Online monitoring method for surface microstructure and roughness in superfinishing turning

Abstract

The invention relates to the technical field of ultra-precision cutting machining, in particular to an online monitoring method for surface microtopography and roughness in ultra-precision turning. And data processing and result prediction are performed through a processing surface microstructure geometric simulation model, a deep learning model and a prediction result drawing program of the signal analysis processing module, and a simulation roughness distribution diagram of the processing surface and a three-dimensional microstructure diagram of a specified local area are drawn. According to the method, monitoring of the vibration condition between the cutter and the workpiece in the machining process is achieved only by means of the internal signals generated in the machining process of the machine tool without the help of external sensing devices. Meanwhile, a geometric simulation model and a deep learning model are combined, and the monitoring function of the surface microstructure and roughness in the machining process considering the relative vibration, plastic side flow and elastic recovery effect between the tool and the workpiece is achieved.

Description

Technical Field

[0001] The invention relates to the technical field of ultra-precision cutting processing, and in particular to an online monitoring method for surface micro-morphology generation and surface roughness during ultra-precision single-point turning processing. Background Art

[0002] Online monitoring of the machining process is an important area in industrial production. It involves the use of various sensors, data acquisition systems and data analysis methods to monitor the status of the production process. Ultra-precision single-point diamond turning technology has a wide range of applications in the fields of machining high-precision and high-surface quality optical free-form surface components, and is an important technology in modern manufacturing. Since the cutting volume of ultra-precision single-point diamond turning is very small, it is difficult to directly observe the quality of the machining process. Therefore, it is usually only possible to use special measuring instruments to measure the machining surface of the workpiece after the machining is completed to evaluate the machining quality. This limits the improvement of machining efficiency and cannot reflect the quality of the machining in a timely manner. Therefore, in the process of ultra-precision single-point diamond turning, online monitoring of product surface quality can ensure product quality, improve production efficiency and reduce production costs, which is of great significance. In order to achieve this goal, it is necessary to deeply analyze the various factors affecting the surface quality during the machining process. Summary of the invention

[0003] The purpose of this technical solution is to provide an online monitoring method for surface micromorphology and roughness in ultra-precision single-point turning processing with high accuracy and without the need for external sensing equipment.

[0004] The purpose of this technical solution is achieved in this way:

[0005] An online monitoring method for surface micromorphology and roughness in ultra-precision turning comprises the following steps:

[0006] (a) During the turning process of the machine tool, the turning parameters, the geometric parameters of the tool and the internal signals of the machine tool are obtained; the internal signals collected include the reference position signal of the X-axis , C-axis reference position signal and the Z-axis position error signal ; Reference position signal Used to determine the radial position of the collected cutting point, reference position signal Used to determine the circumferential position of the collected cutting point, position error signal Used to reflect the relative vibration displacement between the tool and the workpiece in the cutting depth direction;

[0007] (b) inputting turning processing parameters, geometric parameters of the tool and internal signals of the machine tool into a micro-morphology geometry simulation program of the machined surface in the host computer to realize micro-morphology geometry simulation of the machined surface;

[0008] (c) predicting the effects of plastic lateral flow and elastic recovery on the micromorphology of the machined surface by inputting the turning processing parameters and the geometric simulation results obtained in step (b) into a pre-trained deep learning model, and obtaining the prediction results;

[0009] (d) Using the prediction results obtained in step (c), a prediction result drawing program in the host computer is used to calculate and draw a simulated roughness distribution map of the machined surface.

[0010] Preferably, in step (a), the reference position signal , reference position signal and position error signal According to the set sampling time and sampling frequency Collect and complete each After sampling the duration, the next The sampling of the time length; the signal data after sampling is sent to the micro-morphology and geometry simulation program of the machined surface.

[0011] Preferably, the machining surface micro-morphology geometry simulation program receives in real time every The internal signal data of the machine tool transmitted at any time dynamically updates the results of the geometric simulation; the steps of the program to realize the dynamic geometric simulation of the micro-morphology of the machining surface include:

[0012] (1) According to the internal signal sampling frequency and the machine tool spindle rotation frequency , calculate the number of sampling points for each circle ; According to the spindle rotation frequency and sampling duration , calculate the previous The number of revolutions of the spindle in a given time ; According to the feed speed and sampling duration , calculate the previous Radial feed distance in time ;

[0013] (2) In the previous During this time, the processing area is an annular area, and the three-dimensional surface of the annular processing area is discretized into Two-dimensional radial sections, each containing In the same radial section, the C-axis reference position of all cutting sampling points is the same, and the X-axis reference position increases in equal steps; when the tool is fed to the workpiece rotation center, the X-axis reference position is , then the radius of the cutting sampling point ,radius It is the difference between the current X-axis reference position of the cutting sampling point and the X-axis reference position when the tool is fed to the rotation center of the workpiece;

[0014] (3) In each radial section, all cutting sampling points are arranged according to the radius Sort them in ascending order to form a set representing the tool tip position in the radial section. ,

[0015] ;

[0016] When the tool is a single-point tool with an arc blade, its blade profile can be approximately expressed as ,in Indicates the radius of the tool tip arc; according to the tool tip position set in the radial section and the blade profile formula, the two-dimensional profile in the radial section can be obtained;

[0017] The two-dimensional contour is discretized into radial spacing The discrete point set ,

[0018] , where the size of the point set is ;

[0019] The height of the contour at the discrete point is ,

[0020] ;

[0021] According to the above steps, the coordinates of all discrete points in the radial section are calculated to obtain a two-dimensional profile in the radial section that only considers kinematic factors;

[0022] (4) According to step (3), calculate all The two-dimensional profile in the radial section is obtained from the previous The micro-morphology and geometric simulation results of the annular area processed in time.

[0023] Preferably, in step (c), the deep learning model used is a deep long short-term memory network, which accepts sequence input and outputs a sequence of equal length; the model is used to transform the radial profile obtained in step (b) that only considers kinematic factors into a radial profile that considers plastic lateral flow and elastic recovery; the input sequence length of the network is equal to the number of discrete points of the radial profile Each step of the input sequence is a contour height value composed of discrete points , cutting depth of turning , Feed speed Cutting line speed corresponding to discrete points A four-dimensional vector composed of ; Each step of the output sequence is a scalar value , that is, the height value of the discrete point contour after considering plastic lateral flow and elastic recovery; the network structure is a sequential structure, which is composed of an input layer, four bidirectional LSTM layers, a Leaky ReLU activation layer, two fully connected layers and an output layer in sequence; among them, the dimensions of the hidden state and cell state of the four bidirectional LSTM layers are set to 150, and a dropout layer with a probability of 0.2 is added between adjacent bidirectional LSTM layers to alleviate the overfitting of the network; the number of neurons in the two fully connected layers is 300 and 30 respectively, and each fully connected layer also uses Leaky ReLU as the activation function; the training data collection and training process steps of the network are as follows:

[0024] (1) For a specific target workpiece material, several groups of experiments with different cutting parameters are conducted, and the radial profile of the machined surface is obtained using a measuring instrument such as a white light interferometer or a profilometer. Multiple radial profiles are extracted from each machined surface along different radial directions; the cutting parameters include cutting depth, feed speed, and spindle speed;

[0025] (2) Extracting a set of tool tip position coordinates from the measured radial profile; since the elastic recovery amount at the tool tip position is approximately the same and the plastic side flow is approximately zero in the same machining, the relative positions of all tool tip position points remain unchanged before and after the plastic side flow and elastic recovery occur; based on the extracted tool tip position coordinate set and the theoretical profile of the cutting edge, reversely restore the ideal profile before the plastic side flow and elastic recovery occur;

[0026] (3) According to the ideal profile and corresponding processing parameters obtained in step (2), an input sequence is prepared according to the above-mentioned input data format requirements, and the corresponding measured profile is used as an output sequence to constitute a training sample of the network; the above operation is performed on all measured radial profiles to obtain a training set of the network;

[0027] (4) Train the network. Before training, standardize the training data to speed up the convergence of the network. Set the batch size to 20. Use the Adam optimizer as the optimizer, and set the learning rate to 0.001 and the weight decay rate to 0.001. Use MSE as the loss function to describe the difference between the network output and the target value. The Pytorch deep learning framework is used to build and train the network, and the number of training rounds is 100.

[0028] Preferably, in step (d), the prediction result drawing program in the host computer will The annular area processed during the time period is divided into several sub-areas at equal intervals in the radial and circumferential directions. Then, according to the three-dimensional micro-morphology prediction results of the processed surface obtained in step (c), the surface roughness values ​​of all sub-areas are calculated respectively. Then, the calculation results are plotted using a pseudo-color map to obtain a simulated roughness distribution map, which is used to intuitively display the predicted roughness distribution law of different local areas on the entire processed surface. The simulated roughness distribution map is plotted every Refresh once a second to display the past The latest processing results within seconds can be obtained, thereby realizing online monitoring of the processing quality during the processing; the prediction result drawing program can draw the local micro-morphology prediction results in the specified area in the form of a three-dimensional surface diagram, and give the roughness index calculation results in the area.

[0029] Preferably, the tool used is a single-point diamond tool.

[0030] Preferably, the kinematic factors include tool tip radius, feed speed and relative vibration.

[0031] Compared with the prior art, the technical solution has the following outstanding and beneficial technical effects:

[0032] The present invention realizes the monitoring of the relative vibration of the tool and the workpiece in the process of ultra-precision single-point turning by online acquisition and analysis of the internal signals generated by the machine tool during the processing, and on this basis, establishes a geometric simulation model of the micro-morphology of the processing surface considering the relative vibration, simplifies the vibration into a superposition method of several constant simple harmonic vibrations of different frequencies, and can more accurately reflect the complex and changeable vibration conditions in the processing process; further, the method predicts the influence of plastic lateral flow and elastic recovery effect on the micro-morphology of the processing surface by building and training a deep learning model, and this method has higher prediction accuracy and generalization performance than the traditional method of predicting the plastic lateral flow and elastic recovery effect using empirical formulas, and the network training process is simple, and the number of experiments required for collecting training data is less than the number of experiments required by the traditional method to calibrate many coefficients, which effectively reduces the implementation cost of the method. The present invention realizes the online monitoring of the surface micro-morphology and roughness in the process of ultra-precision single-point turning by comprehensively considering the tool parameters, processing parameters, the relative vibration between the tool and the workpiece during processing, and the plastic measurement and elastic recovery effect, and the present invention does not need to use additional sensor equipment, and has low implementation cost and high production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 Schematic diagram of cutting sampling points and radial section division.

[0034] Figure 2 Schematic diagram of the generation and discretization of the two-dimensional profile considering only the kinematic factors in the radial section.

[0035] Figure 3 Schematic diagram of the effects of plastic lateral flow and elastic recovery on the micromorphology of the machined surface.

[0036] Figure 4 Schematic diagram of the network structure of the deep learning model.

[0037] Figure 5 Schematic diagram of obtaining network training data.

[0038] Figure 6 A diagram showing the network's prediction performance on the validation set.

[0039] Figure 7 A schematic diagram showing a complete online monitoring process. DETAILED DESCRIPTION

[0040] The following is combined with Figure 1-Figure 7 The specific implementation methods of this technical solution are further described in detail.

[0041] like Figure 1As shown in the figure, taking the end face turning on a three-axis ultra-precision single-point diamond lathe as an example, during the processing, the motion controller sends command signals to the lathe according to the executed NC program to guide the movement of each feed axis of the machine tool. At the same time, the sensor inside the lathe measures the actual position of the feed axis and sends feedback signals to the motion controller.

[0042] The online monitoring method of surface micromorphology and roughness in ultra-precision turning includes the following steps:

[0043] Step (a), during the turning process of the machine tool, the turning processing parameters, the geometric parameters of the tool and the internal signals of the machine tool are obtained; the internal signals during the processing (the reference position signal of the X-axis , C-axis reference position signal and the Z-axis position error signal ) are collected, and each set of sampling values ​​constitutes a cutting sampling point, representing the state of the cutting at that moment. Among them, the reference position signal of the X-axis and C-axis reference position signal They are used to determine the radial and circumferential positions of the collected cutting points and the position error signal of the Z axis. Used to reflect the relative vibration displacement between the tool and the workpiece in the cutting depth direction. The C axis refers to the rotating axis corresponding to the machine tool spindle. It is a common term in the mechanical profession. Those skilled in the art can understand and clarify the relevant definition. Establish communication between the motion controller and the host computer, and transmit the internal signal data of the machine tool collected in the motion controller to the host computer.

[0044] Reference position signal , reference position signal and position error signal According to the set sampling time and sampling frequency Collect and complete each After sampling the duration, the next Duration sampling, completed simultaneously The sampling data of the time length is sent to the micro-morphology geometry simulation program of the processing surface on the host computer through the communication program. Set to 5 seconds. Set to 2000Hz.

[0045] Step (b), in the host computer, the micro-morphology geometry simulation program of the machined surface can be based on the turning processing parameters, the geometric parameters of the tool, and the time interval during the machining process. The internal signal data of the machine tool is transmitted from time to time, and the result of geometric simulation is dynamically updated. The steps of realizing dynamic geometric simulation of micro-morphology of machining surface by the program include: the geometric simulation program of micro-morphology of machining surface is run in MATLAB software.

[0046] (1) According to the internal signal sampling frequency and spindle rotation frequency , calculate the number of sampling points for each circle In this example, the spindle speed is 1000RPM, that is, the spindle rotation frequency ≈ 1000 / 60 Hz, so ; According to the spindle rotation frequency and sampling duration , calculate the previous The number of revolutions of the spindle in a given time , in this case, ; According to the feed speed and sampling duration , calculate the previous Radial feed distance in time In this case, , so It should be noted that at the beginning of the cutting The last one of inner and end cutting The actual feed distance is smaller than the above calculated value, because the cutting is generally done in the first and the last one The actual cutting feed time is insufficient because the cutting feed time starts and ends in the middle. , which needs to be determined based on the time points when cutting starts and ends.

[0047] (2) In the last 5 seconds, the processing area is an annular area. The three-dimensional surface of the annular processing area is discretized into 120 two-dimensional radial sections, each of which contains 83 or 84 cutting sampling points. In the same radial section, the C-axis reference position of all cutting sampling points is the same, the X-axis reference position increases in equal steps (assuming the feed direction is X+), and the Z-axis position error signal is different due to the existence of vibration. When the tool is fed to the workpiece rotation center, the reference position of the X-axis is , in this example , then the radius of the cutting sampling point , that is, the difference between the current X-axis reference position of the cutting sampling point and the X-axis reference position when the tool is fed to the rotation center of the workpiece.

[0048] (3) If Figure 2 As shown, in each radial section, all cutting sampling points are divided into Sort them in ascending order to form a set representing the tool tip position in the radial section. ,

[0049] .

[0050] For a single-point tool with an arc edge, the blade profile can be approximately expressed as ,in Indicates the radius of the tool tip arc. In this example, , so the blade profile formula is According to the tool tip position set in the radial section And the blade profile formula can be used to obtain the two-dimensional profile in the radial section. In order to facilitate storage and calculation, the two-dimensional profile needs to be discretized into radial spacings of The discrete point set ,

[0051] , where the size of the point set is , where [] is the rounding symbol. In this example, the discrete spacing , so the point set size Similarly, in the first 5 seconds of the cutting process and the last 5 seconds of the cutting process, the size of the point set is smaller than the above calculated value. The specific size of the point set is determined by the time points of the cutting process. Located at two adjacent tool tip positions and Between , then the height of the contour at the discrete point is ,

[0052] , that is, take the smaller value of the two adjacent blade profiles. In fact, interference may also occur between non-adjacent blade profiles, especially when the vibration amplitude is large, the tool tip radius is large, and the feed rate per revolution is small. Therefore, when calculating the profile height at a discrete point, it is necessary to consider the coordinate information of multiple adjacent tool tip position points. Taking the two front and back as an example, the profile height calculation formula is corrected to:

[0053] .

[0054] In this case, the number is 5. According to the above steps, the coordinates of all discrete points in the radial section are calculated to obtain a two-dimensional profile in the radial section that only considers the kinematic factors (i.e., tool tip radius, feed speed, and relative vibration).

[0055] (4) According to step (3), the two-dimensional profiles in all 120 radial sections are calculated, thereby obtaining the microscopic geometric simulation results of the annular area processed in the last 5 seconds.

[0056] Step (c) predicts the influence of plastic lateral flow and elastic recovery on the micromorphology of the machined surface by inputting the turning processing parameters and the geometric simulation results obtained in step (b) into a pre-trained deep learning model, and obtains the prediction results; the prediction results are in the form of data as a series of coordinate values ​​of discrete points. These discrete points together form a three-dimensional surface, which is the three-dimensional morphology prediction result of the machined surface. In fact, the output result of the "geometric simulation program" mentioned in step (b) above is also a series of discrete point coordinate values. The difference is that the result obtained in step (b) only considers geometric factors, while step (c) considers the influence of plastic lateral flow and elastic recovery on this basis, and this function is realized by the deep learning model in step (c). Figure 3 As shown in Figure 2, in addition to the contribution of the above-mentioned kinematic factors to the machined surface morphology, Figure 3 The plastic lateral flow and Figure 3 The elastic recovery shown in (b) will also change the machined surface morphology. Plastic lateral flow will increase the residual height of the surface, while elastic recovery will reduce the residual height of the surface. In this method, a deep learning model is used to predict the degree of change of the surface caused by plastic lateral flow and elastic recovery under specific machining conditions. The network accepts sequence input and outputs a sequence of equal length. Each step of the input sequence is the profile height value of a discrete point of the radial profile when only kinematic factors are considered. , cutting depth of turning , Feed speed Cutting line speed corresponding to discrete points A four-dimensional vector composed of Each step of the output sequence is a scalar value , that is, the height value of the discrete point contour after considering plastic flow measurement and elastic recovery. The constructed deep learning model is a deep long short-term memory (LSTM) network, and the network structure is a sequential structure, such as Figure 4 As shown in the figure, it consists of an input layer, four bidirectional LSTM layers, a Leaky ReLU activation layer, two fully connected layers, and an output layer. The dimensions of the hidden state and cell state of the four bidirectional LSTM layers are all set to 150, and a dropout layer with a probability of 0.2 is added between adjacent bidirectional LSTM layers to alleviate the overfitting of the network; the number of neurons in the two fully connected layers is 300 and 30 respectively, and each fully connected layer also uses Leaky ReLU as the activation function. The steps of network training data collection and training process are as follows:

[0057] (1) For a specific target workpiece material, several groups of experiments with different cutting parameters (cutting depth, feed speed, spindle speed) are conducted, and the radial profile of the machined surface is obtained using measuring instruments such as white light interferometers or profilometers. Multiple radial profiles are extracted from each machined surface along different radial directions. In this example, the workpiece is a square copper sheet with a size of 10 mm × 10 mm, and the processing area is a circle with a diameter of 9 mm with the center of the copper sheet as the center. The 15 groups of cutting parameter combinations listed in Table 1 below are used as the training set data source. After each cutting, the three-dimensional morphology of the machined surface is measured with a white light interferometer, and the radial profile is extracted from it. The measurement results of the three-dimensional morphology of the machined surface are shown in the figure below. Figure 5 shown.

[0058] Table 1. Training set data table

[0059] Serial number Spindle speed (RPM) Feed speed (μm / min) Cutting depth(μm) 1 1000 4 2 2 1000 4 4 3 1000 4 6 4 1000 6 2 5 1000 6 4 6 1000 6 6 7 1000 8 2 8 1000 8 4 9 1000 8 6 10 1000 10 2 11 1000 10 4 12 1000 10 6 13 1000 12 2 14 1000 12 4 15 1000 12 6

[0060] (2) Extract the set of tool tip position coordinates from the measured radial profile (generally, all the minimum points of the measured profile). Since the elastic recovery amount at the tool tip position is approximately the same and the plastic flow measurement amount is approximately zero in the same machining, the relative positions of all tool tip position points remain unchanged before and after the plastic flow measurement and elastic recovery. Based on this, the ideal profile before the plastic flow measurement and elastic recovery can be reversely restored based on the extracted tool tip position coordinate set and the theoretical profile of the blade, that is, the radial profile result of the machined surface when only kinematic factors such as tool parameters, machining parameters and relative vibration are considered.

[0061] (3) According to the ideal profile and corresponding processing parameters obtained in step (2), an input sequence is prepared according to the above input data format requirements, and the corresponding measured profile is used as an output sequence (label value) to form a training sample of the network. The above operation is performed on all measured radial profiles to obtain the training set of the network.

[0062] (4) Train the network. Before training, standardize the training data to speed up the convergence of the network. Set the batch size (the number of training samples used in a single training) to 20. Use the Adam optimizer as the optimizer, with the learning rate set to 0.001 and the weight decay rate set to 0.001. Use MSE as the loss function to describe the difference between the network output and the target value. The Pytorch deep learning framework is used to build and train the network, and the number of training rounds is 100.

[0063] Figure 6 The performance of the trained model on the validation set is shown. Figure 6 (a) is a sample from Experiment 2. Figure 6(b) is a sample of Experiment 7, which shows the radial profile obtained by actual measurement, the ideal radial profile reversely restored according to the above method, and the profile predicted by the deep learning model based on the input of the ideal profile. From the results shown in the figure, it can be seen that the trained deep learning model can well predict the effects of plastic lateral flow and elastic recovery based on the input ideal profile information, thus approaching the actual measured profile morphology.

[0064] In this embodiment, the machining surface micro-morphology geometry simulation program obtains the prediction results when only geometric factors are considered; based on the above results, the deep learning model considers the influence of plastic deformation and elastic recovery to obtain the final prediction results; the original data of the prediction results obtained in the previous step is a series of data points, and then the prediction result drawing program is used to calculate, analyze and visualize these data.

[0065] In step (d), the prediction result obtained in step (c) is used to calculate and draw a simulated roughness distribution map of the machined surface through a prediction result drawing program in the host computer; the prediction result drawing program in the host computer converts the previous The annular area processed within a time period (5 seconds in this embodiment) is divided into several sub-areas at equal intervals in the radial and circumferential directions, and then the three-dimensional micro-morphology prediction results of the processed surface are obtained according to the above steps, and the surface roughness values ​​of all sub-areas are calculated respectively (such as arithmetic roughness value Sa, root mean square roughness value Sq, etc.), and then the calculation results are plotted using a pseudo-color map to obtain a simulated roughness distribution map, which is used to intuitively display the predicted roughness distribution law of different local areas on the entire processed surface. The simulated roughness distribution map is refreshed every 5 seconds to display the latest processing results in the past 5 seconds, thereby realizing online monitoring of the processing quality during the processing process. In addition, the prediction result drawing program can draw the local micro-morphology prediction results in the specified area in the form of a three-dimensional surface map, and give the roughness index calculation results in the area, which is convenient for operators to observe the specific morphology of a local area of ​​interest. Figure 7The application results of this method in the one-end turning experiment are given. The cutting parameters of the experiment are set as: spindle speed 1000RPM, feed speed 5mm / min, cutting depth 5μm. The left column shows the refreshing process of the simulated roughness distribution map (taking the surface roughness index Sa as an example) as the machining process proceeds. The right side lists the comparison results of the simulated predicted three-dimensional morphology and the measured three-dimensional morphology of three local areas and the corresponding local area roughness calculation values. It can be seen that the simulated predicted morphology and the measured morphology have a good correspondence, and the local prediction results of the roughness also have a high accuracy. It can be seen from the simulated roughness distribution map that although the cutting parameters remain unchanged during the cutting process, the change in the relative vibration between the tool and the workpiece caused by the change in the cutting environment significantly affects the roughness value of different machining areas, which shows the advantage of the online monitoring of vibration state adopted by this method: compared with the traditional method of simplifying the vibration into the superposition of several constant simple harmonic motions, it can more accurately reflect the changes in vibration during the machining process. In addition, this method also provides the possibility of monitoring abnormal conditions during the machining process. This is because machining anomalies are usually accompanied by abnormal vibrations, which lead to a significant deterioration in the roughness of the machined surface. Therefore, the occurrence of machining anomalies can be effectively monitored by observing the simulated roughness distribution map. In addition, the prediction results of micromorphology and roughness also prove that the deep learning model used in this method can effectively predict the effects of plastic lateral flow and elastic recovery on surface micromorphology and surface roughness.

[0066] The present invention realizes the monitoring of the tool-workpiece relative vibration in the ultra-precision single-point turning process by online acquisition and analysis of the internal signals generated by the machine tool during the machining process, and on this basis establishes a geometric simulation model of the micro-morphology of the machining surface taking into account the relative vibration. Compared with the traditional method of measuring the relative vibration between the tool and the workpiece before actual machining and simplifying the vibration into the superposition of several constant simple harmonic vibrations of different frequencies, this method can more accurately reflect the complex and changeable vibration conditions in the machining process; further, the method predicts the influence of plastic lateral flow and elastic recovery effect on the micro-morphology of the machining surface by building and training a deep learning model. Compared with the traditional method of using empirical formulas to predict plastic lateral flow and elastic recovery effect, this method has higher prediction accuracy and generalization performance, and the network training process is simple, and the number of experiments required for collecting training data is less than the number of experiments required for the traditional method to calibrate many coefficients, which effectively reduces the implementation cost of the method. The present invention realizes online monitoring of surface micromorphology and roughness during ultra-precision single-point turning by comprehensively considering tool parameters, processing parameters, relative vibration between the tool and the workpiece during processing, as well as plasticity measurement and elastic recovery effect. The present invention does not require the use of additional sensor equipment and has low implementation cost and high production efficiency.

[0067] The above shows and describes the basic principle and main features of the technical solution and the advantages of the technical solution. The technicians in this industry should understand that the technical solution is not limited by the above embodiments, and the above embodiments and descriptions are only for explaining the principle of the technical solution. Without departing from the spirit and scope of the technical solution, the technical solution will have various changes and improvements, and these changes and improvements fall within the scope of the technical solution to be protected.

Claims

1. An online monitoring method for surface micromorphology and roughness in ultra-finishing turning, characterized in that: The following steps are involved: (a) During the turning process of the machine tool, the turning parameters, the geometric parameters of the tool and the internal signals of the machine tool are obtained; the internal signals collected include the reference position signal of the X-axis , C-axis reference position signal and the Z-axis position error signal ; Reference position signal Used to determine the radial position of the collected cutting point, reference position signal Used to determine the circumferential position of the collected cutting point, position error signal Used to reflect the relative vibration displacement between the tool and the workpiece in the cutting depth direction; (b) inputting turning processing parameters, geometric parameters of the tool and internal signals of the machine tool into a micro-morphology geometry simulation program of the machined surface in the host computer to realize micro-morphology geometry simulation of the machined surface; (c) predicting the effects of plastic lateral flow and elastic recovery on the micromorphology of the machined surface by inputting the turning processing parameters and the geometric simulation results obtained in step (b) into a pre-trained deep learning model, and obtaining the prediction results; (d) Using the prediction results obtained in step (c), a prediction result drawing program in the host computer is used to calculate and draw a simulated roughness distribution map of the machined surface.

2. The method for online monitoring of surface micromorphology and roughness in superfinishing turning according to claim 1, characterized in that: In step (a), the reference position signal , reference position signal and position error signal According to the set sampling time and sampling frequency Collect and complete each After sampling the duration, the next Duration sampling; The sampled signal data is sent to the machining surface micro-morphology geometry simulation program.

3. The method for online monitoring of surface micromorphology and roughness in superfinishing turning according to claim 2, characterized in that: The machining surface micro-morphology geometry simulation program receives the The internal signal data of the machine tool transmitted at any time dynamically updates the results of the geometric simulation; the steps of the program to realize the dynamic geometric simulation of the micro-morphology of the machining surface include: (1) According to the internal signal sampling frequency and the machine tool spindle rotation frequency , calculate the number of sampling points for each circle ; According to the spindle rotation frequency and sampling duration , calculate the previous The number of revolutions of the spindle in a given time ; According to the feed speed and sampling duration , calculate the previous Radial feed distance in time ; (2) In the previous During this time, the processing area is an annular area, and the three-dimensional surface of the annular area is discretized into Two-dimensional radial sections, each containing cutting sampling points; in the same radial section, the C-axis reference position of all cutting sampling points is the same, and the X-axis reference position increases in equal steps; when the tool is fed to the workpiece rotation center, the X-axis reference position is , then the radius of the cutting sampling point ,radius It is the difference between the current X-axis reference position of the cutting sampling point and the X-axis reference position when the tool is fed to the rotation center of the workpiece; (3) In each radial section, all cutting sampling points are arranged according to the radius Sort them in ascending order to form a set T representing the tool tip position in the radial section. ； When the tool is a single-point tool with an arc blade, its blade profile can be approximately expressed as ,in Indicates the radius of the tool tip arc; according to the tool tip position set in the radial section and the blade profile formula, the two-dimensional profile in the radial section can be obtained; The two-dimensional contour is discretized into radial spacing The discrete point set , , where the size of the point set is ; The height of the contour at the discrete point is , ； According to the above steps, the coordinates of all discrete points in the radial section are calculated to obtain a two-dimensional profile in the radial section that only considers kinematic factors; (4) According to step (3), calculate all The two-dimensional profile in the radial section is obtained from the previous The micro-morphology and geometric simulation results of the annular area processed in time.

4. The method for online monitoring of surface micromorphology and roughness in superfinishing turning according to claim 1, characterized in that: In the step (c), the deep learning model used is a deep long short-term memory network, which accepts sequence input and outputs a sequence of equal length; the model is used to transform the radial profile obtained in step (b) that only considers kinematic factors into a radial profile that considers plastic lateral flow and elastic recovery; the input sequence length of the network is equal to the number of discrete points of the radial profile Each step of the input sequence is a contour height value composed of discrete points , cutting depth of turning , Feed speed Cutting line speed corresponding to discrete points A four-dimensional vector composed of ; Each step of the output sequence is a scalar value , that is, the height value of the discrete point contour after considering plastic lateral flow and elastic recovery; the network structure is a sequential structure, which is composed of an input layer, four bidirectional LSTM layers, a Leaky ReLU activation layer, two fully connected layers and an output layer in sequence; among them, the dimensions of the hidden state and cell state of the four bidirectional LSTM layers are set to 150, and a dropout layer with a probability of 0.2 is added between adjacent bidirectional LSTM layers to alleviate the overfitting of the network; the number of neurons in the two fully connected layers is 300 and 30 respectively, and each fully connected layer also uses Leaky ReLU as the activation function; the training data collection and training process steps of the network are as follows: (1) For a specific target workpiece material, several groups of experiments with different cutting parameters are conducted, and the radial profile of the machined surface is obtained using a measuring instrument such as a white light interferometer or a profilometer. Multiple radial profiles are extracted from each machined surface along different radial directions; the cutting parameters include cutting depth, feed speed, and spindle speed; (2) Extracting a set of tool tip position coordinates from the measured radial profile; since the elastic recovery amount at the tool tip position is approximately the same and the plastic side flow is approximately zero in the same machining, the relative positions of all tool tip position points remain unchanged before and after the plastic side flow and elastic recovery occur; based on the extracted tool tip position coordinate set and the theoretical profile of the cutting edge, reversely restore the ideal profile before the plastic side flow and elastic recovery occur; (3) According to the ideal profile and corresponding processing parameters obtained in step (2), an input sequence is prepared according to the above-mentioned input data format requirements, and the corresponding measured profile is used as an output sequence to constitute a training sample of the network; the above operation is performed on all measured radial profiles to obtain a training set of the network; (4) The network was trained. Before training, the training data was standardized to speed up the convergence of the network. The batch size was set to 20. The Adam optimizer was used as the optimizer, and its parameters were set to a learning rate of 0.001 and a weight decay rate of 0.

001. MSE was used as the loss function to describe the difference between the network output and the target value. The Pytorch deep learning framework was used to build and train the network, and the number of training rounds was 100.

5. The method for online monitoring of surface micromorphology and roughness in superfinishing turning according to claim 1, characterized in that: In step (d), the prediction result drawing program in the host computer will The annular area processed during the time period is divided into several sub-areas at equal intervals in the radial and circumferential directions. Then, according to the three-dimensional micro-morphology prediction results of the processed surface obtained in step (c), the surface roughness values ​​of all sub-areas are calculated respectively. Then, the calculation results are plotted using a pseudo-color map to obtain a simulated roughness distribution map, which is used to intuitively display the predicted roughness distribution law of different local areas on the entire processed surface. The simulated roughness distribution map is plotted every Refresh once a second to display the past The latest processing results within seconds can be obtained, thereby realizing online monitoring of the processing quality during the processing; the prediction result drawing program can draw the local micro-morphology prediction results in the specified area in the form of a three-dimensional surface diagram, and give the roughness index calculation results in the area.

6. The method for online monitoring of surface micromorphology and roughness in superfinishing turning according to claim 1, characterized in that: The tool used is a single-point diamond tool.

7. The method for online monitoring of surface micromorphology and roughness in superfinishing turning according to claim 3, characterized in that: The kinematic factors include tool nose radius, feed rate and relative vibration.

Images