A predictive maintenance method and device for industrial valve failure
By building a predictive maintenance model through real-time monitoring and machine learning algorithms, the problems of low efficiency and insufficient safety of traditional valve maintenance methods are solved, and efficient and safe fault prediction and prevention are achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SUPCON TECH CO LTD
- Filing Date
- 2024-12-31
- Publication Date
- 2026-06-30
AI Technical Summary
Traditional valve maintenance methods rely on periodic inspections or repairs after a failure, which leads to low efficiency, high costs, and an inability to effectively prevent sudden failures, potentially causing safety accidents and affecting production progress and equipment lifespan.
By monitoring valve status in real time, using data acquisition servers and machine learning algorithms to eliminate outliers, a predictive maintenance model is built to identify potential faults and issue alarm messages and maintenance recommendations.
It enables forward-looking and efficient valve maintenance, reduces production losses, and improves the safety and stability of the production process and system.
Smart Images
Figure CN122305306A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of industrial automation technology, and in particular to a method and apparatus for predictive maintenance of industrial valve failures. Background Technology
[0002] In modern industrial production, valves are widely used as crucial control components in various fields. Effective valve maintenance is essential to ensure production continuity and safety. Traditional maintenance methods often rely on periodic inspections or post-failure repairs. This passive approach is not only inefficient and costly, but also fails to effectively prevent sudden malfunctions and may lead to unnecessary production interruptions.
[0003] Traditional valve maintenance methods have certain limitations, mainly falling into two categories: periodic inspection and post-failure repair. Periodic inspection refers to checking and maintaining valves at predetermined intervals. This method means that even when valves are in good working order, inspection is still required, consuming significant human and material resources and increasing maintenance costs. While periodic inspections help identify potential problems, the unpredictable timing and cause of sudden failures make effective prevention difficult. Furthermore, periodic inspections typically require temporarily shutting down related equipment or production lines, directly impacting production schedules and efficiency.
[0004] Post-failure repair refers to immediately halting production for maintenance after a valve malfunctions. This approach leads to production interruptions and increases costs. More importantly, if a valve malfunctions under high temperature or pressure, it can cause serious safety incidents such as leaks or fires. Furthermore, long-term reliance on post-failure repair can shorten the lifespan of valves and other related equipment, increasing replacement costs. Therefore, developing an intelligent system capable of predictive valve maintenance is of significant practical importance. Summary of the Invention
[0005] The purpose of this invention is to address the shortcomings of existing industrial valve maintenance methods by proposing a predictive maintenance method and device for industrial valve failures. This method reduces unnecessary periodic maintenance by predicting and preventing failures in advance, thereby saving costs, reducing unplanned downtime, improving the overall efficiency of the production line, and enhancing production process safety by identifying and resolving potential problems early. The technical solution is as follows:
[0006] A predictive maintenance method for industrial valve failures includes:
[0007] Real-time monitoring of valve operating status; valve data is collected using a data acquisition server.
[0008] The collected valve data is preprocessed to remove outliers;
[0009] Analyze the preprocessed valve data to build a predictive maintenance model, identify abnormal situations, and predict potential valve failures;
[0010] For predicted valve malfunctions, an alarm message is issued and maintenance suggestions are provided.
[0011] Optionally, the data acquisition server is used to collect key parameters of industrial valves. The data acquisition server includes an OPC server and an IDM server. The parameters collected by the OPC server include at least loop data and upstream and downstream pressures of the valve. The IDM server is configured with a communication protocol compatible with predictive maintenance systems, and the parameters collected include at least valve stroke feedback values, air supply pressure, and actuation pressure.
[0012] Optionally, the preprocessing of the collected valve data to remove outliers includes using a machine learning algorithm to remove valve data outliers. The machine learning algorithm includes at least the OCSVM algorithm: first, a high-dimensional feature space is constructed, the valve data is mapped to the high-dimensional space using the RBF kernel function, the range of the data point set inside the high-dimensional feature space is determined as the first hyperplane, and the boundary of the first hyperplane is set. When the measured data point is outside the boundary of the first hyperplane, the measured data point is considered to be an outlier.
[0013] The RBF kernel function is in the following form:
[0014] K(x,x')=exp(-γ||x-x'|| 2 )
[0015] Where x and x' represent two input vectors; ||x-x'|| represents the Euclidean distance between the two vectors; K(x,x') is the similarity between the two input vectors; γ is a positive number that controls the width of the kernel function. The smaller the value of γ, the wider the kernel function, and the larger the value of γ, the narrower the kernel function.
[0016] Optionally, the predictive maintenance model includes using machine learning algorithms to establish various diagnostic indicators for the valve's operating status. These diagnostic indicators include at least: valve dead zone, valve viscosity, valve deviation, valve tracking performance, valve blockage coefficient, and valve scouring coefficient.
[0017] Optionally, the analysis of the preprocessed valve data to construct a predictive maintenance model, identify abnormal situations, and predict potential valve failures also includes analyzing the real-time operating status data of the valve based on the valve diagnostic indicators output by the predictive maintenance model, including at least cumulative stroke, changes in the number of actions, and frequency of small opening operations, to predict valve wear trends and potential failures.
[0018] Optionally, the maintenance recommendations include automatically generating preventative maintenance tasks, which at least include inspection, lubrication, and component replacement, and prioritizing them according to the importance of the valves.
[0019] Optionally, a predictive maintenance method for industrial valve failure further includes: remotely viewing the real-time status and maintenance reports of the valve using a remote monitoring platform. The real-time status of the valve refers to its key performance indicators, including at least the actuation pressure, supply pressure, valve position setpoint, valve position stroke value, and loop data. The maintenance report includes at least a valve status overview, performance trend graph, alarm summary, and maintenance suggestions. The maintenance report is generated and pushed out periodically according to a preset cycle.
[0020] An industrial valve failure predictive maintenance device, comprising:
[0021] The data acquisition module is used to monitor the valve's operating status in real time and collect valve data using the data acquisition server.
[0022] The data processing module is used to preprocess the collected valve data and remove outliers;
[0023] The data analysis module is used to analyze the preprocessed valve data, build predictive maintenance models, identify abnormal situations, and predict potential valve failures.
[0024] The fault early warning module is used to issue alarm information and provide maintenance suggestions for predicted valve faults.
[0025] An electronic device includes a memory and a processor, the memory storing computer-readable instructions that, when executed by the processor, cause the processor to perform the industrial valve failure predictive maintenance method as described in any of the preceding claims.
[0026] A storage medium storing computer-readable instructions, which, when executed by one or more processors, cause the one or more processors to perform the industrial valve failure predictive maintenance method as described in any of the preceding claims.
[0027] Compared with the prior art, this application has the following beneficial effects: significantly improving the foresight and efficiency of valve maintenance work, reducing production losses caused by equipment failure, improving the safety of the production process, and ensuring the stable operation of the system. Attached Figure Description
[0028] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.
[0029] Figure 1 A flowchart illustrating a predictive maintenance method for industrial valve failures provided in an embodiment of the present invention;
[0030] Figure 2 A status report for a predictive maintenance method for industrial valve failures, provided in another embodiment of the present invention;
[0031] Figure 3 A schematic diagram of an industrial valve fault predictive maintenance device provided in an embodiment of the present invention;
[0032] Figure 4 This is a schematic diagram of a computer device provided in an embodiment of this application. Detailed Implementation
[0033] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.
[0034] Example 1
[0035] like Figure 1 As shown in the figure, an industrial valve fault predictive maintenance method proposed in this embodiment of the invention includes the following steps:
[0036] Step S101: Monitor the valve's operating status in real time and collect valve data using a data acquisition server.
[0037] The data acquisition server is used to collect key parameters of industrial valves. The data acquisition server includes an OPC server and an IDM server. The OPC server mainly collects parameters such as loop data and upstream and downstream pressures of the valve; the IDM server is configured with a communication protocol compatible with predictive maintenance systems and collects parameters such as valve stroke feedback values, supply pressure, and actuation pressure.
[0038] Step S102: Preprocess the collected valve data and remove outliers.
[0039] The preprocessing includes using machine learning algorithms to remove outliers from the valve data. In this embodiment, the machine learning algorithm used is OCSVM (One-Class Support Vector Machine). Unlike traditional supervised learning, OCSVM only requires training data from one class to build a model to identify outliers. The goal of this algorithm is to find a hyperplane that can enclose normal data within a region as much as possible, while making the region as small as possible; data points falling outside this region are considered outliers. The specific method is as follows:
[0040] First, a high-dimensional feature space is constructed. The valve data is mapped to the high-dimensional space using the RBF kernel function (radial basis function). The range of the data point set inside the high-dimensional feature space is defined as the first hyperplane. The data point set should contain as many data points as possible inside the hyperplane, and the distance between the nearest data point and the hyperplane should be as large as possible. The boundary of the first hyperplane is set. When the measured data point is outside the boundary of the first hyperplane, it is considered an outlier.
[0041] More specifically, the Radial Basis Function (RBF) of the One-Class Support Vector Machine (OCSVM) algorithm can map low-dimensional, non-linearly separable data to a high-dimensional space, making it linearly separable. Furthermore, the RBF kernel function exhibits smoothness and good robustness when dealing with noisy data. Simultaneously, the value of the RBF kernel function decreases rapidly as the distance between two vectors increases, thus limiting the influence range of each training sample and helping to avoid overfitting. The RBF kernel function takes the following form:
[0042] K(x,x')=exp(-γ||x-x'|| 2 )
[0043] Where x and x' represent the input vectors of two training samples; ||x-x'|| represents the Euclidean distance between the two vectors; K(x,x') is the similarity between the two input vectors; γ is a positive number that controls the width of the kernel function. The smaller the value of γ, the wider the kernel function, and the larger the value of γ, the narrower the kernel function.
[0044] Step S103: Analyze the preprocessed valve data, construct a predictive maintenance model, identify abnormal situations, and predict potential valve failures.
[0045] The predictive maintenance model, based on machine learning algorithms, establishes various diagnostic indicators for valve operating status. Based on embedded valve characteristic curves and combined with upstream and downstream correlation data of normal valves, it automatically determines the likelihood of a fault. Diagnostic indicators include: valve dead zone, valve viscosity, valve deviation, valve tracking performance, valve blockage coefficient, and valve erosion coefficient.
[0046] Specifically, valve dead zone refers to the range of signals during control valve operation where the actual position or response of the control valve does not change immediately when the input signal (usually an electrical or pneumatic signal from the controller) changes. Valve dead zone estimation uses the dead zone estimate as a diagnostic characteristic value, estimating the valve's dead zone value (S) through valve stroke feedback and valve setpoint.
[0047] Valve viscosity measures the degree of obstruction in valve operation caused by friction, rust, deposit accumulation, or other factors. For example, the valve viscosity index is set to 0-1, with 0.25 as the warning threshold and 0.5 as the alarm threshold. 0 represents the optimal value, indicating no viscosity, while 1 represents the worst value, indicating significant viscosity. This index is calculated by weighting the significant oscillation index and the slip index.
[0048] Valve deviation refers to the difference between the actual working state of a valve and its ideal state. For example, the valve deviation index is set to 0-1, with 0.25 as the warning threshold and 0.5 as the alarm threshold, where 0 is the optimal value, indicating no deviation; and 1 is the worst value, indicating a large deviation. This index is calculated by weighting the characteristic values of the average stroke deviation and the integral stroke deviation.
[0049] Valve tracking performance refers to the ability of a valve to follow a given signal (usually the output signal of the controller) in a control system. For example, the valve tracking performance index is set to 0-1, with 0.25 as the warning threshold and 0.5 as the alarm threshold, where 0 is the best, indicating fast and accurate valve stroke tracking; 1 is the worst, indicating poor stroke tracking. This index is calculated by weighting the deviation timeout percentage and the absolute integral deviation.
[0050] The valve blockage coefficient describes the performance of a valve under partial or complete blockage conditions, particularly the degree of efficiency or flow reduction when fluid flow is restricted. The valve scouring coefficient is a quantitative indicator of the degree of erosion or wear to the internal structure of a valve when high-pressure fluid or fluid containing abrasive particles passes through it. The process of obtaining blockage and scouring indices includes: first, selecting the initial valve state and setting valve baseline data; establishing a model between the current valve feedback stroke, upstream and downstream pressures, and loop flow; periodically predicting the current loop flow and comparing it with the current real-time measurement value; obtaining the blockage and scouring indices through deviation analysis of the measured and predicted values; issuing warnings or fault alarms when the indices exceed limits. For example, the valve blockage coefficient index is set to 0-1, with 0.25 as the warning threshold and 0.5 as the alarm threshold, where 0 is optimal, indicating no blockage; 1 is worst, indicating significant blockage. This index is mainly calculated through changes in the valve characteristic model. The valve erosion coefficient index is set to 0-1, with 0.25 as the early warning threshold and 0.5 as the alarm threshold. 0 is the optimal value, indicating no erosion phenomenon, and 1 is the worst value, indicating a significant erosion phenomenon. This index is mainly calculated through the change in the valve characteristic model.
[0051] Furthermore, based on the valve diagnostic indicators output by the predictive maintenance model, the real-time operating status data of the valve is analyzed, including data such as cumulative stroke, changes in the number of actions, and frequency of small opening operations, to predict valve wear trends and potential failures.
[0052] Step S104: For the predicted valve failure, issue an alarm message and provide maintenance suggestions.
[0053] The predictive maintenance model sets dynamic alarm thresholds to provide early warnings for key parameters such as dead zone, viscosity, deviation, tracking performance, and blockage / erosion coefficient. For example, the scoring criteria are set to 0-1, divided into three ranges: normal, abnormal, and faulty. Alarms are issued for valves that are abnormal or faulty. Furthermore, based on predicted valve wear trends and potential faults, preventative maintenance tasks are automatically generated, including inspection, lubrication, and component replacement, with priorities assigned according to valve importance.
[0054] In an optional embodiment, step S105 is further included, which involves remotely viewing the real-time status and maintenance reports of the valve using a remote monitoring platform.
[0055] The remote monitoring platform is equipped with an intuitive dashboard that displays key valve performance indicators, facilitating real-time monitoring by operators. This includes monitoring execution pressure, supply pressure, valve position setpoint, valve position stroke, and circuit data.
[0056] Furthermore, the remote monitoring platform automatically generates and standardizes maintenance reports, facilitating management and decision-making. The standardized reports include a valve status overview, performance trend charts, alarm summaries, and maintenance recommendations. The automated reporting system can be set to perform scheduled tasks, automatically generating and pushing valve health reports to relevant departments and personnel weekly / monthly; the frequency can be increased for key valves.
[0057] Example 2
[0058] This embodiment combines the industrial valve 131_LV_12004B with intelligent locator function with the industrial valve fault prediction maintenance method described in this application to further illustrate the present invention.
[0059] The main components of an intelligent positioner valve include: valve body, valve core, actuator, intelligent positioner, feedback device, connectors, and accessories. The valve body is the basic component, including the inlet, outlet, and intermediate flow passage. The valve core is the internal part of the valve used to control fluid flow, and its size varies depending on the valve type. The actuator drives the valve core to move, changing the valve opening. The intelligent positioner receives control signals and converts them into corresponding output pressure, thereby controlling the actuator's movement. The feedback device monitors the valve's actual position and feeds this position information back to the intelligent positioner. Connectors and accessories include pipe fittings, flanges, bolts, etc., used to connect the valve to the piping system.
[0060] The intelligent positioner consists of a signal interface, a microprocessor, a communication module, sensors, and a drive unit. The intelligent positioner receives input signals from the control system. The microprocessor calculates the required output pressure based on the received signals and controls the actuator's movement by changing the output pressure, thereby adjusting the valve opening. Simultaneously, the intelligent positioner supports the HART protocol, using built-in sensors to monitor the valve's actual position and feeding back the position information to the intelligent positioner, achieving two-way communication.
[0061] The industrial valve fault predictive maintenance method described in this application includes, in the first step, data acquisition, which involves using an OPC server to obtain circuit data of the industrial valve, such as the pressure before and after the valve; and using an IDM server to obtain the stroke feedback value of the industrial valve, the valve positioner setting value, the air supply pressure, and the actuation pressure from the intelligent positioner.
[0062] The second step is data processing, which uses a single-class support vector machine (SVM) algorithm to remove outliers from the valve data. Specifically, the collected valve data is divided into training data and test data, and the radial basis function is used as the kernel function of the single-class SVM. The data processing model is initialized, and key parameters are specified. The trained model is used to predict whether the test data points are outliers. Decision boundaries are drawn to distinguish between normal and outlier data points, and outlier data points are removed.
[0063] The third step is data analysis, which involves building an online diagnostic and fault prediction model for industrial valves. This model analyzes indicators such as valve dead zone, viscosity, deviation, tracking performance, blockage coefficient, and scouring coefficient. Based on the embedded valve characteristic curves and combined with normal upstream and downstream valve correlation data, it automatically determines the likelihood of a fault.
[0064] The steps to obtain the valve dead zone estimate include:
[0065] S1-1, using valve operation data and a clustering algorithm, a graph showing the relationship between valve stroke and controller output data is obtained;
[0066] S1-2, Calculate the value of each local dead zone from the relationship diagram;
[0067] S1-3, make necessary corrections to the calculated local dead zone values to improve accuracy;
[0068] S1-4, By combining the correction values of all local dead zones, the overall valve dead zone value is estimated;
[0069] S1-5 compares the overall dead zone estimate with the dead zone fault threshold to determine whether the valve has a dead zone fault.
[0070] In this embodiment, the estimated dead zone value of the valve is calculated to be 8%*S (S is the valve dead zone value).
[0071] To obtain the viscosity, deviation, and tracking performance, it is necessary to analyze the valve deviation value based on the valve positioner data and through feature extraction algorithms.
[0072] Specifically, the formula for calculating the above valve data is as follows:
[0073] Viscosity = W1*I1 + W6*I6 + W st *TD_std
[0074] Tracking performance = W4*I4 + W5*I5 + W2*I2
[0075] Degree of deviation = W4*I4 + W5*I5 + W3*I3
[0076] Where W represents the weight, which is customized according to the specific requirements of the valve application scenario;
[0077] I1 represents the significant oscillation index: the average number of times the travel deviation exceeds the acceptable range (normalized to 1 hour);
[0078] I2 indicates timeout percentage: the percentage of time TD exceeds its acceptable range;
[0079] I3 represents the average travel deviation: the average value of TD;
[0080] I4 indicates integral travel deviation: TD integral (normalized to 1 hour);
[0081] I5 represents the absolute integral travel deviation: the integral of the absolute value of TD (standardized for 1 hour);
[0082] I6 represents the blockage index: the number of times the valve sticks and slides, excluding peak values caused by changes in the setpoint (normalized to 1 hour);
[0083] TD represents the valve stroke deviation (the set value minus the valve stroke feedback), and the valve stroke feedback is read directly from the intelligent positioner; TD_std is the variance of the stroke deviation.
[0084] In this embodiment, the six key indicators based on the deviation value include: significant oscillation index of 3.29; timeout percentage of 47.95%; average stroke deviation of -1.60%; integral stroke deviation of -5765.56%*s; absolute integral stroke deviation of 10632.28%*s; and blockage index of 0.
[0085] Therefore, according to the above formula, the valve viscosity index is calculated by weighting the significant oscillation index and the blockage index, and is 0.54. The valve deviation index is calculated by weighting the average stroke deviation and the integral stroke deviation, and is 0.34; the valve tracking performance index is calculated by weighting the absolute integral stroke deviation and the timeout percentage, and is 0.88.
[0086] The clogging coefficient is used to diagnose valve clogging. The scouring coefficient is used to diagnose valve scouring. In this embodiment, both the clogging index and the scouring index are 0.
[0087] Fourthly, in this embodiment, after analyzing the above indicators, it is found that the valve viscosity index of 0.54 exceeds the fault limit, and a fault alarm is issued. Furthermore, based on the alarm information, a preventative maintenance task is automatically generated. Analysis determines that the valve has a viscosity fault, and the maintenance recommendations are to check the valve core movement, disassemble the valve, and refritch the sealing surfaces of the valve core and valve seat.
[0088] Furthermore, the remote monitoring platform generates valve status reports based on the aforementioned modules, as detailed below. Figure 3 This facilitates remote management and decision-making by users.
[0089] Example 3
[0090] This application provides an embodiment of an industrial valve fault prediction maintenance device capable of implementing Embodiment 1, such as... Figure 3 As shown, it includes:
[0091] The data acquisition module is used to monitor the valve's operating status in real time and collect valve data using the data acquisition server.
[0092] The data processing module is used to preprocess the collected valve data and remove outliers;
[0093] The data analysis module is used to analyze the preprocessed valve data, build predictive maintenance models, identify abnormal situations, and predict potential valve failures.
[0094] The fault early warning module is used to issue alarm information and provide maintenance suggestions for predicted valve faults.
[0095] Example 4
[0096] This application also provides a computer device, including a memory and a processor. The memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the processor performs the industrial valve failure predictive maintenance method as described above.
[0097] Example 5
[0098] This application also provides a storage medium storing computer-readable instructions, which, when executed by one or more processors, cause the one or more processors to perform the industrial valve failure predictive maintenance method as described above.
[0099] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings, but the present invention is not limited to the above embodiments. Even if various changes are made to the present invention, if these changes fall within the scope of the claims of the present invention and their equivalents, they shall still fall within the protection scope of the present invention.
Claims
1. A predictive maintenance method for industrial valve failures, characterized in that, include: Real-time monitoring of valve operating status; valve data is collected using a data acquisition server. The collected valve data is preprocessed to remove outliers; Analyze the preprocessed valve data to build a predictive maintenance model, identify abnormal situations, and predict potential valve failures; For predicted valve malfunctions, an alarm message is issued and maintenance suggestions are provided.
2. The industrial valve fault predictive maintenance method according to claim 1, characterized in that, The data acquisition server is used to collect key parameters of industrial valves. The data acquisition server includes an OPC server and an IDM server. The parameters collected by the OPC server include at least loop data and upstream and downstream pressures of the valve. The IDM server is configured with a communication protocol compatible with predictive maintenance systems and collects at least valve stroke feedback values, air supply pressure, and actuation pressure.
3. The industrial valve fault predictive maintenance method according to claim 1, characterized in that, The preprocessing of the collected valve data to remove outliers includes using machine learning algorithms to remove valve data outliers. The machine learning algorithm includes at least the OCSVM algorithm: First, a high-dimensional feature space is constructed, and the valve data is mapped to the high-dimensional space using the RBF kernel function. The range of the data point set inside the high-dimensional feature space is determined as the first hyperplane, and the boundary of the first hyperplane is set. When the measured data point is outside the boundary of the first hyperplane, the measured data point is considered to be an outlier. The RBF kernel function is in the following form: K(x,x')=exp(-γ||x-x'|| 2 ) Where x and x' represent two input vectors; ||x-x'|| represents the Euclidean distance between the two vectors; K(x,x') is the similarity between the two input vectors; γ is a positive number that controls the width of the kernel function. The smaller the value of γ, the wider the kernel function, and the larger the value of γ, the narrower the kernel function.
4. The industrial valve fault predictive maintenance method according to claim 1, characterized in that, The predictive maintenance model includes the use of machine learning algorithms to establish various diagnostic indicators for valve operating status. These diagnostic indicators include at least: valve dead zone, valve viscosity, valve deviation, valve tracking performance, valve blockage coefficient, and valve scouring coefficient.
5. The industrial valve fault predictive maintenance method according to claim 1, characterized in that, The process of analyzing the preprocessed valve data to construct a predictive maintenance model, identify abnormal situations, and predict potential valve failures also includes analyzing the real-time operating status data of the valve based on the valve diagnostic indicators output by the predictive maintenance model, including at least cumulative stroke, changes in the number of actions, and frequency of small opening operations, to predict valve wear trends and potential failures.
6. The industrial valve fault predictive maintenance method according to claim 1, characterized in that, The maintenance recommendations include automatically generating preventative maintenance tasks, which at least include inspection, lubrication, and component replacement, and prioritizing them according to the importance of the valves.
7. The industrial valve fault predictive maintenance method according to claim 1, characterized in that, Also includes: The real-time status and maintenance reports of valves can be viewed remotely using a remote monitoring platform. The real-time status of the valves refers to their key performance indicators, including at least the actuation pressure, supply pressure, valve position setpoint, valve position stroke, and loop data. The maintenance reports include at least a valve status overview, performance trend graph, alarm summary, and maintenance suggestions. The maintenance reports are generated and pushed out periodically according to a preset cycle.
8. A predictive maintenance device for industrial valve failures, characterized in that, include: The data acquisition module is used to monitor the valve's operating status in real time and collect valve data using the data acquisition server. The data processing module is used to preprocess the collected valve data and remove outliers; The data analysis module is used to analyze the preprocessed valve data, build predictive maintenance models, identify abnormal situations, and predict potential valve failures. The fault early warning module is used to issue alarm information and provide maintenance suggestions for predicted valve faults.
9. An electronic device, characterized in that, The system includes a memory and a processor, wherein the memory stores computer-readable instructions that, when executed by the processor, cause the processor to perform the industrial valve failure predictive maintenance method as described in any one of claims 1 to 7.
10. A storage medium storing computer-readable instructions, characterized in that, When the computer-readable instructions are executed by one or more processors, the one or more processors perform the industrial valve failure predictive maintenance method as described in any one of claims 1 to 7.