Method and apparatus for condition monitoring of a mechanical device

Abstract

The application provides a mechanical equipment state monitoring method and device, the state monitoring method comprises the following steps: collecting monitoring data of each preset monitoring area of the mechanical equipment in real time; for each monitoring area collecting temperature data, judging whether there is temperature abnormality according to the temperature data, and generating a temperature abnormality coefficient if there is; for each monitoring area collecting vibration data, judging whether there is vibration abnormality according to the vibration data, and generating a vibration abnormality coefficient if there is; for each monitoring area collecting video data, judging whether there is safety abnormality according to the video data, and generating a safety abnormality coefficient if there is; calculating comprehensive abnormality data according to the abnormality coefficients of various types generated by each monitoring area, and performing hierarchical alarm according to the comprehensive abnormality data. The application provides a mechanical equipment state monitoring method and device to solve the technical problem that hidden faults are difficult to effectively identify.

Description

Technical Field

[0001] This invention relates to the industrial field, and more particularly to a method and apparatus for monitoring the condition of mechanical equipment. Background Technology

[0002] Dry bulk cargo ports are widely equipped with large-scale machinery such as reclaimers, ship loaders, and tippers. These machines are the core equipment for port loading and unloading operations, directly determining the overall throughput efficiency and continuous operation capacity of the port. Meanwhile, in industrial sectors such as mining, metallurgy, and power, large-scale machinery such as crushers, conveyors, and rotary kilns also play a role in continuous production. These large-scale machines are generally characterized by their massive size, complex transmission mechanisms, long-term high-load continuous operation, high dust concentration, and strong environmental interference. The working condition of the equipment's operating points (such as motors, reducers, and roller bearing housings) directly affects the equipment's operational safety and production stability.

[0003] Currently, the industry generally adopts a traditional operation and maintenance model that combines manual on-site inspections with regular preventative maintenance. This model has significant technical shortcomings: the frequency of inspections is limited and there are many blind spots, making it difficult to effectively identify progressive, latent faults such as early wear, structural fatigue, and minor electrical anomalies. Furthermore, fixed-cycle maintenance strategies cannot match the actual operating conditions of the equipment, resulting in both over-maintenance and under-maintenance. These shortcomings mean that minor early-stage equipment degradation is difficult to detect in a timely manner, often only coming to the attention of maintenance personnel after the degradation has accumulated to a significant level, thus missing the optimal time for maintenance. Therefore, there are areas for improvement. Summary of the Invention

[0004] This invention provides a method and apparatus for monitoring the condition of mechanical equipment, in order to solve the technical problem that latent faults are difficult to identify effectively.

[0005] The present invention provides a method for monitoring the condition of mechanical equipment, comprising: Real-time acquisition of monitoring data from various preset monitoring areas of mechanical equipment; wherein, the monitoring data includes at least one of temperature data, vibration data, and video data; For each monitoring area where the temperature data is collected, it is determined whether there is a temperature anomaly based on the temperature data of that monitoring area. If there is, a temperature anomaly coefficient for that monitoring area is generated. For each monitoring area where the vibration data is collected, it is determined whether there is a vibration anomaly based on the vibration data of that monitoring area. If so, a vibration anomaly coefficient for that monitoring area is generated. For each monitoring area where the video data is collected, determine whether there is a security anomaly based on the video data of that monitoring area. If there is, generate a security anomaly coefficient for that monitoring area. Based on the various anomaly coefficients generated in each monitoring area, comprehensive anomaly data is calculated, and graded alarms are executed based on the comprehensive anomaly data.

[0006] In one embodiment of the present invention, the temperature data is collected by a temperature sensor installed in the monitoring area, and / or the temperature data is obtained by processing a visible light image and an infrared thermal image containing the monitoring area. The processing of the visible light image and infrared thermal image containing the monitoring area includes: Simultaneously, a visible light image and an infrared thermal image containing the monitoring area are acquired, and the visible light image and the infrared thermal image are aligned in position. Identify and segment the pixel mask region corresponding to the monitoring region in the visible light image; The pixel mask area is mapped onto the infrared thermal image, and the temperature value of each pixel within the pixel mask area is extracted. Sort all extracted temperature values ​​from largest to smallest, select the first preset number of temperature values, and use the median of the preset number of temperature values ​​as the temperature data for the monitoring area.

[0007] In one embodiment of the present invention, the step of determining whether a temperature anomaly exists based on the temperature data of the monitoring area, and generating a temperature anomaly coefficient for the monitoring area if an anomaly exists, includes: Based on the time node of the collected temperature data, determine the time period to which the time node belongs, and obtain the historical average temperature and temperature fluctuation threshold corresponding to the time period. Calculate the temperature difference between the temperature data and the historical average temperature; When the temperature difference is greater than the temperature fluctuation threshold, it is determined that there is a temperature anomaly in the monitoring area, and a temperature anomaly coefficient for the monitoring area is generated based on the temperature difference.

[0008] In one embodiment of the present invention, the historical average temperature and temperature fluctuation threshold corresponding to the time period are predetermined by means of the following methods: Based on the start-up and shutdown time nodes of the equipment in the monitoring area, the time is divided into at least one working time period and at least one non-working time period; For each time period: Acquire multiple sets of historical temperature data for the equipment in the monitoring area under normal conditions; The historical average temperature of the monitored area during this time period is determined based on the average of all the historical temperature data. Calculate the standard deviation of temperature fluctuation based on the dispersion of all the historical temperature data relative to the historical average temperature; A temperature fluctuation threshold is set based on the standard deviation of the temperature fluctuation.

[0009] In one embodiment of the present invention, the formula for calculating the temperature anomaly coefficient is as follows:

[0010] in, This is the temperature anomaly coefficient. This is the temperature difference value. This represents the standard deviation of temperature fluctuation.

[0011] In one embodiment of the present invention, the step of determining whether there is a vibration anomaly based on the vibration data of the monitoring area, and generating a vibration anomaly coefficient for the monitoring area if such an anomaly exists, includes: Convert the vibration data of the monitored area into frequency domain data; Extract the amplitude and frequency of at least one main frequency band from the frequency domain data; The amplitude is compared with a preset amplitude threshold, and / or the frequency is compared with a preset frequency range; When the amplitude is greater than the amplitude threshold and / or the frequency is not in the frequency range, it is determined that there is a vibration anomaly in the monitoring area, and a corresponding vibration anomaly coefficient is generated according to the type of equipment to which the monitoring area belongs.

[0012] In one embodiment of the present invention, the step of determining whether there is a security anomaly based on the video data of the monitoring area, and generating a security anomaly coefficient for the monitoring area if such an anomaly exists, includes: Image recognition is performed on the video data of the monitored area to determine whether there are any pre-set violation scenarios within the monitored area; When the violation scenario is identified, it is determined that there is a security anomaly in the monitoring area, and a security anomaly coefficient for the monitoring area is generated based on the risk level of the violation scenario.

[0013] In one embodiment of the present invention, the step of calculating comprehensive anomaly data based on various anomaly coefficients generated in each monitoring area includes: For each monitoring area, if there are multiple anomalous coefficients of the same type, the one with the largest value is retained as the anomalous coefficient of that type for that monitoring area. Based on the similar anomaly coefficients retained in all monitoring areas, calculate the hazard components corresponding to each type of anomaly coefficient; Based on the various hazard components and their corresponding weighting coefficients, comprehensive abnormal data is calculated.

[0014] In one embodiment of the present invention, the calculation formula for the comprehensive abnormal data is as follows:

[0015] in, To integrate abnormal data, , , All are weighting coefficients, and , This represents the total number of monitoring areas with temperature anomalies. For the first Temperature anomaly coefficient of each monitoring area This represents the total number of monitoring areas with vibration anomaly coefficients. For the first Vibration anomaly coefficient of each monitoring area This represents the total number of monitoring areas with anomaly coefficients. For the first The safety anomaly coefficient of each monitoring area.

[0016] The present invention also provides a condition monitoring device for mechanical equipment, comprising: The data acquisition module is used to collect monitoring data from various preset monitoring areas of the mechanical equipment in real time; wherein, the monitoring data includes at least one of temperature data, vibration data, and video data; The temperature judgment module is used to determine whether there is a temperature anomaly for each monitoring area where the temperature data has been collected, and if so, to generate a temperature anomaly coefficient for that monitoring area. The vibration judgment module is used to determine whether there is a vibration anomaly for each monitoring area where the vibration data has been collected, and if so, to generate a vibration anomaly coefficient for that monitoring area. The security judgment module is used to determine whether there is a security anomaly for each monitoring area where the video data has been collected, based on the video data of that monitoring area. If there is, it generates a security anomaly coefficient for that monitoring area. The comprehensive judgment module is used to calculate comprehensive abnormal data based on various abnormal coefficients generated in each monitoring area, and to execute graded alarms based on the comprehensive abnormal data.

[0017] The beneficial effects of this invention are as follows: By collecting real-time monitoring data of temperature, vibration, and video, it replaces traditional manual inspections, solving the problems of limited inspection frequency and numerous blind spots. By identifying and quantifying abnormal temperatures, vibrations, and safety anomalies, it effectively captures latent faults such as early wear and electrical abnormalities, compensating for the difficulty of detecting early defects in manual inspections. By calculating comprehensive abnormal data and executing tiered alarms, it achieves accurate early warning and maintenance decisions based on the actual operating status of the equipment, replacing the fixed-cycle maintenance model, solving the problem of both over-maintenance and under-maintenance, and reducing operation and maintenance costs. Attached Figure Description

[0018] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0019] In the attached diagram: Figure 1 This is a flowchart of a mechanical equipment condition monitoring method provided in one embodiment of the present invention; Figure 2 This is a schematic diagram of a mechanical equipment condition monitoring device provided in one embodiment of the present invention. Detailed Implementation

[0020] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0021] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. The drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0022] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the invention. However, it will be apparent to those skilled in the art that embodiments of the invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the invention.

[0023] Currently, dry bulk cargo ports are widely equipped with large-scale mechanical equipment such as reclaimers, ship loaders, and tippers. These devices are the core equipment for port loading and unloading operations, and their operational stability directly affects the overall throughput efficiency and safety management level of the port. These large-scale mechanical equipment generally have the characteristics of large structure, complex transmission mechanism, long-term high-load continuous operation, high dust concentration on site, and strong environmental interference. The working status of the equipment's operating points (such as motors, reducers, and roller bearing housings) directly determines the equipment's operational safety and production continuity.

[0024] Under the aforementioned operating conditions, progressive, latent faults such as early wear, structural fatigue, and minor electrical anomalies are difficult to identify effectively through conventional methods. Currently, the industry generally adopts a traditional operation and maintenance model that combines manual on-site inspections with regular preventative maintenance. This model heavily relies on the experience and judgment of maintenance personnel and fixed-cycle maintenance plans. Due to limited inspection frequency and numerous blind spots, minor potential problems cannot be detected and addressed in a timely manner. Furthermore, fixed-cycle maintenance strategies cannot match the actual operating status of the equipment, resulting in both over-maintenance and under-maintenance, leading to high operation and maintenance costs and failing to effectively control the risk of downtime due to malfunctions.

[0025] To address this, the present invention provides a method and apparatus for monitoring the condition of mechanical equipment. By collecting temperature data, vibration data, and video data from various monitoring areas of the equipment in real time, the method identifies and quantifies temperature anomalies, vibration anomalies, and safety anomalies in each monitoring area. Based on the anomaly coefficients generated in each monitoring area, the method calculates comprehensive anomaly data and executes graded alarms, thereby achieving comprehensive perception and graded early warning of the equipment's operating status. This effectively improves the intelligence level of equipment operation and maintenance, reduces equipment failure rate and total life cycle maintenance costs, and ensures safe, stable, efficient, and continuous operation of the equipment.

[0026] Please see Figure 1 In a preferred embodiment of the present invention, the condition monitoring method includes the following steps: S10, real-time acquisition of monitoring data of each preset monitoring area of ​​the mechanical equipment; wherein, the monitoring data includes at least one of temperature data, vibration data, and video data.

[0027] Specifically, for large mechanical equipment, multiple monitoring areas are pre-defined. The locations of these monitoring areas include operating points such as motor housings, reducer housings, belt drive rollers, and shaft bearing seats, as well as safety risk areas such as equipment operating areas and areas where clearing and stacking vehicles operate.

[0028] Each monitoring area is configured with at least one monitoring method. The monitoring methods include at least three categories: the first category is for collecting temperature data, the second for collecting vibration data, and the third for collecting video data. A monitoring area can be configured with any one of these three categories, or multiple categories simultaneously. When configuring a particular monitoring method, the number of methods within that category can be one or more. When a monitoring area is configured with multiple monitoring methods simultaneously, it outputs multiple types of monitoring data at the same time point.

[0029] Temperature data is acquired by temperature acquisition devices located in the monitoring area. Vibration data is acquired by vibration acquisition devices located in the monitoring area. Video data is acquired by image acquisition devices located in the monitoring area. All acquisition devices complete data acquisition simultaneously, ensuring that the various monitoring data are temporally corresponding when performing multi-dimensional data fusion analysis on the same monitoring area subsequently.

[0030] Temperature data is used to determine if there are any temperature anomalies in the monitored area. Vibration data is used to determine if there are any vibration anomalies in the monitored area. Video data is used to determine if there are any safety anomalies in the monitored area. Each monitoring area outputs the corresponding type of monitoring data according to its configured monitoring method.

[0031] In a preferred embodiment of the present invention, the temperature data is collected by a temperature sensor located in the monitoring area, and / or the temperature data is obtained by processing a visible light image and an infrared thermogram containing the monitoring area.

[0032] Specifically, the first type of monitoring method is used to collect temperature data, and it has two implementation methods. In the first implementation method, the temperature acquisition device includes a temperature sensor, which directly collects temperature values ​​based on the temperature values ​​set up in the monitoring area. The temperature sensor can be a thermocouple, a resistance temperature detector (RTD), or a semiconductor temperature sensor. The temperature sensor is attached to or embedded in the surface or interior of the monitoring area, and outputs the temperature value of the monitoring area in real time after being powered on. The temperature value output by the temperature sensor is the temperature data of the monitoring area. In the second implementation method, the temperature acquisition device includes a visible light camera and an infrared thermal imager. Temperature data is obtained by fusing the visible light image and the infrared thermal image containing the monitoring area.

[0033] A monitoring area can collect temperature data using only one implementation method or simultaneously using two methods. When a monitoring area uses two implementation methods, it outputs two sets of temperature data at the same time point. Furthermore, when configuring a monitoring area with a particular type of monitoring method, the number of that type of monitoring method can be one or more. For example, a monitoring area can be equipped with multiple temperature sensors, in which case the monitoring area outputs multiple sets of temperature data at the same time point, corresponding to temperature values ​​from different installation locations. Multiple temperature sensors can be located in different sub-areas of the same monitoring area, serving as a redundant configuration for verification and comparison.

[0034] In a preferred embodiment of the present invention, processing the visible light image and infrared thermal image containing the monitoring area includes the following steps: Simultaneously acquire visible light images and infrared thermal images containing the monitoring area, and align the visible light images and infrared thermal images in terms of position; Identify and segment the pixel mask region corresponding to the monitoring area in the visible light image; Map the pixel mask area onto the infrared thermal image and extract the temperature value of each pixel within the pixel mask area; Sort all extracted temperature values ​​from largest to smallest, select the first preset number of temperature values, and use the median of the preset number of temperature values ​​as the temperature data for the monitoring area.

[0035] Specifically, for the monitoring area where temperature data is acquired using the second method, both a visible light image and an infrared thermal image containing the monitoring area need to be acquired simultaneously. The visible light image is captured by a visible light camera, and the infrared thermal image is captured by an infrared thermal imager. The visible light camera and the infrared thermal imager are installed near the monitoring area, and their fields of view cover the same spatial range.

[0036] After acquiring the visible light image and the infrared thermal image, pixel-level position registration is required. Specifically, the registration is performed based on the installation position relationship between the visible light camera and the infrared thermal imager, as well as the mapping parameters obtained from the previous calibration. After pixel-level position registration, a one-to-one correspondence is established between each pixel in the visible light image and the corresponding pixel in the infrared thermal image at the same spatial location.

[0037] Subsequently, the visible light image, after pixel-level location registration, is input into the pre-trained YOLOv8 instance segmentation model. Specifically, the pre-trained YOLOv8 instance segmentation model is able to identify pre-defined monitoring targets from the visible light image; these targets are the device components or spatial regions corresponding to the monitoring area. The YOLOv8 instance segmentation model outputs the pixels occupied by the monitoring area in the visible light image, and these pixels constitute a pixel mask region. The boundary of this pixel mask region coincides with the contour of the monitoring area in the visible light image, accurately delineating the pixel range covered by the monitoring area.

[0038] Then, the obtained pixel mask region is mapped onto the infrared thermal image, and the temperature value of each pixel within the pixel mask region is extracted. Specifically, since the visible light image and the infrared thermal image have already completed pixel-level position registration, the pixel mask region in the visible light image can be directly mapped to the same location in the infrared thermal image. This pixel mask region is then superimposed onto the infrared thermal image, and the temperature value corresponding to each pixel within the pixel mask region in the infrared thermal image is read. The extracted temperature values ​​form a set of temperature values, and the number of temperature values ​​in the set is equal to the total number of pixels within the pixel mask region.

[0039] Finally, all extracted temperature values ​​are sorted from largest to smallest, and a preset number of temperature values ​​are selected. The median of this preset number of temperature values ​​is used as the temperature data for the monitoring area. Specifically, during actual data acquisition, there may be obstructions, recognition gaps, heat reflection interference, etc., causing the temperature values ​​of some pixels to deviate from the normal range. If the average temperature of all pixels within the mask area is directly calculated, these deviations will significantly affect the results. Therefore, all extracted temperature values ​​are first arranged in descending order, and then a preset number of temperature values ​​are selected from the side with higher temperature values. The preset number is less than the total number of temperature values ​​in the temperature value set. Then, the median of these selected temperature values ​​is selected and used as the temperature data for the monitoring area. For example, if there are 2500 pixels in the pixel mask area of ​​a certain monitoring area, 2500 temperature values ​​are extracted. The 2500 temperature values ​​are sorted from largest to smallest, the first 101 temperature values ​​are selected, and then the 51st temperature value among these 101 is selected as the temperature data for the monitoring area.

[0040] Please see Figure 1 In a preferred embodiment of the present invention, the status monitoring method further includes the following steps: S20, for each monitoring area where temperature data has been collected, determine whether there is a temperature anomaly based on the temperature data of the monitoring area, and if so, generate a temperature anomaly coefficient for the monitoring area.

[0041] In a preferred embodiment of the present invention, S20 includes the following steps: S21, determining the time period to which the time node belongs based on the time node of the collected temperature data, and obtaining the historical average temperature and temperature fluctuation threshold corresponding to the time period. The historical average temperature and temperature fluctuation threshold corresponding to the time period are predetermined in the following manner: Based on the start-up and shutdown time nodes of the equipment in the monitoring area, the time is divided into at least one working time period and at least one non-working time period; For each time period: Acquire multiple sets of historical temperature data for the equipment in the monitoring area under normal conditions; The historical average temperature of the monitored area during this time period is determined based on the average of all historical temperature data. Calculate the standard deviation of temperature fluctuation based on the dispersion of all historical temperature data relative to the historical average temperature; Set the temperature fluctuation threshold based on the standard deviation of temperature fluctuation.

[0042] Specifically, after the temperature data is collected, anomaly detection is required for the temperature data of each monitoring area. Before anomaly detection, the historical average temperature and temperature fluctuation threshold for each time period must be determined. The historical average temperature and temperature fluctuation threshold are pre-calculated.

[0043] During the initial commissioning of the equipment or in daily operation, historical temperature data of the equipment in each monitoring area under normal conditions is collected. Since the equipment does not operate continuously 24 / 7, it has working and non-working periods, and the temperature performance varies during these periods. Therefore, it is necessary to segment the time according to the equipment's start-up and shutdown patterns. For example, based on the start-up and shutdown times of the equipment in the monitoring area, the day is divided into at least one working period and at least one non-working period. For instance, if a motor is in working condition from 0:00 to 4:00, in a non-working condition from 4:00 to 6:00, and then enters working condition again from 6:00 to 10:00, repeating this cycle, then 0:00 to 4:00 constitutes one working period, 4:00 to 6:00 constitutes one non-working period, and so on.

[0044] For each defined time period, multiple sets of historical temperature data for that monitoring area are extracted from the database. These historical temperature data originate from temperature records of the monitoring area over several days under normal conditions. The average of these historical temperature data is then calculated and used as the historical average temperature for that monitoring area during that time period. Simultaneously, the dispersion of each historical temperature data point relative to this historical average temperature is calculated to obtain the temperature fluctuation standard deviation. The temperature fluctuation standard deviation reflects the normal fluctuation range of the temperature data in the monitoring area during that time period. Finally, a temperature fluctuation threshold is set based on this temperature fluctuation standard deviation. For example, the temperature fluctuation threshold can be twice the temperature fluctuation standard deviation.

[0045] Once the historical average temperature and temperature fluctuation threshold are determined, they can be used for anomaly detection in the temperature data. Specifically, based on the time node to which the temperature data was collected, the time period to which that time node belongs is determined, and then the corresponding historical average temperature and temperature fluctuation threshold are obtained. The temperature data is collected at a specific time node, which falls within a pre-defined working or non-working time period. Different time periods correspond to different historical average temperatures and different temperature fluctuation thresholds; therefore, it is necessary to first determine the time period to which the time node belongs, and then extract the corresponding historical average temperature and temperature fluctuation threshold for that time period.

[0046] In a preferred embodiment of the present invention, S20 further includes the following step: S22, calculating the temperature difference between the temperature data and the historical average temperature. The temperature difference is calculated by subtracting the historical average temperature from the temperature data and taking the absolute value.

[0047] In a preferred embodiment of the present invention, S20 further includes the following step: S23, when the temperature difference is greater than the temperature fluctuation threshold, it is determined that there is a temperature anomaly in the monitoring area, and a temperature anomaly coefficient for the monitoring area is generated based on the temperature difference. The formula for calculating the temperature anomaly coefficient is as follows:

[0048] in, This is the temperature anomaly coefficient. This is the temperature difference value. This represents the standard deviation of temperature fluctuation.

[0049] Specifically, the temperature difference is compared with a temperature fluctuation threshold. If the temperature difference is greater than the threshold, it indicates that the current temperature data deviates from historical normal levels beyond the normal fluctuation range. In this case, a temperature anomaly is determined to exist in the monitored area, and a temperature anomaly coefficient is generated based on the temperature difference. The temperature anomaly coefficient reflects the severity of the temperature anomaly in the monitored area; the larger the temperature difference, the higher the temperature anomaly coefficient. If the temperature difference is less than or equal to the threshold, it indicates that the current temperature is within the historical normal fluctuation range, and there is no temperature anomaly in the monitored area.

[0050] Please see Figure 1 In a preferred embodiment of the present invention, the state monitoring method further includes the following steps: S30, for each monitoring area where vibration data has been collected, determine whether there is a vibration anomaly based on the vibration data of the monitoring area, and if so, generate a vibration anomaly coefficient for the monitoring area.

[0051] In a preferred embodiment of the present invention, S30 includes the following steps: S31, converting the vibration data of the monitoring area into frequency domain data.

[0052] Specifically, vibration data is acquired by vibration acquisition equipment, which can be a vibration sensor. In the on-site deployment of large mechanical equipment, the vibration sensor and temperature sensor can be integrated into a single design, forming a temperature-vibration integrated sensor. This integrated temperature-vibration sensor integrates both a vibration sensor and a temperature sensor, outputting both vibration and temperature data. This design reduces the number of openings in the equipment casing, lowers installation complexity, and ensures that the vibration and temperature data originate from the same installation location and have the same acquisition time point, facilitating subsequent correlation analysis. In this embodiment, the temperature-vibration integrated sensor is installed on the bearing seat or housing surface of the equipment within the monitoring area. Its vibration sensor is used to acquire vibration data, and its temperature sensor is used to acquire temperature data.

[0053] The vibration sensor's sampling frequency range is between 1Hz and 10kHz. The sensor outputs vibration data, specifically waveform data showing the amplitude changing over time; the specific fault type cannot be directly identified from this waveform data. A Fast Fourier Transform (FFT) is needed to convert the vibration data into frequency domain data. The FFT decomposes the vibration data into a superposition of different frequency components, outputting a spectrum graph where the horizontal axis represents frequency and the vertical axis represents the vibration amplitude at the corresponding frequency. After the conversion, the frequency domain data for the monitored area is obtained.

[0054] In a preferred embodiment of the present invention, S30 further includes the following step: S32, extracting the amplitude and frequency of at least one main frequency band from the frequency domain data.

[0055] Specifically, the dominant frequency band refers to the frequency range in the spectrum with larger vibration amplitudes. These dominant frequency bands correspond to the characteristic frequencies of various mechanical components during equipment operation. In actual extraction, low-frequency interference components in the frequency domain data are first removed. This type of low-frequency interference usually comes from environmental vibration or measurement noise, and removing it can reduce interference with subsequent judgments. Then, the first three dominant frequency bands are extracted from the remaining spectrum, and the amplitude and frequency corresponding to each dominant frequency band are recorded.

[0056] In a preferred embodiment of the present invention, S30 further includes the following steps: S33, comparing the amplitude with a preset amplitude threshold and / or comparing the frequency with a preset frequency range: when the amplitude is greater than the amplitude threshold and / or the frequency is not in the frequency range, it is determined that there is a vibration abnormality in the monitoring area, and a corresponding vibration abnormality coefficient is generated according to the type of equipment to which the monitoring area belongs.

[0057] Specifically, the preset amplitude threshold and preset frequency range are set separately for different types of equipment. Different types of equipment have different operating characteristics, and the corresponding amplitude threshold and frequency range are also different. For example, the rotation frequency of a motor is relatively high, and the meshing frequency of a reducer also has its specific range. The preset amplitude threshold and preset frequency range are pre-calibrated according to the equipment type and operating conditions.

[0058] There are three comparison methods for amplitude and frequency. The first method compares only the amplitude to a preset amplitude threshold. The second method compares only the frequency to a preset frequency range. The third method compares both the amplitude and frequency to the preset amplitude threshold. The specific method used depends on the actual monitoring needs of the monitoring area and the type of equipment. Regardless of the method used, if the amplitude exceeds the amplitude threshold or the frequency is outside the frequency range, the monitoring area is considered to have abnormal vibration.

[0059] After determining that an abnormal vibration exists in the monitored area, a corresponding vibration anomaly coefficient is generated based on the type of equipment in that area. Different equipment types correspond to different vibration anomaly coefficient values ​​for the same abnormal phenomenon. Equipment types include motors, reducers, fans, conveyor belt rollers, etc. Taking a motor as an example, if the motor exhibits a bearing failure characteristic frequency, the vibration anomaly coefficient is 0.7; if the motor exhibits a rotor imbalance characteristic frequency, the vibration anomaly coefficient is 0.5. Different types of equipment have their own corresponding anomaly coefficient value rules, which can be preset according to the importance of the equipment and the risk of failure.

[0060] The difference between the vibration anomaly coefficient and the temperature anomaly coefficient lies in the fact that the vibration anomaly coefficient is discrete, not continuous. The degree of temperature anomaly can be quantified; the greater the deviation from the historical average temperature, the higher the temperature anomaly coefficient, and there is a continuous correlation between the degree of deviation and the temperature anomaly coefficient. However, the judgment of vibration anomalies is essentially a qualitative judgment. It's impossible to calculate a continuous coefficient value based on the magnitude of the deviation; instead, a preset vibration anomaly coefficient can be given based on the detected fault type. For example, a bearing fault corresponds to 0.7, and a gear fault corresponds to 0.5. Therefore, after an anomaly is determined, the vibration anomaly coefficient is directly set based on the equipment type and fault type, without further continuous quantitative calculation.

[0061] Please see Figure 1 In a preferred embodiment of the present invention, the status monitoring method further includes the following steps: S40, for each monitoring area where video data has been collected, determine whether there is a security anomaly based on the video data of the monitoring area, and if so, generate a security anomaly coefficient for the monitoring area.

[0062] In a preferred embodiment of the present invention, S40 includes the following steps: S41, performing image recognition on the video data of the monitoring area to determine whether there is a preset violation scene in the monitoring area.

[0063] Specifically, video data is collected by cameras installed in the monitoring area, with the cameras' field of view covering the entire monitoring area. The collected video data is input frame by frame into an image recognition model. This pre-trained model can identify various pre-defined violation scenarios. These pre-defined violation scenarios include personnel entering prohibited areas, personnel lingering in dangerous areas for more than a preset time, work vehicles illegally entering the monitoring area, equipment being activated when safety interlock conditions are not met, personnel entering the equipment's movement trajectory range during operation, and the number of personnel in the monitoring area exceeding a preset threshold. The image recognition model analyzes each frame in real time. When the recognition result matches any of the above violation scenarios, it determines that a pre-defined violation scenario exists within the monitoring area.

[0064] In a preferred embodiment of the present invention, S40 further includes the following steps: S42, when a violation scene is identified, it is determined that there is a security anomaly in the monitoring area, and a security anomaly coefficient of the monitoring area is generated according to the risk level of the violation scene.

[0065] Specifically, different violation scenarios pose varying degrees of danger to personnel and equipment. Therefore, these scenarios are pre-classified into different risk levels based on their degree of danger. The risk levels include at least three levels: Level 3, Level 2, and Level 1. Level 3 represents minor violations, such as personnel briefly entering a non-dangerous area and then quickly leaving; such violations have a low probability of causing injury to personnel and equipment. Level 2 represents general violations, such as work vehicles illegally driving in unoccupied areas, posing a certain risk of equipment collision. Level 1 represents serious violations, such as personnel entering the movement trajectory of operating parts of equipment or equipment being activated when safety interlock conditions are not met; such violations directly threaten personnel safety or may cause serious damage to equipment.

[0066] Once the risk level is determined, a safety anomaly coefficient for the monitored area is generated based on a preset coefficient value corresponding to that risk level. For example, a level 3 violation scenario has a safety anomaly coefficient of 0.1. A level 2 violation scenario has a safety anomaly coefficient of 0.5. A level 1 violation scenario has a safety anomaly coefficient of 1.0. The value of the safety anomaly coefficient corresponds one-to-one with the risk level of the violation scenario; the higher the risk level, the larger the safety anomaly coefficient value.

[0067] The safety anomaly coefficient differs from the temperature anomaly coefficient and vibration anomaly coefficient in its generation logic. The temperature anomaly coefficient is calculated continuously based on the magnitude of the temperature difference; the larger the temperature difference, the higher the temperature anomaly coefficient. The vibration anomaly coefficient is determined by the equipment type and fault type. The safety anomaly coefficient is determined by the risk level of the violation scenario.

[0068] In addition, for Level 1 violation scenarios, besides generating a safety anomaly coefficient, it is also necessary to coordinate and execute an emergency shutdown operation to stop the operation of the equipment in the monitored area in order to ensure the safety of on-site personnel and equipment.

[0069] Please see Figure 1 In a preferred embodiment of the present invention, the status monitoring method further includes the following steps: S50, calculating comprehensive abnormal data based on various abnormal coefficients generated in each monitoring area, and executing graded alarms based on the comprehensive abnormal data.

[0070] In a preferred embodiment of the present invention, S50 includes the following steps: S51, for each monitoring area, if there are multiple anomalous coefficients of the same type, the one with the largest value is retained as the anomalous coefficient of that type for the monitoring area.

[0071] Specifically, a monitoring area may employ multiple monitoring methods simultaneously. For example, it might be equipped with both temperature sensors and infrared / video fusion for temperature measurement. In this case, the monitoring area might output two sets of temperature data at the same time point, corresponding to the temperature values ​​under each of the two measurement methods. After temperature anomaly detection, these two sets of temperature data may generate two temperature anomaly coefficients. Therefore, the monitoring area may have multiple similar anomaly coefficients at the same time point, i.e., multiple temperature anomaly coefficients. Similarly, if a monitoring area is equipped with multiple temperature sensors, it will also output multiple sets of temperature data, generating multiple temperature anomaly coefficients after anomaly detection.

[0072] In such cases, the largest value should be selected and retained as the anomaly coefficient for that type of monitoring area. For example, if a monitoring area has two temperature anomaly coefficients, 0.4 and 0.7 respectively, then 0.7 should be retained as the temperature anomaly coefficient for that monitoring area. The advantage of this approach is that when a monitoring area has multiple anomaly coefficients of the same type, the largest coefficient represents the most severe anomaly in that monitoring dimension, making it more reasonable to use the most severe anomaly in subsequent calculations. If an average value is taken, the severity of the anomaly in the monitoring area will be underestimated. If the values ​​are accumulated, multiple anomalies in the same monitoring area will be unreasonably amplified, leading to an overestimation of the overall anomaly data.

[0073] The merging of anomaly coefficients of the same type only occurs within the same monitoring area. Anomaly coefficients of the same type between different monitoring areas are not merged and are retained separately. After the above merging process, each monitoring area retains at most one anomaly coefficient for each anomaly type.

[0074] In a preferred embodiment of the present invention, S50 further includes the following step: S52, calculating the hazard component corresponding to each type of abnormal coefficient based on the similar abnormal coefficients retained in all monitoring areas.

[0075] After all monitoring areas have been processed, the temperature anomaly coefficient, vibration anomaly coefficient, and safety anomaly coefficient need to be summarized and the hazard component of each of the three types of anomaly coefficients needs to be calculated.

[0076] The hazard component of the temperature dimension is calculated as follows: The temperature anomaly coefficients of all monitored areas exhibiting temperature anomalies are summed, and this sum is divided by the total number of monitored areas with temperature anomalies to obtain the average value of the temperature anomaly coefficients. This average value is the hazard component of the temperature dimension. For example, if there are four monitored areas with temperature anomalies, and their temperature anomaly coefficients are 0.4, 0.5, 0.6, and 0.7 respectively, then the hazard component of the temperature dimension is the average of these four values, which is 0.55.

[0077] Similarly, the vibration anomaly coefficients of all monitored areas with vibration anomaly coefficients are summed and divided by the total number of monitored areas with vibration anomaly coefficients to obtain the average vibration anomaly coefficient. This average value is the hazard component of the vibration dimension. Similarly, the safety anomaly coefficients of all monitored areas with safety anomaly coefficients are summed and divided by the total number of monitored areas with safety anomaly coefficients to obtain the average safety anomaly coefficient. This average value is the hazard component of the safety dimension.

[0078] In a preferred embodiment of the present invention, S50 further includes the following step: S53, calculating comprehensive abnormal data based on various hazard components and their corresponding weighting coefficients, and executing a graded alarm based on the comprehensive abnormal data. The formula for calculating the comprehensive abnormal data is as follows:

[0079] in, To integrate abnormal data, , , All are weighting coefficients, and , This represents the total number of monitoring areas with temperature anomalies. For the first Temperature anomaly coefficient of each monitoring area This represents the total number of monitoring areas with vibration anomaly coefficients. For the first Vibration anomaly coefficient of each monitoring area This represents the total number of monitoring areas with anomaly coefficients. For the first The safety anomaly coefficient of each monitoring area.

[0080] Temperature, vibration, and safety are each assigned a weighting coefficient, all of which are preset values ​​between 0 and 1, and the sum of the three weighting coefficients equals 1. The values ​​of the weighting coefficients are predetermined based on factors such as equipment type, operating environment, and historical fault statistics. For example, for a piece of machinery where temperature-related faults are the primary risk factor, the temperature weighting coefficient can be set higher. For a piece of machinery where mechanical vibration is the primary risk factor, the vibration weighting coefficient can be set higher.

[0081] The comprehensive abnormal data is calculated by multiplying the hazard component of the temperature dimension by the temperature weighting coefficient, the hazard component of the vibration dimension by the vibration weighting coefficient, and the hazard component of the safety dimension by the safety weighting coefficient, and then adding the three products together to obtain the comprehensive abnormal data.

[0082] After the comprehensive anomaly data calculation is completed, a tiered alarm is executed based on the data values. Once the tiered alarm is executed, the status monitoring process for this monitoring cycle ends, and the next cycle begins again from data acquisition. The tiered alarm includes at least three levels.

[0083] For example, when the overall abnormal data is greater than or equal to 0 and less than 0.3, it is judged as a Level 3 anomaly. A Level 3 anomaly is a minor anomaly, where equipment parameters deviate slightly from the normal range, posing no safety risk. In this case, the anomaly event is automatically recorded and written to the log, and a pop-up notification window appears in the monitoring backend to remind maintenance personnel to pay attention. This anomaly data is included in long-term trend analysis for subsequent tracking of equipment status evolution.

[0084] For example, when the overall abnormal data is greater than or equal to 0.3 and less than 0.7, it is judged as a level 2 abnormality. Level 2 abnormalities are general faults, where parameters such as temperature and vibration of the equipment have shown obvious abnormalities, but the equipment can still operate safely for a short period of time and will not immediately cause an accident. At this time, an audible and visual alarm is triggered to remind on-site personnel to pay attention to the equipment status. At the same time, a maintenance work order is automatically generated and pushed to the operation and maintenance personnel to arrange planned maintenance.

[0085] For example, when the overall abnormal data is greater than or equal to 0.7 and less than or equal to 1.0, it is judged as a Level 1 abnormality. Level 1 abnormalities are high-risk faults, and the corresponding actual operating conditions include electrical equipment overheating, excessive structural stress, severe abnormal noise or vibration of equipment, and unauthorized entry by personnel or vehicles. At this time, the highest level of audible and visual alarm is triggered, and the emergency shutdown operation of the equipment is executed simultaneously to immediately stop the equipment operation and prevent the fault from escalating or causing a safety accident. In addition, automatic cooling of equipment overheating and fire prediction and fire-fighting linkage in the electrical room are initiated simultaneously.

[0086] As can be seen, the above solution replaces traditional manual inspections with comprehensive monitoring of temperature, vibration, and video data, solving the problems of limited inspection frequency and numerous blind spots. It effectively captures latent faults such as early wear and electrical anomalies. Working and non-working time periods are divided according to equipment start-up and shutdown patterns, and temperature benchmarks are established for each, avoiding false alarms caused by equipment state switching. The temperature anomaly coefficient continuously and quantitatively reflects the temperature deterioration trend. The vibration anomaly coefficient is determined by equipment type and fault type, solving the problem of continuous quantitative output in vibration analysis. Unauthorized entry and operation in the work area are assessed through image recognition, and safety anomaly coefficients are set according to risk levels, achieving integrated monitoring of work safety and equipment status. In the calculation of comprehensive anomaly data, the maximum value of the same type of anomaly coefficient in the same monitoring area is merged, and the average anomaly coefficient of each dimension is used as the hazard component, solving the problem of mutual interference of anomaly signals when fusing multi-source data. Based on the comprehensive anomaly data, graded alarms are executed, and high-risk faults trigger emergency shutdowns, realizing a complete closed loop from status monitoring, graded early warning to coordinated response, reducing equipment failure rate and maintenance costs, and ensuring the safe and stable operation of equipment.

[0087] Please see Figure 2 The present invention also discloses a condition monitoring device for mechanical equipment. The above-mentioned condition monitoring method can be applied to the condition monitoring device. The condition monitoring device may include a data acquisition module 100, a temperature judgment module 200, a vibration judgment module 300, a safety judgment module 400, and a comprehensive judgment module 500.

[0088] The data acquisition module 100 is used to collect monitoring data from various preset monitoring areas of the mechanical equipment in real time; the monitoring data includes at least one of temperature data, vibration data, and video data.

[0089] The temperature judgment module 200 is used to determine whether there is a temperature anomaly for each monitoring area where temperature data has been collected, and if so, to generate a temperature anomaly coefficient for that monitoring area.

[0090] The vibration judgment module 300 is used to determine whether there is a vibration anomaly for each monitoring area where vibration data has been collected, based on the vibration data of that monitoring area. If there is, it generates a vibration anomaly coefficient for that monitoring area.

[0091] The security judgment module 400 is used to determine whether there is a security anomaly for each monitoring area where video data has been collected, based on the video data of that monitoring area. If there is, it generates a security anomaly coefficient for that monitoring area.

[0092] The comprehensive judgment module 500 is used to calculate comprehensive abnormal data based on various abnormal coefficients generated in each monitoring area, and to execute graded alarms based on the comprehensive abnormal data.

[0093] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A method for condition monitoring of mechanical equipment, characterized in that, include: Real-time acquisition of monitoring data from various preset monitoring areas of mechanical equipment; wherein, the monitoring data includes at least one of temperature data, vibration data, and video data; For each monitoring area where the temperature data is collected, it is determined whether there is a temperature anomaly based on the temperature data of that monitoring area. If there is, a temperature anomaly coefficient for that monitoring area is generated. For each monitoring area where the vibration data is collected, it is determined whether there is a vibration anomaly based on the vibration data of that monitoring area. If so, a vibration anomaly coefficient for that monitoring area is generated. For each monitoring area where the video data is collected, determine whether there is a security anomaly based on the video data of that monitoring area. If there is, generate a security anomaly coefficient for that monitoring area. Based on the various anomaly coefficients generated in each monitoring area, comprehensive anomaly data is calculated, and graded alarms are executed based on the comprehensive anomaly data.

2. The method for monitoring the condition of mechanical equipment according to claim 1, characterized in that, The temperature data is collected by a temperature sensor installed in the monitoring area, and / or the temperature data is obtained by processing a visible light image and an infrared thermal image containing the monitoring area. The processing of the visible light image and infrared thermal image containing the monitoring area includes: Simultaneously, a visible light image and an infrared thermal image containing the monitoring area are acquired, and the visible light image and the infrared thermal image are aligned in position. Identify and segment the pixel mask region corresponding to the monitoring region in the visible light image; The pixel mask area is mapped onto the infrared thermal image, and the temperature value of each pixel within the pixel mask area is extracted. Sort all extracted temperature values ​​from largest to smallest, select the first preset number of temperature values, and use the median of the preset number of temperature values ​​as the temperature data for the monitoring area.

3. The method for monitoring the condition of mechanical equipment according to claim 1, characterized in that, The step of determining whether there is a temperature anomaly based on the temperature data of the monitored area, and generating a temperature anomaly coefficient for the monitored area if an anomaly exists, includes: Based on the time node of the collected temperature data, determine the time period to which the time node belongs, and obtain the historical average temperature and temperature fluctuation threshold corresponding to the time period. Calculate the temperature difference between the temperature data and the historical average temperature; When the temperature difference is greater than the temperature fluctuation threshold, it is determined that there is a temperature anomaly in the monitoring area, and a temperature anomaly coefficient for the monitoring area is generated based on the temperature difference.

4. The method for monitoring the condition of mechanical equipment according to claim 3, characterized in that, The historical average temperature and temperature fluctuation threshold corresponding to the time period are predetermined in the following ways: Based on the start-up and shutdown time nodes of the equipment in the monitoring area, the time is divided into at least one working time period and at least one non-working time period; For each time period: Acquire multiple sets of historical temperature data for the equipment in the monitoring area under normal conditions; The historical average temperature of the monitored area during this time period is determined based on the average of all the historical temperature data. Calculate the standard deviation of temperature fluctuation based on the dispersion of all the historical temperature data relative to the historical average temperature; A temperature fluctuation threshold is set based on the standard deviation of the temperature fluctuation.

5. The method for monitoring the condition of mechanical equipment according to claim 4, characterized in that, The formula for calculating the temperature anomaly coefficient is as follows: in, This is the temperature anomaly coefficient. This is the temperature difference value. This represents the standard deviation of temperature fluctuation.

6. The method for monitoring the condition of mechanical equipment according to claim 1, characterized in that, The step of determining whether there is a vibration anomaly based on the vibration data of the monitoring area, and generating a vibration anomaly coefficient for the monitoring area if an anomaly is found, includes: Convert the vibration data of the monitored area into frequency domain data; Extract the amplitude and frequency of at least one main frequency band from the frequency domain data; The amplitude is compared with a preset amplitude threshold, and / or the frequency is compared with a preset frequency range; When the amplitude is greater than the amplitude threshold and / or the frequency is not in the frequency range, it is determined that there is a vibration anomaly in the monitoring area, and a corresponding vibration anomaly coefficient is generated according to the type of equipment to which the monitoring area belongs.

7. The method for monitoring the condition of mechanical equipment according to claim 1, characterized in that, The step of determining whether there is a security anomaly based on the video data of the monitored area, and generating a security anomaly coefficient for the monitored area if it exists, includes: Image recognition is performed on the video data of the monitored area to determine whether there are any pre-set violation scenarios within the monitored area; When the violation scenario is identified, it is determined that there is a security anomaly in the monitoring area, and a security anomaly coefficient for the monitoring area is generated based on the risk level of the violation scenario.

8. The method for monitoring the condition of mechanical equipment according to claim 1, characterized in that, The calculation of comprehensive abnormal data based on various abnormal coefficients generated in each monitoring area includes: For each monitoring area, if there are multiple anomalous coefficients of the same type, the one with the largest value is retained as the anomalous coefficient of that type for that monitoring area. Based on the similar anomaly coefficients retained in all monitoring areas, calculate the hazard components corresponding to each type of anomaly coefficient; Based on the various hazard components and their corresponding weighting coefficients, comprehensive abnormal data is calculated.

9. The method for monitoring the condition of mechanical equipment according to claim 8, characterized in that, The formula for calculating the comprehensive anomaly data is as follows: in, To integrate abnormal data, , , All are weighting coefficients, and , This represents the total number of monitoring areas with temperature anomalies. For the first Temperature anomaly coefficient of each monitoring area This represents the total number of monitoring areas with vibration anomaly coefficients. For the first Vibration anomaly coefficient of each monitoring area This represents the total number of monitoring areas with anomaly coefficients. For the first The safety anomaly coefficient of each monitoring area.

10. A condition monitoring device for mechanical equipment, characterized in that, include: The data acquisition module is used to collect monitoring data from various preset monitoring areas of the mechanical equipment in real time; wherein, the monitoring data includes at least one of temperature data, vibration data, and video data; The temperature judgment module is used to determine whether there is a temperature anomaly for each monitoring area where the temperature data has been collected, and if so, to generate a temperature anomaly coefficient for that monitoring area. The vibration judgment module is used to determine whether there is a vibration anomaly for each monitoring area where the vibration data has been collected, and if so, to generate a vibration anomaly coefficient for that monitoring area. The security judgment module is used to determine whether there is a security anomaly for each monitoring area where the video data has been collected, based on the video data of that monitoring area. If there is, it generates a security anomaly coefficient for that monitoring area. The comprehensive judgment module is used to calculate comprehensive abnormal data based on various abnormal coefficients generated in each monitoring area, and to execute graded alarms based on the comprehensive abnormal data.

Images