Device and method for evaluating grinding skill of safety valve checking personnel

Abstract

The invention relates to a grinding skill evaluation device and method for safety valve verification personnel. A simulation grinding device comprises a pressing module, a moving module and a grinding plane. The pressing module is arranged on the grinding plane through the moving module, and after force is applied to the pressing module, the moving module is pushed to move on the grinding plane for grinding; the data acquisition module comprises a pressure sensor, a light-emitting element and a light sensation imager, and the pressure sensor is arranged between the pressing module and the moving module and used for detecting pressure data in the grinding process; the light-emitting element and the light sensation imager are both arranged on the moving module and used for detecting displacement data of the moving module relative to the grinding plane. The data processing module is connected with the pressure sensor and the light sensation imager and used for receiving and processing the pressure data and the displacement data so as to evaluate the grinding skill. The display module is connected with the data processing module and used for displaying the grinding track and the skill evaluation conclusion and simulating grinding deformation and roughness of the safety valve.

Description

Technical Field

[0001] This invention relates to the field of safety valve calibration technology, specifically to a device and method for assessing the grinding skills of safety valve calibration personnel. Background Technology

[0002] Safety valves are critical devices for ensuring the safe operation of pressure systems. Their sealing surfaces are easily damaged due to various reasons, leading to failure. Regular grinding and maintenance by calibration personnel are necessary to ensure sealing performance. Current grinding skill assessments mainly rely on manual experience and simple testing tools, lacking scientific, accurate, and quantitative assessment methods.

[0003] For example, traditional assessment methods often rely on the sealing performance of the safety valve's sealing surface after grinding to roughly judge the skill level of the personnel. This method is heavily influenced by subjective factors, resulting in a lack of objectivity. Furthermore, it fails to effectively analyze key parameters such as pressure changes and trajectory rationality during the grinding process, making it difficult to pinpoint the shortcomings of the personnel and hindering skill improvement.

[0004] In existing technologies, such as patent document CN107097146A, which relates to a grinding device, it determines whether the grinding tool is in contact with the substrate by monitoring motor current and travel distance, and is mainly used for wafer grinding. The problem is that it focuses on contact detection and position control, without addressing skill assessment or processing grinding trajectory, pressure distribution, or skill quantification. Patent document CN108319962A describes tool wear monitoring based on convolutional neural networks, analyzing wear through vibration signals and spectrum diagrams. However, it targets tool wear rather than grinding skill assessment, and does not involve comprehensive analysis of pressure and displacement, nor does it provide real-time feedback or trajectory evaluation. JP11126765A describes a grinding simulation method that calculates pressure and grinding amount through elasticity division, used for grinding simulation in semiconductor manufacturing. The problem is that it is a simulation method, does not involve personnel skill assessment, and lacks hardware devices for data acquisition and analysis.

[0005] Therefore, it is necessary to propose a device and method for evaluating the grinding skills of safety valve calibration personnel, so as to scientifically evaluate the grinding skills of calibration personnel. Summary of the Invention

[0006] This invention proposes a device and method for evaluating the grinding skills of safety valve calibration personnel, which is used to scientifically evaluate the grinding skills of calibration personnel.

[0007] To achieve the above objectives, the technical solution of the present invention is: a grinding skill assessment device for safety valve calibration personnel, comprising a simulated grinding device, a data acquisition module, a data processing module, and a display module. The simulated grinding device includes a pressing module, a moving module, and a grinding surface. The pressing module is placed on the grinding surface via the moving module, and grinding is performed by applying force to the pressing module to push the moving module to move on the grinding surface. The data acquisition module includes a pressure sensor, a light-emitting element, and a photosensitive imager. The pressure sensor is disposed between the pressing module and the moving module to detect pressure data during the grinding process. The light-emitting element and the photosensitive imager are both disposed on the moving module to detect displacement data of the moving module relative to the grinding surface. The data processing module is connected to the pressure sensor and the photosensitive imager respectively to receive and process pressure data and displacement data to assess grinding skills. The display module is connected to the data processing module to display the grinding trajectory, skill assessment conclusion, and the grinding deformation and roughness of the simulated safety valve.

[0008] Furthermore, multiple springs are provided between the pressing module and the moving module, and the pressure sensor is set corresponding to each spring to detect the pressure on each spring.

[0009] Furthermore, the light-emitting element emits light towards the grinding surface, and the photosensitive imager receives the light reflected by the grinding surface. By using high frequency to image the light source reflected by the grinding surface, and comparing the interval imaging information, the relative displacement between the moving module and the grinding surface is obtained. After parsing, the displacement data is uploaded to the data processing module, and after signal conversion, a grinding trajectory is formed and generated on the canvas.

[0010] A method for assessing the grinding skills of safety valve calibration personnel, using the aforementioned assessment device, includes the following steps:

[0011] S1: Pressure data during the grinding process is collected by a pressure sensor, and displacement data of the moving module relative to the grinding plane is collected by a light-emitting element and a photosensitive imager.

[0012] S2: Transmit pressure and displacement data to the data processing module for processing;

[0013] S3: Evaluate grinding skills based on processing results.

[0014] Furthermore, the data processing module employs a convolutional neural network (CNN) recognition model, which processes pressure and displacement data and outputs skill assessment conclusions. The CNN recognition model includes multiple convolutional layers, pooling layers, and fully connected layers. The convolutional layers are used to extract data features, the pooling layers are used for data dimensionality reduction, and the fully connected layers are used to generate assessment conclusions.

[0015] Furthermore, the operation formula for the convolutional layer is y = wx + b, where x is the input data of the convolutional layer, y is the output data of the convolutional layer, w is the selected convolutional kernel, and b is the bias after training; the pooling layer uses max pooling; the fully connected layer reduces the output data of the pooling layer to a one-dimensional vector and then maps it to the evaluation conclusion, and its operation formula is:

[0016] Output 评定结论 = f(a1x1+a2x2+a3x3+…+a n x n +b)

[0017] Where x n a is the input to the fully connected layer. n is the weight after training, and b is the bias after training.

[0018] Furthermore, in step S2, the data processing module generates a grinding trajectory based on the pressure data and displacement data; the data processing module inputs the pressure data and displacement data into the convolutional neural network recognition model, and the convolutional neural network recognition model outputs a skill assessment conclusion.

[0019] Furthermore, in step S3, the grinding skill is evaluated based on the skill assessment conclusion.

[0020] Furthermore, the method also includes a step of real-time calculation of the grinding deformation of the simulated safety valve: calculating the pressure, grinding force and theoretical grinding amount at each point on the sealing surface of the simulated safety valve based on pressure data and displacement data; calculating the thickness of each point after grinding based on the theoretical grinding amount to obtain the ground simulated safety valve, and calculating its roughness.

[0021] Furthermore, the pressure at a point i on the sealing surface of the simulated safety valve The calculation formula is:

[0022]

[0023] in, The pressure detected by the pressure sensor. The distance from the spring to the center of the simulated grinding device. Let i be the distance from point i to the center of the simulated grinding device;

[0024] The theoretical grinding force f at point i i The calculation formula is:

[0025]

[0026] in, Let i be the modulus of the velocity at point i. Let μ be the modulus of pressure at point i, and μ be the empirical coefficient relating pressure and velocity to grinding force.

[0027] The theoretical grinding amount at point i l i The calculation formula is:

[0028]

[0029] in, The empirical coefficient representing the relationship between grinding force, abrasive particle size, and grinding amount is derived from experiments using abrasive brands. C is a random number related to abrasive particle size.

[0030] C = C1·C2

[0031] Where C1 follows a binomial distribution with a value of 0 or 1, and the probability is based on the density of the abrasive particles; C2 follows a normal distribution with a probability density function of:

[0032]

[0033] Where ρ is the abrasive particle size determined by the abrasive brand, the length of point i after abrasion is:

[0034]

[0035] in, Let i be the thickness after the nth grinding. Let i be the thickness before the nth grinding. The thickest part of the sealing surface, l i Let i be the theoretical grinding amount.

[0036] The thickness of all points on the sealing surface after grinding is calculated using the above formula to obtain the simulated safety valve after grinding. The surface roughness Ra value is then calculated to obtain the surface roughness, which is displayed on the monitor for observation by safety valve calibration personnel.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] 1. By setting up a pressure sensor and a displacement detection component, this invention can collect pressure and displacement data in real time during the grinding process, providing an objective and accurate quantitative basis for the evaluation of grinding skills, and overcoming the subjectivity and inaccuracy of traditional evaluations that rely on human experience.

[0039] 2. The grinding trajectory is generated using the trajectory generation unit, which can intuitively display the path and pressure changes during the grinding process, making it easier to analyze the rationality of the grinding trajectory and helping calibration personnel to find problems in their own operation.

[0040] 3. By using a convolutional neural network recognition model to process the data, it can automatically extract data features and generate skill assessment conclusions, which improves the efficiency and accuracy of the assessment and has good generalization ability.

[0041] 4. It can calculate the grinding deformation and roughness of the simulated safety valve in real time, allowing calibration personnel to understand the grinding effect in real time and adjust the operation in a timely manner, which helps to improve their grinding skills. Attached Figure Description

[0042] Figure 1 This is a schematic diagram of a grinding skill assessment device for safety valve calibration personnel;

[0043] Figure 2 In Example 1, the pressure and displacement data are uploaded and then converted into a grinding trajectory, which is then generated as a schematic diagram on the canvas.

[0044] Figure 3 This is a schematic diagram of the input / output data of the convolutional neural network recognition model in Example 2;

[0045] Figure 4 This is a schematic diagram illustrating the composition of the convolutional neural network recognition model in Example 2, as well as its feature extraction and evaluation conclusion generation.

[0046] Figure 5 This is a schematic diagram of the evaluation device and method in Example 3 calculating the grinding deformation of the safety valve in real time during the simulated grinding process;

[0047] Figure 6 This is the solution diagram from Example 3;

[0048] In the diagram: 1-Pressing module, 2-Spring, 3-Pressure sensor, 4-Moving module, 5-Grinding plane, 6-Guide post, 7-Light emitting diode, 8-Photosensitive micro-imager, 9-Simulated grinding device, 10-Grinding trajectory, 11-Canvas, 12-Grinding data, 13-Convolutional neural network recognition model, 14-Skill assessment conclusion, 15-Multi-layer convolutional layer, 16-Pooling layer, 17-Fully connected layer, 18-Assessment conclusion, 19-Display, 20-Simulated safety valve, 21-Table. Detailed Implementation

[0049] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0050] Example 1:

[0051] refer to Figure 1 , 2A device and method for assessing the grinding skills of safety valve testers, based on a simulated grinding device 9, which consists of a pressing module 1, a spring 2, a pressure sensor 3, a moving module 4, a guide column 6, a light-emitting diode 7, and a photosensitive micro-imager 8.

[0052] When the grinder uses the simulated grinding device 9, he applies force to the pressing module 1 and pushes the moving module 4 to move on the grinding surface 5.

[0053] The pressing module 1 and the guide post 6 are in clearance fit. When the pressing module 1 is subjected to force, it generates different pressures on multiple springs 2 and transmits them to the pressure sensor 3. The pressure sensor 3 senses the pressure, analyzes it, and uploads the data.

[0054] The moving module 4 is embedded with a light-emitting diode 7 and a photosensitive micro-imager 8. The grinding surface 5 reflects the light source of the light-emitting diode 7 to the photosensitive micro-imager 8. The photosensitive micro-imager 8 uses a higher frequency to image the light source reflected by the grinding surface 5. By comparing the interval imaging information, the relative displacement between the moving module 4 and the grinding surface 5 is realized. After analysis, the displacement data is uploaded.

[0055] After the pressure and displacement data are uploaded, they are converted into a grinding trajectory 10 and generated on the canvas 11.

[0056] Example 2:

[0057] refer to Figure 3 , 4 A device and method for assessing the grinding skills of safety valve testers obtains grinding data 12 from simulated grinding, inputs the grinding data 12 into a convolutional neural network recognition model 13, and outputs a skill assessment conclusion 14 after the model calculates.

[0058] The convolutional neural network recognition model 13 consists of multiple convolutional layers 15 and pooling layers 16. It reduces the image of the polished data 12 to a certain size and extracts features, then maps them to a fully connected layer 17. After dimensionality reduction, it generates an evaluation conclusion 18.

[0059] The formula for convolutional layer 15 is y = wx + b, where x is the input data of convolutional layer 15, y is the output data of convolutional layer 15, w is the selected convolutional kernel, and b is the bias after training.

[0060] Pooling layer 16 uses max pooling to take the maximum value of the output generated by convolutional layer 15 within a certain matrix region, thereby reducing the dimensionality of the data.

[0061] Fully connected layer 17 reduces the output data of pooling layer 16 to a one-dimensional vector, and then maps it to the evaluation conclusion 18, as shown in the formula Output. 评定结论 = f(a1x1+a2x2+a3x3+…+a n x n+b), where x n a is the input to the fully connected layer. n is the weight after training, and b is the bias after training.

[0062] The transfer function between the convolutional layer 15, the pooling layer 16, and the fully connected layer 17 is linked by the ReLU function, where ReLU(x) = max(0,x), and x is the result of the layer operation, thereby making the convolutional neural network recognition model 13 a non-linear model.

[0063] Example 3:

[0064] refer to Figure 5 , Figure 6 A device and method for assessing the grinding skills of safety valve calibration personnel can calculate the grinding deformation of a safety valve in real time during simulated safety valve grinding. Before grinding begins, a simulated safety valve 20 with a non-compliant sealing surface is randomly generated on the display 19. The calibration personnel select a suitable grinding compound on the display 19 based on the state of the simulated safety valve 20. When the simulated grinding device 9 performs simulated grinding on the grinding plane 5, the display 19 shows the grinding deformation and roughness of the simulated safety valve 20 in real time. The derivation process is as follows:

[0065] The grinding deformation of the simulated safety valve 20 is calculated based on the pressure and displacement signals output by the grinding device. In this embodiment, four springs 2 are used, and their distance from the center of the simulated grinding device 9 is [missing information]. The pressure measured by pressure sensor 3 is

[0066] Solution Reference Figure 6 As shown, point i is a point on the sealing surface of the safety valve, then the pressure at point i is:

[0067]

[0068] in, Let i be the distance from point i to the center of the simulated grinding device 9. Then the theoretical grinding force acting on point i is:

[0069]

[0070] in, Let i be the modulus of the velocity at point i. Let μ be the modulus of pressure at point i, and μ be the empirical coefficient relating pressure and velocity to grinding force, derived from experiments using the abrasive grade. Then, the theoretical grinding amount generated at point i is:

[0071]

[0072] in, This is an empirical coefficient relating grinding force, abrasive particle size, and grinding amount, derived from experiments using abrasive brands. C is a random number related to abrasive particle size.

[0073] C = C1·C2

[0074] C1 follows a binomial distribution with a value of 0 or 1, and the probability is based on the density of the abrasive particles. C2 follows a normal distribution with a probability density function of:

[0075]

[0076] Wherein, ρ is the abrasive particle size determined by the abrasive grade.

[0077] The length of point i after grinding is:

[0078]

[0079] in, Let i be the thickness after the nth grinding. Let i be the thickness before the nth grinding. The thickest part of the sealing surface, l i Let i be the theoretical grinding amount.

[0080] The thickness of all points on the sealing surface after grinding is calculated using the above formula to obtain the simulated safety valve 20 after grinding. The surface roughness Ra value is then calculated to obtain the surface roughness, which is displayed on the display 19 for observation by safety valve calibration personnel.

Claims

1. A device for assessing the grinding skills of safety valve calibration personnel, characterized in that, The device includes a simulated grinding apparatus, a data acquisition module, a data processing module, and a display module. The simulated grinding apparatus comprises a pressing module, a moving module, and a grinding surface. The pressing module is placed on the grinding surface via the moving module. Applying force to the pressing module pushes the moving module to move on the grinding surface for grinding. The data acquisition module includes a pressure sensor, a light-emitting element, and a photosensitive imager. The pressure sensor is positioned between the pressing module and the moving module to detect pressure data during the grinding process. The light-emitting element and the photosensitive imager are both located on the moving module to detect displacement data of the moving module relative to the grinding surface. The data processing module is connected to both the pressure sensor and the photosensitive imager to receive and process the pressure and displacement data to evaluate grinding skills. The display module is connected to the data processing module and is used to display the grinding trajectory, skill assessment conclusion, and grinding deformation and roughness of the simulated safety valve.

2. The safety valve calibration personnel grinding skill assessment device according to claim 1, characterized in that, Multiple springs are also provided between the pressing module and the moving module, and the pressure sensor is set corresponding to the spring to detect the pressure on each spring.

3. The safety valve calibration personnel grinding skill assessment device according to claim 1, characterized in that, The light-emitting element emits light towards the grinding surface, and the photosensitive imager receives the light reflected by the grinding surface. It then uses high frequency to image the light source reflected by the grinding surface, compares the interval imaging information, obtains the relative displacement between the moving module and the grinding surface, analyzes the data, uploads the displacement data to the data processing module, converts the signal to form a grinding trajectory, and generates it on the canvas.

4. A method for assessing the grinding skills of safety valve calibration personnel, using the grinding skill assessment device for safety valve calibration personnel as described in any one of claims 1-3, characterized in that, Includes the following steps: S1: Pressure data during the grinding process is collected by a pressure sensor, and displacement data of the moving module relative to the grinding plane is collected by a light-emitting element and a photosensitive imager. S2: Transmit pressure and displacement data to the data processing module for processing; S3: Evaluate grinding skills based on processing results.

5. The method for assessing the grinding skills of safety valve calibration personnel according to claim 4, characterized in that, The data processing module employs a convolutional neural network (CNN) recognition model, which processes pressure and displacement data and outputs skill assessment conclusions. The CNN recognition model includes multiple convolutional layers, pooling layers, and fully connected layers. The convolutional layers are used to extract data features, the pooling layers are used for data dimensionality reduction, and the fully connected layers are used to generate assessment conclusions.

6. The method for assessing the grinding skills of safety valve calibration personnel according to claim 5, characterized in that, The convolutional layer's operation formula is y = wx + b, where x is the input data of the convolutional layer, y is the output data of the convolutional layer, w is the selected convolutional kernel, and b is the bias after training. The pooling layer uses max pooling. The fully connected layer reduces the pooling layer's output data to a one-dimensional vector and then maps it to the evaluation conclusion; its operation formula is: Output 评定结论 =f(a1x1+a2x2+a3x3+…+a n x n +b) Where, x n a is the input to the fully connected layer. n is the weight after training, and b is the bias after training.

7. The method for assessing the grinding skills of safety valve calibration personnel according to claim 5, characterized in that, In step S2, the data processing module generates a grinding trajectory based on pressure data and displacement data; the data processing module inputs the pressure data and displacement data into a convolutional neural network recognition model, and the convolutional neural network recognition model outputs a skill assessment conclusion.

8. The method for assessing the grinding skills of safety valve calibration personnel according to claim 7, characterized in that, In step S3, the grinding skill is evaluated based on the skill assessment conclusion.

9. The method for assessing the grinding skills of safety valve calibration personnel according to claim 8, characterized in that, It also includes the step of real-time calculation of the grinding deformation of the simulated safety valve: calculating the pressure, grinding force and theoretical grinding amount at each point on the sealing surface of the simulated safety valve based on the pressure data and displacement data; calculating the thickness of each point after grinding based on the theoretical grinding amount to obtain the ground simulated safety valve, and calculating its roughness.

10. The method for assessing the grinding skills of safety valve calibration personnel according to claim 9, characterized in that, The pressure at a point i on the sealing surface of the simulated safety valve The calculation formula is: in, The pressure detected by the pressure sensor. The distance from the spring to the center of the simulated grinding device. Let i be the distance from point i to the center of the simulated grinding device; The theoretical grinding force f at point i i The calculation formula is: in, Let i be the modulus of the velocity at point i. Let μ be the modulus of pressure at point i, and μ be the empirical coefficient relating pressure and velocity to grinding force. The theoretical grinding amount at point i l i The calculation formula is: in, The empirical coefficient representing the relationship between grinding force, abrasive particle size, and grinding amount is derived from experiments using abrasive brands. C is a random number related to abrasive particle size. C = C1·C2 Where C1 follows a binomial distribution with a value of 0 or 1, and the probability is based on the density of the abrasive particles; C2 follows a normal distribution with a probability density function of: Where ρ is the abrasive particle size determined by the abrasive brand, the length of point i after abrasion is: in, Let i be the thickness after the nth grinding. Let i be the thickness before the nth grinding. The thickest part of the sealing surface, l i Let i be the theoretical grinding amount. The thickness of all points on the sealing surface after grinding is calculated using the above formula to obtain the simulated safety valve after grinding. The surface roughness Ra value is then calculated to obtain the surface roughness, which is displayed on the monitor for observation by safety valve calibration personnel.

Images