Pilot operated safety valve state monitoring method and device with helium online leakage detection function

Abstract

The invention discloses a pilot operated safety valve state monitoring method and device with helium online leakage detection, and relates to the technical field of industrial process safety and equipment state monitoring, and the method comprises the steps: installing a microcavity leakage collection module on a to-be-detected pilot operated safety valve, connecting a helium online analysis subsystem and a tracing gas source through a pipeline, and carrying out the detection of the helium online leakage detection of the to-be-detected pilot operated safety valve; starting a vacuum pump set to vacuumize to a working vacuum degree; pure nitrogen is introduced to eliminate background helium, sampling analysis is performed after an interface is closed, and a helium background signal is recorded as a base line; low-concentration helium mixed gas is injected at a controlled speed, a tracing environment is constructed, and helium concentration, pressure and temperature multi-dimensional time sequence data are synchronously collected and preprocessed; inputting the data into a composite artificial intelligence model to judge a leakage state and divide alarm grades; the leakage rate is analyzed, the remaining life is calculated, monitoring information is uploaded to an upper-layer system, multi-channel notification is triggered, and monitoring is completed. Through the helium mass spectrum and the edge AI technology, accurate real-time monitoring is achieved, and the problem of off-line lag misinformation is solved.

Description

Technical Field

[0001] This invention relates to the field of industrial process safety and equipment condition monitoring technology, and in particular to a method and device for monitoring the condition of a pilot-operated safety valve with online helium leak detection. Background Technology

[0002] As a critical overpressure protection device, the reliability of a pilot-operated safety valve directly affects the safe and stable operation of the entire industrial system. It mainly consists of a main valve and a pilot valve. The pilot valve accurately senses the system pressure and controls the opening and closing of the main valve to achieve rapid and accurate pressure relief protection. The valve's sealing performance is its core performance indicator. Leakage not only causes material loss and environmental pollution but may also lead to the failure of the protective function due to the inability to effectively seal at the set pressure, potentially causing a catastrophic accident.

[0003] Currently, there are several main methods for condition monitoring of pilot-operated safety valves and their limitations: 1. Offline periodic calibration: This is the most traditional method, which involves removing the safety valve from the pipeline and sending it to a calibration bench for sealing tests and pressure calibration. This method has problems such as long calibration cycles, inability to reflect valve operating conditions in real time, and the potential for secondary damage during disassembly and assembly.

[0004] 2. Online Set Pressure Monitoring: Some advanced pilot-operated safety valves allow for online monitoring of their set pressure without interrupting the process. However, this technology primarily focuses on the valve's opening performance and is ineffective for minor leaks (internal leakage) when the valve seat is closed.

[0005] 3. Acoustic / Vibration Monitoring Method: This method involves installing acoustic emission or vibration sensors external to the valve body to capture high-frequency signals generated when the medium leaks, thereby determining the leakage status. While this method enables online monitoring, it is highly susceptible to interference from ambient noise and pipeline fluid noise, resulting in a low signal-to-noise ratio, difficulty in accurately quantifying the leakage rate, and a high false alarm rate.

[0006] 4. Conventional Tracer Gas Leak Detection Method: Helium mass spectrometry (HMS) leak detection technology is widely recognized as the most reliable method for detecting minute leaks due to its extremely high sensitivity. Helium, as a tracer gas, has advantages such as small atomic weight, stable chemical properties, and extremely low content in air. However, existing HMS leak detection applications are mostly limited to quality control before product delivery (such as back pressure method, jet purging method) or offline detection when the system is shut down. Currently, there is no mature technical solution on the market that can combine HMS leak detection technology with the structural characteristics of pilot-operated safety valves to achieve long-term, continuous, and intelligent online monitoring.

[0007] Therefore, there is an urgent need to develop a new monitoring method that can not only detect minute leaks in pilot-operated safety valves in real time with high sensitivity, but also intelligently analyze the leak status and predict its development trend, thereby achieving a leapfrog upgrade from "passive maintenance" to "predictive maintenance". Summary of the Invention

[0008] The main objective of this invention is to provide a method for monitoring the status of a pilot-operated safety valve with online helium leak detection.

[0009] Another objective of this invention is to provide a pilot-operated safety valve status monitoring device with online helium leak detection.

[0010] The third objective of this invention is to provide an electronic device.

[0011] A fourth objective of this invention is to provide a non-transitory computer-readable storage medium.

[0012] To achieve the above objectives, a first aspect of the present invention provides a method for monitoring the status of a pilot-operated safety valve with online helium leak detection, comprising: S1, install the micro-chamber leakage acquisition model on the pilot-operated safety valve to be tested, connect the helium online analysis subsystem and the tracer gas source through the pipeline, and start the vacuum pump group to evacuate the analysis chamber and sampling pipeline to the working vacuum level; S2, pure nitrogen gas is introduced into the microchamber through the purge port to remove background helium gas. After the interface is sealed, the online helium analysis subsystem continuously samples and analyzes the helium background signal as a dynamic baseline. S3 uses a mass flow controller to precisely control the flow rate of injecting a low-concentration helium mixture into the microchamber, creating a stable tracer environment, and simultaneously collecting and preprocessing multi-dimensional time series data of helium concentration, pressure and temperature. S4. Input the preprocessed multidimensional time series data into the composite artificial intelligence model. The leakage status is determined by a weighted comprehensive analysis of the reconstruction error of the long short-term memory network autoencoder and the anomaly score of the isolated forest. The alarm level is then classified according to the leakage rate. S5 uses leakage rate data analysis and time-series prediction algorithms to calculate remaining service life and generate maintenance warnings. Through industrial communication interfaces, it uploads monitoring data, leakage levels, and health status prediction results to the upper-level management system, simultaneously triggering audible and visual alarms, SMS or email notifications, thus realizing a complete monitoring process.

[0013] Optionally, the step of installing the micro-chamber leak acquisition model on the pilot-operated safety valve under test, connecting the helium online analysis subsystem and the tracer gas source through pipelines, and starting the vacuum pump group to evacuate the analysis chamber and sampling pipeline to the working vacuum level also includes: The microchamber leakage acquisition model is installed on the pilot-operated safety valve to be tested, ensuring a reliable seal between the microchamber leakage acquisition model and the valve body of the pilot-operated safety valve; Each interface of the microchamber leakage acquisition model is connected to the helium online analysis subsystem and the tracer gas source via pipelines, wherein the tracer gas source is a mixed gas cylinder containing a low concentration of helium. Start the vacuum pump unit to evacuate the analysis chamber and sampling pipeline of the helium mass spectrometer leak detector in the online helium analysis subsystem until the vacuum level of the analysis chamber and sampling pipeline reaches the preset working vacuum threshold.

[0014] Optionally, the step of introducing pure nitrogen gas into the microchamber through the purge port to remove background helium, and then having the interface sealed and the helium online analysis subsystem continuously sample and analyze the gas, recording the helium background signal as a dynamic baseline, further includes: Pure nitrogen gas is introduced into the acquisition microcavity through the purge / exhaust port and the cavity is thoroughly purged to remove air adsorbed inside the acquisition microcavity and the inner wall of the pipeline, especially to remove background helium mixed in the air. After purging is complete, stop the flow of pure nitrogen and seal all interfaces of the acquisition microcavity; Start the helium online analysis subsystem to continuously sample and analyze the gas in the acquisition microcavity; The edge intelligent computing unit collects and records the helium background signal output by the helium online analysis subsystem. The helium background signal is used as the zero-point baseline for subsequent leakage judgment. The zero-point baseline must meet the requirement of stable low background to ensure the detection sensitivity of the system.

[0015] Optionally, the step of precisely controlling the flow rate of the low-concentration helium mixture injected into the microchamber via a mass flow controller to construct a stable tracer environment, and simultaneously acquiring and preprocessing multidimensional time-series data of helium concentration, pressure, and temperature, further includes: By using a mass flow controller to precisely control the injection rate through the tracer gas inlet, a low-concentration helium gas mixture is continuously or periodically injected into the acquisition microcavity to create a stable ultra-low concentration helium tracer environment within the acquisition microcavity. The online helium analysis subsystem is activated to continuously extract gas from the acquisition microcavity through the sampling outlet and perform composition analysis to obtain continuous time series data of helium concentration changes over time. Simultaneously, the helium concentration signal output by the helium mass spectrometer leak detector, the pressure signal output by the pressure sensor, and the temperature signal output by the temperature sensor are acquired at a high sampling rate, and normalization and noise reduction preprocessing operations are performed on the multidimensional signal data.

[0016] Optionally, the weighted comprehensive determination of the leakage state through the reconstruction error of the Long Short-Term Memory Network autoencoder and the anomaly score of the isolated forest further includes: The Long Short-Term Memory Network Autoencoder Model learns the complex temporal dynamic pattern of the helium concentration sequence under normal operating conditions. During real-time monitoring, it reconstructs the input signal sequence. When there is no leakage in the valve, the reconstruction error is in the minimum range. When the valve leaks gradually and slowly, the continuous abnormal fluctuation or rise of the helium concentration will cause the input sequence to deviate from the normal pattern, which will lead to a significant increase in the reconstruction error. The isolated forest model identifies mutations or outliers in the data stream, enabling rapid detection of instantaneous, pulse-like leaks caused by pressure fluctuations. The edge intelligent computing unit performs a weighted calculation on the reconstruction error of the long short-term memory network autoencoder and the anomaly score of the isolated forest to obtain a comprehensive anomaly score. When the comprehensive anomaly score continuously exceeds a preset dynamic threshold, it is determined that the valve has actually leaked. Optionally, the method of classifying alarm levels based on leakage rate also includes: The system is based on the stable rise value of helium concentration, the known volume of the microcavity and the sampling rate. It substitutes the data into a calibration model that has been calibrated through a standard leak, calculates the equivalent standard leakage rate in real time and classifies the leakage status into alarm levels such as minor leakage, warning leakage, and serious leakage according to the preset leakage rate threshold range.

[0017] Optionally, the process of using leakage rate data analysis, employing time-series prediction algorithms to calculate remaining service life and generate maintenance warnings, and uploading monitoring data, leakage levels, and health status prediction results to the upper-level management system via an industrial communication interface to simultaneously trigger audible and visual alarms, SMS or email notifications, thus realizing a complete monitoring process, also includes: The edge intelligent computing unit stores historical leakage rate data in a local storage model. By analyzing the trend characteristics of the leakage rate over time, it selects a regression algorithm or an autoregressive integral moving average model time series prediction algorithm to construct a valve sealing performance degradation trajectory prediction model, calculates the remaining service life of the valve, and generates corresponding maintenance warnings and maintenance suggestions based on the threshold range of the remaining service life. The edge intelligent computing unit uploads the valve's real-time status information, quantified leakage rate data, and health status prediction results to the factory's distributed control system, data acquisition and monitoring control system, or equipment health management platform through an industrial communication interface. When the system detects a leak, it simultaneously triggers an audible and visual alarm and sends alarm notifications to relevant maintenance personnel via SMS and email, completing the full monitoring process.

[0018] To achieve the above objectives, a second aspect of the present invention provides a pilot-operated safety valve status monitoring device with online helium leak detection, comprising: The initialization module is used to install the micro-chamber leak acquisition model on the pilot-operated safety valve under test, connect the helium online analysis subsystem and the tracer gas source through pipelines, and start the vacuum pump group to evacuate the analysis chamber and sampling pipeline to the working vacuum level. The baseline establishment module is used to introduce pure nitrogen into the microchamber through the purge port to remove background helium. After the interface is sealed, the online helium analysis subsystem continuously samples and analyzes the helium background signal as a dynamic baseline. The data acquisition and preprocessing module is used to precisely control the flow rate of the mass flow controller to inject a low-concentration helium mixture into the microchamber, build a stable tracer environment, and simultaneously acquire and preprocess multi-dimensional time series data of helium concentration, pressure and temperature. The leak diagnosis and grading module is used to input preprocessed multidimensional time series data into a composite artificial intelligence model. The leak status is determined by a weighted comprehensive analysis of the reconstruction error of the long short-term memory network autoencoder and the anomaly score of the isolated forest, and the alarm level is divided according to the leak rate. The lifespan prediction and information interaction module is used to calculate the remaining lifespan and generate maintenance warnings by analyzing leakage rate data and using time-series prediction algorithms. Through the industrial communication interface, it uploads monitoring data, leakage level and health status prediction results to the upper management system, and simultaneously triggers audible and visual alarms, SMS or email notifications to realize the complete monitoring process.

[0019] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0020] To achieve the above objectives, a third aspect of this application provides an electronic device, including a processor and a memory; wherein the processor reads executable program code stored in the memory to run a program corresponding to the executable program code, for implementing a pilot-operated safety valve status monitoring method with online helium leak detection as described in the first aspect embodiment.

[0021] To achieve the above objectives, a fourth aspect of this application provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements a pilot-operated safety valve status monitoring method with online helium leak detection as described in the first aspect embodiment.

[0022] The embodiments of the present invention have the following beneficial effects: 1. It has higher detection sensitivity and accuracy, inheriting the core advantages of helium mass spectrometry leak detection technology. It can detect low-level micro-leakage and other online monitoring methods such as ultrasonics, and can effectively detect sealing failure problems in their early stages.

[0023] 2. Through the micro-chamber structure and online analysis system, continuous 24 / 7 monitoring of in-service valves is achieved, completely solving the problem of lag in offline detection and ensuring the timely detection of faults.

[0024] 3. Through a composite diagnostic model based on edge AI, it is possible to intelligently distinguish between real leakage signals and environmental noise and operational disturbances, thereby reducing the false alarm rate and improving the reliability of the monitoring system.

[0025] 4. Through intelligent analysis of leakage development trends, it "predicts" future failure risks, providing strong data support for enterprises to implement condition-based maintenance and predictive maintenance, which helps to optimize maintenance strategies, reduce maintenance costs, and reduce unplanned downtime. Attached Figure Description

[0026] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 A flowchart illustrating a method for monitoring the status of a pilot-operated safety valve with online helium leak detection, provided as an embodiment of the present invention; Figure 2 A flowchart illustrating the overall process of a pilot-operated safety valve status monitoring method with online helium leak detection, as provided in an embodiment of the present invention. Figure 3 This is a structural diagram of a pilot-operated safety valve status monitoring device with online helium leak detection, provided in an embodiment of the present invention. Detailed Implementation

[0027] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0028] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0029] The following description, with reference to the accompanying drawings, describes a method and apparatus for monitoring the condition of a pilot-operated safety valve with online helium leak detection according to an embodiment of the present invention.

[0030] Example 1 This invention provides a method for monitoring the status of a pilot-operated safety valve with online helium leak detection. Figure 1 This is a schematic flowchart illustrating a method for monitoring the status of a pilot-operated safety valve with online helium leak detection, provided in an embodiment of the present invention. Figure 2 This is an overall flowchart of a pilot-operated safety valve status monitoring method with online helium leak detection, provided as an embodiment of the present invention. Figure 1 , Figure 2 As shown, the method includes the following steps: Step S1: Install the micro-chamber leakage acquisition model on the pilot-operated safety valve to be tested, connect the helium online analysis subsystem and the tracer gas source through pipelines, and start the vacuum pump group to evacuate the analysis chamber and sampling pipeline to the working vacuum level.

[0031] In this embodiment, the operator needs to assemble the microchamber leak acquisition model onto the pilot-operated safety valve to be tested. The microchamber leak acquisition model is made of 316L stainless steel and designed with a split clamp structure, with a Viton O-ring embedded inside. This structural design ensures the sealing of the connection between the model and the safety valve body, avoiding data deviation caused by installation gaps from the outset. After assembly, the various interfaces of the microchamber leak acquisition model are connected to the helium online analysis subsystem and the tracer gas source via dedicated pipelines. The tracer gas source is specifically a low-concentration helium mixture cylinder, with an exemplary mixture of 1% helium and 99% nitrogen by volume. After the pipeline connections are completed, the vacuum pump unit is started to evacuate the analysis chamber and sampling pipeline of the helium mass spectrometer leak detector in the helium online analysis subsystem until the chamber and pipeline reach the preset working vacuum level.

[0032] In this embodiment, the vacuum pump group is configured with two levels of vacuum pumping components: a backing pump and a high-vacuum pump. The backing pump is an Edwards nXDS10i dry vortex pump, which can fundamentally eliminate the generation of oil vapor and prevent oil vapor from entering the analysis environment and causing pollution. The high-vacuum pump is a Pfeiffer HiPace 80 turbomolecular pump. Relying on the strong pumping capacity of the high-vacuum pump, a stable ultra-high vacuum environment can be constructed for the analysis chamber of the helium mass spectrometer leak detector, thereby providing a reliable environmental guarantee for the high-precision leak detection operation in the subsequent stage.

[0033] In step S2, pure nitrogen gas is introduced into the microchamber through the purge port to remove background helium gas. After the interface is sealed, the helium gas online analysis subsystem continuously samples and analyzes the helium gas background signal as a dynamic baseline.

[0034] In this embodiment, to ensure the complete removal of background helium adsorbed within the microcavity and on the inner walls of the pipeline, the purging time with pure nitrogen is set to be no less than 10 minutes. After the purging process is completed, the operator must completely seal all interfaces of the microcavity, and then start the online helium analysis subsystem, which continuously samples and analyzes the gas inside the microcavity. In this embodiment, the helium mass spectrometer leak detector used in the online helium analysis subsystem is specifically the Pfeiffer Vacuum ASM 340 series product. This model of leak detector has excellent detection performance, with a minimum detectable helium leak rate better than 5 × 10⁻⁶. - ¹²Pa With a speed of m³ / s and a response time of less than 1 second, it can quickly and accurately capture subtle changes in helium concentration within a microcavity.

[0035] Furthermore, while the helium online analysis subsystem performs sampling and analysis, the edge intelligent computing unit simultaneously collects and records the current helium background signal, and sets this signal as the dynamic baseline for subsequent leak detection. It is important to emphasize that obtaining a stable low background signal is a crucial prerequisite for ensuring the detection sensitivity of the entire monitoring system. Therefore, in this embodiment, the stability criterion for the helium background signal is defined as follows: the fluctuation range of the helium concentration is controlled within 0.01 × 10⁻⁶ for 5 consecutive minutes. -6 Only background signals that meet this condition can be used as a valid dynamic baseline for subsequent applications.

[0036] Step S3: A low-concentration helium gas mixture is injected into the microchamber by precisely controlling the flow rate through a mass flow controller to construct a stable tracer environment. Simultaneously, multi-dimensional time series data of helium concentration, pressure, and temperature are collected and preprocessed.

[0037] In this embodiment, the injection flow rate of the low-concentration helium gas mixture is precisely controlled within the range of 1~10 mL / min. This flow rate parameter has been verified through multiple experiments, enabling a steady increase in the helium concentration within the microcavity and maintaining dynamic equilibrium while avoiding gas turbulence disturbances. This, in turn, constructs a uniform and stable ultra-low concentration helium tracer environment within the microcavity. This embodiment specifies strict criteria for determining this tracer environment, specifically that the fluctuation range of the helium concentration within the microcavity does not exceed ±0.02 × 10⁻⁶ mL / min. -6 Furthermore, the duration of this stable state is no less than 10 minutes.

[0038] Meanwhile, the helium online analysis subsystem continuously collects gas samples from the microcavity at a constant extraction rate through a preset sampling outlet and performs real-time component analysis, thereby obtaining a continuous time-series data stream of helium concentration changes over time. Simultaneously, a multi-source sensing and data acquisition model acquires three key monitoring signals in parallel at a high sampling rate: the core helium concentration signal output by the helium mass spectrometer leak detector, the pressure auxiliary signal fed back by the pressure sensor, and the temperature compensation signal acquired by the temperature sensor. The pressure sensor uses the Rosemount 3051 series pressure transmitter, which features strong anti-interference capabilities and high measurement accuracy, enabling precise capture of subtle changes in inlet pressure. The data acquisition card is the National Instruments NI-6211, equipped with a multi-channel high-precision analog input interface, enabling simultaneous acquisition and rapid transmission of multiple signals, effectively ensuring the timeliness and integrity of the data.

[0039] After completing the acquisition of the multidimensional time series data, this embodiment of the application performs preprocessing operations such as normalization and denoising on the acquired raw data. Specifically, wavelet denoising algorithm is used to accurately remove high-frequency interference noise in the signal while retaining effective signal features. At the same time, based on the min-max normalization method, multidimensional data with different dimensions and different numerical ranges are uniformly normalized to the [0,1] interval to eliminate the dimensional differences between data, improve the analysis accuracy of subsequent artificial intelligence models, and provide high-quality data support for the stable operation of the entire monitoring system.

[0040] Step S4: Input the preprocessed multidimensional time series data into the composite artificial intelligence model. The leakage status is determined by a weighted comprehensive analysis of the reconstruction error of the long short-term memory network autoencoder and the anomaly score of the isolated forest. The alarm level is then classified according to the leakage rate.

[0041] In this embodiment, the hardware carrier of the edge intelligent computing and diagnostic unit is the NVIDIA Jetson AGX Orin Developer Kit. This hardware platform integrates a high-performance GPU computing core, and its powerful parallel computing capability can provide sufficient computing power support for the real-time inference and computation of the composite artificial intelligence model, ensuring that millisecond-level data analysis and diagnostic response can still be achieved under complex working conditions in industrial sites.

[0042] Specifically, the composite artificial intelligence model used in this application is built on a hybrid architecture of Long Short-Term Memory Network Autoencoder (LSTM-Autoencoder) and Isolation Forest. These two algorithms complement each other, working together to accurately identify valve leakage states. The LSTM-Autoencoder model, with its strong ability to fit time-series data, can deeply learn the complex temporal dynamic evolution of helium concentration sequences under normal operating conditions and reconstruct the real-time input signal sequence based on the learned normal pattern. When the valve is in a leak-free normal state, the input real-time signal sequence closely matches the normal pattern learned by the model, and the reconstruction error of the model output remains within a very small range. However, when the valve experiences a gradual, slow leak, the helium concentration in the microcavity exhibits continuous abnormal fluctuations or a step-like increase, causing the real-time input sequence to deviate significantly from the normal pattern, thereby triggering a significant increase in the model's reconstruction error. This achieves highly sensitive identification of slow leaks. In contrast, the isolated forest model focuses on capturing abrupt changes and outliers in the data stream. Its anomaly detection mechanism based on random forest can quickly distinguish between normal and anomalous data. For instantaneous and pulse-like leaks caused by factors such as system pressure fluctuations and instantaneous operating condition disturbances, it can quickly identify them as anomalies, making up for the shortcomings of single time series models in responding to abrupt leaks with lag.

[0043] Based on this, the edge computing unit performs weighted fusion calculations on the reconstruction error output by the LSTM-Autoencoder and the anomaly score output by the isolated forest to obtain a comprehensive anomaly score that fully reflects the valve status. In this embodiment, when the comprehensive anomaly score continuously exceeds a preset dynamic threshold, the system immediately determines that the valve has experienced a real leak. It is worth noting that the aforementioned dynamic threshold is not a fixed value, but is dynamically updated based on 3 to 5 times the standard deviation of the comprehensive anomaly score under historical normal operating conditions, thereby adapting to the differences in operating conditions at different stages and improving the accuracy of leak detection. Once the system confirms a leak, it retrieves a calibration model pre-calibrated using a standard leak hole, and, combined with the stable rise in helium concentration, the known volume of the acquisition microcavity, and the constant sampling rate of the sampling pipeline, calculates the equivalent standard leakage rate in real time, with units of Pa. m³ / s; at the same time, the system classifies the leakage status into three levels: “minor leakage”, “attention leakage”, and “serious leakage” according to the preset leakage rate threshold range, providing operation and maintenance personnel with an intuitive reference for the severity of leakage.

[0044] Step S5: Using leakage rate data analysis, the remaining service life is calculated using a time-series prediction algorithm and a maintenance warning is generated. The monitoring data, leakage level, and health status prediction results are uploaded to the upper-level management system through an industrial communication interface, and audible and visual alarms, SMS or email notifications are triggered simultaneously to achieve a complete monitoring process.

[0045] In one embodiment of the present invention, the edge computing unit stores the leakage rate data obtained from each monitoring session in a locally configured power-loss protected solid-state drive. The data storage period is no less than 12 months to ensure the integrity and traceability of historical data. By retrieving the stored historical leakage rate data and analyzing its trend characteristics over time, this embodiment of the application selects mature time-series prediction models such as regression algorithms or autoregressive integral moving average models (ARIMA) to construct a degradation trajectory prediction model for valve sealing performance. This allows for accurate assessment of the valve's remaining service life (RUL) and, based on the length of the RUL, generates tiered maintenance warnings and targeted maintenance recommendations in advance. This design achieves a technological leap from traditional post-fault diagnosis to pre-fault prediction, providing data-driven decision-making basis for equipment lifecycle management.

[0046] In this embodiment, the edge intelligent computing unit is equipped with a standardized industrial communication interface that natively supports the OPCUA industrial communication protocol, enabling seamless integration and data interaction with the factory's upper-level management system. To address data transmission needs under different operating conditions, this embodiment sets differentiated data upload frequencies: when the valve is in normal operating condition, the data upload frequency is set to 5-15 minutes per upload, thus ensuring data integrity while reducing communication bandwidth usage; when the system detects abnormal conditions such as leakage and triggers an alarm, the data upload frequency automatically increases to 1-5 seconds per upload, ensuring that the factory's DCS, SCADA, or EAM upper-level management systems can accurately and in real-time grasp the abnormal valve status and related monitoring data.

[0047] When the system detects a valve leak and triggers an alarm, a tiered early warning response mechanism will be activated simultaneously. Firstly, the system activates the tiered audible and visual alarm devices on-site. For "minor leaks," a yellow audible and visual warning signal is triggered; for "attention leaks," an orange signal; and for "serious leaks," a red signal. This intuitive color and audible / visual differentiation helps on-site maintenance personnel quickly assess the severity of the leak. Secondly, the system automatically generates alarm information and pushes it to relevant personnel via SMS, email, and other remote communication methods. Alarm notification recipients can be configured in tiers: Level 1 maintenance warnings are pushed to the work team's maintenance personnel for timely on-site investigation and initial handling; Level 2 emergency maintenance warnings are simultaneously pushed to the equipment management department head, ensuring management can grasp the fault situation immediately and coordinate subsequent maintenance resources. Through tiered response and rapid handling, the risk of unplanned equipment downtime is minimized, ensuring the stable operation of the industrial system.

[0048] Example 2 This invention provides a pilot-operated safety valve status monitoring device with online helium leak detection. Figure 3 This is a schematic flowchart of a pilot-operated safety valve status monitoring device with online helium leak detection provided in an embodiment of the present invention. Figure 3 As shown, the device includes: The initialization module 100 is used to install the micro-chamber leakage acquisition model on the pilot-operated safety valve under test, connect the helium online analysis subsystem and the tracer gas source through pipelines, and start the vacuum pump group to evacuate the analysis chamber and sampling pipeline to the working vacuum level. The baseline establishment module 200 is used to introduce pure nitrogen into the microchamber through the purge port to remove background helium. After the interface is sealed, the helium online analysis subsystem continuously samples and analyzes the helium background signal as a dynamic baseline. The data acquisition and preprocessing module 300 is used to precisely control the flow rate of injecting a low-concentration helium mixture into the microchamber through a mass flow controller, to build a stable tracer environment, and to simultaneously acquire and preprocess multi-dimensional time series data of helium concentration, pressure and temperature. The Leakage Diagnosis and Grading Module 400 is used to input preprocessed multidimensional time series data into a composite artificial intelligence model. The leakage status is determined by a weighted comprehensive analysis of the reconstruction error of the Long Short-Term Memory Network Autoencoder and the anomaly score of the Isolated Forest, and the alarm level is divided according to the leakage rate. The 500-level lifespan prediction and information interaction module is used to calculate the remaining lifespan and generate maintenance warnings by analyzing leakage rate data and using time-series prediction algorithms. Through the industrial communication interface, it uploads monitoring data, leakage level and health status prediction results to the upper management system, and simultaneously triggers audible and visual alarms, SMS or email notifications to realize the complete monitoring process.

[0049] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0050] Example 3 To implement the methods of the above embodiments, the present invention also provides an electronic device, which includes a memory and a processor; wherein the processor reads executable program code stored in the memory to run a program corresponding to the executable program code, so as to implement the various steps of the methods described above.

[0051] Example 4 To implement the above embodiments, this application also proposes a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the method described in the foregoing embodiments.

[0052] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0053] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0054] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

Claims

1. A method for monitoring the condition of a pilot-operated safety valve with online helium leak detection, characterized in that, include: S1, install the micro-chamber leakage acquisition model on the pilot-operated safety valve to be tested, connect the helium online analysis subsystem and the tracer gas source through the pipeline, and start the vacuum pump group to evacuate the analysis chamber and sampling pipeline to the working vacuum level; S2, pure nitrogen gas is introduced into the microchamber through the purge port to remove background helium gas. After the interface is sealed, the online helium analysis subsystem continuously samples and analyzes the helium background signal as a dynamic baseline. S3 uses a mass flow controller to precisely control the flow rate of injecting a low-concentration helium mixture into the microchamber, creating a stable tracer environment, and simultaneously collecting and preprocessing multi-dimensional time series data of helium concentration, pressure and temperature. S4. Input the preprocessed multidimensional time series data into the composite artificial intelligence model. The leakage status is determined by a weighted comprehensive analysis of the reconstruction error of the long short-term memory network autoencoder and the anomaly score of the isolated forest. The alarm level is then classified according to the leakage rate. S5 uses leakage rate data analysis and time-series prediction algorithms to calculate remaining service life and generate maintenance warnings. Through industrial communication interfaces, it uploads monitoring data, leakage levels, and health status prediction results to the upper-level management system, simultaneously triggering audible and visual alarms, SMS or email notifications, thus realizing a complete monitoring process.

2. The method according to claim 1, characterized in that, The process of installing the microchamber leak acquisition model on the pilot-operated safety valve under test, connecting the helium online analysis subsystem and the tracer gas source through pipelines, and starting the vacuum pump group to evacuate the analysis chamber and sampling pipeline to the working vacuum level also includes: The microchamber leakage acquisition model is installed on the pilot-operated safety valve to be tested, ensuring a reliable seal between the microchamber leakage acquisition model and the valve body of the pilot-operated safety valve; Each interface of the microchamber leakage acquisition model is connected to the helium online analysis subsystem and the tracer gas source via pipelines, wherein the tracer gas source is a mixed gas cylinder containing a low concentration of helium. Start the vacuum pump unit to evacuate the analysis chamber and sampling pipeline of the helium mass spectrometer leak detector in the online helium analysis subsystem until the vacuum level of the analysis chamber and sampling pipeline reaches the preset working vacuum threshold.

3. The method according to claim 2, characterized in that, The process of introducing pure nitrogen gas into the microchamber through a purge port to remove background helium, followed by continuous sampling and analysis by an online helium analysis subsystem after sealing the interface, recording the helium background signal as a dynamic baseline, also includes: Pure nitrogen gas is introduced into the acquisition microcavity through the purge / exhaust port and the cavity is thoroughly purged to remove air adsorbed inside the acquisition microcavity and the inner wall of the pipeline, especially to remove background helium mixed in the air. After purging is complete, stop the flow of pure nitrogen and seal all interfaces of the acquisition microcavity; Start the helium online analysis subsystem to continuously sample and analyze the gas in the acquisition microcavity; The edge intelligent computing unit collects and records the helium background signal output by the helium online analysis subsystem. The helium background signal is used as the zero-point baseline for subsequent leak judgment. The zero-point baseline must meet the requirement of stable low background to ensure the detection sensitivity of the system.

4. The method according to claim 3, characterized in that, The process of injecting a low-concentration helium mixture into a microchamber using a mass flow controller to precisely control the flow rate, thereby creating a stable tracer environment, and simultaneously acquiring and preprocessing multidimensional time-series data on helium concentration, pressure, and temperature, also includes: By using a mass flow controller to precisely control the injection rate through the tracer gas inlet, a low-concentration helium gas mixture is continuously or periodically injected into the acquisition microcavity to create a stable ultra-low concentration helium tracer environment within the acquisition microcavity. The online helium analysis subsystem is activated to continuously extract gas from the acquisition microcavity through the sampling outlet and perform composition analysis to obtain continuous time series data of helium concentration changes over time. Simultaneously, the helium concentration signal output by the helium mass spectrometer leak detector, the pressure signal output by the pressure sensor, and the temperature signal output by the temperature sensor are acquired at a high sampling rate, and normalization and noise reduction preprocessing operations are performed on the multidimensional signal data.

5. The method according to claim 4, characterized in that, The method of determining the leakage status by weighted comprehensive analysis of the reconstruction error of the Long Short-Term Memory Network autoencoder and the anomaly score of the isolated forest also includes: The Long Short-Term Memory Network Autoencoder Model learns the complex temporal dynamic pattern of the helium concentration sequence under normal operating conditions. During real-time monitoring, it reconstructs the input signal sequence. When there is no leakage in the valve, the reconstruction error is in the minimum range. When the valve leaks gradually and slowly, the continuous abnormal fluctuation or rise of the helium concentration will cause the input sequence to deviate from the normal pattern, which will lead to a significant increase in the reconstruction error. The isolated forest model identifies mutations or outliers in the data stream, enabling rapid detection of instantaneous, pulse-like leaks caused by pressure fluctuations. The edge intelligent computing unit performs a weighted calculation on the reconstruction error of the long short-term memory network autoencoder and the anomaly score of the isolated forest to obtain a comprehensive anomaly score. When the comprehensive anomaly score continuously exceeds a preset dynamic threshold, it is determined that the valve has actually leaked.

6. The method according to claim 5, characterized in that, The method of classifying alarm levels based on leakage rate also includes: The system is based on the stable rise value of helium concentration, the known volume of the microcavity and the sampling rate. It substitutes the data into a calibration model that has been calibrated through a standard leak, calculates the equivalent standard leakage rate in real time and classifies the leakage status into alarm levels such as minor leakage, warning leakage, and serious leakage according to the preset leakage rate threshold range.

7. The method according to claim 6, characterized in that, The process of using leakage rate data analysis and time-series prediction algorithms to calculate remaining service life and generate maintenance warnings, and uploading monitoring data, leakage levels, and health status prediction results to the upper-level management system via an industrial communication interface to simultaneously trigger audible and visual alarms, SMS or email notifications, thus realizing a complete monitoring process, also includes: The edge intelligent computing unit stores historical leakage rate data in a local storage model. By analyzing the trend characteristics of the leakage rate over time, it selects a regression algorithm or an autoregressive integral moving average model time series prediction algorithm to construct a valve sealing performance degradation trajectory prediction model, calculates the remaining service life of the valve, and generates corresponding maintenance warnings and maintenance suggestions based on the threshold range of the remaining service life. The edge intelligent computing unit uploads the valve's real-time status information, quantified leakage rate data, and health status prediction results to the factory's distributed control system, data acquisition and monitoring control system, or equipment health management platform through an industrial communication interface. When the system detects a leak, it simultaneously triggers an audible and visual alarm and sends alarm notifications to relevant maintenance personnel via SMS and email, completing the full monitoring process.

8. A pilot-operated safety valve status monitoring device with online helium leak detection, characterized in that, include: The initialization module is used to install the micro-chamber leak acquisition model on the pilot-operated safety valve under test, connect the helium online analysis subsystem and the tracer gas source through pipelines, and start the vacuum pump group to evacuate the analysis chamber and sampling pipeline to the working vacuum level. The baseline establishment module is used to introduce pure nitrogen into the microchamber through the purge port to remove background helium. After the interface is sealed, the online helium analysis subsystem continuously samples and analyzes the helium background signal as a dynamic baseline. The data acquisition and preprocessing module is used to precisely control the flow rate of the mass flow controller to inject a low-concentration helium mixture into the microchamber, build a stable tracer environment, and simultaneously acquire and preprocess multi-dimensional time series data of helium concentration, pressure and temperature. The leak diagnosis and grading module is used to input preprocessed multidimensional time series data into a composite artificial intelligence model. The leak status is determined by a weighted comprehensive analysis of the reconstruction error of the long short-term memory network autoencoder and the anomaly score of the isolated forest, and the alarm level is divided according to the leak rate. The lifespan prediction and information interaction module is used to calculate the remaining lifespan and generate maintenance warnings by analyzing leakage rate data and using time-series prediction algorithms. Through the industrial communication interface, it uploads monitoring data, leakage level and health status prediction results to the upper management system, and simultaneously triggers audible and visual alarms, SMS or email notifications to realize the complete monitoring process.

9. An electronic device, characterized in that, Including processor and memory; The processor reads executable program code stored in the memory to run a program corresponding to the executable program code, so as to implement the method as described in any one of claims 1-7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1-7.

Images