multiple consecutive trend data points in the real-time acceleration trend data are all greater than the alarm threshold, it is determined that the target device is in an abnormal state. If the values ​​of multiple consecutive trend data points in the real-time acceleration trend data show an upward trend and each value is greater than the product of the alarm threshold and the preset alarm threshold coefficient, the target device is determined to be in an abnormal state. Determine the number of trend data points in the real-time acceleration trend data whose values ​​exceed the alarm threshold within the preset alarm time threshold. If the number of exceedances is greater than or equal to the preset alarm point number threshold, the target device is determined to be in an abnormal state.

9. A computer device, comprising: include: The memory is configured to store instructions; as well as A processor configured to retrieve the instructions from the memory and, when executing the instructions, to implement the method according to any one of claims 1 to 8.

10. A machine-readable storage medium, characterized in that, The machine-readable storage medium stores instructions for causing the machine to perform the method according to any one of claims 1 to 8.

Images