Valve quality detection method and detection system thereof

Abstract

The invention relates to the technical field of detection, in particular to a valve quality detection method and system. According to the method, a valve is divided into sub-blocks, ultrasonic thickness, eddy current pitting corrosion signals, temperature, pH value, Cl <-> concentration and other environmental parameters are synchronously collected, a comprehensive corrosion coefficient is generated through a formula, a laser displacement meter is used for conducting three-dimensional scanning on a valve rod to obtain crack point cloud data, the volume density is calculated, and a health index is generated through fusion for decision making. The system comprises a multi-mode sensor, an active scanning module, a data analysis module and a decision output module, and the detection precision is improved through an anti-shake mechanism and a transfer learning model. According to the method, the problems of insufficient local corrosion capture, crack depth deficiency, complex structure shielding leak detection and the like in traditional detection are solved, accurate evaluation of the health state of the valve is realized, a scientific basis is provided for predictive maintenance of valves in petrochemical engineering, electric power and other industries, and the risks of production shutdown and equipment damage are effectively reduced.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of detection, and in particular to a valve quality detection method and a detection system thereof. BACKGROUND

[0002] Valve quality detection refers to various tests and evaluations on valves. Through detection and evaluation, the quality of the valve can be ensured to meet the design and manufacturing standards, so that the reliability, stability and safety of the valve in the use process can be ensured, installation problems can be found and corrected as soon as possible, and production stoppage and equipment damage caused by valve failure can be avoided.

[0003] A patent application with the publication number CN118641631B discloses a valve quality detection method and a detection system thereof, which is used to solve the problem of high labor and financial cost caused by overall replacement of the valve when the valve rod is stuck. The method includes the following steps: screening out the valve sub-blocks prone to corrosion, detecting the corrosion degree of the surface of the screened valve sub-blocks in real time, calculating the average corrosion coefficient, calculating the crack characteristic coefficient of each valve rod sub-region, clustering the crack characteristic coefficient, calculating the surface damage coefficient of the valve rod according to the clustering result, collecting the repair cost and the replacement cost of the valve, calculating the replacement cost coefficient, comprehensively evaluating the valve replacement index, and judging whether to replace or repair the valve. The progress speed of the project is effectively improved, the replacement and repair of the valve can be judged in time according to the valve condition when the valve rod is damaged, and the project cost is effectively saved.

[0004] However, the above-mentioned scheme still has some problems in actual application. The corrosion detection of the above-mentioned scheme has limitations. The corrosion degree formula used only depends on the thickness change, but in actual corrosion, it will show local forms such as pitting and stress corrosion cracking. The ultrasonic thickness measurement is difficult to accurately capture such non-uniform corrosion, and the crack characteristic coefficient formula used only calculates the area ratio of the surface crack, ignoring the crack depth, which will directly affect the remaining strength of the valve rod. The visual detection is also easily blocked by the complex structure such as the valve rod thread and the groove, resulting in missed detection.

[0005] Therefore, an innovative valve quality detection method and a detection system thereof are needed. SUMMARY

[0006] The application provides a valve quality detection method and a detection system thereof, which are fused with multi-source damage quantification and three-dimensional scanning, and are suitable for valve health state evaluation in the petroleum chemical industry, the electric power industry and the like.

[0007] The technical scheme of the application is as follows: a valve quality detection method, comprising the following steps: S1, divide the valve into n sub-blocks, synchronously collect ultrasonic thickness data, eddy current pitting corrosion signals and medium environment parameters of each sub-block, including temperature, pH value, Cl - concentration; S2, generate a comprehensive corrosion coefficient according to DC = [ (h - h_c) / (h_0 - h_c) + α × (A_p / A_0) ] × (1 + β ×E_env), wherein A_p is the pitting corrosion area detected by eddy current, A_0 is the reference area of the sub-block, E_env=k1T+k2pH+k3C_Cl - is the environmental corrosion factor, and α and β are calibration coefficients, which are calibrated by material electrochemical corrosion test; S3, use laser scanning to make the laser displacement meter perform three-dimensional scanning on the surface of the valve stem to obtain crack point cloud data; S4, perform crack depth detection, and calculate crack volume density CV = Σ(L_i × W_i ×D_i) / V_total based on the point cloud data, wherein L_i, W_i and D_i are the length, width and depth of a single crack, and V_total is the volume of the sub-region; S5, generate a valve health index by fusing DC and CV and make a decision.

[0008] Further, the collection of the environmental parameters uses a multi-parameter sensor array built in the flow channel of the valve.

[0009] Further, the laser scanning drives the laser displacement meter to rotate 360° around the valve stem axis through a rotating mechanism.

[0010] Further, the rotating mechanism includes an adaptive positioning clamp that can dynamically adjust the scanning radius according to the diameter of the valve stem.

[0011] Further, the crack depth detection uses a triangulation method, the laser displacement meter emits a linear laser to the surface of the valve stem, a CCD sensor captures the deformed laser stripe, and the depth D_i is calculated based on the stripe distortion.

[0012] Further, the fusion process in S5 includes establishing a health index HI = (DC^w1 × CV^w2) ^ [1 / (w1+w2)], wherein the weights w1 and w2 are determined by the type of the valve material, and the weights w1 and w2 are called from a pre-built material database, and the database stores the corrosion-crack correlation weights of different materials in fatigue tests.

[0013] A valve quality detection system, comprising: A multi-modal sensor module for collecting valve thickness, pitting corrosion and environmental parameters; An active scanning module, physically integrated with the multi-modal sensor module, is used to perform a valve stem three-dimensional scan; A data analysis module, electrically connected to the multi-modal sensor module and the active scanning module, performs comprehensive corrosion coefficient and crack volume density calculation; A decision output module, electrically connected to the data analysis module, generates a visual health index report.

[0014] Further, the active scanning module is built-in with an anti-shake mechanism, which includes a gyroscope to detect vibration amplitude in real time and a piezoelectric ceramic driver to dynamically compensate for optical path deviation.

[0015] Further, the data analysis module deploys a transfer learning model to identify corrosion-crack coupling failure modes through a pre-trained crack feature library, the transfer learning model uses ResNet34 as a backbone network, is pre-trained on an ImageNet dataset, fine-tunes the fully connected layer using 100,000 valve crack images, and outputs a corrosion-crack coupling failure probability P, which triggers an alarm when P>0.9.

[0016] The present application has the following advantages: The present application significantly improves the accuracy and reliability of valve quality detection through multi-source sensor fusion and three-dimensional quantitative analysis: integrating ultrasonic, eddy current and environmental parameter sensors, synchronously collecting thickness, pitting area, temperature, pH value, Cl - Concentration data, combined with the comprehensive corrosion coefficient formula DC, solves the problem of missed detection of non-uniform corrosion (such as pitting and stress corrosion) by traditional ultrasonic thickness measurement; laser displacement meter three-dimensional scanning of valve stem combined with triangulation method is used to obtain crack length, width and depth three-dimensional features, and the volume density CV is used to quantify the damage, breaking through the limitations of traditional two-dimensional features for valve stem strength evaluation; the system realizes 360° unobstructed scanning of valve stem threads, grooves and other complex structures through self-adaptive positioning clamps and gyro-piezoelectric ceramic anti-shake mechanism, and cooperates with the transfer learning model to identify corrosion-crack coupling failure modes; finally, based on the material type, the weight is dynamically adjusted to fuse DC and CV to generate a health index HI, realizing quantitative guidance for maintenance and replacement decision-making, providing a scientific basis for predictive maintenance of valves in petroleum and chemical, power and other industries, effectively reducing the risk of production downtime and equipment damage caused by valve failure. BRIEF DESCRIPTION OF DRAWINGS

[0017] The present application will be further described in detail below in conjunction with the drawings and specific embodiments.

[0018] Figure 1 The flowchart of the present application embodiment; Figure 2 The system block diagram of the valve quality detection system of the present application.

[0019] In the figure: 1, multi-modal sensor module; 2, active scanning module; 3, data analysis module; 4, decision output module. DETAILED DESCRIPTION

[0020] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0021] Embodiment 1

[0022] As shown in Figure 1 , 2 , the present embodiment proposes a valve quality detection method, comprising the following steps: S1, divide the valve into n sub-blocks, synchronously collect the ultrasonic thickness data, eddy current pitting corrosion signal and medium environment parameters of each sub-block, including temperature, pH value, Cl - concentration; S2, generate a comprehensive corrosion coefficient according to DC = [ (h - h_c) / (h_0 - h_c) + α × (A_p / A_0) ] × (1 + β ×E_env), wherein A_p is the pitting corrosion area detected by eddy current, A_0 is the reference area of the sub-block, E_env=k1T + k2pH + k3C_Cl - is the environmental corrosion factor, and α, β are calibration coefficients; It should be added that h is the current thickness measurement value, h_c is the minimum safe thickness of the material (taken according to the ASME B16.34 standard), k1, k2, k3 are environmental corrosion weight coefficients based on the NACE SP0193 standard, and are calibrated by electrochemical corrosion rate test; S3, use laser scanning to make the laser displacement meter perform three-dimensional scanning on the surface of the valve stem to obtain crack point cloud data; It should be added that a laser displacement meter with a wavelength of 650 nm is used to equally divide m sub-regions along the axial direction of the valve stem; S4, crack depth detection, based on point cloud data, calculate crack volume density CV = Σ(L_i × W_i ×D_i) / V_total, wherein L_i, W_i, D_i are the length, width and depth of a single crack, and V_total is the volume of the sub-region; It should be added that Σ is the number of cracks, L_i is the crack length calculated based on the laser triangulation method, W_i is the crack width calculated based on the laser triangulation method, and D_i is the crack depth calculated based on the laser triangulation method; S5, fusion DC and CV generate valve health index and decision.

[0023] Further, the collection of environmental parameters uses a multi-parameter sensor array built into the valve flow channel.

[0024] Further, the laser scanning is driven by a rotating mechanism to rotate the laser displacement meter 360° around the valve stem axis at a scanning speed of 10 rpm.

[0025] Further, the rotating mechanism includes an adaptive positioning clamp that can dynamically adjust the scanning radius according to the diameter of the valve stem.

[0026] Further, the crack depth detection uses a triangulation method, in which the laser displacement meter emits a linear laser to the surface of the valve stem, and the deformed laser stripe is captured by a CCD sensor, and the depth D_i is calculated based on the stripe distortion. It should be noted that the formula for calculating the depth is D_i=Δx×f / M×sinθ, where Δx is the spot displacement, f is the lens focal length, M is the magnification, and θ=45°±2° is the laser incidence angle.

[0027] Further, the fusion process in S5 includes establishing a health index HI = (DC^w1 × CV^w2) ^ [1 / (w1+w2)], where the weights w1 and w2 are determined by the valve material type, and when HI≥0.15, output a replacement suggestion. It should be noted that the weights w1 and w2 are called from the material database, and the transfer learning model uses a ResNet34 backbone network, which is pre-trained on ImageNet and fine-tuned on 100,000 crack images.

[0028] A valve quality detection system, comprising: A multi-modal sensor module 1 for collecting valve thickness, pitting and environmental parameters; An active scanning module 2 physically integrated with the multi-modal sensor module 1 for performing three-dimensional scanning of the valve stem; A data analysis module 3 electrically connected to the multi-modal sensor module 1 and the active scanning module 2 for performing comprehensive corrosion coefficient and crack volume density calculation; A decision output module 4 electrically connected to the data analysis module 3 for generating a visual health index report.

[0029] Further, the active scanning module 2 is built-in with an anti-shake mechanism, which includes a gyroscope to detect the vibration amplitude in real time and a piezoelectric ceramic driver to dynamically compensate for the optical path offset.

[0030] Further, the data analysis module 3 deploys a transfer learning model to identify corrosion-crack coupling failure modes through a pre-trained crack feature library.

[0031] Working principle: This method first divides the valve structure into n regular sub-blocks according to the valve structure. Through the integrated composite sensing module of ultrasonic probe, eddy current sensor and multi-parameter environmental probe, the thickness data, pitting signal and environmental parameters such as medium temperature, pH value and Cl - concentration of each block are synchronously collected. Among them, the ultrasonic thickness measurement module uses a 5MHz probe to obtain the current thickness h, and combines the initial thickness h0 and the minimum safe thickness h_c of the material to calculate the basic corrosion amount; the eddy current sensor detects the pitting area A_p through a 10kHz alternating magnetic field to quantify the degree of local corrosion; after the environmental parameters are collected by the built-in sensor array in real time, the corrosion promoting factor E_env = k1T + k2pH + k3C_Cl - is constructed, where k1, k2, k3 are environmental influence coefficients based on NACE standard fitting.

[0032] In the damage quantification link, the method generates a comprehensive corrosion coefficient DC = [(h-h_c) / ( h0-h_c) + α×(A_p / A0)]×(1+β×E_env), where α and β are calibration coefficients calibrated according to the corrosion resistance of the material, realizing the joint evaluation of uniform corrosion and local corrosion. For valve rod detection, a 650nm laser displacement meter is used to cooperate with a rotating mechanism for 360° three-dimensional scanning, and the crack depth D_i is calculated by triangulation method (D_i = Δx × f / M × sinθ), combined with point cloud data to extract crack length L_i and width W_i, and then the volume density CV = Σ(L_i × W_i × D_i) / V_total is obtained, breaking through the limitations of traditional two-dimensional features on strength evaluation.

[0033] In the health assessment stage, the system fuses the damage indicators based on HI = (DC^w1 × CV^w2)^[1 / (w1+w2)], where w1 and w2 are automatically retrieved from the database according to the valve material type (such as 316L stainless steel or ductile cast iron), and a replacement suggestion is triggered when HI≥0.15. The detection system includes four modules: the multi-modal sensor module 1 synchronously collects multi-source signals through time division multiplexing technology; the active scanning module 2 uses a motor-driven rotating bracket (scanning speed 10rpm) combined with a gyroscope-piezoelectric ceramic anti-shake mechanism (accuracy ±0.05mm) to realize full coverage of complex structures; the data analysis module 3 deploys a pre-trained ResNet34 transfer learning model to recognize corrosion-crack coupling failure modes through fine-tuning of 100,000 crack images; the decision output module 4 generates a visual report containing a three-dimensional damage cloud map.

Claims

1. A valve quality detection method, characterized by, The method comprises the following steps: S1, divide the valve into n sub-blocks, synchronously collect ultrasonic thickness data, eddy current pitting corrosion signals and medium environmental parameters of each sub-block, including temperature, pH value, Cl - concentration; S2, generating a comprehensive corrosion coefficient according to DC = [(h - h_c) / (h_0 - h_c) + α × (A_p / A_0)] × (1 + β × E_env), wherein A_p A_p is a pitting area of eddy current detection, A_0 is a reference area of a sub-block, E_env = k1T + k2pH + k3C_Cl - E_env is an environmental corrosion factor, and α and β are calibration coefficients; S3, laser scanning is adopted to make the laser displacement meter perform three-dimensional scanning on the surface of the valve stem to obtain crack point cloud data; S4, crack depth detection is performed, and crack volume density CV = Σ(L_i × W_i × D_i) / V_total is calculated based on the point cloud data, where L_i, W_i, and D_i are length, width, and depth of a single crack, and V_total is a sub-region volume; S5, a valve health index is generated by fusing DC and CV, and a decision is made.

2. The method of claim 1, wherein, The acquisition of the environmental parameters adopts a multi-parameter sensor array built in a valve flow channel.

3. The method of claim 1, wherein, The laser scanning drives the laser displacement meter to rotate 360° around the valve stem axis through a rotating mechanism.

4. The method of claim 3, wherein, The rotating mechanism comprises an adaptive positioning clamp that can dynamically adjust the scanning radius according to the valve stem diameter.

5. The method of claim 1, wherein The crack depth detection adopts a triangulation method, the laser displacement meter emits a linear laser to the surface of the valve stem, a CCD sensor captures the deformed laser stripe, and the depth D_i is calculated based on the stripe distortion variable, and the formula for calculating the depth is D_i = Δx × f / M × sinθ, where Δx is the spot displacement amount, f is the lens focal length, M is the magnification, and θ = 45° ± 2° is the laser incidence angle.

6. The method of claim 1, wherein The fusion process in S5 includes establishing a health index HI = (DC^w1 × CV^w2 ) ^ [1 / (w1+w2)], where the weights w1 and w2 are determined by the valve material type.

7. A valve quality detection system, characterized by, It comprises: A multi-modal sensor module for acquiring valve thickness, pitting, and environmental parameters; An active scanning module physically integrated with the multi-modal sensor module for performing three-dimensional scanning of the valve stem; A data analysis module electrically connected to the multi-modal sensor module and the active scanning module for performing comprehensive corrosion coefficient and crack volume density calculation; A decision output module electrically connected to the data analysis module for generating a visual health index report.

8. The valve quality detection system of claim 7, wherein, The active scanning module is built-in with an anti-shake mechanism, and the anti-shake mechanism comprises a gyroscope for real-time detection of vibration amplitude and a piezoelectric ceramic driver for dynamic compensation of optical path deviation.

9. The valve quality detection system of claim 7, wherein, The data analysis module deploys a transfer learning model to identify corrosion-crack coupling failure modes through a pre-trained crack feature library.

Images