An intelligent valve adaptive sealing method and system based on operating condition perception

Abstract

An intelligent valve adaptive sealing method based on operating condition perception, belonging to the field of industrial process control and fluid transportation technology, includes the following steps: First, multi-dimensional operating condition perception and data preprocessing; second, sealing requirement modeling based on physical models and data-driven approaches; third, sealing force setpoint optimization based on intelligent decision-making strategies; fourth, closed-loop control and execution; and fifth, sealing status health diagnosis and early warning. The intelligent valve adaptive sealing method is implemented through an intelligent valve adaptive sealing system, including an operating condition perception and preprocessing module, a sealing requirement calculation module, a sealing setpoint optimization module, an intelligent control and precision execution module, and a status diagnosis and early warning module. This invention can adjust the sealing force setpoint in advance according to predicted changes in operating conditions, realizing dynamic, real-time adaptive setting, prediction, and closed-loop optimization of valve sealing force; significantly extending valve life, while simultaneously achieving real-time transparency and diagnostic early warning of valve sealing status.

Description

Technical Field

[0001] This invention belongs to the field of industrial process control and fluid transport technology, and relates to an intelligent valve adaptive sealing method and system based on operating condition perception. Background Technology

[0002] As a key control unit in fluid transport systems, valves' core functions are to cut off, regulate, guide, or prevent backflow of media. They are widely used in almost all industrial fields, including petroleum, natural gas, chemical, power, metallurgy, water treatment, pharmaceuticals, and aerospace. The sealing performance of a valve, as its most important technical performance indicator, directly determines the safety, reliability, energy efficiency, and environmental friendliness of the entire production system. Valve leakage not only causes media loss, energy waste, and environmental pollution, but in environments with flammable, explosive, toxic, or radioactive media, it can also lead to catastrophic accidents.

[0003] Valve sealing technology has been extensively studied. Gu Changming et al. proposed a double-acting throttling valve sealing structure (Chinese Invention Patent CN202410015604.7). However, their invention method is clearly passive and static; the compensation mechanism is fixed and unadjustable. Its sealing effect heavily depends on the pressure and flow of the medium, making dynamic adjustment impossible based on actual operating conditions. Furthermore, it lacks the ability to perceive its own sealing state and cannot handle complex situations. Zhang Zepeng et al. proposed a cloud computing-based valve status monitoring system (Chinese Invention Patent CN202510476648.4). However, their invention method suffers from a disconnect between perception and decision-making. Its decision-making is based on indirect, macroscopic data, making it difficult to accurately model, predict, and adjust the valve sealing state, and it lacks adaptive execution capabilities.

[0004] Based on existing research, valve sealing technology has the following significant drawbacks: lack of sensing capability or superficial application, or even direct over-sealing treatment, making it impossible to directly and online obtain the microscopic physical information of the working surface of the sealing pair; insufficient decision-making capability or disconnection from the physical process, making it impossible to achieve prevention and adaptive adjustment before valve leakage; passive, preset or fixed execution mechanism, limited compensation capability, and inability to form online, accurate, and dynamic closed-loop control based on real-time sensing status.

[0005] Therefore, in order to overcome the shortcomings of existing technologies, it is necessary to propose an intelligent valve adaptive sealing method and system based on working condition perception. Summary of the Invention

[0006] To address the problems and shortcomings of existing technologies, this invention provides an intelligent valve adaptive sealing method and system based on operating condition perception. By deploying a multi-sensor array on the valve body to perceive and preprocess multi-dimensional data of the valve's operating conditions in real time, and dynamically calculating the minimum and maximum allowable sealing forces required to meet the current sealing level based on an established sealing demand model, the optimal sealing force setpoint is determined using an intelligent decision-making strategy. A force servo closed-loop control system is then constructed to precisely control the actuator, ensuring that the actual sealing force accurately tracks the setpoint and initiating health diagnostics and leakage warnings. This constructs a complete intelligent valve adaptive sealing closed loop integrating perception, decision-making, execution, and early warning, enabling the valve to adaptively adjust the sealing force according to changes in operating conditions, ultimately significantly extending valve life while ensuring absolute sealing reliability.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0008] An adaptive sealing method for intelligent valves based on operating condition perception, the adaptive sealing method for intelligent valves includes the following steps:

[0009] The first step is multi-dimensional operating condition perception and data preprocessing. A multi-sensor array is deployed to collect various parameters in real time, including valve stem axial force. The collected data is then optimized and fused to obtain the operating condition state vector. The details are as follows:

[0010] Step 1.1, Real-time Data Acquisition. Since the valve's operating conditions constitute a multi-dimensional parameter space, sensor arrays deployed at key locations on the valve are used to acquire parameters in real time, obtaining sensor data. These key locations include the valve body, valve chamber, valve stem, and actuator, as detailed below:

[0011] A high-precision pressure transmitter is installed on the valve body at the upstream inlet and downstream outlet ends to calculate the pressure difference across the valve seat and determine the leakage of the medium.

[0012] Temperature sensors are installed inside the valve chamber and on the actuator to measure the temperature of the medium and the ambient temperature.

[0013] A strain gauge force sensor is installed at the connection between the actuator push rod and the valve stem to accurately measure the axial force of the valve stem.

[0014] Installing acoustic emission sensors on the outside of the valve body and as close as possible to the monitored sealing surfaces to monitor the micro-friction of the sealing surfaces, turbulent noise generated by leakage flow, and impact signals generated by particles passing through the sealing surfaces is key to achieving early leakage warning and wear assessment.

[0015] Step 1.2, Data Optimization. Based on the sensor data obtained in Step 1.1, the sensor data is optimized through time synchronization and linear interpolation.

[0016] Step 1.3, Data Fusion. Based on the optimized data obtained in Step 1.2, the Kalman filter algorithm is used to output the optimal state estimate, generating a high-confidence, unified working condition state vector. :

[0017] (1)

[0018] in, It is the real-time pressure difference across the valve seat. It is the temperature of the medium. It is the valve stem axial force collected in real time. It is the root mean square value of the acoustic emission signal. For any time.

[0019] The second step involves modeling the sealing requirements based on a physical model and data-driven approaches. This is based on the operating condition vector obtained in the first step. Dynamically calculate the minimum theoretical sealing force required to achieve the desired sealing level (such as zero leakage) under the current operating conditions. And the maximum permissible sealing force to ensure that the seals do not suffer permanent damage. The details are as follows:

[0020] Step 2.1, Modeling the minimum theoretical sealing force. The minimum theoretical sealing force... The lower limit required to achieve a reliable seal under current operating conditions :

[0021] (2)

[0022] (3)

[0023] (4)

[0024] (5)

[0025] (6)

[0026] in, It is the component force required to overcome the pressure of the medium. It is the component force required to overcome the surface micro-roughness. It is the friction force of valve opening and closing. It is thermal compensation force. It is the effective actual contact area of ​​the sealing pair. It is the sealing pressure coefficient. It is surface bonding stress. It is the coefficient of friction. It is the thermal compensation coefficient. It is the temperature at which the valve is initially pre-tightened (reference temperature).

[0027] Step 2.2, Modeling the maximum permissible sealing force. The maximum permissible sealing force... Used to protect seals from plastic flow, creep, or structural damage caused by excessive compression:

[0028] (7)

[0029] in, The yield strength of the sealing material at the current temperature. This is for the safety factor.

[0030] Step 2.3, Set the sealing window. Based on the minimum theoretical sealing force obtained in Step 2.1. The maximum permissible sealing force obtained in step 2.2 This provides a safe sealing working window for dynamically tracking the ever-updating sealing force. :

[0031] (8)

[0032] The third step is to optimize the sealing force setpoint based on an intelligent decision-making strategy. This is based on the minimum theoretical sealing force obtained in the second step. With maximum permissible sealing force By selecting optimization targets and employing intelligent decision-making strategies, and through continuous optimization of the sealing force setpoint, adaptive setting and optimization of the sealing force are achieved. Specifically:

[0033] Step 3.1, Determine the optimization objectives. Selecting the sealing force setpoint within the safe sealing working window obtained in Step 2 is a multi-objective optimization problem. The optimization objectives are determined, including reliability, lifespan, energy consumption, and control stability objectives. Specifically:

[0034] The reliability objective is to ensure that the sealing force is always greater than the minimum theoretical sealing force. To prevent leakage.

[0035] The lifespan target is to minimize sealing force in order to reduce wear and plastic deformation of the sealing pair and extend valve life.

[0036] The energy consumption target is to minimize the sealing force in order to reduce the energy consumption of the actuator.

[0037] The control stability objective is to avoid drastic changes in the sealing force setpoint to ensure the stability of the control system.

[0038] Because there are conflicting objectives, such as maximizing sealing force for high reliability and minimizing sealing force for long life and low energy consumption, the key to the multi-objective optimization problem is to find a balance point that is the optimal sealing force setpoint.

[0039] Step 3.2, Select an intelligent decision-making strategy. Based on the optimization objective determined in Step 3.1, a hierarchical decision-making structure is adopted to select an appropriate strategy for calculation, in order to determine the optimal sealing force setpoint under different operating states and performance requirements. Specifically:

[0040] The hierarchical decision structure includes two types: reactive basic rule-based decision-making and predictive advanced optimization decision-making, as detailed below:

[0041] The first type: Reactive basic rule decision-making is the most direct and computationally least computationally intensive decision-making method, reflecting the system's fast response logic. Its strategy is a dynamic margin adaptive strategy, mainly based on the dynamic margin method of acoustic emission feedback.

[0042] (9)

[0043] in, It is a dynamic safety margin. It is the sealing force set point. It is the minimum theoretical sealing force; its adaptive law is as follows:

[0044] First, by conducting long-term statistical analysis of the acoustic emission (AE) signal of a healthy valve under good sealing conditions, an ideal range for acoustic emission intensity can be established. ,in As the lower limit of the ideal range, This represents the upper limit of the ideal range. Then, the adaptive strategy is iterated repeatedly until the optimal sealing state is achieved, at which point the corresponding dynamic safety margin is output.

[0045] like This indicates that the system may be over-sealed, and the system should be reduced at a certain rate. :

[0046] (10)

[0047] like This indicates that the sealing condition is optimal. Just leave it as is.

[0048] like This indicates a risk of leakage or excessive friction, and the system should be increased at a certain rate. :

[0049] (11)

[0050] in, The step size ranges from 0.1 kN to 0.5 kN.

[0051] The dynamic margin adaptive strategy realizes closed-loop optimization based on sealing state feedback, which can automatically find the minimum dynamic safety margin required to maintain the seal, and finally determine an optimal sealing force setpoint at the current moment.

[0052] The second type, predictive advanced optimization decision-making, is mainly used in scenarios with drastic changes in operating conditions or high performance requirements. It has a more advanced and complex model predictive control strategy, capable of predicting changes in operating conditions over a period of time and solving optimization problems to obtain the optimal sealing force setpoint at the current moment, as detailed below:

[0053] First, a discrete-time model describing the system's dynamics is established. State variables... Control input Interference input Therefore, the discrete-time model can be simplified to:

[0054] (12)

[0055] Where k represents a certain moment; A is the state variable setting parameter; and B is the control input conversion parameter, with dimensions [missing information]. C is the interference input transformation matrix.

[0056] Then, the optimization problem is solved. Essentially, this involves solving a quadratic programming problem at each time step k to obtain a future control sequence. As shown in formula (13):

[0057] (13)

[0058] Satisfy constraints:

[0059] (14)

[0060] (15)

[0061] (16)

[0062] in, Optimize the function for the objective; The axial force of the valve stem at time k+i is predicted at time k. The minimum theoretical sealing force predicted at time k+i is given at time k. The maximum allowable sealing force at time k+i is predicted at time k. The weighting coefficients used to balance the various optimization objectives have the following dimensions: ; The predicted time domain length in a single optimization; , These are the minimum and maximum constraints of the input, respectively; Input variation constraints.

[0063] The optimal control sequence was obtained after solving the problem. Take the first value in the sequence and convert it with the control input parameter. Multiply, and use as the current sealing force setpoint. As shown in formula (17):

[0064] (17)

[0065] Then, the process proceeds to the next sampling time k+1 for remeasurement and optimization, repeating this cycle and performing rolling optimization until it is forcibly interrupted by an external command or system failure.

[0066] Model predictive control strategies can adjust the sealing force setpoint in advance based on predicted changes in operating conditions, achieving dynamic and real-time adaptive prediction and optimization, and ultimately determining an optimal sealing force setpoint for the current moment.

[0067] The fourth step is closed-loop control and execution. Based on the sealing force setpoint obtained in the third step and the valve stem axial force acquired in real time in the first step, the force control deviation between the optimal sealing force setpoint and the actual value of the valve stem axial force is calculated. A closed-loop control algorithm is then used to obtain the control quantity, which drives the actuator to perform actions. This ultimately constructs a fast and accurate force servo closed-loop control system, ensuring that the intelligent decision-making strategy selected in the third step is accurately and reliably translated into the actual sealing effect, achieving real-time, adaptive adjustment of the sealing force. Specifically:

[0068] Step 4.1: Based on the sealing force setpoint obtained in step 3, receive the sealing force setpoint at the optimized stopping time and use it as the optimal sealing force setpoint. .

[0069] Step 4.2: Based on the valve stem axial force collected in real time in the first step, obtain the actual value of the valve stem axial force at this time. .

[0070] Step 4.3, calculate the force control deviation. Based on the optimal sealing force setpoint obtained in Step 4.1. And the actual value of the valve stem axial force obtained in step 4.2 Calculate the force control deviation between the optimal sealing force setpoint and the actual value of the valve stem axial force:

[0071] (18)

[0072] Step 4.4: Execute the closed-loop control algorithm. Based on the force control deviation obtained in Step 4.3, the required control quantity is calculated using a PID control algorithm.

[0073] Step 4.5, Action Execution. Based on the control quantity obtained in Step 4.4, it is converted into an analog signal to drive the actuator to perform actions, generating a corresponding axial force, which is transmitted to the sealing pair through the valve stem to generate sealing force.

[0074] Step 4.6, Adaptive Adjustment and Closed-Loop Formation. Based on the sealing force generated in Step 4.5, the force tracking error and system status are continuously monitored, and adjustments are made as necessary. Simultaneously, since the output sealing force affects the valve status, this change is detected again by the sensor array from Step 1, forming a closed loop, thus obtaining the force servo closed-loop control system.

[0075] Step 5: Sealing condition health diagnosis and early warning. Based on the force servo closed-loop control system constructed in Step 4 and the parameters collected by the multi-sensor array in Step 1, online assessment of the wear condition of the sealing pair and leakage early warning through multi-source information fusion are performed through active excitation monitoring and wear quantification, as detailed below:

[0076] Step 5.1, Active Excitation Monitoring. Based on the force servo closed-loop control system constructed in step four, as the number of valve opening and closing and the number of times the valve stem transmits sealing force increases, the sealing pair will wear, its mechanical properties will change, leading to an increase in the equivalent contact stiffness of the sealing pair. Therefore, the force servo closed-loop control system should periodically execute a low-amplitude, high-frequency main force excitation. For ease of analysis and control, a sine wave with an amplitude of 3 kN and a frequency of 10 Hz is used for the main force excitation. The excitation is then continued for 1 second, during which the valve stem displacement signal is acquired to obtain the valve stem displacement response amplitude. This excitation cycle is repeated 4-5 times, and the average value of the valve stem displacement response amplitude is taken. Finally, the current equivalent contact stiffness is obtained by calculating the ratio of the main force excitation amplitude to the averaged valve stem displacement response amplitude. .

[0077] Step 5.2, Wear Quantification. Based on the data obtained in Step 5.1 Define a wear index :

[0078] (19)

[0079] in, It is the reference value of contact stiffness when the valve is put into operation and is in a brand new condition.

[0080] Step 5.3, Online assessment of the wear condition of the sealing pair. Based on the wear index defined in Step 5.2. It can perform health status classification and early warning for sealing components:

[0081] when At this time, the wear is slight, it can operate normally, and the status level is "healthy".

[0082] when When wear enters a stable period, monitor the wear trend, prepare spare parts, and the status level is "Caution".

[0083] when When wear and tear accelerates, it is recommended to plan a shutdown for maintenance, and the status level is "warning".

[0084] when When the sealing pair is about to fail, immediate maintenance is required, and the status level is "alarm".

[0085] in, For custom parameters, For health threshold, To pay attention to the threshold; This is the failure threshold for the sealing pair.

[0086] Step 5.4, Leakage Warning. Based on the parameters collected by the multi-sensor array in Step 1, the DS evidence theory is used to perform multi-data fusion, output the final comprehensive confidence level, and initiate a tiered warning response: low-confidence warnings only display prompt information to attract attention, while high-confidence warnings require immediate intervention for confirmation and handling.

[0087] An intelligent valve adaptive sealing system based on operating condition perception is disclosed. This system implements the aforementioned intelligent valve adaptive sealing method based on operating condition perception. The intelligent valve adaptive sealing system includes an operating condition perception and preprocessing module, a sealing requirement calculation module, a sealing setting optimization module, an intelligent control and precision execution module, and a status diagnosis and early warning module. Specifically:

[0088] The operating condition sensing and preprocessing module consists of a multi-sensor array and a data acquisition and processing system. By deploying the multi-sensor array at key locations on the valve, the data acquisition and processing system synchronously collects physical monitoring data and performs preprocessing to ensure data integrity and accuracy. At the same time, it provides high-quality input data for subsequent modules and serves as the sensing hub of the entire intelligent valve adaptive sealing system.

[0089] The sealing requirement calculation module is connected to the working condition perception and preprocessing module. It consists of a high-performance computing engine and has built-in calculation models for the minimum theoretical sealing force and the maximum allowable sealing force for dynamic calculation. It is the calculation core of the entire intelligent valve adaptive sealing system.

[0090] The sealing setting optimization module is connected to the sealing demand calculation module and the working condition perception and preprocessing module. It consists of a high-performance industrial computing platform and is used to run optimization algorithms and select decision strategies to calculate the optimal sealing force setpoint. It is the decision center of the entire intelligent valve adaptive sealing system.

[0091] The intelligent control and precision execution module is connected to the sealing setting optimization module, the sealing demand calculation module, and the working condition perception and preprocessing module. It consists of a servo controller and a force servo actuator, and is used to implement the closed-loop control algorithm and convert the control quantity into physical sealing force. It is the action realization of the entire intelligent valve adaptive sealing system.

[0092] The status diagnosis and early warning module is connected to the intelligent control and precision execution module, the sealing setting optimization module, the sealing demand calculation module, and the working condition perception and preprocessing module. It consists of a dedicated high-performance server and is used to realize online assessment of the wear status of the sealing pair and leakage early warning based on multi-source information fusion. It is the safety barrier of the entire intelligent valve adaptive sealing system.

[0093] The beneficial effects of this invention are as follows:

[0094] (1) This invention uses a multi-sensor array deployed on the valve body to perceive multi-dimensional data of valve operating conditions in real time and preprocess it. Based on the established sealing demand model, it dynamically calculates the minimum and maximum allowable sealing force required to meet the current sealing level. It uses intelligent decision-making strategy to determine the optimal sealing force setting point and can adjust the sealing force setting point in advance according to the predicted changes in operating conditions. This realizes the dynamic and real-time adaptive setting, prediction and closed-loop optimization of valve sealing force.

[0095] (2) The present invention constructs a force servo closed-loop control system to perform precise closed-loop control of the actuator, so that the actual sealing force accurately tracks the set point and initiates health diagnosis and leakage warning, thereby constructing a complete intelligent valve adaptive sealing closed loop integrating perception, decision-making, execution and warning. Ultimately, it significantly extends the valve life under the premise of ensuring absolute sealing reliability, and realizes real-time transparency and diagnostic warning of valve sealing status, providing a guarantee for predictive maintenance and management, so that the valve changes from "passive sealing" to "active sealing", and from "blind over-sealing" to "precise on-demand sealing", achieving a unified improvement in reliability and life. Attached Figure Description

[0096] Figure 1 This is a schematic diagram of a smart valve adaptive sealing method based on working condition perception according to the present invention.

[0097] Figure 2 This is a schematic diagram of an intelligent valve adaptive sealing system based on working condition perception according to the present invention. Detailed Implementation

[0098] The invention will be further illustrated below with a specific implementation example of a high-pressure, small-diameter, fully welded ball valve used in a long-distance natural gas pipeline.

[0099] An adaptive sealing method for intelligent valves based on operating condition perception includes the following steps:

[0100] The first step is multi-dimensional condition sensing and data preprocessing. A multi-sensor array is deployed to collect various parameters in real time, including valve stem axial force, and the collected data is then optimized and fused. Specifically:

[0101] Step 1.1, Real-time Data Acquisition. Since the valve operates in a multi-dimensional parameter space, sensor arrays deployed at key locations on the valve are used to acquire parameters in real time, obtaining sensor data. These key locations mainly include the valve body, valve chamber, valve stem, and actuator, as detailed below:

[0102] A high-precision pressure transmitter is installed on the valve body at the upstream inlet and downstream outlet ends to calculate the pressure difference across the valve seat and determine the leakage of the medium.

[0103] Temperature sensors are installed inside the valve chamber and on the actuator to measure the temperature of the medium and the ambient temperature.

[0104] A strain gauge force sensor is installed at the connection between the actuator push rod and the valve stem to accurately measure the axial force of the valve stem.

[0105] Installing acoustic emission sensors on the outside of the valve body and as close as possible to the monitored sealing surfaces to monitor the micro-friction of the sealing surfaces, turbulent noise generated by leakage flow, and impact signals generated by particles passing through the sealing surfaces is key to achieving early leakage warning and wear assessment.

[0106] Step 1.2, Data Optimization. Based on the sensor data obtained in Step 1.1, the sensor data is optimized through time synchronization and linear interpolation.

[0107] Step 1.3, Data Fusion. Based on the optimized data obtained in Step 1.2, the Kalman filter algorithm is used to output the optimal state estimate, generating a high-confidence, unified working condition state vector. :

[0108] (1)

[0109] in, It is the real-time pressure difference across the valve seat. It is the temperature of the medium. It is the valve stem axial force collected in real time. It is the root mean square value of the acoustic emission signal. For any time.

[0110] The second step involves modeling the sealing requirements based on a physical model and data-driven approaches. This is based on the operating condition vector obtained in the first step. Dynamically calculate the minimum theoretical sealing force required to achieve the desired sealing level (such as zero leakage) under the current operating conditions. And the maximum permissible sealing force to ensure that the seals do not suffer permanent damage. The details are as follows:

[0111] Step 2.1, Modeling the minimum theoretical sealing force. The minimum theoretical sealing force... The lower limit required to achieve a reliable seal under current operating conditions :

[0112] (2)

[0113] (3)

[0114] (4)

[0115] (5)

[0116] (6)

[0117] in, It is the component force required to overcome the pressure of the medium. It is the component force required to overcome the surface micro-roughness. It is the friction force of valve opening and closing. It is thermal compensation force. It is the effective actual contact area of ​​the sealing pair. It is the sealing pressure coefficient. It is surface bonding stress. It is the coefficient of friction. It is the thermal compensation coefficient. It is the temperature at which the valve is initially pre-tightened (reference temperature).

[0118] In this embodiment, , , , , .

[0119] Step 2.2, Modeling the maximum permissible sealing force. The maximum permissible sealing force... Used to protect seals from plastic flow, creep, or structural damage caused by excessive compression:

[0120] (7)

[0121] in, The yield strength of the sealing material at the current temperature. This is for the safety factor.

[0122] In this embodiment, , .

[0123] Step 2.3, Set the sealing window. Based on the minimum theoretical sealing force obtained in Step 2.1. The maximum permissible sealing force obtained in step 2.2 This provides a safe sealing working window for dynamically tracking the ever-updating sealing force. :

[0124] (8)

[0125] The third step is to optimize the sealing force setpoint based on an intelligent decision-making strategy. This is based on the minimum theoretical sealing force obtained in the second step. With maximum permissible sealing force By selecting optimization targets and employing intelligent decision-making strategies, and through continuous optimization of the sealing force setpoint, adaptive setting and optimization of the sealing force are achieved. Specifically:

[0126] Step 3.1, Determine the optimization objectives. Selecting the sealing force setpoint within the safe sealing working window obtained in Step 2 is a multi-objective optimization problem. The optimization objectives are determined, including reliability, lifespan, energy consumption, and control stability objectives. Specifically:

[0127] The reliability objective is to ensure that the sealing force is always greater than the minimum theoretical sealing force. To prevent leakage.

[0128] The lifespan target is to minimize sealing force in order to reduce wear and plastic deformation of the sealing pair and extend valve life.

[0129] The energy consumption target is to minimize the sealing force in order to reduce the energy consumption of the actuator.

[0130] The control stability objective is to avoid drastic changes in the sealing force setpoint to ensure the stability of the control system.

[0131] Because there are conflicting objectives, such as maximizing sealing force for high reliability and minimizing sealing force for long life and low energy consumption, the key to the multi-objective optimization problem is to find a balance point that is the optimal sealing force setpoint.

[0132] Step 3.2, Select an intelligent decision-making strategy. Based on the optimization objective determined in Step 3.1, a hierarchical decision-making structure is adopted to select an appropriate strategy for calculation, in order to determine the optimal sealing force setpoint under different operating states and performance requirements. Specifically:

[0133] The hierarchical decision structure includes two types: reactive basic rule-based decision-making and predictive advanced optimization decision-making, as detailed below:

[0134] The first type: Reactive basic rule decision-making is the most direct and computationally least computationally intensive decision-making method, reflecting the system's fast response logic. Its strategy is a dynamic margin adaptive strategy, mainly based on the dynamic margin method of acoustic emission feedback.

[0135] (9)

[0136] in, It is a dynamic safety margin. It is the sealing force set point. It is the minimum theoretical sealing force; its adaptive law is as follows:

[0137] First, by conducting long-term statistical analysis of the acoustic emission (AE) signal of a healthy valve under good sealing conditions, an ideal range for acoustic emission intensity is established. ,in As the lower limit of the ideal range, This represents the upper limit of the ideal range. Then, the adaptive strategy is iterated repeatedly until the optimal sealing state is achieved, at which point the corresponding dynamic safety margin is output.

[0138] like This indicates that the system may be over-sealed, and the system should be reduced at a certain rate. :

[0139] (10)

[0140] like This indicates that the sealing condition is optimal. Just leave it as is.

[0141] like This indicates a risk of leakage or excessive friction, and the system should be increased at a certain rate. :

[0142] (11)

[0143] in, The step size ranges from 0.1 kN to 0.5 kN.

[0144] The dynamic margin adaptive strategy realizes closed-loop optimization based on sealing state feedback, which can automatically find the minimum dynamic safety margin required to maintain the seal, and finally determine an optimal sealing force setpoint at the current moment.

[0145] The second type, predictive advanced optimization decision-making, is mainly used in scenarios with drastic changes in operating conditions or high performance requirements. It has a more advanced and complex model predictive control strategy, capable of predicting changes in operating conditions over a period of time and solving optimization problems to obtain the optimal sealing force setpoint at the current moment, as detailed below:

[0146] First, a discrete-time model describing the system's dynamics is established. State variables... Control input Interference input Therefore, the discrete-time model can be simplified to:

[0147] (12)

[0148] Where k represents a certain moment; A is the state variable setting parameter; and B is the control input conversion parameter, with dimensions [missing information]. C is the interference input transformation matrix.

[0149] Then, the optimization problem is solved. Essentially, this involves solving a quadratic programming problem at each time step k to obtain a future control sequence. As shown in formula (13):

[0150] (13)

[0151] Satisfy constraints:

[0152] (14)

[0153] (15)

[0154] (16)

[0155] in, Optimize the function for the objective; The axial force of the valve stem at time k+i is predicted at time k. The minimum theoretical sealing force predicted at time k+i is given at time k. The maximum allowable sealing force at time k+i is predicted at time k. The weighting coefficients used to balance the various optimization objectives have the following dimensions: ; The predicted time domain length in a single optimization; , These are the minimum and maximum constraints of the input, respectively; Input variation constraints.

[0156] The optimal control sequence was obtained after solving the problem. Take the first value in the sequence and convert it with the control input parameter. Multiply, and use as the current sealing force setpoint. As shown in formula (17):

[0157] (17)

[0158] Then, the process proceeds to the next sampling time k+1 for remeasurement and optimization, repeating this cycle and performing rolling optimization until it is forcibly interrupted by an external command or system failure.

[0159] Model predictive control strategies can adjust the sealing force setpoint in advance based on predicted changes in operating conditions, achieving dynamic and real-time adaptive prediction and optimization, and ultimately determining an optimal sealing force setpoint for the current moment.

[0160] In this embodiment, a predictive advanced optimization decision-making and model predictive control strategy is selected. , , , , (Since the sampling period is 100ms, the prediction time domain is 2s). , , .

[0161] The fourth step is closed-loop control and execution. Based on the sealing force setpoint obtained in the third step and the valve stem axial force acquired in real time in the first step, the force control deviation between the optimal sealing force setpoint and the actual value of the valve stem axial force is calculated. A closed-loop control algorithm is then used to obtain the control quantity, which drives the actuator to perform actions. This ultimately constructs a fast and accurate force servo closed-loop control system, ensuring that the intelligent decision-making strategy selected in the third step is accurately and reliably translated into the actual sealing effect, achieving real-time, adaptive adjustment of the sealing force. Specifically:

[0162] Step 4.1: Based on the sealing force setpoint obtained in step 3, receive the sealing force setpoint at the optimized stopping time and use it as the optimal sealing force setpoint. .

[0163] Step 4.2: Based on the valve stem axial force collected in real time in the first step, obtain the actual value of the valve stem axial force at this time. .

[0164] Step 4.3, calculate the force control deviation. Based on the optimal sealing force setpoint obtained in Step 4.1. And the actual value of the valve stem axial force obtained in step 4.2 Calculate the force control deviation between the optimal sealing force setpoint and the actual value of the valve stem axial force:

[0165] (18)

[0166] Step 4.4: Execute the closed-loop control algorithm. Based on the force control deviation obtained in Step 4.3, the required control quantity is calculated using a PID control algorithm.

[0167] Step 4.5, Action Execution. Based on the control quantity obtained in Step 4.4, it is converted into an analog signal to drive the actuator to perform actions, generating a corresponding axial force, which is transmitted to the sealing pair through the valve stem to generate sealing force.

[0168] Step 4.6, Adaptive Adjustment and Closed-Loop Formation. Based on the sealing force generated in Step 4.5, the force tracking error and system status are continuously monitored, and adjustments are made as necessary. Simultaneously, since the output sealing force affects the valve status, this change is detected again by the sensor array from Step 1, forming a closed loop, thus obtaining the force servo closed-loop control system.

[0169] In this embodiment, when a stable state is reached at a certain time period... , , Therefore, it is necessary to drive the actuator to move and then compensate. The sealing force.

[0170] Step 5: Sealing condition health diagnosis and early warning. Based on the force servo closed-loop control system constructed in Step 4 and the operating parameters collected by the multi-sensor array in Step 1, online assessment of the wear condition of the sealing pair and leakage early warning through multi-source information fusion are performed through active excitation monitoring and wear quantification, as detailed below:

[0171] Step 5.1, Active Excitation Monitoring. Based on the force servo closed-loop control system constructed in step four, as the number of valve opening and closing and the number of times the valve stem transmits sealing force increases, the sealing pair will wear, its mechanical properties will change, leading to an increase in the equivalent contact stiffness of the sealing pair. Therefore, the force servo closed-loop control system should periodically execute a low-amplitude, high-frequency main force excitation. For ease of analysis and control, a sine wave with an amplitude of 3 kN and a frequency of 10 Hz is used for the main force excitation. The excitation is then continued for 1 second, during which the valve stem displacement signal is acquired to obtain the valve stem displacement response amplitude. This excitation cycle is repeated 4-5 times, and the average value of the valve stem displacement response amplitude is taken. Finally, the current equivalent contact stiffness is obtained by calculating the ratio of the main force excitation amplitude to the averaged valve stem displacement response amplitude. .

[0172] Step 5.2, Wear Quantification. Based on the data obtained in Step 5.1 Define a wear index :

[0173] (19)

[0174] in, It is the reference value of contact stiffness when the valve is put into operation and is in a brand new condition.

[0175] Step 5.3, Online assessment of the wear condition of the sealing pair. Based on the wear index defined in Step 5.2. It can perform health status classification and early warning for sealing components:

[0176] when At this time, the wear is slight, it can operate normally, and the status level is "healthy".

[0177] when When wear enters a stable period, monitor the wear trend, prepare spare parts, and the status level is "Caution".

[0178] when When wear and tear accelerates, it is recommended to plan a shutdown for maintenance, and the status level is "warning".

[0179] when When the sealing pair is about to fail, immediate maintenance is required, and the status level is "alarm".

[0180] in, For custom parameters, For health threshold, To pay attention to the threshold; This is the failure threshold for the sealing pair.

[0181] In this embodiment, , , .

[0182] During a certain execution of low-amplitude, high-frequency main power excitation, the calculation was... Then, according to formula (19), we can obtain According to the condition classification, the sealing pair of the high-pressure small-diameter fully welded ball valve is currently only slightly worn and can still operate normally.

[0183] Step 5.4, Leakage Warning. Based on the operating parameters collected by the multi-sensor array in Step 1, the DS evidence theory is used to perform multi-data fusion, output the final comprehensive confidence level, and initiate a tiered warning response: low-confidence warnings only display prompt information to attract attention, while high-confidence warnings require immediate intervention for confirmation and handling.

[0184] An intelligent valve adaptive sealing system based on operating condition perception is disclosed. This system implements the aforementioned intelligent valve adaptive sealing method based on operating condition perception. The intelligent valve adaptive sealing system includes an operating condition perception and preprocessing module, a sealing requirement calculation module, a sealing setting optimization module, an intelligent control and precision execution module, and a status diagnosis and early warning module. Specifically:

[0185] The operating condition sensing and preprocessing module consists of a multi-sensor array and a data acquisition and processing system. By deploying the multi-sensor array at key locations on the valve, the data acquisition and processing system synchronously collects physical monitoring data and performs preprocessing to ensure data integrity and accuracy. At the same time, it provides high-quality input data for subsequent modules and serves as the sensing hub of the entire intelligent valve adaptive sealing system.

[0186] The sealing requirement calculation module is connected to the working condition perception and preprocessing module. It consists of a high-performance computing engine and has built-in calculation models for the minimum theoretical sealing force and the maximum allowable sealing force for dynamic calculation. It is the calculation core of the entire intelligent valve adaptive sealing system.

[0187] The sealing setting optimization module is connected to the sealing demand calculation module and the working condition perception and preprocessing module. It consists of a high-performance industrial computing platform and is used to run optimization algorithms and select decision strategies to calculate the optimal sealing force setpoint. It is the decision center of the entire intelligent valve adaptive sealing system.

[0188] The intelligent control and precision execution module is connected to the sealing setting optimization module, the sealing demand calculation module, and the working condition perception and preprocessing module. It consists of a servo controller and a force servo actuator, and is used to implement the closed-loop control algorithm and convert the control quantity into physical sealing force. It is the action realization of the entire intelligent valve adaptive sealing system.

[0189] The status diagnosis and early warning module is connected to the intelligent control and precision execution module, the sealing setting optimization module, the sealing demand calculation module, and the working condition perception and preprocessing module. It consists of a dedicated high-performance server and is used to realize online assessment of the wear status of the sealing pair and leakage early warning based on multi-source information fusion. It is the safety barrier of the entire intelligent valve adaptive sealing system.

[0190] The above embodiments are merely illustrative of the implementation methods of the present invention, but should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the protection scope of the present invention.

Claims

1. A method of smart valve self-adaptive sealing based on working condition perception, characterized in that, The intelligent valve self-adaptive sealing method comprises the following steps: The first step is multi-dimensional working condition perception and data preprocessing; various parameters including the axial force of the valve stem are collected in real time through a deployed multi-sensor array, and the collected data is optimized and fused to obtain a working condition state vector ; Second step, modeling of sealing requirement based on physical model and data driven model; based on the working condition state vector obtained in the first step , dynamically calculating the minimum theoretical sealing force required to achieve the desired sealing level under the current working condition and the maximum allowable sealing force to ensure that the sealing element does not suffer permanent damage ; Specifically: Step 2.1, Minimum theoretical seal force modeling; minimum theoretical seal force : lower limit required to achieve reliable sealing for the current operating condition : （2）； （3）； （4）； （5）； （6）； wherein, is the force component required to overcome the medium pressure, is the force component required to overcome the surface micro roughness, is the valve opening and closing friction force, is the thermal compensation force, is the effective actual contact area of the sealing pair, is the sealing specific pressure coefficient, is the surface adhesion stress, is the friction coefficient, is the thermal compensation coefficient, is the temperature at the initial pre-tightening of the valve, as reference temperature; Step 2.2, Maximum allowable sealing force modeling; maximum allowable sealing force To protect the seal from plastic flow, creep or structural damage due to excessive compression: （7）； wherein is the yield strength of the sealing material at the current temperature, is the safety factor; Step 2.

3. Set the sealing window; based on the minimum theoretical sealing force obtained in step 2.1 and the maximum allowed sealing force obtained in step 2.2 to obtain a safe sealing working window to dynamically track the real-time updated sealing force : （8）； Third step, sealing force set point optimization based on intelligent decision strategy; based on the minimum theoretical sealing force obtained in the second step With the maximum allowable sealing force Optimization target and intelligent decision strategy are selected, and sealing force adaptive setting and optimization are realized through continuous optimization of the sealing force set point; specifically: Step 3.1, determining the optimization target; based on the safe sealing working window obtained in the second step, selecting the sealing force set point in it is a multi-objective optimization problem; determine the optimization target, the optimization target includes reliability target, life target, energy consumption target and control stability target; Step 3.2, selecting intelligent decision strategy; based on the optimization target determined in step 3.1, taking a hierarchical decision structure, selecting appropriate strategy for calculation, determining the optimal sealing force set point under different operating conditions and performance requirements; The fourth step is closed loop control and execution; based on the sealing force set point obtained in the third step and the axial force of the valve rod collected in the first step, the force control deviation between the optimal sealing force set point and the actual value of the valve rod axial force is calculated, and the control amount is obtained by adopting closed loop control algorithm, and then the actuator is driven to act, finally a force servo closed loop control system is constructed, and the real-time and self-adaptive adjustment of the sealing force is realized; The fifth step is sealing state health diagnosis and early warning; based on the force servo closed loop control system constructed in the fourth step and the parameters collected in the first step, through active excitation monitoring and wear quantification, the sealing pair wear state online evaluation and multi-source information fusion leakage early warning are carried out.

2. A method of intelligent valve self-adaptive sealing based on working condition perception according to claim 1, characterized in that, The first step is specifically: Step 1.1, real-time data acquisition; through the sensor array deployed in the key position of the valve, real-time acquisition of parameters is realized, and sensor data is obtained, wherein the key position includes valve body, valve cavity, valve rod and actuator; Step 1.2, data optimization; based on the sensor data obtained in step 1.1, the sensor data is optimized through time synchronization and linear interpolation; Step 1.3, data fusion; based on the optimized data obtained in step 1.2, a Kalman filter algorithm is used to output the optimal state estimation value, generating a high-confidence, unified working condition state vector : （1）； Wherein, is the real-time pressure difference on both sides of the valve seat, is the medium temperature, is the real-time collected valve rod axial force, is the acoustic emission signal root mean square value, is any time.

3. The method of claim 2, wherein the method further comprises: Step 1.1 is specifically as follows: High-precision pressure transmitters are installed on the valve body at the upstream inlet end and the downstream outlet end to calculate the pressure difference on both sides of the valve seat and judge the medium leakage; Temperature sensors are installed inside the valve cavity and the actuator to measure the medium temperature and the ambient temperature; Strain type force sensors are installed at the connection between the push rod of the actuator and the valve rod to accurately measure the axial force of the valve rod; Acoustic emission sensors are installed outside the valve body as close as possible to the monitored sealing pair to monitor the micro friction of the sealing pair, the turbulent noise generated by the leakage flow and the impact signal generated by the particles passing through the sealing surface, and realize early leakage warning and wear evaluation.

4. The method of claim 3, wherein the method further comprises: In step 3.1: The reliability target is to ensure that the sealing force is always greater than the minimum theoretical sealing force , preventing leaks; The life target is to minimize the sealing force to reduce the wear and plastic deformation of the sealing pair and prolong the service life of the valve; The energy consumption target is to minimize the sealing force to reduce the energy consumption of the actuator; The control stability target is that the sealing force set point should avoid sharp change to ensure the stability of the control system.

5. A method of intelligent valve self-adaptive sealing based on working condition perception according to claim 4, characterized in that, In step 3.2, the hierarchical decision structure includes two kinds of reactive basic rule decision and predictive advanced optimization decision, which are specifically as follows: The first kind: the reactive basic rule decision is a dynamic margin adaptive strategy, which is a dynamic margin method based on acoustic emission feedback: （9）； wherein, is a dynamic safety margin, is a seal force setpoint, is a minimum theoretical seal force; whose adaptive law is as follows: First, by long-term statistics of AE signals of healthy valves in good sealing state, a desired range of acoustic emission intensity is set wherein is the lower limit of the desired range, is the upper limit of the desired range; then a self-adaptive strategy is repeatedly cycled until the corresponding dynamic safety margin is output when the sealing state is optimal: If , indicates over-tightening, rate decrease needed : （10）； If , indicates that the sealing state is optimal, remains unchanged; If , indicates a risk of leakage or high friction, rate increase needed : （11）； wherein, is a step size, and has a value ranging from 0.1 KN to 0.5 KN; The dynamic margin adaptive strategy realizes closed-loop optimization based on sealing state feedback, determines the minimum dynamic safety margin required to maintain sealing, and further determines the optimal sealing force set point at the current time; The second kind: the predictive advanced optimization decision can predict the working condition changes in the future period of time, and solve the optimization problem to obtain the optimal sealing force set point at the current time, which is as follows: First, a discrete-time model describing the system dynamics is established; the state variable , the control input , the disturbance input , the discrete-time model is simplified as: （12）； Wherein, k is a certain time; A is a state variable setting parameter, B is a control input conversion parameter, and the dimension is ; C is an interference input conversion matrix; Then, the optimization problem is solved to obtain a future control sequence As shown in equation (13): （13）； Satisfy the constraint: （14）； （15）； （16）； wherein, is the target optimization function; is the predicted k+i time valve stem axial force at k time; is the predicted k+i time minimum theoretical seal force at k time; is the predicted k+i time maximum allowable seal force at k time; is the weight coefficient for balancing each optimization target, dimensionless ; is the prediction time domain length in the first optimization; , are the input minimum and maximum constraints, respectively; is the input variation constraint; The optimal control sequence is obtained by solving The first value in the sequence is taken and multiplied by the control input conversion parameter to give the seal force set point at the current time as shown in equation (17): （17）； Then re-measure and optimize at the next sampling time k+1, and repeatedly cycle to perform rolling optimization until it is forcibly interrupted by an external instruction or failure; The model predictive control strategy can adjust the sealing force set point in advance according to the predicted working condition changes, realize dynamic and real-time adaptive prediction and optimization, and ultimately determine the optimal sealing force set point at the current time.

6. A method of intelligent valve self-adaptive sealing based on working condition perception according to claim 5, characterized in that, The fourth step is specifically: Step 4.1, based on the sealing force setpoint obtained in the third step, receive an optimized stop-time sealing force setpoint, which is used as the optimal sealing force setpoint ; Step 4.2, based on the first step of real-time acquisition of the valve stem axial force, to obtain the actual value of the valve stem axial force at this time ; Step 4.3, calculate the force control deviation; based on the optimal seal force setpoint obtained in step 4.1 and the valve stem axial force actual value obtained in step 4.2 a force control deviation between the optimal seal force setpoint and the valve stem axial force actual value is calculated: （18）； Step 4.4, execute the closed-loop control algorithm; based on the force control deviation obtained in step 4.3, use the PID control algorithm to calculate the required control amount; Step 4.5, action execution; based on the control amount obtained in step 4.4, convert it into an analog signal to drive the actuator to act, generate a corresponding axial force, and transmit it to the sealing pair through the valve stem to generate a sealing force; Step 4.6, adaptive adjustment and closed-loop formation; based on the sealing force generated in step 4.5, continuously monitor the force tracking error and system state for adjustment; at the same time, the output sealing force affects the valve state, which is detected by the sensor array in the first step, forming a closed loop to obtain a force servo closed-loop control system.

7. A method of intelligent valve self-adaptive sealing based on working condition perception according to claim 6, characterized in that, The fifth step is specifically: Step 5.1, active excitation monitoring; the force servo closed-loop control system constructed in the fourth step performs an active force excitation of low amplitude and high frequency periodically, specifically: the active force excitation adopts a sine wave, the amplitude is 3KN, and the frequency is 10Hz; then, the valve rod displacement signal at this time is collected, and then the valve rod displacement response amplitude is obtained, and the valve rod displacement response amplitude is averaged for 4-5 excitation cycles; finally, the ratio of the active force excitation amplitude and the averaged valve rod displacement response amplitude is calculated to obtain the current equivalent contact stiffness ; Step 5.2, wear quantification; Based on the obtained in step 5.1 , a wear indicator is defined : （19）； wherein is the contact stiffness reference value at the initial stage of operation of the valve, when it is new. Step 5.3, on-line evaluation of the sealing pair wear state; based on the wear index defined in step 5.2 , performing sealing pair health state classification and early warning; When wear is light, normal operation, and the status level is healthy; When wear enters a steady state, the wear trend is monitored, spare parts are prepared, and the status level is attention; When wear is accelerated, a planned shutdown maintenance is recommended, the status level is warning; When the sealing pair is about to fail, maintenance is immediately required, and the status level is alarm. wherein, is a custom parameter, is a health threshold, is an attention threshold; is a seal pair failure threshold; Step 5.4, leakage warning; based on the parameters collected by the multi-sensor array in the first step, use D-S evidence theory for multi-data fusion to output the final comprehensive confidence, and start a hierarchical warning response: low confidence warning only displays a prompt message to attract attention, while high confidence warning requires immediate intervention confirmation and processing.

8. A condition-aware based intelligent valve self-adaptive sealing system, characterized in that, The intelligent valve adaptive sealing system realizes the intelligent valve adaptive sealing method based on working condition sensing according to any one of claims 1-7, and the intelligent valve adaptive sealing system comprises a working condition sensing and preprocessing module, a sealing demand calculation module, a sealing setting optimization module, an intelligent control and precise execution module, and a state diagnosis and warning module; specifically: The working condition sensing and preprocessing module is composed of a multi-sensor array and a data acquisition and processing system, which synchronously acquires physical monitoring data and pre-processes them by deploying a multi-sensor array at key positions of the valve and using a data acquisition and processing system to provide input data; The sealing demand calculation module accesses the working condition sensing and preprocessing module and is used to calculate the minimum theoretical sealing force and the maximum allowable sealing force; The sealing setting optimization module accesses the sealing demand calculation module and the working condition sensing and preprocessing module, and is used to run optimization algorithms and decision strategy selection to calculate the optimal sealing force set point; The intelligent control and precise execution module accesses the sealing setting optimization module, the sealing demand calculation module and the working condition sensing and preprocessing module, is composed of a servo controller and a force servo execution mechanism, is used for realizing a closed loop control algorithm, and converts a control quantity into a physical sealing force. The state diagnosis and early warning module accesses the intelligent control and precise execution module, the sealing setting optimization module, the sealing demand calculation module and the working condition sensing and preprocessing module, and is used for realizing online evaluation of a sealing pair wear state and leakage early warning based on multi-source information fusion.

Images