An IoT-based valve condition monitoring system

Abstract

This invention discloses a valve status monitoring system based on the Internet of Things (IoT). The invention relates to the field of monitoring technology. The valve status monitoring system includes an information acquisition module, an interception condition module, a sensing module, a status extraction module, and a management and warning module. The advantages of this invention are: the interception condition module optimizes storage space while ensuring event integrity, enabling the system to operate efficiently even in resource-constrained environments; the sensing module effectively suppresses signal aliasing, ensuring high-fidelity extraction of vibration characteristics; the alternating arrangement of thermoelectric units achieves high-density monitoring of the spatial distribution of the temperature field, avoiding missed detection of local hotspots; the status extraction module ensures that the intercepted video contains key status information, and the addition of a connection threshold ensures the integrity and continuity of the video, facilitating the management of valve status information; and the system achieves an innovative closed loop of flexible sensing, intelligent interception, multi-source fusion, and precise early warning.

Description

Technical Field

[0001] This invention relates to the field of monitoring technology, specifically to a valve status monitoring system based on the Internet of Things. Background Technology

[0002] Valves are control components in pipeline fluid transport systems, used to change the cross-sectional area of ​​the passage and the direction of medium flow. They have functions such as guiding, stopping, throttling, checking back, diverting, or overflowing and relieving pressure. Valves used for fluid control range from the simplest shut-off valves to various valves used in extremely complex automatic control systems. There are many types and specifications of valves. The nominal diameter of valves ranges from the smallest instrument valves to industrial pipeline valves with a diameter of up to 10m. They can be used to control the flow of various types of fluids such as water, steam, oil, gas, mud, various corrosive media, liquid metals, and radioactive fluids.

[0003] Existing valve condition monitoring systems suffer from signal aliasing, resulting in low fidelity of vibration feature extraction. Furthermore, insufficient thermoelectric unit density makes it easy to miss local hot spots. The unreasonable arrangement of thermoelectric units also prevents local hot spots from being detected in a timely manner, thus failing to effectively reflect the valve condition and hindering valve condition management. To address this, we propose an IoT-based valve condition monitoring system. Summary of the Invention

[0004] The purpose of this invention is to provide a valve status monitoring system based on the Internet of Things.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a valve status monitoring system based on the Internet of Things, the valve status monitoring system comprising;

[0006] The information acquisition module is used to collect the valve model, obtain all valve information from the Internet of Things, monitor the valve's status changes during operation, obtain valve monitoring video, and analyze the valve's status changes under abnormal conditions to obtain abnormal change information.

[0007] The interception condition module extracts the valve's status during operation and obtains the interception condition information. Then, it calculates the difference in status changes, sorts the values, and obtains a change table. Finally, it extracts Y interception condition information from the change table from top to bottom to obtain the monitoring interception information.

[0008] The sensing module acquires an X-shaped flexible substrate and alternately arranges piezoelectric and thermoelectric units on it, with the spacing between the piezoelectric units satisfying 0.5D, where D is the wavelength corresponding to the main vibration frequency of the valve. This results in a flexible sensor array, which is then mounted to the curved surface of the valve to obtain the valve's sensing state information. Simultaneously, a micro-MEMS microphone array is integrated on the surface of the flexible sensor array to construct a vibration-acoustic dual-modal sensing network. The vibration information and sound are registered using a wavelet packet decomposition algorithm to establish a valve motion feature fingerprint database. A transfer learning algorithm is then used to perform domain adaptation matching, with normal operating condition data as the source domain and real-time data as the target domain.

[0009] The status extraction module extracts monitoring intercept information and valve monitoring video. It then analyzes the start and end times of the valve monitoring video corresponding to the monitoring intercept information, and analyzes the valve monitoring that needs to be intercepted based on the start and end times to obtain the original intercepted video. The module then analyzes the video duration of the original intercepted video to obtain the original trimming time. At the same time, a connection threshold is established, and the duration of the original trimming time is increased using the connection threshold. Based on the original trimming time and the connection threshold, the actual trimming time in the original intercepted video is calculated, thereby obtaining the status sub-video.

[0010] As a further embodiment of the present invention: the valve status monitoring system also includes a management and warning module;

[0011] The management alert module extracts the interception condition information and creates management folders according to the interception condition information. Then, it records the status sub-videos into the corresponding management folders, calculates the status change index of the sensed status information, establishes a basic interval time for the status change index, and establishes an early warning threshold. When the status change index exceeds the early warning threshold, it sends an early warning signal to the user.

[0012] As a further aspect of the present invention: after obtaining abnormal change information in the information acquisition module, a 2 R table, then record all abnormal change names and abnormal change values ​​in the abnormal change information into 2. In table R, an abnormal change table is obtained. Then, the valve's state changes during operation are recorded according to the format of the abnormal change table to obtain a worksheet. Here, R represents the number of different abnormal change names and the number of different abnormal change values. After obtaining the worksheet, the information acquisition module extracts the abnormal change values ​​from the abnormal change table and the corresponding values ​​from the worksheet, calculating the abnormality index of both the abnormal change table and the worksheet. Let the abnormal change values ​​in the abnormal change table be... Let the corresponding values ​​in the worksheet be... Let the abnormal change value be :

[0013] ;

[0014] The abnormal change values ​​are calculated using the above formula, and a work anomaly table is created simultaneously. When this happens, the corresponding value in the worksheet is considered an abnormal value, and the corresponding value and name in the worksheet are recorded in the worksheet's abnormality table. When this happens, the corresponding value in the worksheet is considered a non-abnormal value.

[0015] As a further aspect of the present invention: when calculating the state change difference in the interception condition module, the state values ​​of the valve at different time points are analyzed to establish a difference period, where the difference period refers to the interval time of the valve state. This ensures that the state change difference is calculated according to the difference period to obtain the valve state at different times. Let the difference period be P, and let the valve state value at time X be... Let the valve state value at time X+1 be... Let the difference in state change be . :

[0016] ;

[0017] The difference in state change is calculated using the formula above.

[0018] As a further aspect of the present invention: when the interception condition information in the interception condition module is extracted, an editing unit is simultaneously established. The user has the right to edit the number of interception condition information extracted through the editing unit, so that the number of extracted interception condition information better meets the user's needs.

[0019] As a further aspect of the present invention: the connection threshold in the state extraction module includes proportional increase and time increase. Proportional increase involves the user setting a value to obtain a proportional index, then calculating the product of the original cropping time and the proportional index, and increasing the cropping time according to the value of the product to obtain cropping time number one. Time increase involves the user directly inputting the time to be added to obtain cropping time number two. The user selects between cropping time number one and cropping time number two to obtain the actual cropping time. Proportional increase is suitable for scenarios that need to dynamically adapt to different video durations, while time increase is suitable for scenarios where the user explicitly needs a fixed delay.

[0020] As a further aspect of the present invention: when calculating the actual cutting time in the state extraction module, the original cutting time is set to be... Let the proportionality index be... Let the user directly input the additional time required. Let the cutting time for the first cut be... Let the cutting time for the second cut be... :

[0021] ;

[0022] ;

[0023] The first cutting time and the second cutting time can be calculated using the two formulas mentioned above.

[0024] As a further aspect of the present invention: after obtaining the actual cropping time in the state extraction module, the actual cropping time is added to the starting time of the original cropped video, thereby obtaining the actual cropping starting time. Let the starting time of the original cropped video be... Let the actual cutting time be... Let the actual cutting start time be... The units for the original video capture start time, the actual cropping time, and the actual cropping start time are all in seconds.

[0025] ;

[0026] The actual cutting start time can be calculated using the above formula.

[0027] As a further aspect of the present invention: when calculating the status change index in the management alert module, the sensing status information is extracted based on the base interval time. Then, a line graph is plotted based on the extracted sensing status information and the base interval time. The values ​​of the sensing status information in the line graph are extracted and numbered sequentially. Let the value of the Xth sensing status information be... Let the value of the sensing state information at time X+1 be... Let the state change index be... :

[0028] ;

[0029] The state change index is calculated using the formula above.

[0030] Compared with the prior art, the beneficial effects of the present invention by adopting the above technical solution are as follows:

[0031] 1. This invention optimizes storage space while ensuring event integrity through the interception condition module, enabling the system to operate efficiently even in resource-constrained environments. The sensing module effectively suppresses signal aliasing, ensuring high-fidelity extraction of vibration characteristics. The alternating arrangement of thermoelectric units enables high-density monitoring of the spatial distribution of the temperature field, avoiding missed detection of local hotspots. The state extraction module ensures that the intercepted video contains key state information, and the addition of a connection threshold ensures the integrity and continuity of the video, thus facilitating the management of valve state information. Through the innovative closed loop of flexible sensing, intelligent interception, multi-source fusion, and accurate early warning, this invention solves the pain points of traditional valve monitoring, such as data isolation, high false alarm rate, and poor installation adaptability.

[0032] 2. This invention improves the efficiency of data management through the information acquisition module and provides a unified data format for subsequent anomaly detection, reducing the complexity of data processing, reducing the risk of human error, helping to detect abnormalities in valve operation in a timely manner, and improving the accuracy and real-time performance of the monitoring system. The interception condition module ensures that there is a unified standard when acquiring different valve states at different times, improving the accuracy of filtering key information from the valve's operating state and enhancing the system's personalization and adaptability.

[0033] 3. This invention balances flexibility and ease of use through its status extraction module, ensuring the integrity of key video clips, avoiding the omission of important status information, facilitating accurate calculation and reasonable selection by users according to actual needs, and guaranteeing the scientific and accurate nature of the calculation process. The management and warning module optimizes the utilization of storage resources, avoids the accumulation of redundant data, and intuitively reflects the changing trend of sensing status information over time, enabling timely detection of abnormal fluctuations in valve status. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the system flow in an embodiment of the present invention. Detailed Implementation

[0035] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0036] Furthermore, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0037] Please see the appendix Figure 1 The present invention provides a valve status monitoring system based on the Internet of Things, the valve status monitoring system comprising:

[0038] The information acquisition module is used to collect the valve model, obtain all valve information from the Internet of Things, monitor the valve's status changes during operation, obtain valve monitoring video, and analyze the valve's status changes under abnormal conditions to obtain abnormal change information.

[0039] The interception condition module extracts the valve's status during operation and obtains the interception condition information. Then, it calculates the difference in status changes, sorts the values, and obtains a change table. Finally, it extracts Y interception condition information from the change table from top to bottom to obtain the monitoring interception information.

[0040] The sensing module acquires an X-shaped flexible substrate and alternately arranges piezoelectric and thermoelectric units on the substrate, with the spacing between the piezoelectric units satisfying 0.5D, where D is the wavelength corresponding to the main vibration frequency of the valve. A flexible sensor array is then obtained and mounted on the curved surface of the valve to obtain the valve's sensing state information. Simultaneously, a micro-MEMS microphone array is integrated on the surface of the flexible sensor array to construct a vibration-acoustic dual-modal sensing network. The vibration information and sound are registered using a wavelet packet decomposition algorithm to establish a valve motion feature fingerprint database. A transfer learning algorithm is used to use normal operating condition data as the source domain and real-time data as the target domain for domain adaptation matching.

[0041] Vibration information of the valve during operation is obtained by a spectrum analyzer, its dominant frequency component is extracted, and the wavelength is calculated according to the formula D=v / f, where v is the propagation speed of the vibration wave in the valve material and f is the dominant vibration frequency.

[0042] A miniature MEMS microphone array is an array obtained by arranging multiple miniature MEMS microphones onto an X-shaped flexible substrate according to the arrangement steps of piezoelectric and thermoelectric units.

[0043] The vibration-acoustic dual-mode sensor network needs to analyze the sound information corresponding to different vibrations of the valve through experiments, so as to construct a vibration-acoustic matching table. Then, when the valve emits different sounds, the corresponding vibration information can be found through the matching table.

[0044] The wavelet packet decomposition algorithm is based on the multi-resolution analysis idea of ​​wavelet transform, which decomposes the signal at different scales and frequencies. It not only decomposes the low-frequency part of the signal, but also further subdivides the high-frequency part, thereby more comprehensively describing the detailed information of the signal. The signal is decomposed into sub-signals of different frequency bands through a series of filter banks. Each sub-signal represents the characteristics of the original signal in a specific frequency range.

[0045] The valve motion feature fingerprint database refers to the various characteristic signals generated by a valve during its movement, such as vibration, sound, and current. These characteristic signals are closely related to the valve's working state and fault type. By collecting and analyzing these characteristic signals, characteristic parameters that can represent the valve's motion state are extracted and then stored in the fingerprint database. When it is necessary to monitor and diagnose the valve's state, it is only necessary to compare the real-time collected valve motion feature parameters with the data in the fingerprint database to determine whether the valve is in normal working condition and whether there are potential faults.

[0046] Transfer learning algorithms aim to transfer knowledge learned from one or more source domains to a target domain, thereby improving the learning effect in the target domain. In many real-world scenarios, the number of samples in the target domain is limited, making it difficult to train a high-performance model. Transfer learning, on the other hand, can leverage the rich data and knowledge in the source domain to enable the model to perform better in the target domain.

[0047] The source domain refers to a field or dataset that has abundant data and knowledge and has been thoroughly researched or studied.

[0048] The target domain is the specific domain or dataset that needs to be solved in transfer learning. It is usually related to the source domain but has certain differences. These differences may be reflected in data distribution, feature representation, task type, etc. For example, in equipment fault diagnosis, normal operating condition data is the source domain, and real-time data when the equipment has different faults is the target domain.

[0049] Domain adaptation matching is a key step in transfer learning. It aims to find a method or transformation that makes the data in the source domain and the target domain as close as possible in a certain feature space, thereby reducing the degradation of model performance caused by inter-domain differences.

[0050] The status extraction module extracts monitoring intercept information and valve monitoring video, then analyzes the start and end times corresponding to the monitoring intercept information in the valve monitoring video, and analyzes the valve monitoring that needs to be intercepted based on the start and end times to obtain the original intercepted video. Then, it analyzes the video length of the original intercepted video to obtain the original trimming time. At the same time, it establishes a connection threshold and uses the connection threshold to increase the duration of the original trimming time. Based on the original trimming time and the connection threshold, it calculates the actual trimming time in the original intercepted video, thereby obtaining the status sub-video.

[0051] The valve condition monitoring system also includes a management alert module;

[0052] The management alert module extracts the interception condition information and creates management folders according to the interception condition information. Then, it records the status sub-videos into the corresponding management folders, calculates the status change index of the sensed status information, establishes a basic interval time for the status change index, and establishes an early warning threshold. When the status change index exceeds the early warning threshold, it sends an early warning signal to the user.

[0053] In one embodiment of the present invention: after obtaining abnormal change information in the information acquisition module, a 2 R table, then record all abnormal change names and abnormal change values ​​in the abnormal change information into 2. In table R, an abnormal change table is obtained. Then, the valve's state changes during operation are recorded according to the format of the abnormal change table, resulting in a worksheet. Here, R represents the number of different abnormal change names and the number of different abnormal change values. After obtaining the worksheet, the information acquisition module extracts the abnormal change values ​​from the abnormal change table and the corresponding values ​​from the worksheet, calculating the abnormality index for both the abnormal change table and the worksheet. Let the abnormal change values ​​in the abnormal change table be... Let the corresponding values ​​in the worksheet be... Let the abnormal change value be :

[0054] ;

[0055] The abnormal change values ​​are calculated using the above formula, and a work anomaly table is created simultaneously. When this happens, the corresponding value in the worksheet is considered an abnormal value, and the corresponding value and name in the worksheet are recorded in the worksheet's abnormality table. When this happens, the corresponding value in the worksheet is considered a non-abnormal value.

[0056] In one embodiment of the present invention: when calculating the state change difference in the condition module, the state values ​​of the valve at different time points are analyzed to establish a difference period, where the difference period refers to the interval time of the valve state. This ensures that the state change difference is calculated according to the difference period to obtain the valve state at different times. Let the difference period be P, and let the valve state value at time X be... Let the valve state value at time X+1 be... Let the difference in state change be . :

[0057] ;

[0058] The difference in state change is calculated using the formula above.

[0059] In one embodiment of the present invention: when the interception condition information in the interception condition module is extracted, an editing unit is simultaneously established. The user has the right to edit the number of interception condition information extracted through the editing unit, so that the number of interception condition information extracted is more in line with the user's needs.

[0060] In one embodiment of the present invention: the connection threshold in the state extraction module includes proportional increase and time increase. Proportional increase is achieved by the user setting a value to obtain a proportional index, then calculating the product of the original cropping time and the proportional index, and increasing the cropping time according to the value of the product to obtain cropping time number one. Time increase is achieved by the user directly inputting the time to be increased to obtain cropping time number two. The user selects cropping time number one and cropping time number two to obtain the actual cropping time. Proportional increase is suitable for scenarios that need to dynamically adapt to different video durations, while time increase is suitable for scenarios where the user explicitly needs a fixed delay.

[0061] In one embodiment of the present invention: when calculating the actual trimming time in the state extraction module, the original trimming time is assumed to be... Let the proportionality index be... Let the user directly input the additional time required. Let the cutting time for No. 1 be... Let the cutting time for the second cut be... :

[0062] ;

[0063] ;

[0064] The first cutting time and the second cutting time can be calculated using the two formulas mentioned above.

[0065] In one embodiment of the present invention: after obtaining the actual cropping time in the state extraction module, the actual cropping time is added to the starting time of the original cropped video to obtain the actual cropping starting time. Let the starting time of the original cropped video be... Let the actual cutting time be... Let the actual cutting start time be... The units for the original video capture start time, the actual cropping time, and the actual cropping start time are all in seconds.

[0066] ;

[0067] The actual cutting start time can be calculated using the above formula.

[0068] In one embodiment of the present invention: when calculating the status change index in the management alert module, the sensed status information is extracted based on the base interval time. Then, a line graph is plotted based on the extracted sensed status information and the base interval time. The values ​​of the sensed status information in the line graph are extracted and numbered sequentially. Let the value of the Xth sensed status information be... Let the value of the sensing state information at time X+1 be... Let the state change index be... :

[0069] ;

[0070] The state change index is calculated using the formula above.

[0071] Example 1, please refer to the appendix. Figure 1 The system obtains all valve information, monitors valve status changes during operation, acquires interception condition information, calculates state change differences, and sorts them. Y interception condition information items are extracted sequentially from top to bottom from the change table. Piezoelectric and thermoelectric units are alternately arranged on an X-shaped flexible substrate, and the flexible sensor array is fitted to the valve's curved surface to obtain the valve's sensing state information. The system analyzes the start and end times of the monitoring interception information in the valve monitoring information, identifies the valve monitoring data that needs to be intercepted based on the start and end times, analyzes the original intercepted video duration, increases the original trimming time using a connection threshold, calculates the actual trimming time in the original intercepted video based on the original trimming time and connection threshold, establishes management folders according to the interception condition information, and records the status sub-videos into the corresponding management folders. A base interval time is established for the status change index. When the status change index exceeds the warning threshold, a warning signal is sent to the user.

[0072] Example 2, please refer to the appendix. Figure 1 Obtain all valve information, monitor valve status changes during operation, and establish 2 R table, then record all abnormal change names and abnormal change values ​​in the abnormal change information into 2. In table R, record the valve's status changes during operation according to the format of the abnormal change table. Extract the abnormal change values ​​from the abnormal change table and the corresponding values ​​from the worksheet. Calculate the abnormality index for both the abnormal change table and the worksheet. When this happens, the corresponding value in the worksheet is considered an abnormal value, and the corresponding value and name in the worksheet are recorded in the worksheet's abnormality table. When the corresponding value in the worksheet is considered non-abnormal, the interception condition information is obtained, the state change difference is calculated, and the values ​​are sorted. Y interception condition information are extracted from the change table from top to bottom. The state values ​​of the valve at different time points are analyzed, and a difference period is established. When calculating the state change difference, the valve's state at different times will be uniformly obtained according to the difference period. An editing unit is established simultaneously. Users have the right to edit the number of interception condition information extracted through the editing unit, so that the number of extracted interception condition information is more in line with user needs.

[0073] Example 3, please refer to the appendix. Figure 1 The transition threshold includes two time-increase methods: proportional increase and time increase. Proportional increase involves the user setting a value to obtain a proportional index, then calculating the product of the original cropping time and the proportional index, and increasing the cropping time according to the product value to obtain cropping time #1. Time increase involves the user directly inputting the desired increase time to obtain cropping time #2. The user then selects between cropping time #1 and cropping time #2, adding the actual cropping time to the original starting time of the video capture to obtain the actual cropping starting time. Sensing state information is extracted based on the base interval time. A line graph is then plotted based on the extracted sensing state information and the base interval time, and the values ​​of the sensing state information in the line graph are extracted and numbered sequentially.

[0074] Specifically, the X-shaped flexible substrate can be made of polyimide, polyester, polydimethylsiloxane, polyvinyl alcohol, and polyethylene naphthalate. All of these materials are lightweight, thin, flexible, and stretchable, and can be bent, folded, twisted, and stretched. This allows the resulting X-shaped flexible substrate to adapt to different shapes and working conditions, with excellent transparency, without affecting optical properties, insulation, corrosion resistance, ensuring stable operation of electronic devices, and resistance to external environmental erosion.

[0075] Specifically, 2 The R table can also be adjusted based on the more detailed classification of abnormal change information, making 2 The R table is converted to L during generation. The R table contains a number of L values ​​that vary depending on the level of detail in the anomaly information.

[0076] Specifically, based on different operating conditions such as valve startup and shutdown, the sensitivity of the state change difference and the amount of intercepted condition information are dynamically adjusted. For example, during the valve startup and shutdown phases, the sensitivity of the state change difference is increased, and the amount of intercepted condition information is increased to capture key changes. Using video analysis technology, key scenes in the original intercepted video are automatically identified, such as valve opening and closing actions and abnormal deformation of components. Combined with the connection threshold, the actual trimming time and state sub-video content are intelligently determined. It supports multi-video source fusion editing, integrating valve monitoring videos with videos from other related equipment to generate more comprehensive state sub-videos.

[0077] Working principle:

[0078] First, obtain all valve information, monitor valve status changes during operation, and establish a 2 R table, then record all abnormal change names and abnormal change values ​​in the abnormal change information into 2. In table R, record the valve's status changes during operation according to the format of the abnormal change table. Extract the abnormal change values ​​from the abnormal change table and the corresponding values ​​from the worksheet. Calculate the abnormality index for both the abnormal change table and the worksheet. When this happens, the corresponding value in the worksheet is considered an abnormal value, and the corresponding value and name in the worksheet are recorded in the worksheet's abnormality table. At that time, the corresponding values ​​in the worksheet are considered non-abnormal values. The interception condition information is obtained, the state change difference is calculated, and sorted. Y interception condition information are extracted sequentially from top to bottom from the change table. The valve's state values ​​at different time points are analyzed, and a difference period is established. This ensures that the state change difference is calculated according to the difference period, allowing the valve's state at different times to be obtained uniformly. An editing unit is simultaneously established, giving users the authority to edit the number of interception condition information extracted, making the extracted interception condition information more aligned with user needs. Piezoelectric and thermoelectric units are alternately arranged on an X-shaped flexible substrate, and the flexible sensor array is mounted against the valve's curved surface to obtain the valve's sensing state information. The start and end times of the monitoring interception information in the valve monitoring information are analyzed, and the valve monitoring data to be intercepted based on the start and end times is analyzed. The video duration of the original intercepted video is analyzed, and the duration of the original trimming time is increased using a joining threshold. Based on the original trimming time and joining threshold, the duration of the original intercepted video is calculated. The actual video trimming time, including the transition threshold, involves both proportional and time increases. Proportional increases are achieved by the user setting a value to obtain a proportional index, then calculating the product of the original trimming time and the proportional index, and increasing the trimming time accordingly to obtain trimming time #1. Time increases are achieved by the user directly inputting the desired increase time to obtain trimming time #2. The user then selects between trimming time #1 and trimming time #2, adding the actual trimming time to the original starting time of the video capture, thus obtaining the actual trimming starting time. A management folder is created based on the capture condition information, and status sub-videos are recorded in the corresponding folders. A base interval time is established for the status change index. When the status change index exceeds the warning threshold, a warning signal is sent to the user. Sensing status information is extracted using the base interval time as a standard. A line graph is then plotted based on the extracted sensing status information and the base interval time, and the values ​​of the sensing status information in the line graph are extracted and numbered sequentially. This completes the entire workflow.

[0079] While the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the invention. Any variations and modifications can be made by those skilled in the art without departing from the spirit and scope of the invention. Therefore, any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention, without departing from the scope of the invention, fall within the protection scope defined by the claims of the present invention.

Claims

1. A valve status monitoring system based on the Internet of Things, characterized in that: The valve status monitoring system includes: The information acquisition module is used to collect the valve model, obtain all valve information from the Internet of Things, monitor the valve's status changes during operation, obtain valve monitoring video, and analyze the valve's status changes under abnormal conditions to obtain abnormal change information. The interception condition module extracts the valve's status during operation and obtains the interception condition information. Then, it calculates the difference in status changes, sorts the values, and obtains a change table. Finally, it extracts Y interception condition information from the change table from top to bottom to obtain the monitoring interception information. The sensing module acquires an X-shaped flexible substrate and alternately arranges piezoelectric and thermoelectric units on the substrate, with the spacing between the piezoelectric units satisfying 0.5D, where D is the wavelength corresponding to the main vibration frequency of the valve. A flexible sensor array is then obtained and mounted on the curved surface of the valve to obtain the valve's sensing state information. Simultaneously, a micro-MEMS microphone array is integrated on the surface of the flexible sensor array to construct a vibration-acoustic dual-modal sensing network. The vibration information and sound are registered using a wavelet packet decomposition algorithm to establish a valve motion feature fingerprint database. A transfer learning algorithm is used to perform domain adaptation matching with normal operating condition data as the source domain and real-time data as the target domain. The status extraction module extracts monitoring intercept information and valve monitoring video. It then analyzes the start and end times of the valve monitoring video corresponding to the monitoring intercept information, and analyzes the valve monitoring that needs to be intercepted based on the start and end times to obtain the original intercepted video. The module then analyzes the video duration of the original intercepted video to obtain the original trimming time. At the same time, a connection threshold is established, and the duration of the original trimming time is increased using the connection threshold. Based on the original trimming time and the connection threshold, the actual trimming time in the original intercepted video is calculated, thereby obtaining the status sub-video.

2. The valve status monitoring system based on the Internet of Things according to claim 1, characterized in that: The valve status monitoring system also includes a management and alert module; The management alert module extracts the interception condition information and creates management folders according to the interception condition information. Then, it records the status sub-videos into the corresponding management folders, calculates the status change index of the sensed status information, establishes a basic interval time for the status change index, and establishes an early warning threshold. When the status change index exceeds the early warning threshold, it sends an early warning signal to the user.

3. The valve status monitoring system based on the Internet of Things according to claim 1, characterized in that: After obtaining abnormal change information, the information acquisition module establishes 2 R table, then record all abnormal change names and abnormal change values ​​in the abnormal change information into 2. In table R, an abnormal change table is obtained. Then, the valve's state changes during operation are recorded according to the format of the abnormal change table to obtain a worksheet. Here, R represents the number of different abnormal change names and the number of different abnormal change values. After obtaining the worksheet, the information acquisition module extracts the abnormal change values ​​from the abnormal change table and the corresponding values ​​from the worksheet, calculating the abnormality index of both the abnormal change table and the worksheet. Let the abnormal change values ​​in the abnormal change table be... Let the corresponding values ​​in the worksheet be... Let the abnormal change value be : ； The abnormal change values ​​are calculated using the above formula, and a work anomaly table is created simultaneously. When this happens, the corresponding value in the worksheet is considered an abnormal value, and the corresponding value and name in the worksheet are recorded in the worksheet's abnormality table. When this happens, the corresponding value in the worksheet is considered a non-abnormal value.

4. The valve status monitoring system based on the Internet of Things according to claim 1, characterized in that: The state change difference in the interception condition module analyzes the valve's state values ​​at different time points during calculation, establishing a difference period. The difference period refers to the interval between valve states. This ensures that the state change difference calculation consistently uses the difference period to obtain the valve's state at different times. Let the difference period be P, and let the valve state value at time X be... Let the valve state value at time X+1 be... Let the difference in state change be . : ； The difference in state change is calculated using the formula above.

5. A valve status monitoring system based on the Internet of Things according to claim 4, characterized in that: When the interception condition information in the interception condition module is extracted, an editing unit is created simultaneously. Users have the right to edit the number of interception condition information extracted through the editing unit, so that the number of interception condition information extracted is more in line with user needs.

6. The valve status monitoring system based on the Internet of Things according to claim 1, characterized in that: The connection thresholds in the state extraction module include proportional increase and time increase. Proportional increase involves the user setting a value to obtain a proportional index, then calculating the product of the original cropping time and the proportional index, and increasing the cropping time according to the product value to obtain cropping time number one. Time increase involves the user directly inputting the time to be added to obtain cropping time number two. The user selects between cropping time number one and cropping time number two to obtain the actual cropping time. Proportional increase is suitable for scenarios that need to dynamically adapt to different video lengths, while time increase is suitable for scenarios where the user explicitly needs a fixed delay.

7. A valve status monitoring system based on the Internet of Things according to claim 6, characterized in that: In the calculation of the actual cutting time in the state extraction module, the original cutting time is assumed to be... Let the proportionality index be... Let the user directly input the additional time required. Let the cutting time for No. 1 be... Let the cutting time for the second cut be... : ； ； The first cutting time and the second cutting time can be calculated using the two formulas mentioned above.

8. A valve status monitoring system based on the Internet of Things according to claim 6, characterized in that: After obtaining the actual cropping time, the state extraction module adds the actual cropping time to the starting time of the original video clipping, thus obtaining the actual cropping starting time. Let the starting time of the original video clipping be... Let the actual cutting time be... Let the actual cutting start time be... The units for the original video capture start time, the actual cropping time, and the actual cropping start time are all in seconds. ； The actual cutting start time can be calculated using the above formula.

9. A valve status monitoring system based on the Internet of Things according to claim 2, characterized in that: The status change index in the management alert module is calculated by extracting the sensed status information based on a base interval time. Then, a line graph is plotted based on the extracted sensed status information and the base interval time. The values ​​of the sensed status information in the line graph are extracted and numbered sequentially. Let the value of the Xth sensed status information be... Let the value of the sensing state information at time X+1 be... Let the state change index be... : ； The state change index is calculated using the formula above.

Images