Standard valve leakage rate calibration method and system based on ultrasonic sensor

Abstract

The invention relates to the technical field of valve detection, in particular to a valve detection method and system based on an ultrasonic sensor, and aims to apply actively emitted sound waves to valve inner leakage detection, so that the anti-interference capability is higher, and the internal structure information of a valve can be detected. Ultrasonic array detection is carried out on the valve, the signal transmitting and receiving modes are variable, the device adapts to various working modes of one-transmitting-multiple-receiving-multiple-transmitting-multiple-receiving, detection information can be reflected or transmitted sound waves, and the obtained information is rich. According to the method, a conventional detection method is avoided, the pressure boundary of a detected object does not need to be damaged, and the method has online detection capability and has very high practicability in inner leakage detection and analysis of the in-use valve. The method is not only suitable for valve inner leakage detection analysis in the nuclear power field, but also suitable for valve inner leakage detection analysis in the fields of wind power, thermal power, chemical engineering, petroleum and the like.

Description

Technical Field

[0001] The present invention relates to the field of valve detection technology, and in particular to a valve detection method and system based on an ultrasonic sensor. Background Art

[0002] At present, major projects such as thermal power, nuclear power, hydropower, petrochemical, oil and gas pipelines, coal liquefaction and metallurgy around the world have put forward high requirements for the reliability of valves. However, internal leakage failures are difficult to avoid during the use of valves. Especially in the field of nuclear power, valves in the primary circuit system have internal leakage, and the valve internal leakage cannot be isolated and eliminated during operation, which increases the safety risk of the system and seriously affects the overall reliability of the primary circuit system. At present, the main methods for detecting valve internal leakage are: water pressure sealing test, temperature field, pressure field monitoring method, etc.

[0003] The water pressure sealing test has no online detection capability and requires disassembly of the valve and conduction of a high-pressure water pressure sealing test on a specific tooling according to the standard. This has a significant impact on the experimental system and is not suitable for normally operating devices. The temperature field and pressure field monitoring method requires the arrangement of sensors before and after the valve, which will destroy the pressure boundary and increase the risk of leakage. It is also only suitable for situations where there is a large pressure difference and temperature difference before and after the valve. Summary of the invention

[0004] The object of the present invention is to provide a valve detection method and system based on an ultrasonic sensor to solve the above-mentioned problems in the prior art.

[0005] The embodiments of the present invention are implemented by the following technical solutions:

[0006] In a first aspect, the present invention provides a valve detection method based on an ultrasonic sensor, comprising:

[0007] Collect data on the valve to be inspected, obtain the ultrasonic signal of the inspection object, and establish a medium stress elastic model based on the ultrasonic signal;

[0008] The detection propagation speed of the sound wave in three directions in the space coordinates is obtained through the medium stress elastic model;

[0009] Through experiments, the use data of standard valves is obtained, and the ultrasonic detection results of standard valves are extracted. Through the system input pressure, the detected material stress and the actual measured leakage, a leakage detection model based on the ultrasonic detection results and the pressure difference before and after the valve is established. The leakage of the valve to be detected is output through the leakage model for internal leakage detection of valves of the same type;

[0010] The standard sound wave propagation velocity test results and standard leakage amount of the valve to be tested are extracted. The detection model is used to predict the leakage amount of the valve to be tested through the valve front-end system pressure and the sound wave test results. The propagation velocity is compared with the standard sound wave propagation velocity. If they are inconsistent, the valve to be tested fails the test.

[0011] According to the internal leakage requirements of the system to which the valve belongs, a detection comparison value is set. If the leakage of the valve to be detected exceeds the detection comparison value, the test fails. If it does not exceed the detection comparison value, the test passes.

[0012] If they are consistent, the leakage of the valve to be tested is compared with the standard leakage. If they are consistent, the valve to be tested passes the test. If they are inconsistent, the valve to be tested fails the test.

[0013] Preferably, the data collection for the valve to be inspected includes:

[0014] Set valve monitoring points. When there is no leakage in the valve, the pipeline pressure changes within the set pressure threshold range, and the echo signals of each measuring point are recorded;

[0015] Within the set pressure threshold range, the leakage amount is adjusted by adjusting the preload force, and when adjusting the preload force, the torque used for each adjustment is recorded;

[0016] A defective valve to be inspected is used, and data of the defective valve to be inspected is collected, including the echo signal of each measuring point and the torque used for each adjustment.

[0017] Preferably, the medium stress elastic model includes:

[0018]

[0019]

[0020]

[0021] Where ρ0 is the initial density of the medium, v 11 、v 12 、v 13 are the sound wave velocities propagating along the x-axis with vibration directions in the x, y, and z directions, λ and μ are second-order elastic constants, l, m, and n are third-order elastic constants, α1, α2, and α3 are the components of the homogeneous triaxial strain field in the x, y, and z directions, θ is the sum of α1, α2, and α3. It can be found that the wave velocity and strain magnitude are positively correlated.

[0022] Preferably, the valve material stress obtained through detection and the corresponding valve leakage during signal detection are used for proxy model training to establish a leakage prediction model.

[0023] Preferably, the establishing of the leakage prediction model comprises:

[0024] A set of leakage data is obtained through a set of uniformly increasing preload forces, and an initial model fitting is performed;

[0025] Sampling is performed between the maximum preload and 0 preload, and all samples are predicted using the initial model. The prediction variance is obtained through the Kriging model, and the value of the U function corresponding to all sample points is calculated.

[0026] Select the point with the smallest value as the new training point, and obtain the data closest to this point by adjusting the experimental system. Add this set of data to the training set and refit the Kriging model, and repeat this process until the value of the U function of all prediction points is greater than 2.

[0027] Preferably, the Kriging model includes:

[0028] Linear weighting function:

[0029]

[0030] In the formula, is the response value of any point in space, ω (i) is the weighting coefficient value, y (i) is the input value, n and i are natural numbers;

[0031] Static random process definition function:

[0032] Y(x)=β0+Z(x)

[0033] In the formula, β0 is an unknown constant, also known as the global trend model, Y(x) is the mathematical expectation, and Z(x) is the mean value x and the variance σ 2 A static random process.

[0034] Covariance function expressing correlation:

[0035] Cov[z(x), Z(x′)]=σ 2 R(x, x′)

[0036] In the above formula: R(x, x′) is the relationship between random variables x and x′, σ 2 is the variance value, Z(x′) is the mean x′ and the variance σ 2 A static random process.

[0037] Preferably, the U function includes:

[0038]

[0039] Where U is the output value of the U function, GK (x) is the predicted mean, σ K (x) is the standard deviation.

[0040] In a second aspect, the present invention further provides a valve detection system based on an ultrasonic sensor, comprising:

[0041] An ultrasonic transmitting device, which is used to be arranged on the valve to be inspected and transmit an ultrasonic signal;

[0042] A data acquisition device, the data acquisition device is used to be arranged on the valve to be detected, and collect the ultrasonic signal reflected by the valve to be detected;

[0043] A storage device, wherein the storage device is used to receive the ultrasonic signal collected by the data collection device and execute the above-mentioned valve detection method based on the ultrasonic sensor.

[0044] In a third aspect, an electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned valve detection method based on an ultrasonic sensor when executing the computer program.

[0045] In a fourth aspect, a computer-readable storage medium is provided, wherein a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the above-mentioned valve detection method based on an ultrasonic sensor is implemented.

[0046] The technical solution of the embodiment of the present invention has at least the following advantages and beneficial effects:

[0047] By adopting the method provided by the present invention, actively emitted sound waves are used for valve internal leakage detection, which has stronger anti-interference ability and can detect the internal structure information of the valve. The ultrasonic array detection of the valve has a variety of signal transmission and reception modes, which can adapt to various working modes from one-send-multiple-receive to multiple-send-multiple-receive. The detection information can be reflected or transmitted sound waves, and the information obtained is rich. It avoids the use of conventional detection methods, does not need to destroy the pressure boundary of the detection object, and has online detection capabilities, which is very practical for the internal leakage detection and analysis of valves in use. It is not only suitable for valve internal leakage detection and analysis in the nuclear power field, but also for valve internal leakage detection and analysis in the fields of wind power, thermal power, chemical industry, petroleum, etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0049] Figure 1 It is a schematic diagram of the process of the present invention;

[0050] Figure 2 It is a structural schematic diagram of the cold leak detection and calibration device of the present invention;

[0051] Figure 3 It is a structural schematic diagram of the hot leak detection and calibration device of the present invention;

[0052] Figure 4 Schematic diagram of the system of the present invention. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0054] The modules or submodules described independently may or may not be physically separated: they may be implemented by software or hardware, and some modules or submodules may be implemented by software, and the processor calls the software to implement the functions of these modules or submodules, while other modules or submodules may be implemented by hardware, such as by hardware circuits. In addition, some or all of the modules may be selected according to actual needs to achieve the purpose of the present application. Please refer to Figure 1 The present invention provides a valve detection method based on an ultrasonic sensor, comprising:

[0055] S101: collecting data of the valve to be inspected, obtaining ultrasonic signals of the inspection object, and establishing a medium stress elasticity model according to the ultrasonic signals;

[0056] Changes in valve seal preload and seal surface damage will affect the propagation / reflection of the detected acoustic signal, causing changes in the echo. The theory of acoustic elastic effect establishes a mathematical model for the effect of medium stress on the medium elastic coefficient, which can describe the change in the propagation speed of sound in the medium when the medium stress changes (strain is generated).

[0057] S102: Obtaining the detection propagation speeds of the sound wave in three directions in the space coordinates through the medium stress elastic model;

[0058] S103: Acquire standard valve usage data through experiments, extract the acoustic wave detection results of the standard valves, and establish a detection model of leakage volume related to the acoustic wave detection results and the pressure difference before and after the valve through the system input pressure, the detected material stress and the actual measured leakage volume, which is used for internal leakage detection of valves of the same type;

[0059] S104: extracting the acoustic wave detection result of the valve to be detected, and using the detection model to predict the leakage amount of the valve to be detected through the valve front-end system pressure and the acoustic wave detection result.

[0060] Extract the standard sound wave propagation velocity test results and standard leakage of the valve to be tested. Use the test model to predict the leakage propagation velocity of the valve to be tested through the valve front-end system pressure and the sound wave test results and compare them with the standard sound wave propagation velocity. If they are inconsistent, the output is that the valve to be tested fails the test.

[0061] S105: Obtain a detection comparison value. If the leakage of the valve to be detected exceeds the detection comparison value, the detection fails. If it does not exceed the detection comparison value, the detection passes.

[0062] By adopting the method provided by the present invention, actively emitted sound waves are used for valve internal leakage detection, which has stronger anti-interference ability and can detect the internal structure information of the valve. The ultrasonic array detection of the valve has a variety of signal transmission and reception modes, which can adapt to various working modes from one-send-multiple-receive to multiple-send-multiple-receive. The detection information can be reflected or transmitted sound waves, and the information obtained is rich. It avoids the use of conventional detection methods, does not need to destroy the pressure boundary of the detection object, and has online detection capabilities, which is very practical for the internal leakage detection and analysis of valves in use. It is not only suitable for valve internal leakage detection and analysis in the nuclear power field, but also for valve internal leakage detection and analysis in the fields of wind power, thermal power, chemical industry, petroleum, etc.

[0063] In addition, if Figure 2-3 As shown, for a new valve with different structural features, the equipment is calibrated. The diagram shows the piping and equipment connections and working methods of the circuit used in the experiment. Pressure, temperature and other measuring points are also arranged at typical positions in the circuit. The sensors of the ultrasonic array detection equipment are arranged at typical parts of the experimental valve, and ultrasonic emission is performed under multiple working conditions, and signal acquisition, storage and analysis are performed. It includes a cold leak detection calibration device and a hot leak detection calibration device.

[0064] Among them, the cold leakage detection calibration device includes a test valve and a manual pressure test pump, the test valve is connected to the manual pressure test pump and the leakage interface, and the manual pressure test pump is connected to the test valve and the water source.

[0065] The thermal leakage detection and calibration device includes a test valve and an electric heating element arranged on the test valve, and its pipeline is also connected with a leak detection interface, a voltage stabilizer, a drain outlet and a water replenishment and pressurization system.

[0066] In an exemplary embodiment of the present invention, the data collection of the valve to be detected includes:

[0067] Set valve monitoring points. When there is no leakage in the valve, the pipeline pressure changes within the set pressure threshold range, and the echo signals of each measuring point are recorded;

[0068] Within the set pressure threshold range, the leakage amount is adjusted by adjusting the preload force, and when adjusting the preload force, the torque used for each adjustment is recorded;

[0069] A defective valve to be inspected is used, and data of the defective valve to be inspected is collected, including the echo signal of each measuring point and the torque used for each adjustment.

[0070] Specifically, a no-leakage pressure benchmark experiment is carried out. When the valve has no leakage, the pipeline pressure changes between 1-21MPa (determined according to the operating conditions of the target valve), and the echo signals of each measuring point are recorded to obtain a data set for feature extraction and analysis. This experiment is used to calibrate the effect of pressure on sealing stress;

[0071] Carry out valve leakage experiments, adjust the leakage by adjusting the pre-tightening force under different pressure conditions of 1-21MPa, use a torque wrench to close the valve, control the valve closing torque, classify the catalog, store the subsequent data, record the echo signals of each measuring point according to different leakage ranges, and name the collected data.

[0072] A leakage experiment on sealing surface damage is carried out using a valve plate with prefabricated defects on the sealing surface. The data set is collected referring to the experimental design method in the second step. The correlation between the characteristic source and the leakage amount and the degree of valve plate damage is analyzed, and the excellent features are used for sealing surface damage detection of this type of valve.

[0073] In an exemplary embodiment of the present invention, the medium stress elastic model includes:

[0074]

[0075]

[0076]

[0077] Where ρ0 is the initial density of the medium, v 11 、v 12 、v 13are the sound wave velocities propagating along the x-axis with vibration directions in the x, y, and z directions, λ and μ are second-order elastic constants, l, m, and n are third-order elastic constants, α1, α2, and α3 are the components of the homogeneous triaxial strain field in the x, y, and z directions, θ is the sum of α1, α2, and α3. It can be found that the wave velocity and strain magnitude are positively correlated.

[0078] The valve material stress detected and the corresponding valve leakage during signal detection are used for proxy model training to obtain a leakage prediction model. When using the global training method, the amount of training samples required for accurate model prediction is too large, which is almost impossible for the calibration of the detection software. Therefore, it is necessary to transform the problem to introduce an adaptive model training method. The adaptive Kriging model is applicable here. The Kriging model is an interpolation model, and its interpolation result is the linear weighting of the known sample response values, that is:

[0079]

[0080] is the response value of any point in space, ω (i) is the weighting coefficient value, y (i) is the input value, n and i are natural numbers, so as long as the value of the weighting coefficient is given, the response value of any point in the design space can be obtained. In order to calculate the weighting coefficient, the Kriging model introduces a statistical assumption: the unknown function is regarded as a specific realization of a Gaussian static random process, which is defined as:

[0081] Y(x)=β0+Z(x)

[0082] Where β0 is an unknown constant, also known as the global trend model, which represents the mathematical expectation of Y(x); Z(·) is a model with a mean of zero and a variance of σ 2 A static random process. At different locations in the design space, these random variables have a certain correlation (covariance), which can be expressed as:

[0083] Cov[Z(x),Z(x′)]=σ 2 R(x, x′)

[0084] In the above formula: R(x, x′) is the relationship between random variables x and x′, σ 2 is the variance value, Z(x′) is the mean x′ and the variance σ 2 The static random process, that is, R(x, x′) is a correlation function, which is only related to the spatial distance of the sample and satisfies that it is equal to 1 when the distance is zero and equal to 0 when the distance is infinite. The correlation decreases as the distance increases. The model code implementation and its derivation process can be found in the relevant literature, which will not be repeated here.

[0085] For nuclear-grade valves, the first category leakage level standard applies, which is no more than 10mL / min. The valve leakage (unit: mL / min) is subtracted from 10. When the result is greater than 0, the leakage exceeds the standard, and when it is less than 0, the leakage meets the standard. In this way, the valve internal leakage detection and judgment is converted into a problem of predicting the positive and negative value of the function. As long as the sign of the model prediction result is correct, the correct judgment result can be given.

[0086] The Kriging model contains a Gaussian process, whose output is a Gaussian field. When predicting the output for each sample point, the information that can be obtained is the predicted mean G K (x) and standard deviation σ K (x). If G K (x)>0, when limiting the probability P of accurate judgment of sample output sign, it is only necessary to ensure that the output value bσ interval [G K (x)-b·σ K (x), G K (x)+b·σ K (x)] has a lower boundary greater than or equal to 0. If G K (x)<0, it is necessary to ensure that the upper boundary of the bσ interval is less than or equal to 0, b is the random coefficient, that is:

[0087]

[0088] The two situations can be summarized as:

[0089] |G K (x)|-b·σ K (x)≥0

[0090] The definition of U function is:

[0091]

[0092] When U = 2, that is, |G K (x)| / σ K (x) = 2, then G K (x)-2σ K (x) = 0. If the sign of the output value at this point is wrong, its value should fall into G K (x)-2σ K (x) has a probability of Φ(-2), so when the corresponding U=2, the probability of the sign judgment of the point is accurate is P=1-Φ(-2)=97.72%. In summary, when the corresponding U>2, the probability of the sign judgment of a sample point is accurate is P≥97.72%. On this theoretical basis, if the point adding strategy used in training the Kriging model is to add the point with the smallest U learning function value, that is, the point with the highest probability of wrong sign judgment of the output value, until U is satisfied.min >2, it can be guaranteed that the sign accuracy of the predicted values ​​of all points in the sample pool meets P≥97.72%.

[0093] Based on the above model training ideas, we first obtain a set of leakage data according to a set of uniformly increasing preloads, and perform initial model fitting. The maximum preload should not cause the valve to become stuck when opening. Sampling is performed between the maximum preload and 0 preload, and all samples are predicted using the initial model. The prediction variance is obtained through the Kriging model, and the value of the U function corresponding to all sample points is calculated. The point with the smallest value is selected as the new training point. By adjusting the experimental system, the data closest to this point is obtained. This set of data is added to the training set and the Kriging model is refitted. This process is repeated until U>2 for all prediction points.

[0094] When the same sensor arrangement is used, the judgment result is obtained by substituting the detection signal into the model. If its value is greater than 0, the leakage of the valve to be tested exceeds the standard and the test fails. If its value is less than zero, the valve to be tested passes.

[0095] In an exemplary embodiment of the present invention, the material stress can be measured by quantifying the change in the reception time of the reflected wave or the transmitted wave (time domain operation), or by quantifying the change in the transmission frequency of the pulse phase-locked loop ultrasonic device when the phase is locked (frequency domain operation). Depending on the different valve structures (the angle between the direction of the compressive stress on the sealing surface and the direction that is easy to detect is different), the transverse wave or the longitudinal wave can be selected for measurement to obtain the size of the valve sealing force.

[0096] In addition, when detecting the damage of the sealing surface, different degrees of damage will cause different reflected waves or transmitted waves. The RF waveform time domain characteristics of the signal, the power spectrum of the signal in the frequency domain, the analysis spectrum of the signal in the frequency domain, the phase angle of the RF signal in the frequency domain and the transfer function domain based on the initial pulse can be selected, combined with the system pressure and leakage data before the valve to build a quantitative model.

[0097] The probability density function method, spatial scatter diagram and other methods are used for feature screening and analysis. At different detection levels, inferior features, general features and excellent features can be distinguished. Excellent features are the preferred features when conducting quantitative modeling, and general features are selected according to specific circumstances.

[0098] like Figure 4 As shown, the present invention also provides a valve detection system based on an ultrasonic sensor, comprising:

[0099] An ultrasonic transmitting device, which is used to be arranged on the valve to be inspected and transmit an ultrasonic signal;

[0100] A data acquisition device, the data acquisition device is used to be arranged on the valve to be detected, and collect the ultrasonic signal reflected by the valve to be detected;

[0101] A storage device, wherein the storage device is used to receive the ultrasonic signal collected by the data collection device and execute the above-mentioned valve detection method based on the ultrasonic sensor.

[0102] The ultrasonic transmitting device includes an ultrasonic transmitter and an ultrasonic transmitting probe, and the data acquisition device includes a high-speed data acquisition card and an ultrasonic receiving probe. The working modes of the transmitting probe include only transmitting and self-transmitting and self-receiving. The number is not limited to one, and the probe form is not limited to a straight probe. The signals received by the receiving probe include reflection signals and transmission signals. By processing and analyzing the array signal, the internal information of the valve is restored.

[0103] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0104] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. The computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0105] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A standard valve leakage calibration method based on ultrasonic sensor, characterized in that: include: Collect data on the valve to be inspected, obtain the ultrasonic signal of the inspection object, and establish a medium stress elastic model based on the ultrasonic signal; The detection propagation speed of the sound wave in three directions in the space coordinates is obtained through the medium stress elastic model; Through experiments, the use data of standard valves are obtained, the acoustic wave detection results of standard valves are extracted, and a leakage prediction model based on the acoustic wave detection results and the pressure difference before and after the valve is established through the system input pressure, the detected material stress and the actual measured leakage; Extract the acoustic wave detection results of the valve to be detected, and use the detection model to predict the leakage of the valve to be detected based on the valve front-end system pressure and the acoustic wave detection results; Obtain a detection comparison value. If the leakage of the valve to be detected exceeds the detection comparison value, the detection fails. If it does not exceed the detection comparison value, the detection passes.

2. The standard valve leakage calibration method based on ultrasonic sensor according to claim 1 is characterized in that: The use data of the standard valve obtained through experiments include: Set valve monitoring points. When there is no leakage in the valve, the pipeline pressure changes within the set pressure threshold range, and the echo signals of each measuring point are recorded; Within the set pressure threshold range, the leakage amount is adjusted by adjusting the preload force, and when adjusting the preload force, the torque used for each adjustment is recorded; A defective valve to be inspected is used, and data of the defective valve to be inspected is collected, including the echo signal of each measuring point and the torque used for each adjustment.

3. The standard valve leakage calibration method based on ultrasonic sensor according to claim 2 is characterized in that: The medium stress elastic model includes: Where ρ0 is the initial density of the medium, v 11 、v 12 、v 13 are the sound wave velocities propagating along the x-axis with vibration directions in the x, y, and z directions, respectively; λ and μ are the second-order elastic constants; l, m, and n are the third-order elastic constants; α1, α2, and α3 are the components of the homogeneous triaxial strain field in the x, y, and z directions, respectively; θ is the sum of α1, α2, and α3.

4. The standard valve leakage calibration method based on ultrasonic sensor according to claim 2 is characterized in that: It also includes the valve material stress obtained through detection and the corresponding valve leakage during signal detection for proxy model training to establish a leakage prediction model.

5. The standard valve leakage calibration method based on ultrasonic sensor according to claim 2 is characterized in that: The leakage prediction model is established including: A set of leakage data is obtained through a set of uniformly increasing preload forces, and an initial model fitting is performed; Sampling is performed between the maximum preload and 0 preload, and all samples are predicted using the initial model. The prediction variance is obtained through the Kriging model, and the value of the U function corresponding to all sample points is calculated. Select the point with the smallest value as the new training point, and obtain the data closest to this point by adjusting the experimental system. Add this set of data to the training set and refit the Kriging model, and repeat this process until the value of the U function of all prediction points is greater than 2.

6. The standard valve leakage calibration method based on ultrasonic sensor according to claim 5 is characterized in that: The Kriging model includes: Linear weighting function: In the formula, is the response value of any point in space, ω (i) is the weighting coefficient value, y (i) is the input value, n and i are natural numbers; Static random process definition function: Y(x)=β0+Z(x) In the formula, β0 is an unknown constant, also known as the global trend model, Y(x) is the mathematical expectation, and Z(x) is the mean value x and the variance σ 2 A static random process. Covariance function expressing correlation: Cov[Z(x),Z(x′)]=σ 2 R(x,x′) In the above formula: R(x,x′) is the relationship between random variables x and x′, σ 2 is the variance value, Z(x′) is the mean x′ and the variance σ 2 A static random process.

7. The standard valve leakage calibration method based on ultrasonic sensor according to claim 5 is characterized in that: The U function includes: Where U is the output value of the U function, G K (x) is the predicted mean, σ K (x) is the standard deviation.

8. The standard valve leakage calibration system based on ultrasonic sensor is characterized by: include; An ultrasonic transmitting device, which is used to be arranged on the valve to be inspected and transmit an ultrasonic signal; A data acquisition device, the data acquisition device is used to be arranged on the valve to be detected, and collect the ultrasonic signal reflected by the valve to be detected; A storage device, wherein the storage device is used to receive the ultrasonic signal collected by the data acquisition device and execute the standard valve leakage calibration method based on the ultrasonic sensor as described in claim 1.

Images