Intelligent equipment fault detection method

Abstract

The invention relates to an equipment fault intelligent detection method, which comprises the following steps of: when equipment has a fault, transmitting a fault code back to a central server in real time; matching corresponding machine fault code information based on the fault code, analyzing a fault reason, and generating a preliminary diagnosis report; fault causes and processing schemes are pushed to an equipment PLC panel according to priorities, so that maintenance personnel can carry out troubleshooting and maintenance according to a scheme sequence; the step of acquiring the fault code in the step S1 comprises the following steps: S1, acquiring video stream data of an equipment operation area in real time; s2, performing image analysis on the video stream data, and identifying key component information of equipment; s3, detecting the running state of the equipment based on an equipment posture and action recognition algorithm; s4, according to a preset equipment fault behavior rule, judging whether the running state of the current equipment is faulty or not; and S5, if it is judged that the current equipment operation state is a fault, a fault code generation mechanism is triggered.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of equipment fault detection, and particularly relates to an equipment fault intelligent detection method. BACKGROUND

[0002] In modern industrial production and many complex operation scenes, stable operation of equipment is crucial. Traditional equipment fault detection mainly relies on manual inspection, single parameter monitoring of sensors or simple threshold alarm. Manual inspection has strong subjectivity, low efficiency and is difficult to cover comprehensively in real time, and subtle fault hidden dangers are easily missed. Although sensor monitoring can obtain some key parameters, it is difficult to grasp the overall operation posture of the equipment, complex actions and the cooperative state of multiple components, and it is difficult to intuitively reflect the real scene when the fault occurs. With the rapid development of computer vision technology, image processing algorithm and intelligent data analysis, using video stream data for equipment fault intelligent detection has become a new research hotspot and trend.

[0003] A fault detection method and a fault detection device are disclosed in Chinese Patent CN115932709A. The method includes: establishing a communication connection with a smart meter according to a preset communication interface, obtaining meter reading data based on the communication connection, and detecting a fault event according to a meter reading data threshold and the meter reading data. Although the power consumption inspection and fault event detection of the smart meter are solved, the detection range is relatively narrow, and only specific fault types related to the meter reading can be judged. At the same time, the detection method based on the meter reading data threshold is relatively single. In the actual running environment, the normal operation state of the equipment may be affected by multiple dynamic factors. Relying only on the fixed meter reading data threshold to judge the fault may result in misjudgment or omission. For example, under some special time periods or special load conditions, normal data fluctuations may be misjudged as faults, or some potential faults cannot be detected in time when the data does not exceed the threshold in the early stage.

[0004] Therefore, an equipment fault intelligent detection method is proposed herein. SUMMARY

[0005] The main purpose of the present application is to provide an equipment fault intelligent detection method, which aims to solve the problem that the detection method based on data values alone in the prior art cannot provide intuitive on-site information such as the appearance state of the equipment and the surrounding environment when a fault occurs, which is not conducive to the maintenance personnel to quickly and accurately locate the fault cause and take targeted maintenance measures.

[0006] To achieve the above object, the application provides a kind of equipment failure intelligent detection method, comprising the following steps: when equipment fails, fault code is returned to central server in real time;Based on fault code, corresponding machine fault code information is matched, fault reason is analyzed, and preliminary diagnosis report is generated;Fault reason and treatment scheme are pushed to equipment PLC panel according to priority, and maintenance personnel are arranged according to scheme sequence Maintenance;The acquisition step of the fault code comprises the following steps: S1, real-time video stream data of equipment running area is acquired; S2, image analysis is carried out on the video stream data, and equipment key component information is identified; S3, the running state of the equipment is detected based on equipment posture and action recognition algorithm; S4, according to the preset equipment failure behavior rule, whether the current equipment running state is failure is judged; S5, if the current equipment running state is failure, the fault code generation mechanism is triggered.

[0007] Preferably, the equipment posture and action recognition algorithm is specifically: the key feature point coordinates of the equipment are extracted in real time by using OpenPose algorithm, and the key feature point coordinates include feature point motion trajectories of key connection parts, moving parts and the like of the equipment;The key feature point coordinate data extracted above is preprocessed, and a space-time motion trajectory model of each part of the equipment is constructed based on the preprocessed key feature point coordinates;The equipment posture and action features in the video stream data are extracted through the constructed space-time motion trajectory model, and whether the equipment has failure is judged.

[0008] Preferably, the judgment of equipment failure information in step S3 is specifically: When the instantaneous speed of the feature point coordinates of the key moving parts of the equipment exceeds the threshold value and the motion trajectory shows abnormal fluctuation, it is judged that the equipment has unstable running failure in the running process; When the acceleration mutation value of the key feature points of the equipment exceeds the preset threshold value, it is judged that there is an abnormal situation of too large equipment running impact; When the relative angle change frequency of the key connection parts of the equipment exceeds the preset threshold value, it is judged that there is a failure situation of too large equipment vibration or loose connection.

[0009] Preferably, the preset threshold value of the instantaneous speed threshold value, the acceleration mutation value and the preset threshold value of the relative angle change frequency are all set based on the basic specification parameters and historical running data of the detected equipment.

[0010] Preferably, the judgment of equipment failure information in step S3 further comprises: When the angle between the feature point of the key support part of the device and the ground projection is less than the first preset angle threshold, it is judged that there is an excessive tilt fault posture of the device; When the difference between the motion activity of the feature points of the symmetrical components on both sides of the device exceeds the second preset threshold, it is judged that there is an irregular state of unbalanced operation of the device; When the motion trajectory of the key component of the device to the target operating position deviates from the optimal path and the offset exceeds the third preset threshold, it is judged that there is an inefficient state of device operation detour or roundabout.

[0011] Preferably, the first preset angle threshold, the second preset threshold and the third preset threshold are determined by the following method: The first preset angle threshold is set according to the normal tilt range and safety standards of the device; The second preset threshold is determined based on the statistical results of the feature point motion activity difference of the symmetrical components on both sides of the device in the normal operating state; The third preset threshold is calculated according to the optimal path of the key component of the device to the target operating position and the acceptable deviation range.

[0012] Preferably, the maintenance scheme pushed to the PLC panel includes priority sorting, operation steps, required tools and fault image evidence.

[0013] Preferably, the maintenance personnel can upload image or video evidence of the fault scene through the PLC panel, bind the maintenance scheme for storage to the database, and form a multi-dimensional fault case library.

[0014] Preferably, the violation image evidence includes: N frames of time sequence images before the violation action occurs, local close-up images of the violation action, and same frame images including express delivery numbers and operator ID cards.

[0015] Preferably, when the central server calls the database through the fault code, the device model, the operating environment and the historical maintenance record are further associated.

[0016] The technical scheme of the present application has the following advantages: At the moment of equipment failure, the fault code is transmitted to the central server in real time, ensuring the timeliness and accuracy of the fault information, so that the central server can obtain the key information of the equipment failure in the first time, which saves valuable time for subsequent fault analysis and processing, and avoids further expansion of the fault or long-term shutdown of the equipment due to information delay. At the same time, after receiving the fault code, the central server immediately retrieves all historical fault information, equipment parameters and solution from the database according to the code, so that the maintenance personnel can obtain rich reference information in a short time after the fault occurs, quickly understand the possible causes and processing methods of the fault, and thus quickly start fault troubleshooting and repair work, greatly shortening the equipment downtime.

[0017] Through the integration and analysis of a large amount of historical fault information, the central server can provide more comprehensive and in-depth diagnostic basis for the current fault. Not only can the processing methods and lessons learned under the same fault code in the past be referred to, but also the specific parameters of the equipment can be combined to more accurately determine the cause of the fault. For example, some faults may exhibit different characteristics and impact ranges under different equipment parameters. Based on the comprehensive analysis of historical data and equipment parameters, false judgments caused by a single factor can be avoided, and the accuracy of fault diagnosis can be improved. Based on the matching of fault codes and corresponding machine fault code information, the causes of the fault are analyzed and a preliminary diagnosis report is generated to help maintenance personnel quickly locate the fault point in complex equipment systems. Even when facing some rare or complex faults, the preliminary diagnosis report can serve as an important reference to guide maintenance personnel to gradually troubleshoot, reduce blindness, and improve maintenance efficiency.

[0018] By associating with the equipment model, the central server can accurately match the equipment structure, performance parameters and fault feature library of different models. For example, the same fault code may correspond to different fault causes in different models of equipment (such as differences in sensor models, differences in mechanical structure design, etc.). By combining equipment models, false judgments caused by "one-size-fits-all" can be avoided, ensuring that the diagnosis results are highly consistent with the actual characteristics of the equipment.

[0019] The introduction of operating environment data (such as temperature, humidity, vibration, load, etc.) can real-time correct fault judgment threshold and diagnosis logic. For example, the allowable vibration value of the equipment bearing in a high-temperature environment may be lower than that in a normal-temperature environment. By combining environmental parameters, abnormal judgment standards can be dynamically adjusted to avoid false alarms caused by environmental interference. In addition, environmental data can reveal the correlation between faults and external conditions (such as short circuits caused by humid environments), providing key evidence for root cause analysis. DETAILED DESCRIPTION

[0020] The embodiments of the present application are described in detail below, examples of which are shown in the schemes, wherein the same or similar reference signs represent the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the schemes are exemplary and are intended to explain the present application, and cannot be understood as a limitation of the present application, all other embodiments obtained by those skilled in the art without creative labor on the basis of the embodiments in the present application are within the scope of protection of the present application.

[0021] In addition, if the description involving "first", "second" and the like in the present application is only for the purpose of description, such as for distinguishing the same or similar elements, and cannot be understood as indicating or implying the relative importance of the indicated technical features or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include at least one of the features. In addition, the technical solutions between various embodiments can be combined with each other, but it must be based on the fact that the technical solutions can be realized by those skilled in the art, when the combination of technical solutions appears contradictory or cannot be realized, it should be considered that the combination of technical solutions does not exist, nor within the scope of protection required by the present application.

[0022] The present application provides an intelligent equipment fault detection method, comprising the following steps: when the equipment fails, the fault code is returned to the central server in real time; the central server extracts all historical fault information, equipment parameters and solutions associated with the fault code from the database according to the received fault code; based on the matching of the fault code and the corresponding machine fault code information, the fault reason is analyzed, and a preliminary diagnosis report is generated; the fault reason and the treatment scheme are pushed to the PLC panel of the equipment according to the priority, and the maintenance personnel can check and maintain according to the scheme order; after the maintenance is completed, the maintenance personnel selects the specific maintenance scheme through the PLC panel, and the execution result is fed back to the database.

[0023] Among them, the acquisition step of the fault code is specifically: S1, real-time acquisition of video stream data of equipment running area; S2, image analysis is performed on the video stream data to identify equipment key component information; S3, detecting the running state of the equipment based on equipment posture and action recognition algorithm; S4, judging whether the current equipment running state is faulty according to the preset equipment fault behavior rule; S5, if the current equipment running state is judged to be faulty, triggering the fault code generation mechanism.

[0024] Furthermore, the device posture and motion recognition algorithm is specifically as follows: using the OpenPose algorithm to extract the coordinates of the key feature points of the device in real time, the key feature point coordinates include the motion trajectories of the feature points of the key connection parts, moving parts, etc. of the device; preprocessing the above-extracted key feature point coordinate data, and constructing a spatiotemporal motion trajectory model of each part of the device based on the preprocessed key feature point coordinates; extracting the device posture and motion features in the video stream data through the above-constructed spatiotemporal motion trajectory model to determine whether the device has a fault.

[0025] Specifically, the displacement change rate of key feature points between adjacent frames is Calculate; where, Expressed as time Coordinates of key points; is the video frame time interval; Expressed as The instantaneous velocity vector at time t; The rate of change of the velocity of the key feature points is calculated by Calculate; where, Expressed as the instantaneous velocity vector at time t; Expressed as The instantaneous acceleration vector at the moment; the rate of change of the joint bending angle is expressed by calculate; In this embodiment, the fault code is transmitted back to the central server in real time at the moment of a device failure, ensuring the timeliness and accuracy of fault information. This allows the central server to immediately obtain key information about the equipment failure, saving valuable time for subsequent fault analysis and resolution, and preventing further escalation of the failure or prolonged equipment downtime due to information delays. Furthermore, upon receiving the fault code, the central server immediately retrieves all associated historical fault information, equipment parameters, and solutions from the database based on the code. This rapid response mechanism enables maintenance personnel to obtain a wealth of reference information within a short period of time after the failure occurs, quickly understanding the possible causes and solutions for the failure, and thus rapidly initiating troubleshooting and repair work, significantly reducing equipment downtime.

[0026] In addition, by integrating and analyzing a large amount of historical fault information, the central server can provide more comprehensive and in-depth diagnostic basis for the current fault. Not only can the processing method and experience under the same fault code in the past be referred to, but also the specific parameters of the device can be combined to more accurately determine the cause of the fault. For example, some faults may exhibit different characteristics and impact ranges under different device parameters. Based on the comprehensive analysis of historical data and device parameters, misjudgment caused by a single factor can be avoided, and the accuracy of fault diagnosis can be improved. Based on the matching of fault codes and corresponding machine fault code information, the cause of the fault is analyzed and a preliminary diagnosis report is generated to help maintenance personnel quickly locate the fault point in complex device systems. Even when facing some rare or complex faults, the preliminary diagnosis report can serve as an important reference to guide maintenance personnel to gradually troubleshoot, reduce blindness, and improve maintenance efficiency.

[0027] In one of the embodiments, the central server further associates the device model, operating environment and historical maintenance records when calling the database through the fault code.

[0028] In this embodiment, by associating the device model, the central server can accurately match the device structure, performance parameters and fault feature library of different models. For example, the same fault code may correspond to different fault causes in different models of devices (such as differences in sensor models, differences in mechanical structure design, etc.). By combining the device model, misjudgment caused by "one-size-fits-all" can be avoided, and the diagnostic results can be highly consistent with the actual characteristics of the device.

[0029] On the other hand, introducing operating environment data (such as temperature, humidity, vibration, load, etc.) can real-time correct fault judgment threshold and diagnostic logic. For example, the allowable vibration value of the device bearing in a high-temperature environment may be lower than that in a normal-temperature environment. By combining environmental parameters, abnormal judgment criteria can be dynamically adjusted to avoid false positives caused by environmental interference. In addition, environmental data can reveal the correlation between faults and external conditions (such as short circuits caused by humid environments), providing key evidence for root cause analysis.

[0030] Further, the judgment of the device fault information is as follows: When the instantaneous speed of the feature point coordinates of the key moving parts of the device exceeds the threshold value and the motion trajectory shows abnormal fluctuations, it is judged that there is a running instability fault in the running process of the device; When the acceleration mutation value of the key feature points of the device exceeds the preset threshold value, it is judged that there is an abnormal situation of excessive impact of the device; When the relative angle change frequency of the key connection parts of the device exceeds the preset threshold value, it is judged that there is a fault situation of excessive vibration or loose connection of the device; When the angle between the feature points of the key support parts of the device and the ground projection is less than a first preset angle threshold, it is judged that there is a fault posture of excessive tilting of the device; When the difference between the motion activity of the feature points of the symmetric components on both sides of the device exceeds a second preset threshold value, it is determined that there is an irregular state of unbalanced operation of the device; When the motion trajectory of the key components of the device to the target operating position deviates from the optimal path and the offset exceeds a third preset threshold value, it is determined that there is an inefficient state of detour or bypass of the device operation.

[0031] The first preset angle threshold value, the second preset threshold value, and the third preset threshold value are determined by the following method: The first preset angle threshold value is set according to the normal inclination range and safety standards of the device; The second preset threshold value is determined based on the statistical results of the difference between the motion activity of the feature points of the symmetric components on both sides of the device in the normal operating state; The third preset threshold value is calculated according to the optimal path and the acceptable deviation range of the key components of the device to the target operating position.

[0032] In this embodiment, the instantaneous speed threshold value and the motion trajectory feature of the key motion components are used to determine the unstable state of the device in operation, which avoids single parameter misjudgment (such as short and fast movement but not abnormal fluctuation), accurately identifies the unstable state of the device in operation, and reduces production accidents caused by device failure. At the same time, the sudden impact is judged based on the mutation value of the acceleration of the feature points (such as the sharp change of acceleration in a short time), which can identify the situation that the impact force is too large caused by sudden external force or internal failure of the device in operation, prevent the device from being damaged due to too large impact, and improve the safety of the device in operation. The relative angle change frequency of the key connection position is used to quantify the vibration of the device, and the high frequency may indicate that the vibration of the device is too large (such as high frequency vibration caused by loose parts) or the connection position is loose (such as loose connection caused by long-term accumulation of small vibration), which monitors the stability of the device in operation and warns potential device failure risk in advance. The angle between the feature points of the key support position and the ground projection is used to judge the inclination degree of the device, and the small angle (such as close to 90 degrees) indicates that the inclination is too large, which corrects the inclination problem caused by improper installation of the device or foundation settlement, and avoids damage and production safety accidents caused by inclination. On the other hand, by comparing the motion activity difference (such as displacement distance, speed, and trajectory complexity) of the feature points of the symmetric components on both sides of the device, the large difference indicates that the device is not balanced, which avoids uneven stress, accelerated wear and tear, and other safety hazards caused by unbalanced operation of the device, and forces the balanced and stable operation of the device.

[0033] In one of the embodiments, the repair scheme pushed to the PLC panel includes priority, operation steps, required tools, and fault image evidence.

[0034] In this embodiment, by setting priorities for the maintenance solutions (such as emergency repair, temporary relief, root cause treatment, etc.), maintenance personnel can quickly identify the most critical operation steps, and avoid the deterioration of the fault or secondary damage to the equipment due to incorrect solution selection. For example, performing "cut off the power supply of the fault circuit" instead of directly replacing the component can prevent the spread of short circuit risk. At the same time, automatically listing the required tools (such as torque wrench, multimeter) and spare parts models (such as bearing specifications, relay types) can reduce the maintenance preparation time and avoid rework due to tool mismatch or spare parts errors.

[0035] On the other hand, the fault image evidence pushed to the PLC panel (such as internal photos of the equipment, sensor data visualization charts, thermal imaging maps, etc.) can help maintenance personnel quickly locate the fault location and reduce troubleshooting time. For example, by comparing the vibration spectrum graphs under normal and fault conditions, the degree of bearing wear can be directly judged.

[0036] In one embodiment, the maintenance personnel can upload image or video evidence of the fault site through the PLC panel, bind it with the maintenance solution, and store it in the database to form a multi-dimensional fault case library.

[0037] Further, the fault image evidence includes: N frames of time sequence images before the fault occurs, close-up images of the fault site, and images of the same frame of the equipment number and the surrounding environment identification.

[0038] In this embodiment, by recording the video pictures of the previous N frames (for example, N=5 or 10) before the fault occurs, the spatio-temporal context of the equipment running scene is completely restored, the brewing process of the fault, the equipment running state and the on-site environment are displayed, and objective basis for determining the continuity of the fault is provided. At the same time, the equipment number and the surrounding environment identification information are recorded in the same frame picture, ensuring the strong correlation between the fault and the specific equipment and the surrounding environment.

[0039] In one embodiment, the method further comprises: generating a multi-dimensional statistical report according to the execution result of step S5, including a device fault frequency heat map, a device component fault ranking, and a safety hazard early warning list.

[0040] It should be noted that in this paper, the term "comprising", "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, device, article or equipment failure intelligent detection method including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or includes elements inherent to such process, device, article or equipment failure intelligent detection method. Without more limitations, the element defined by the sentence "including a…" does not exclude the presence of another identical element in the process, device, article or equipment failure intelligent detection method including the element.

[0041] The above merely provides the preferred embodiments of the present application, and is not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation, or direct or indirect application in other related technical fields, which is made according to the content of the present application, shall be included in the patent protection scope of the present application.

Claims

1. A method for intelligently detecting equipment failures, comprising the following steps: When a device fails, the fault code is transmitted back to the central server in real time; based on the fault code matching the corresponding machine fault code information, the cause of the fault is analyzed and a preliminary diagnosis report is generated; the fault cause and treatment plan are pushed to the device PLC panel according to priority, so that maintenance personnel can investigate and repair according to the plan sequence; the step of obtaining the fault code includes the following steps: S1. Real-time acquisition of video stream data of the device operation area; S2. Perform image analysis on the video stream data to identify key component information of the device; S3, detecting the operating status of the device based on the device posture and motion recognition algorithm; S4. Determine whether the current device operation status is faulty based on the preset device fault behavior rules; S5. If the current device operation status is determined to be a fault, a fault code generation mechanism is triggered.

2. The intelligent detection method for equipment failure according to claim 1, characterized in that: The device posture and motion recognition algorithm specifically includes: using the OpenPose algorithm to extract the coordinates of the device's key feature points in real time, wherein the key feature point coordinates include the motion trajectories of the feature points of the device's key connection parts, moving parts, etc.; preprocessing the extracted key feature point coordinate data, and constructing a spatiotemporal motion trajectory model of each part of the device based on the preprocessed key feature point coordinates; and extracting the device posture and motion features in the video stream data through the constructed spatiotemporal motion trajectory model to determine whether the device is faulty.

3. The intelligent detection method for equipment failure according to claim 2, characterized in that: The determination of the equipment failure information in step S3 is specifically as follows: When the instantaneous speed of the characteristic point coordinates of the key moving parts of the equipment exceeds the threshold and the motion trajectory fluctuates abnormally, it is determined that the equipment has an unstable operation fault during operation; When the acceleration mutation value of the key characteristic point of the equipment exceeds the preset threshold, it is judged that there is an abnormal situation of excessive equipment operation impact; When the frequency of relative angle changes at key connection points of the device exceeds a preset threshold, it is determined that there is a fault situation of excessive device vibration or loose connection.

4. The intelligent detection method for equipment failure according to claim 3, characterized in that: In step S3, the instantaneous speed threshold, the preset threshold of the acceleration mutation value, and the preset threshold of the relative angle change frequency are all set based on the basic specification parameters and historical operation data of the detection equipment.

5. The intelligent detection method for equipment failure according to claim 1, characterized in that: The determination of the equipment fault information in step S3 further includes: When the angle between the feature point of the key support part of the device and the projection on the ground is less than a first preset angle threshold, it is determined that there is a fault posture of excessive tilt of the device; When the difference in the motion activity of the characteristic points of the symmetrical components on both sides of the equipment exceeds a second preset threshold, it is determined that the equipment is in an irregular state of unbalanced operation; When the motion trajectory of the key component of the equipment to the target operating position deviates from the optimal path and the offset exceeds a third preset threshold, it is determined that there is an inefficient state of equipment operation bypassing or detour.

6. The intelligent device failure detection method according to claim 5, characterized in that: The first preset angle threshold, the second preset threshold, and the third preset threshold are all determined in the following manner: The first preset angle threshold is set according to the normal tilt range of the device and safety standards; The second preset threshold is determined based on the statistical results of the difference in motion activity of the feature points of the symmetrical components on both sides of the device under normal operating conditions; The third preset threshold is calculated based on the optimal path from the key components of the equipment to the target operating position and the acceptable deviation range.

7. The intelligent device failure detection method according to claim 1, characterized in that: The maintenance plan pushed to the PLC panel includes: priority sorting, operation steps, required tools and fault image evidence.

8. The intelligent device failure detection method according to claim 1, characterized in that: The maintenance personnel can upload images or video evidence of the fault scene through the PLC panel, and store them in the database together with the maintenance plan to form a multi-dimensional fault case library.

9. The intelligent device failure detection method according to claim 8, characterized in that: The fault image evidence includes: time-series images of N frames before the fault occurs, local close-up images of the fault location, and same-frame images of the device number and surrounding environment identification.

10. The intelligent device failure detection method according to claim 9, characterized in that: When the central server calls the database through the fault code, it further associates the equipment model, operating environment and historical maintenance records.