A method for reconstructing internal morphology of an interface

Abstract

The present invention discloses a method for reconstructing the internal morphology of an interface. First, the interface morphology and the interface friction force experimental results are obtained through a rough interface friction experiment; then, the direction angle of the surface morphology is divided into different characteristic directions, and the interface morphology is analyzed by using a deep learning technology based on computer vision to obtain the distribution ratio of the interface morphology corresponding to the characteristic direction, and the morphology orientation ratio recognition result is obtained based on this, and then the "friction force-orientation ratio" scaling rate relationship is jointly constructed with the interface friction force experimental result, and the morphology orientation ratio of the target sample is calculated by combining the relationship between the spring mass block model and the friction force; finally, the reconstructed morphology is obtained according to the morphology orientation ratio and the interface internal morphology reconstruction formula based on the characteristic direction. The present invention can realize the morphology reconstruction and prediction at any time during the friction process of the target sample, and the power spectrum density of the reconstructed interface morphology and the actual morphology are well matched.

Description

Technical Field

[0001] The invention belongs to the technical field of rough surface micromorphology research, and in particular relates to a method for reconstructing the internal morphology of an interface. Background Art

[0002] There are a large number of contact interfaces inside major engineering equipment. Contact interfaces are the key parts for force transmission, electrical conduction and heat transfer between components. Due to the complexity, randomness and opacity of rough interfaces, the rough interface morphology is difficult to detect and observe in real time, which makes it difficult to effectively characterize the interface contact characteristics. This poses a huge challenge to the evaluation and design of the interface performance of related engineering equipment. Therefore, the development of interface internal morphology reconstruction methods is of great value in practical applications.

[0003] The existing rough surface 3D topography reconstruction methods mainly include rough surface statistical methods that assume the shape of micro-convex bodies, digital rough surface methods based on statistical parameters, fractal description methods, and simulation methods based on real measurement data. These methods provide means for evaluating interface contact performance to a certain extent, but have the following shortcomings:

[0004] 1. The statistical method that assumes the shape of micro-convex bodies simplifies the real geometric characteristics of the rough surface morphology and cannot reflect the complexity and non-uniformity of the micro-convex body shape. In addition, when friction occurs between components, it is impossible to reconstruct the dynamic evolution of the interface morphology;

[0005] 2. Based on statistical parameters, such as morphological roughness, skewness and kurtosis, it can only reflect the overall characteristics of the rough surface and lacks accurate description of local details. The generated digital rough surface lacks the physical basis of actual measurement in dynamic prediction, and it is difficult to truly simulate the evolution of interface morphology;

[0006] 3. Fractal methods are suitable for describing rough surfaces with self-similar properties, but the actual interface morphology may not fully conform to the fractal assumption, especially for multi-scale or non-uniform surfaces. In addition, although the interface fractal dimension can provide the overall characteristics of the morphology, it lacks the ability to describe time-related and local deformation behaviors when predicting dynamic evolution, which limits the prediction of the dynamic evolution of the interface morphology;

[0007] 4. The simulation method based on real measurement data can provide a high-precision description of the surface morphology, but it still cannot achieve dynamic prediction of the interface morphology, but relies on static measurement results. Summary of the invention

[0008] In order to solve the problems existing in the above-mentioned background technology, the present invention provides a method for reconstructing the internal morphology of an interface.

[0009] In order to achieve the above object, the present invention adopts the following technical solution:

[0010] A method for reconstructing the internal morphology of an interface comprises the following steps:

[0011] S1. By conducting rough interface friction experiments on different materials, the evolution characteristics of the rough surface morphology and friction force during the friction process are measured, and the experimental results of the interface morphology and rough interface friction force at different friction moments are obtained;

[0012] S2. Use surface direction as a feature, characterize the surface morphology by defining different surface directions, divide the direction angles of the surface morphology into different characteristic directions, use deep learning technology based on computer vision to analyze the interface morphology, obtain the distribution ratio of the interface morphology corresponding to the characteristic direction at different friction moments, and obtain the recognition results of the morphology orientation ratio at different friction moments according to the distribution ratio;

[0013] S3, plotting the time, the rough interface friction test results and the morphology orientation ratio identification results as x variables, y variables and z variables in the spatial coordinate system, fitting the relationship between normalized friction and morphology orientation ratio and time, determining the fitting parameters according to the interface performance parameters, and obtaining the "friction-orientation ratio" scaling relationship:

[0014]

[0015] In the formula, is the normalized friction force, is the proportion of morphological orientation, is the friction coefficient, e is a natural constant, is the static-to-dynamic transition time, For time, is the static friction coefficient, is the coefficient of kinetic friction;

[0016] S4. Equivalent the block interface friction to a spring mass block model. The friction force is calculated based on the spring mass block model. The friction force satisfies the following formula:

[0017]

[0018] In the formula, represents the friction force, Indicates The position of the mass relative to the initial position, express About Time The first derivative of is the static friction of the i-th mass block, and the static friction is based on the static equilibrium Calculated, represents the spring stiffness coefficient, The value does not exceed the maximum static friction , is the static friction coefficient, F i It is i Normal load on the mass, It is The kinetic friction of the block, pass Calculated, is the coefficient of kinetic friction;

[0019] S5. Calculate the friction force of the target sample at any time according to the spring mass block model and the friction force formula, and insert the obtained friction force into the "friction force-orientation ratio" scaling ratio relationship after normalization to calculate the morphology orientation ratio of the target sample;

[0020] S6. Establish the interface internal morphology reconstruction formula based on the characteristic direction:

[0021]

[0022] In the formula, Reconstruct the shape of the target, is the morphology orientation ratio, is the interface morphology distribution characteristics before friction, , is the weight ratio, =0.6, =0.4, The interface morphology generated by machine learning is interpolated to the morphology closest to the input morphology orientation ratio. is the average interface morphology generated by machine learning based on the input morphology data;

[0023] The rough surface morphology image dataset of the target sample is processed using machine learning methods. and , and the calculated morphology orientation ratio of the target sample are substituted into the interface internal morphology reconstruction formula to obtain the target reconstructed morphology.

[0024] Furthermore, in S1, the rough interface friction test result is obtained by multiplying the deformation value of the force ring at different times by its stiffness coefficient, and the rough interface friction test result can be obtained by a rough interface friction test system.

[0025] Furthermore, in S2, the method of analyzing the interface morphology using the deep learning technology based on computer vision is as follows: inputting regular surface morphology data samples into the deep learning model based on computer vision, learning different test samples for each characteristic direction, using an artificial neural network to generate a classifier for the rough surface and its characteristic direction, and obtaining the judgment criteria for the corresponding characteristic direction by training the sample groups of different characteristic directions, and determining the distribution ratio of the characteristic direction corresponding to the interface morphology at different friction moments according to the judgment criteria. Therefore, the interface morphology at different friction moments can be obtained by the MicroXAM-800 in the friction experiment, and the distribution ratio of various characteristic directions of the interface morphology can be determined according to the judgment criteria.

[0026] Furthermore, in S2, the direction angles of the surface morphology include horizontal, vertical, 30°, 60°, 105° and 120°.

[0027] Further, in S4, when the block interface friction is equivalent to a spring mass block model, the established block motion equation is:

[0028]

[0029] In the formula, Indicates The displacement of the mass relative to the initial position, n represents the number of mass blocks, m i Indicates The mass of the mass block, represents the spring stiffness coefficient, express About Time The second-order derivative of is the tangential load, ,in, x 0 is the displacement of the driving block relative to the initial position, Indicates that the first mass moves at a constant speed The driving spring stiffness when driving, It's friction.

[0030] Compared with the shortcomings and deficiencies of the prior art, the present invention has the following beneficial effects:

[0031] The interface internal morphology reconstruction method provided by the present invention can realize the morphology reconstruction and prediction at any time during the friction process of the target sample, and the power spectrum density of the reconstructed interface morphology is well consistent with the actual morphology. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 11 is the interface morphology of different samples before the friction test provided by the embodiment of the present invention, and (a), (b), and (c) respectively represent the interface morphology of stainless steel, cast iron, and aluminum before the friction test;

[0033] Figure 2 1 is a diagram of the interface morphology evolution of a stainless steel sample provided by an embodiment of the present invention during the friction process, wherein (a), (b), and (c) are the sample morphologies at the friction moments of 0s, 10s, and 30s in the static friction stage, respectively; (d) is the sample morphology at the moment when the static friction stage transitions to the dynamic friction stage; (e) and (f) are the sample morphologies at the friction moments of 300s and 420s in the dynamic friction stage, respectively;

[0034] Figure 3 The friction force of the rough interface between stainless steel and cast iron provided by the embodiment of the present invention varies with time;

[0035] Figure 4 is the recognition result of the morphology orientation ratio of stainless steel and cast iron provided by the embodiment of the present invention;

[0036] Figure 5 is the “friction force-orientation ratio” scaling ratio of stainless steel ((a) in the figure) and cast iron ((b) in the figure) provided by the embodiment of the present invention;

[0037] Figure 6 Schematic diagram of a block interface friction equivalent to a spring mass block model provided by an embodiment of the present invention, wherein (a) is a block interface friction model, and (b) is a spring mass block model;

[0038] Figure 7 yes Figure 6 The enlarged view of point A in the middle;

[0039] Figure 8 (a) and (b) are comparison diagrams of the original interface morphology and the reconstructed morphology at different friction stages provided by an embodiment of the present invention. Figure 8 (c) and (d) are power spectrum density diagrams of the original interface morphology and the reconstructed morphology at different friction stages provided by the embodiments of the present invention;

[0040] Fig. 9 It is the prediction result of reconstructing the rough interface morphology at any time during the friction process provided by an embodiment of the present invention. In the figure, (a), (c), (e), (f), and (h) are the morphologies measured at the friction times of 0s, 10s, 30s, 79s, and 300s, respectively, and (b), (d), (g), and (i) are the morphologies reconstructed at the friction times of 5s, 20s, 200s, and 360s, respectively. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0042] S1. Conducting rough interface friction experiments

[0043] Conduct friction experiments on the rough interfaces of stainless steel, cast iron, and aluminum. The interface morphology of the samples before the friction experiments is as follows: Figure 1 As shown in the figure, (a), (b), and (c) respectively represent the interface morphology of stainless steel, cast iron, and aluminum before the friction experiment. The evolution characteristics of the rough surface morphology and friction force during the friction process are measured. The interface morphology at different friction moments is measured by MicroXAM-800 non-contact optical profilometer to obtain the interface morphology at different friction moments. The interface morphology evolution of the stainless steel sample during the friction process is shown in Figure 2 As shown in the figure, (a), (b), and (c) are the sample morphologies at 0s, 10s, and 30s of friction in the static friction stage, respectively; (d) is the sample morphology at the transition from the static friction stage to the dynamic friction stage; (e) and (f) are the sample morphologies at 300s and 420s of friction in the dynamic friction stage, respectively, and the surface shows typical orientation characteristics; the experimental parameters are normal load F=1000N and shear rate V=0.8mm / min. The friction force is calculated as the product of the deformation value of the force ring at different times and its stiffness coefficient. The rough interface friction force experimental results are obtained through the rough interface friction force test system. The rough interface friction force of stainless steel and cast iron varies with time as shown in the following figure. Figure 3 As shown, the dotted line in the figure represents the transition moment from the static friction stage to the dynamic friction stage. The stage before the dotted line is the static friction stage, and the stage after the dotted line is the dynamic friction stage.

[0044] S2. Extracting directional features of surface morphology based on computer vision recognition methods

[0045] The rough surface evolves continuously during the friction process and shows different morphological features at different times. In order to define the change of surface morphology during the friction process, the surface direction is used as the feature. The surface morphology is characterized by defining different surface directions. According to the Cartesian coordinate system, the direction angles of the surface morphology are divided into different characteristic directions, including horizontal, vertical, 30°, 60°, 105° and 120°. The deep learning technology based on computer vision is used to input regular surface morphology data samples, and at least 20 groups of different test samples are learned for each characteristic direction. To ensure the credibility and rationality of the learning results, each group of samples has different morphological features while having the same characteristic direction.

[0046] The classifier of rough surface and its characteristic direction is generated by artificial neural network. By training the sample groups with different characteristic directions, the judgment criteria of the corresponding characteristic directions are obtained. This process establishes the relevant discrimination model under different surface characteristic direction trends, which can judge whether the input morphology of the test sample contains a specific characteristic direction and provide the corresponding percentage composition. Therefore, for the interface morphology obtained by MicroXAM-800 at different friction moments in the friction experiment, the distribution ratio of various characteristic directions can be determined according to the judgment criteria, and the recognition results of the morphology orientation ratio at different friction moments can be obtained. The recognition results of the morphology orientation ratio of stainless steel and cast iron are as follows: Figure 4 shown.

[0047] S3. Constructing the scaling relationship between friction force and orientation ratio

[0048] According to the experimental results of rough interface friction and the results of morphology orientation ratio identification, the curves of the two changing with time are drawn respectively. Analysis of the curves shows that the changing trends of the two are highly consistent. Therefore, the time, the experimental results of rough interface friction and the results of morphology orientation ratio identification are plotted as x variables, y variables and z variables in the spatial coordinate system respectively, and the relationship between normalized friction and morphology orientation ratio and time is fitted. The fitting parameters are determined according to the interface performance parameters, so as to obtain the "friction-orientation ratio" scaling relationship, such as Figure 5 As shown in the figure, (a) and (b) are the "friction force-orientation ratio" scaling rates of stainless steel and cast iron, respectively. The morphology orientation identification result is adjusted by formula (1) and is consistent with the evolution of friction force.

[0049] The obtained "friction force-orientation ratio" scaling relationship is:

[0050] (1)

[0051] In the formula, is the normalized friction force, is the proportion of morphological orientation, is the friction coefficient, e is a natural constant, is the static-to-dynamic transition time, For time, is the static friction coefficient, is the coefficient of kinetic friction.

[0052] S4. Establishing a dynamic coupling friction model

[0053] The morphology orientation ratio can be calculated using the scaling relationship of “friction force-orientation ratio”, but the friction force of the friction interface needs to be obtained. In order to calculate the friction force at any time instead of experimentally measuring it, a dynamic coupling friction model is established, and the block interface friction is equivalent to a spring mass block model. The friction force at any time (including the static friction stage and the dynamic friction stage) is obtained by equivalent calculation of the spring mass block model, such as Figure 6 As shown in the figure, (a) is the block interface friction model, (b) is the spring mass block model, Figure 6 See the enlarged view of A in the middle. Figure 7 , the equation of motion of the block is:

[0054] (2)

[0055] In the formula, Indicates The displacement of the mass relative to the initial position, n represents the number of mass blocks, m i Indicates The mass of the mass block, represents the spring stiffness coefficient, express About Time The second-order derivative of is the tangential load, ,in, x 0 is the displacement of the driving block relative to the initial position, Indicates that the first mass moves at a constant speed The driving spring stiffness when driving, is the friction force, then the friction force satisfies the following formula:

[0056] (3)

[0057] In the formula, express About Time The first derivative of is the static friction of the i-th mass block, and the static friction is based on the static equilibrium Calculated, The value does not exceed the maximum static friction , is the static friction coefficient, F i It is i Normal load on the mass, It is The kinetic friction of the block, pass Calculated, is the coefficient of kinetic friction.

[0058] S5. Calculate the morphology orientation ratio of the target sample

[0059] For the target sample under study, the friction force of the target sample at any time is first calculated according to the friction force formula of S4, and then the obtained friction force is normalized and inserted into formula (1) to obtain the morphological orientation ratio of the target sample.

[0060] S6. Reconstruction of rough interface morphology based on feature orientation

[0061] The rough surface morphology is closely related to the initial characteristic direction and the morphology of the sample before friction. Combined with the morphology orientation ratio, the formula for reconstructing the internal morphology of the interface based on the characteristic direction is constructed:

[0062] (4)

[0063] In the formula, Reconstruct the shape of the target, is the proportion of morphological orientation, is the distribution characteristics of the interface morphology before friction, , is the weight ratio, =0.6, =0.4, The interpolated interface morphology is generated by machine learning for the morphology interpolation that is closest to the input morphology orientation ratio. is the average interface morphology generated by machine learning based on the input morphology data.

[0064] The rough surface morphology image dataset of the target sample was processed using a machine learning method. 20 rough surface morphology images before and after friction were input, and the size of all images was standardized to 256×256 pixels to ensure the uniformity of the image and compatibility with the machine learning model. The average rough surface morphology was generated based on the image dataset. , which represents the overall characteristics of the rough surface in the data set; the interpolated interface morphology is obtained by inputting the orientation ratio of the target morphology , will get and , and the calculated morphology orientation ratio of the target sample is substituted into the interface internal morphology reconstruction formula to obtain the target reconstructed morphology, as Figure 8As shown, (a) is a comparison between the original morphology (left) and the reconstructed morphology (right) in the static friction stage, (b) is a comparison between the original morphology (left) and the reconstructed morphology (right) in the dynamic friction stage, (c) is a power spectrum density diagram of the original morphology and the reconstructed morphology in the static friction stage, and (d) is a power spectrum density diagram of the original morphology and the reconstructed morphology in the dynamic friction stage. The quality of the interface morphology reconstruction is determined by the comparison of the power spectrum density. Therefore, the power spectrum density of the reconstructed interface morphology is compared with the actual morphology, and the agreement is good, which verifies the accuracy of the interface internal morphology reconstruction method of the present invention.

[0065] The rough interface morphology at any time during the friction process of stainless steel material is reconstructed by the interface internal morphology reconstruction method of the present invention, and the prediction results are as follows: Fig. 9 As shown, (a), (c), (e), (f), and (h) are the morphologies measured at the friction time of 0s, 10s, 30s, 79s, and 300s, respectively, and (b), (d), (g), and (i) are the morphologies reconstructed at the friction time of 5s, 20s, 200s, and 360s, respectively. Fig. 9 It shows that the interface internal morphology reconstruction method of the present invention can realize the reconstruction and prediction of the rough interface morphology at any time during the friction process, making it possible to detect the interface morphology during the friction process of opaque solids, and providing a new method for revealing the friction mechanism and predicting friction behavior.

[0066] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for reconstructing the internal morphology of an interface, characterized in that: The steps include: S1. By conducting rough interface friction experiments on different materials, the evolution characteristics of the rough surface morphology and friction force during the friction process are measured, and the experimental results of the interface morphology and rough interface friction force at different friction moments are obtained; S2. Use surface direction as a feature, characterize the surface morphology by defining different surface directions, divide the direction angles of the surface morphology into different characteristic directions, use deep learning technology based on computer vision to analyze the interface morphology, obtain the distribution ratio of the interface morphology corresponding to the characteristic direction at different friction moments, and obtain the recognition results of the morphology orientation ratio at different friction moments according to the distribution ratio; S3. Using the rough interface friction force experimental results and the morphology orientation ratio identification results, a "friction force-orientation ratio" scaling relationship is constructed: In the formula, is the normalized friction force, D c is the ratio of morphology orientation, μ is the friction coefficient, e is the natural constant, t c is the static-to-dynamic transition time, t is time, μ s is the static friction coefficient, μ k is the coefficient of kinetic friction; S4. Equivalent the block interface friction to a spring mass block model. The friction force is calculated based on the spring mass block model. The friction force satisfies the following formula: In the formula, f(x i ) represents the friction force, x i represents the displacement of the ith mass block relative to the initial position, Represents x i The first derivative with respect to time t is is the static friction of the i-th mass block, and the static friction is based on the static equilibrium Calculated, k s,i represents the spring stiffness coefficient connecting the ith mass block, The value does not exceed the maximum static friction μ s is the static friction coefficient, F i is the normal load of the ith mass block, is the kinetic friction of the ith block, pass Calculated, μ k is the coefficient of kinetic friction; S5. Calculate the friction force of the target sample at any time according to the spring mass block model and the friction force formula, and insert the obtained friction force into the "friction force-orientation ratio" scaling ratio relationship after normalization to calculate the morphology orientation ratio of the target sample; S6. Establish the interface internal morphology reconstruction formula based on the characteristic direction: Where Y Re Reconstruct the shape of the target, D c is the proportion of morphological orientation, Y0 is the distribution characteristic of interface morphology before friction, β1 and β2 are weight proportions, β1=0.6, β2=0.4, The interface morphology generated by machine learning is interpolated to the morphology closest to the input morphology orientation ratio. is the average interface morphology generated by machine learning based on the input morphology data; The rough surface morphology image dataset of the target sample is processed using machine learning methods. and The calculated morphology orientation ratio of the target sample is substituted into the interface internal morphology reconstruction formula to obtain the target reconstructed morphology.

2. The method for reconstructing the internal morphology of an interface according to claim 1, characterized in that: In S1, the rough interface friction test result is obtained by measuring the product of the deformation value of the force ring at different times and its stiffness coefficient.

3. The method for reconstructing the internal morphology of an interface according to claim 1, characterized in that: In S2, the method of analyzing interface morphology using deep learning technology based on computer vision is as follows: inputting regular surface morphology data samples into the deep learning model based on computer vision, learning different test samples for each characteristic direction, using an artificial neural network to generate a classifier for the rough surface and its characteristic direction, and obtaining the judgment criteria for the corresponding characteristic directions by training sample groups with different characteristic directions, and determining the distribution ratio of the characteristic directions corresponding to the interface morphology at different friction moments according to the judgment criteria.

4. The method for reconstructing the internal morphology of an interface according to claim 1 or 3, characterized in that: In S2, the direction angles of the surface morphology include horizontal, vertical, 30°, 60°, 105° and 120°.

5. The method for reconstructing the internal morphology of an interface according to claim 1, characterized in that: In S4, when the block interface friction is equivalent to a spring mass block model, the established block motion equation is: In the formula, x i represents the displacement of the i-th mass block relative to the initial position, n represents the number of mass blocks, m i represents the mass of the ith mass block, k s,1 represents the spring stiffness coefficient of the first mass block, k s,i represents the spring stiffness coefficient connecting the i-th mass block, k s,n represents the spring stiffness coefficient connecting the nth mass block, Represents x i The second derivative with respect to time t, F X is the tangential load, F X =K(x0-x1)=K(Vt-x1), where x0 is the displacement of the drive block relative to the initial position, K is the stiffness of the drive spring when the first mass block is driven at a constant speed V, and f(x i ) is the friction force.

Images