Compensation method and device for bottom dead center position of punch press sliding block

Abstract

The invention provides a compensation method and device for the bottom dead center position of a punch press sliding block, and relates to the technical field of punching equipment.The method comprises the steps that first temperature data of a punch press workbench at the current moment, second temperature data of the sliding block, abrasion loss data of a sensor sensing head and first position data of the sliding block are obtained, the first position data is bottom dead center position data of the sliding block at the current moment; the first temperature data of the punching machine workbench, the second temperature data of the sliding block and the abrasion loss data of the sensing head of the sensor are processed to determine target feature data; inputting the target feature data into a trained target network model, and determining a prediction compensation value through the processing of the target network model; and controlling the sliding block according to the predicted compensation value and the first position data so as to realize the compensation of the bottom dead center position of the sliding block, thereby effectively improving the accuracy and reliability of the bottom dead center position compensation.

Description

Technical Field

[0001] The present application relates to the technical field of stamping equipment, and in particular to a method and device for compensating the bottom dead point position of a punch press slider. Background Art

[0002] Stamping equipment is essential manufacturing equipment for key industries such as automotive, military, aerospace, and rail transit. Its technological level has become a key indicator of a country's overall manufacturing strength. Slide bottom dead center accuracy, a key performance indicator for high-speed stamping equipment, directly impacts the precision of stamped parts, die life, and the yield rate of stamped parts.

[0003] In related art, when the bottom dead center position of a slider exceeds the allowable error, the oil chamber oil temperature is typically adjusted to bring the adjusted bottom dead center position within the allowable error range. This means that dynamic compensation of the bottom dead center position is only based on the oil chamber oil temperature, which can result in low accuracy in the determined dynamic compensation of the bottom dead center position. Therefore, improving the accuracy of dynamic compensation of the slider's bottom dead center position is crucial. Summary of the Invention

[0004] The present application provides a method and device for compensating the bottom dead point position of a punch press slider.

[0005] According to a first aspect of the present application, a method for compensating the bottom dead center position of a punch slide is provided, the method comprising:

[0006] Acquire first temperature data of the punching machine worktable, second temperature data of the slider, wear data of the sensor head, and first position data of the slider at the current moment, wherein the first position data is bottom dead center position data of the slider at the current moment, and the sensor is disposed on a guide plate of the punching machine worktable;

[0007] Processing the first temperature data of the punching machine worktable, the second temperature data of the slider, and the wear data of the sensor head to determine target characteristic data;

[0008] Inputting the target feature data into the trained target network model to determine a predicted compensation value after being processed by the target network model;

[0009] The slider is controlled according to the predicted compensation value and the first position data to achieve compensation for the bottom dead center position of the slider.

[0010] Optionally, the processing of the first temperature data of the punching table, the second temperature data of the slider, and the wear data of the sensor head to determine the target characteristic data includes:

[0011] Performing feature extraction on the first temperature data to obtain first feature data of the punching machine worktable;

[0012] Performing feature extraction on the second temperature data of the slider to obtain second feature data of the slider;

[0013] performing feature extraction on the wear data of the sensor head to obtain third feature data of the sensor head;

[0014] The first feature data, the second feature data and the third feature data are fused to determine target feature data.

[0015] Optionally, controlling the slider according to the predicted compensation value and the first position data to achieve compensation for the bottom dead point position of the slider includes:

[0016] determining a temperature difference correction value according to the first temperature data and the second temperature data;

[0017] Determining a target compensation value based on the predicted compensation value and the temperature difference correction value;

[0018] The slider is controlled according to the target compensation value and the first position data to achieve compensation for the bottom dead point position of the slider.

[0019] Optionally, determining the temperature difference correction value according to the first temperature data and the second temperature data includes:

[0020] Determining mean temperature data of the punching worktable according to the first temperature data of the punching worktable and the corresponding weight value;

[0021] A temperature difference correction value is determined according to the second temperature data, the mean temperature data, and the temperature difference deformation coefficient.

[0022] Optionally, controlling the slider according to the target compensation value and the first position data to achieve compensation for the bottom dead point position of the slider includes:

[0023] determining a control variable of a controller according to the target compensation value;

[0024] determining the output thrust of the voice coil motor according to the control amount;

[0025] The motion state of the slider is controlled according to the output thrust of the voice coil motor and the first position data, so as to compensate for the bottom dead point position of the slider.

[0026] Optionally, after controlling the motion state of the slider according to the output thrust of the voice coil motor and the first position data, the method further includes:

[0027] Acquiring second position data of the slider, wherein the second position data is bottom dead center position data of the slider after being controlled by the output thrust of the voice coil motor;

[0028] determining a difference between the second position data and the first position data as an actual displacement difference;

[0029] The control parameters of the controller are adjusted according to the absolute value of the error between the actual displacement difference and the target compensation value.

[0030] Optionally, before inputting the target feature data into the trained target network model to determine the predicted compensation value through processing by the target network model, the method further includes:

[0031] Acquire a training data set, wherein the training data set includes third temperature data of the punching machine worktable, fourth temperature data of the slider, historical wear data of the sensor head, historical bottom dead center position data of the slider, and label data;

[0032] Processing the third temperature data of the punching machine worktable, the fourth temperature data of the slider, and the historical wear data of the sensor head to determine training feature data;

[0033] Inputting the training feature data into an initial network model to determine a training compensation value after being processed by the initial network model;

[0034] Determining a loss value based on a difference between the training compensation value and the label data;

[0035] The initial network model is trained according to the loss value to generate a trained target network model.

[0036] According to a second aspect of the present application, a compensation device for the bottom dead center position of a punch slide is provided, comprising:

[0037] a first acquisition module, configured to acquire first temperature data of the punching machine worktable, second temperature data of the slider, wear data of the sensor head, and first position data of the slider at a current moment, wherein the first position data is bottom dead center position data of the slider at a current moment, and the sensor is disposed on a guide plate of the punching machine worktable;

[0038] a first processing module, configured to process the first temperature data of the punching machine worktable, the second temperature data of the slider, and the wear data of the sensor head to determine target characteristic data;

[0039] a first determination module, configured to input the target feature data into a trained target network model, so as to determine a predicted compensation value after processing by the target network model;

[0040] A control module is used to control the slider according to the predicted compensation value and the first position data to achieve compensation for the bottom dead point position of the slider.

[0041] Optionally, the first processing module is specifically configured to:

[0042] Performing feature extraction on the first temperature data to obtain first feature data of the punching machine worktable;

[0043] Performing feature extraction on the second temperature data of the slider to obtain second feature data of the slider;

[0044] performing feature extraction on the wear data of the sensor head to obtain third feature data of the sensor head;

[0045] The first feature data, the second feature data and the third feature data are fused to determine target feature data.

[0046] Optionally, the control module includes:

[0047] a first determining unit, configured to determine a temperature difference correction value according to the first temperature data and the second temperature data;

[0048] a second determining unit, configured to determine a target compensation value based on the predicted compensation value and the temperature difference correction value;

[0049] A control unit is used to control the slider according to the target compensation value and the first position data to achieve compensation for the bottom dead point position of the slider.

[0050] Optionally, the first determining unit is specifically configured to:

[0051] Determining mean temperature data of the punching worktable according to the first temperature data of the punching worktable and the corresponding weight value;

[0052] A temperature difference correction value is determined according to the second temperature data, the mean temperature data, and the temperature difference deformation coefficient.

[0053] Optionally, the control unit is specifically configured to:

[0054] determining a control variable of a controller according to the target compensation value;

[0055] determining the output thrust of the voice coil motor according to the control amount;

[0056] The motion state of the slider is controlled according to the output thrust of the voice coil motor and the first position data, so as to compensate for the bottom dead point position of the slider.

[0057] Optionally, the control module is further configured to:

[0058] Acquiring second position data of the slider, wherein the second position data is bottom dead center position data of the slider after being controlled by the output thrust of the voice coil motor;

[0059] determining a difference between the second position data and the first position data as an actual displacement difference;

[0060] The control parameters of the controller are adjusted according to the absolute value of the error between the actual displacement difference and the target compensation value.

[0061] Optionally, the device further includes:

[0062] a second acquisition module, configured to acquire a training data set, wherein the training data set includes third temperature data of the punching machine worktable, fourth temperature data of the slider, historical wear data of the sensor head, historical bottom dead center position data of the slider, and label data;

[0063] a second processing module, configured to process the third temperature data of the punching machine worktable, the fourth temperature data of the slider, and the historical wear data of the sensor head to determine training feature data;

[0064] a second determining module, configured to input the training feature data into an initial network model to determine a training compensation value after processing by the initial network model;

[0065] A third determining module is used to determine a loss value according to a difference between the training compensation value and the label data;

[0066] A generation module is used to train the initial network model according to the loss value to generate a trained target network model.

[0067] According to a third aspect of the present application, an electronic device is provided, comprising: a processor and a memory storing computer program instructions; when the processor executes the computer program instructions, any of the above-mentioned methods for compensating the bottom dead point position of the punch slide is implemented.

[0068] According to a fourth aspect of the present application, a computer-readable storage medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, any of the above-mentioned methods for compensating the bottom dead point position of the punch slide is implemented.

[0069] In summary, the compensation method and device for the bottom dead point position of the punch press slider provided in the present application have at least the following beneficial effects: the first temperature data of the punch press worktable, the second temperature data of the slider, the wear data of the sensor sensing head and the first position data of the slider at the current moment can be obtained first, wherein the first position data is the bottom dead point position data of the slider at the current moment, and the sensor is arranged at the guide plate of the punch press worktable. Then, the first temperature data of the punch press worktable, the second temperature data of the slider, and the wear data of the sensor sensing head can be processed to determine the target feature data, and then the target feature data can be input into the trained target network model to determine the predicted compensation value after processing by the target network model. Then, the slider is controlled according to the predicted compensation value and the first position data to achieve compensation for the bottom dead point position of the slider. Therefore, by processing the first temperature data of the punching machine worktable, the second temperature data of the slider, and the wear data of the sensor sensing head, the target characteristic data is determined, and then the predicted compensation value is obtained in combination with the target network model. Finally, the slider is controlled based on the predicted compensation value. That is, in the process of compensating the bottom dead point position of the slider, the temperature influence of the punching machine worktable and the slider and the wear influence of the sensor sensing head are fully considered, thereby effectively improving the accuracy and reliability of the bottom dead point position compensation and ensuring the smooth operation of the punching machine. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] In order to more clearly illustrate the specific implementation methods of the present application or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the specific implementation methods or the description of the prior art. Obviously, the drawings described below are some implementation methods of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without any creative work.

[0071] Figure 1 A flowchart of a method for compensating the bottom dead center position of a punch press slider provided in an embodiment of the present application;

[0072] Figure 2 A schematic diagram of the location of a sensor provided in an embodiment of the present application;

[0073] Figure 3 A flowchart of another method for compensating the bottom dead center position of a punch slide provided in an embodiment of the present application;

[0074] Figure 4A flowchart of another method for compensating the bottom dead center position of a punch slide provided in an embodiment of the present application;

[0075] Figure 5 A compensation device for the bottom dead point position of a punch press slider provided in an embodiment of the present application;

[0076] Figure 6 A structural diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0077] In order to make the above and other features and advantages of the present application more clear, the present application is further described below with reference to the accompanying drawings. It should be understood that the specific embodiments given herein are for the purpose of explaining to those skilled in the art and are only exemplary and not restrictive.

[0078] In the following description, many specific details are set forth to provide a thorough understanding of the present application. However, it will be apparent to those skilled in the art that it is not necessary to adopt the specific details to practice the present application. In other cases, well-known steps or operations are not described in detail to avoid obscuring the present application.

[0079] The method for compensating the bottom dead point position of the punch press slider provided in the embodiment of the present application can be performed by the compensation device for the bottom dead point position of the punch press slider provided in the embodiment of the present application, and the device can be configured in an electronic device.

[0080] refer to Figure 1 The present application provides a method for compensating the bottom dead point position of a punch press slider, the method comprising:

[0081] Step 101 : obtaining first temperature data of a punching machine worktable, second temperature data of a slider, wear data of a sensor head, and first position data of the slider at a current moment.

[0082] The first position data is the bottom dead point position data of the slider at the current moment.

[0083] The sensor can be set at the guide plate of the punching machine workbench, such as Figure 2 As shown, 21 is a sensor. The sensor can be a high-precision contact grating scale CMOS sensor, or other sensors that can achieve the same function, which is not limited in this application.

[0084] Optionally, four sets of contact scale CMOS sensors can be symmetrically mounted below the four slider guide plates of the punch press, with the sensor heads facing the slider motion surface. Specifically, the four sets of contact scale CMOS sensors can be mounted directly below the four guide plates through rigid alloy bases, ensuring that the perpendicularity error between the sensor head centerline and the slider motion trajectory is within an allowable range, such as within a range of ≤0.005mm / m, to ensure the accuracy and reliability of the entire operation process. This application does not impose any restrictions on this.

[0085] In addition, the first temperature data of the punching worktable can be understood as the temperature data of certain specific locations on the punching worktable, etc. The temperature data can be collected by setting multiple sensors, or by using an infrared thermal imager, etc. The first temperature data fully reflects the temperature condition of the punching worktable in the spatial dimension. This application does not limit this.

[0086] In addition, the temperature of the slider can be monitored by embedding a temperature sensor in the slider. For example, a PT100 temperature sensor can be embedded in the slider to collect the second temperature data of the slider, etc. This application does not limit this.

[0087] Optionally, a PT100 thin film sensor may be used to collect the temperature of the slider at a frequency of 100 Hz to form 10-point sliding window time series data, namely, second temperature data, which fully reflects the temperature condition of the slider in the time dimension.

[0088] In addition, the wear data of the sensor head can be determined based on the cumulative number of touches, real-time contact force and wear coefficient of the sensor head, and this application does not limit this.

[0089] In addition, to minimize deviations caused by wear of the sensor head, the sensor head body can be made of composite ceramics, such as ZrO2-Al2O3, or other wear-resistant materials. The front end can have a hemispherical contact surface, or other shapes, which are not limited in this application.

[0090] Step 102 : Processing the first temperature data of the punching machine worktable, the second temperature data of the slider, and the wear data of the sensor head to determine target characteristic data.

[0091] Among them, any desirable method can be used to perform feature extraction processing on the first temperature data of the punching worktable, the second temperature data of the slider, and the wear data of the sensor head. For example, it can be implemented through a convolutional neural network, or through a feature extractor, etc. This application does not limit this.

[0092] It can be understood that in the process of determining the target characteristic data, the first temperature data of the punching machine worktable in the spatial dimension and the time dimension, the second temperature data of the slider, and the wear data of the sensor head are fully considered, so that the target characteristic data can be made more comprehensive and reliable, thereby providing a basis for using the target characteristic data for subsequent operations.

[0093] Optionally, an infrared thermopile array can be used to monitor the punch press worktable to collect first temperature data of the punch press worktable. The first temperature data can then be feature extracted to obtain first characteristic data of the punch press worktable. The second temperature data of the slider can be feature extracted to obtain second characteristic data of the slider. The wear data of the sensor head can be determined based on the cumulative number of touches of the sensor head, the real-time contact force, the single touch wear coefficient, and the contact force wear coefficient. The wear data of the sensor head can then be feature extracted to obtain third characteristic data of the sensor head. The first characteristic data, the second characteristic data, and the third characteristic data can then be fused to determine the target characteristic data.

[0094] Particularly, since the infrared thermopile array is used to monitor the punch press worktable, the collected first temperature data can have certain spatial characteristics.

[0095] Optionally, when the infrared thermopile array is 8×8 in size, the first temperature data is also in the form of an 8×8 temperature field matrix, which can then be processed by bicubic interpolation to generate a 64×64 standardized temperature grid, which can then be further processed by a graph convolutional network (GCN) to establish a heat conduction relationship between pixels.

[0096] For example, the first feature data H obtained after the above processing spatial It can be expressed as:

[0097] H spatial =ReLU(A·T t W g )

[0098] Among them, the adjacency matrix A reflects the spatial correlation of the temperature field, W g is the trainable weight matrix of the graph convolutional network, T t is the first temperature data.

[0099] In addition, a temporal convolutional network (TCN) can be used to extract the temporal features of the slider temperature, so that the obtained second feature data can be expressed as:

[0100] ht =TCN(T s (t-τ:t),W t )

[0101] Among them, τ is the time window length, t is the current time, T s is the second temperature data, W t is the trainable weight matrix of the temporal convolutional network.

[0102] For the convenience of experiment, τ can be calibrated to 10, or it can be calibrated to other values, etc., which is not limited in this application.

[0103] Optionally, wear data of the sensor head may be calculated based on the cumulative number of touches, the real-time contact force, and the wear coefficient.

[0104] For example, the wear of the sensor head can be expressed as:

[0105]

[0106] Among them, K1 is the single touch wear coefficient, N(t) is the cumulative number of touches, K2 is the contact force wear coefficient, and F(t) is the real-time contact force.

[0107] Among them, K2 can be calibrated through the load-wear rate experimental curve of the sensor head material, or can be calibrated through other methods, which is not limited in this application.

[0108] Furthermore, after calculating the wear data of the sensor head, feature extraction and processing can be performed to obtain corresponding third feature data. It is understood that after obtaining the first, second, and third feature data, they can be fused, and the fusion method can be various. For example, the first, second, and third feature data can be directly concatenated to obtain the target feature data; or the first, second, and third feature data can be weightedly fused according to their corresponding weight values ​​to obtain the target feature data, etc. This application does not limit this.

[0109] It can be understood that since the first characteristic data can reflect the spatial characteristics of the second temperature data of the punching worktable, the second characteristic data can reflect the temporal characteristics of the first temperature data of the slider, and the third characteristic data can reflect the wear characteristics of the sensor sensing head, the target characteristic data obtained by fusing the three takes into account the spatial characteristics, temporal characteristics of the temperature data and the wear characteristics of the sensor sensing head, that is, the target characteristic data fully takes into account the influence of temperature and wear, thereby providing a basis for subsequent guarantee of the accuracy of the dynamic compensation of the bottom dead point position.

[0110] Step 103: input the target feature data into the trained target network model to determine the predicted compensation value after being processed by the target network model.

[0111] The target network module can be any trained prediction regression network model, or a pre-trained model that has been fine-tuned, so that the target network model can be directly used to process the target feature data, and after processing by the target network model, a predicted compensation value can be obtained. This application does not limit this.

[0112] Step 104 : Control the slider according to the predicted compensation value and the first position data to compensate for the bottom dead center position of the slider.

[0113] After the predicted compensation value is determined, the slider may be controlled to move according to the predicted compensation value based on the first position data of the slider at present, thereby achieving compensation for the bottom dead point position of the slider.

[0114] It can be understood that in the embodiment of the present application, since the target characteristic data fully reflects the spatial characteristics of the temperature of the punching worktable, the temporal characteristics of the slider temperature and the wear characteristics of the sensor head, the characteristic information contained in the target characteristic data is relatively rich and comprehensive, so the predicted compensation value determined using the target characteristic data is also more accurate and reliable. In the process of compensating the bottom dead point position of the slider according to the predicted compensation value, the accuracy and reliability of the compensation are effectively improved, thereby ensuring the smooth operation of the punching machine.

[0115] In an embodiment of the present application, the first temperature data of the punching worktable, the second temperature data of the slider, the wear data of the sensor sensing head and the first position data of the slider at the current moment can be first obtained, wherein the first position data is the bottom dead point position data of the slider at the current moment, and the sensor is set at the guide plate of the punching worktable. Then, the first temperature data of the punching worktable, the second temperature data of the slider, and the wear data of the sensor sensing head can be processed to determine the target feature data, and then the target feature data can be input into the trained target network model to determine the predicted compensation value after processing by the target network model. Then, the slider is controlled according to the predicted compensation value and the first position data to achieve compensation for the bottom dead point position of the slider. Therefore, by processing the first temperature data of the punching machine worktable, the second temperature data of the slider, and the wear data of the sensor sensing head, the target characteristic data is determined, and then the predicted compensation value is obtained in combination with the target network model. Finally, the slider is controlled based on the predicted compensation value. That is, in the process of compensating the bottom dead point position of the slider, the temperature influence of the punching machine worktable and the slider and the wear influence of the sensor sensing head are fully considered, thereby effectively improving the accuracy and reliability of the bottom dead point position compensation and ensuring the smooth operation of the punching machine.

[0116] like Figure 3 As shown, the method for compensating the bottom dead point position of the punch slide may include the following steps:

[0117] Step 201 : obtaining first temperature data of the punching machine worktable, second temperature data of the slider, wear data of the sensor head, and first position data of the slider at the current moment.

[0118] The first position data is the bottom dead point position data of the slider at the current moment, and the sensor is set at the guide plate of the punching machine workbench.

[0119] Step 202 : Process the first temperature data of the punching machine worktable, the second temperature data of the slider, and the wear data of the sensor head to determine target characteristic data.

[0120] Step 203: input the target feature data into the trained target network model to determine the predicted compensation value after being processed by the target network model.

[0121] It should be noted that the specific content and implementation methods of steps 201 to 203 can be referred to the description of each embodiment of the present application and will not be repeated here.

[0122] Step 204: Determine a temperature difference correction value based on the first temperature data and the second temperature data.

[0123] Step 205: Determine the target compensation value based on the predicted compensation value and the temperature difference correction value.

[0124] Among them, the second temperature data of the punch worktable and the first temperature data of the slider may be the same or different. When the two are different, there may be a certain temperature deviation. At this time, the predicted compensation value can be further corrected by the temperature difference correction item to obtain a more accurate and reliable target compensation value.

[0125] Optionally, the mean temperature data of the punching worktable can be determined based on the first temperature data of the punching worktable and the corresponding weight value, and then the temperature difference correction value can be determined based on the second temperature data, the mean temperature data and the temperature difference deformation coefficient.

[0126] Among them, since the first temperature data of the punching machine workbench is the temperature data of multiple different positions, and the corresponding weight values ​​may be different for different positions, the average temperature difference data can be determined based on the weight values ​​of each position and the first temperature data of the punching machine workbench. It can be expressed as:

[0127]

[0128] Among them, T ij is the temperature value of the pixel at row i and column j on the punching machine workbench collected by the infrared thermopile array, ω ij is a weight coefficient that can be used to reflect the thermal deformation sensitivity of the region.

[0129] Therefore, in the embodiment of the present application, due to the average temperature data of the punching machine workbench It is the weighted average of 64-point temperature data collected by the infrared thermopile array, which can more comprehensively and accurately characterize the overall thermal state of the workbench.

[0130] Correspondingly, the target compensation value can be expressed as:

[0131]

[0132] Among them, f STGCN-ECM (T s ,T t ,W) is the predicted compensation value, is the average temperature of the workbench, α is the temperature difference-deformation coefficient, is the temperature difference correction value.

[0133] In the embodiment of the present application, the possible temperature effects of the punching machine worktable and the slider are fully taken into consideration, so as to further determine the temperature correction value to update the predicted compensation value to obtain the target compensation value. The target compensation value fully takes into account the possible temperature effects and the wear of the sensor sensing head, thereby further improving the accuracy and reliability of the target compensation value.

[0134] Step 206 : Control the slider according to the target compensation value and the first position data to compensate for the bottom dead point position of the slider.

[0135] It can be understood that after determining the target compensation value, the slider can be controlled to move according to the target compensation value based on the first position data of the slider's current position, thereby achieving compensation for the slider's bottom dead point position, thereby improving the accuracy and reliability of the slider's bottom dead point position compensation.

[0136] Optionally, the control quantity of the controller can be determined based on the determined target compensation value, and then the output thrust of the voice coil motor can be determined based on the control quantity. Then, the motion state of the slider can be controlled based on the output thrust of the voice coil motor and the first position data to achieve compensation for the bottom dead point position of the slider.

[0137] Among them, the control quantity, such as current or voltage instruction, can be calculated through proportional (P), integral (I), and differential (D) links; or the control quantity can be calculated by any other desirable method, etc., and this application does not limit this.

[0138] In addition, the output thrust of the voice coil motor can be determined by adjusting the current through a current driver, such as PWM modulation; or the output thrust can be determined by any other desirable method, which is not limited in this application.

[0139] For example, when the current bottom dead center position information of the slider is the first position data, the target compensation value obtained through the above processing is ΔL obtain = +0.025mm, the controller control quantity u(t) is calculated at this time, and then the output thrust F of the voice coil motor can be further determined. The force F applied by the voice coil motor, the bottom dead point position of the slider after compensation is: the first position data + ΔL obtain .

[0140] It should be noted that the above-mentioned target compensation value, the control quantity u(t) of the controller, the output thrust F of the voice coil motor, etc. are all schematic illustrations and cannot be used as a limitation on the specific numerical values ​​and determination methods in this application.

[0141] It is understood that after controlling the slider's motion state based on the voice coil motor's output thrust and the first position data, second position data of the slider can be further obtained. The difference between the first and second position data is then determined as the actual displacement difference. The controller's control parameters are then adjusted based on the absolute value of the error between the actual displacement difference and the target compensation value.

[0142] The second position data is the bottom dead point position data of the slider after being controlled by the output thrust of the voice coil motor.

[0143] For example, if the current first position data of the slider is 0.1 mm, the target compensation value is ΔL obtain When the displacement is +0.025mm, the controller's control variable is u(t), the voice coil motor's output thrust is F, and the voice coil motor applies force F, resulting in the slider's second position data being: 0.1mm + 0.024mm. Therefore, the difference between the first and second position data is +0.024mm, meaning the actual displacement difference is +0.024mm. The absolute error between the actual displacement difference and the target compensation value is 0.001mm. At this point, the controller's control parameters can be adjusted based on this 0.001mm value, achieving closed-loop control.

[0144] It should be noted that the above examples are merely illustrative and cannot be used as limitations on the first position data, second position data, actual displacement difference, target compensation value, etc. in the embodiments of the present application.

[0145] Optionally, when adjusting the controller's control parameters based on the absolute value of the error, a segmented anti-disturbance method can be used. For example, when the absolute value of the error is greater than a certain threshold, high-gain PID control is used; when the error is less than or equal to a certain threshold, fuzzy PID control is used to dynamically adjust the parameters.

[0146] For example, when the threshold is 0.005mm, if the absolute value of the error currently determined is greater than 0.005mm, a high-gain PID such as K can be used in the coarse adjustment stage. p =2.0,K i =0.5,K d =0.2 to achieve rapid convergence. If the absolute value of the error currently determined is ≤0.005mm, the fine-tuning stage can be switched to fuzzy PID and parameters such as K can be dynamically adjusted. p '=K p μ(e),

[0147] in,

[0148]

[0149] It should be noted that the above examples are merely illustrative and cannot be used as a limitation on the control parameter adjustment method in the embodiments of the present application.

[0150] In an embodiment of the present application, the first temperature data of the punching worktable, the second temperature data of the slider, the wear data of the sensor sensing head and the first position data of the slider at the current moment can be first obtained, and then the first temperature data of the punching worktable, the second temperature data of the slider and the wear data of the sensor sensing head can be processed to determine the target feature data, and then the target feature data can be input into the trained target network model to determine the predicted compensation value after processing by the target network model, and then the temperature difference correction value can be determined according to the first temperature data and the second temperature data, and the target compensation value can be determined according to the predicted compensation value and the temperature difference correction value, and then the slider can be controlled according to the target compensation value and the first position data to achieve compensation for the bottom dead point position of the slider. Therefore, the first temperature data of the punching table, the second temperature data of the slider, and the wear data of the sensor sensing head are first processed to determine the target characteristic data, and then the predicted compensation value is obtained in combination with the target network model. The predicted compensation value is then updated according to the error correction value to obtain the target compensation value, and finally the slider is controlled based on the target compensation value. That is, in the process of compensating the bottom dead point position of the slider, the temperature influence of the punching table and the slider and the wear influence of the sensor sensing head are fully considered, thereby effectively improving the accuracy and reliability of the bottom dead point position compensation and ensuring the smooth operation of the punching machine.

[0151] like Figure 4As shown, the method for compensating the bottom dead point position of the punch slide may include the following steps:

[0152] Step 301: Acquire a training data set, wherein the training data set includes the third temperature data of the punching machine worktable, the fourth temperature data of the slider, the historical wear data of the sensor head, the historical bottom dead point position data of the slider, and label data.

[0153] The third temperature data of the punching worktable can be understood as the historical temperature data of the punching worktable. The temperature data can be collected by setting multiple sensors, or by using an infrared thermal imager, etc. The third temperature data fully reflects the temperature condition of the punching worktable in the spatial dimension. The third temperature data and the first temperature data can be temperature data collected at the same position on the punching worktable. This application is not limited to this.

[0154] In addition, the fourth temperature data of the slider can be collected by embedding a temperature sensor in the slider. For example, a PT100 temperature sensor can be embedded in the slider to collect sliding window time series temperature data at a specific frequency, etc. This application does not limit this.

[0155] In addition, the historical wear data of the sensor head can be calculated from the contact force data monitored in real time by the integrated force sensor and the number of touches recorded by the Hall sensor, and exponential weighted smoothing is used to eliminate noise. This application does not impose any restrictions on this.

[0156] In addition, label data can be obtained by measuring the thermal deformation displacement using a contact scale CMOS sensor.

[0157] Step 302 : Process the third temperature data of the punching machine worktable, the fourth temperature data of the slider, and the historical wear data of the sensor head to determine training feature data.

[0158] Among them, any desirable method can be used to perform feature extraction processing on the third temperature data of the punching machine worktable, the third temperature data of the slider, and the historical wear data of the sensor head. For example, it can be implemented through a convolutional neural network, or through a feature extractor, etc. This application does not limit this.

[0159] Step 303: input the training feature data into the initial network model to determine the training compensation value after being processed by the initial network model.

[0160] The initial network model can be any network model capable of predictive regression, or a pre-trained model. For example, the initial network model can be a spatiotemporal graph convolutional network model, where the graph convolution portion can process spatial feature data and the causal convolution portion can process temporal feature data. This application does not impose any restrictions on this.

[0161] Step 304: Determine a loss value based on the difference between the training compensation value and the label data.

[0162] Among them, the loss value can be calculated by using a loss function, such as the mean squared error (MSE) loss function, the mean absolute error (MAE) loss function, or the Huber Loss function, etc. This application does not limit this.

[0163] Step 305: Train the initial network model according to the loss value to generate a trained target network model.

[0164] It is understood that during the training of the initial network model based on the loss value, the RAdam optimizer can be used in conjunction with the Huber loss function to enhance model robustness. Gradient clipping is implemented to prevent gradient explosion, and an early stopping mechanism is set to avoid overfitting. The number of epochs can be set to 500, or other values, which are not limited in this application.

[0165] Optionally, the cosine annealing strategy can be used to dynamically adjust the learning rate, which can reduce the minimum learning rate to 1×10 -5 To ensure full convergence, the training data is augmented through random temperature field shifting and Gaussian noise injection to improve model generalization. After each round of training, the validation set is evaluated, and the model parameters with the lowest validation loss are retained. After all rounds of training are completed, the model with the best performance is selected as the target network model.

[0166] Therefore, in the embodiment of the present application, during the training process, the initial network model is trained by continuously using the training data set until the model converges, and finally a target network model is obtained. The target network model has good performance, which provides a guarantee for the accuracy of the subsequent use of the target network model to compensate for the bottom dead point position of the slider.

[0167] In an embodiment of the present application, a training data set can be first obtained, wherein the training data set includes the third temperature data of the punching machine worktable, the fourth temperature data of the slider, the historical wear data of the sensor sensing head, the historical bottom dead point position data of the slider, and the label data. Then, the third temperature data of the punching machine worktable, the fourth temperature data of the slider, and the historical wear data of the sensor sensing head can be processed to determine the training feature data, and then the training feature data can be input into the initial network model to determine the training compensation value after processing by the initial network model. Then, based on the difference between the training compensation value and the label data, the loss value is determined, and the initial network model is trained based on the loss value to generate a trained target network model. Thus, by using the training data set to train the initial network model until a trained target network model is obtained, the target network model has good performance, and the accuracy of the slider bottom dead point position compensation using the target network model is also improved.

[0168] According to the present application, a compensation device 400 for the bottom dead point position of a punching machine slider is provided, such as Figure 5 As shown, the device includes a first acquisition module 410 , a first processing module 420 , a first determination module 430 and a control module 440 .

[0169] Among them, the first acquisition module 410 is used to obtain the first temperature data of the punch worktable, the second temperature data of the slider, the wear data of the sensor sensing head and the first position data of the slider at the current moment, wherein the first position data is the bottom dead point position data of the slider at the current moment.

[0170] The first processing module 420 is used to process the first temperature data of the punching machine worktable, the second temperature data of the slider, and the wear data of the sensor head to determine target feature data.

[0171] The first determination module 430 is used to input the target feature data into the trained target network model to determine the predicted compensation value after being processed by the target network model.

[0172] The control module 440 is configured to control the slider according to the predicted compensation value and the first position data, so as to compensate for the bottom dead point position of the slider.

[0173] Optionally, the first processing module 420 is specifically configured to:

[0174] Performing feature extraction on the first temperature data to obtain first feature data of the punching machine worktable;

[0175] Performing feature extraction on the second temperature data of the slider to obtain second feature data of the slider;

[0176] performing feature extraction on the wear data of the sensor head to obtain third feature data of the sensor head;

[0177] The first feature data, the second feature data and the third feature data are fused to determine target feature data.

[0178] Optionally, the control module 440 includes:

[0179] a first determining unit, configured to determine a temperature difference correction value according to the first temperature data and the second temperature data;

[0180] a second determining unit, configured to determine a target compensation value based on the predicted compensation value and the temperature difference correction value;

[0181] A control unit is used to control the slider according to the target compensation value and the first position data to achieve compensation for the bottom dead point position of the slider.

[0182] Optionally, the first determining unit is specifically configured to:

[0183] Determining mean temperature data of the punching worktable according to the first temperature data of the punching worktable and the corresponding weight value;

[0184] A temperature difference correction value is determined according to the second temperature data, the mean temperature data, and the temperature difference deformation coefficient.

[0185] Optionally, the control unit is specifically configured to:

[0186] determining a control variable of a controller according to the target compensation value;

[0187] determining the output thrust of the voice coil motor according to the control amount;

[0188] The motion state of the slider is controlled according to the output thrust of the voice coil motor and the first position data, so as to compensate for the bottom dead point position of the slider.

[0189] Optionally, the control module is further configured to:

[0190] Acquiring second position data of the slider, wherein the second position data is bottom dead center position data of the slider after being controlled by the output thrust of the voice coil motor;

[0191] determining a difference between the second position data and the first position data as an actual displacement difference;

[0192] The control parameters of the controller are adjusted according to the absolute value of the error between the actual displacement difference and the target compensation value.

[0193] Optionally, the device further includes:

[0194] a second acquisition module, configured to acquire a training data set, wherein the training data set includes third temperature data of the punching machine worktable, fourth temperature data of the slider, historical wear data of the sensor head, historical bottom dead center position data of the slider, and label data;

[0195] a second processing module, configured to process the third temperature data of the punching machine worktable, the fourth temperature data of the slider, and the historical wear data of the sensor head to determine training feature data;

[0196] a second determining module, configured to input the training feature data into an initial network model to determine a training compensation value after processing by the initial network model;

[0197] A third determining module is used to determine a loss value according to a difference between the training compensation value and the label data;

[0198] A generation module is used to train the initial network model according to the loss value to generate a trained target network model.

[0199] The compensation device for the bottom dead point position of the punch press slider provided in the present application can first obtain the first temperature data of the punch press worktable, the second temperature data of the slider, the wear data of the sensor sensing head and the first position data of the slider at the current moment, wherein the first position data is the bottom dead point position data of the slider at the current moment, and the sensor is arranged at the guide plate of the punch press worktable. Thereafter, the first temperature data of the punch press worktable, the second temperature data of the slider, and the wear data of the sensor sensing head can be processed to determine the target feature data, and then the target feature data can be input into the trained target network model to determine the predicted compensation value after processing by the target network model. Then, the slider is controlled according to the predicted compensation value and the first position data to realize compensation for the bottom dead point position of the slider. Therefore, by processing the first temperature data of the punching machine worktable, the second temperature data of the slider, and the wear data of the sensor sensing head, the target characteristic data is determined, and then the predicted compensation value is obtained in combination with the target network model. Finally, the slider is controlled based on the predicted compensation value. That is, in the process of compensating the bottom dead point position of the slider, the temperature influence of the punching machine worktable and the slider and the wear influence of the sensor sensing head are fully considered, thereby effectively improving the accuracy and reliability of the bottom dead point position compensation and ensuring the smooth operation of the punching machine.

[0200] It should be understood that the specific features, operations, and details described hereinabove with respect to the method of the present application may also be similarly applied to the apparatus and system of the present application, or vice versa. In addition, each step of the method of the present application described above may be performed by a corresponding component or unit of the apparatus or system of the present application.

[0201] It should be understood that the various modules / units of the apparatus of the present application may be implemented in whole or in part by software, hardware, firmware, or a combination thereof. Each module / unit may be embedded in the processor of the electronic device in the form of hardware or firmware or may be independent of the processor, or may be stored in the memory of the electronic device in the form of software for the processor to call to execute the operation of each module / unit. Each module / unit may be implemented as an independent component or module, or two or more modules / units may be implemented as a single component or module.

[0202] like Figure 6 As shown, the present application provides an electronic device 500, which includes a processor 501 and a memory 502 storing computer program instructions. When the processor 501 executes the computer program instructions, it implements the steps of the above-mentioned method for compensating the bottom dead center position of the punch slide. The electronic device 500 can be broadly defined as a server, a terminal, or any other electronic device with the necessary computing and / or processing capabilities.

[0203] In one embodiment, the electronic device 500 may include a processor, memory, network interface, communication interface, etc. connected via a system bus. The processor of the electronic device 500 may be used to provide necessary computing, processing, and / or control capabilities. The memory of the electronic device 500 may include a non-volatile storage medium and an internal memory. The non-volatile storage medium may store an operating system, computer programs, etc. The internal memory may provide an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface and communication interface of the electronic device 500 may be used to connect to and communicate with external devices via a network. When the computer program is executed by the processor, the steps of the method of the present application are performed.

[0204] The present application provides a computer-readable storage medium having computer program instructions stored thereon. When the computer program instructions are executed by a processor, the above-mentioned method for compensating the bottom dead point position of the punch press slider is implemented.

[0205] Those skilled in the art will appreciate that the method steps of the present application can be performed by instructing relevant hardware such as the electronic device 500 or a processor through a computer program, and the computer program can be stored in a non-transitory computer-readable storage medium, which causes the steps of the present application to be performed when the computer program is executed. Depending on the circumstances, any reference to memory, storage or other media herein may include non-volatile or volatile memory. Examples of non-volatile memory include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, magnetic tape, floppy disk, magneto-optical data storage device, optical data storage device, hard disk, solid-state disk, etc. Examples of volatile memory include random access memory (RAM), external cache memory, etc.

[0206] The various technical features described above can be combined arbitrarily. Although not all possible combinations of these technical features are described, any combination of these technical features should be considered to be covered by this specification as long as such combination does not conflict.

[0207] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for compensating the bottom dead point position of a punch press slide, characterized in that: include: Acquire first temperature data of the punching machine worktable, second temperature data of the slider, wear data of the sensor head, and first position data of the slider at the current moment, wherein the first position data is bottom dead center position data of the slider at the current moment; Processing the first temperature data of the punching machine worktable, the second temperature data of the slider, and the wear data of the sensor head to determine target characteristic data; Inputting the target feature data into the trained target network model to determine a predicted compensation value after being processed by the target network model; The slider is controlled according to the predicted compensation value and the first position data to achieve compensation for the bottom dead center position of the slider.

2. The method according to claim 1, wherein The processing of the first temperature data of the punching machine worktable, the second temperature data of the slider, and the wear data of the sensor head to determine the target characteristic data includes: Performing feature extraction on the first temperature data to obtain first feature data of the punching machine worktable; Performing feature extraction on the second temperature data of the slider to obtain second feature data of the slider; performing feature extraction on the wear data of the sensor head to obtain third feature data of the sensor head; The first feature data, the second feature data and the third feature data are fused to determine target feature data.

3. The method according to claim 1, wherein The controlling of the slider according to the predicted compensation value and the first position data to achieve compensation for the bottom dead center position of the slider includes: determining a temperature difference correction value according to the first temperature data and the second temperature data; Determining a target compensation value based on the predicted compensation value and the temperature difference correction value; The slider is controlled according to the target compensation value and the first position data to achieve compensation for the bottom dead point position of the slider.

4. The method according to claim 3, wherein The determining of the temperature difference correction value according to the first temperature data and the second temperature data includes: Determining mean temperature data of the punching worktable according to the first temperature data of the punching worktable and the corresponding weight value; A temperature difference correction value is determined according to the second temperature data, the mean temperature data, and the temperature difference deformation coefficient.

5. The method according to claim 3, wherein The controlling of the slider according to the target compensation value and the first position data to achieve compensation for the bottom dead point position of the slider includes: determining a control variable of a controller according to the target compensation value; determining the output thrust of the voice coil motor according to the control amount; The motion state of the slider is controlled according to the output thrust of the voice coil motor and the first position data, so as to compensate for the bottom dead point position of the slider.

6. The method according to claim 5, wherein After controlling the motion state of the slider according to the output thrust of the voice coil motor and the first position data, the method further includes: Acquiring second position data of the slider, wherein the second position data is bottom dead center position data of the slider after being controlled by the output thrust of the voice coil motor; determining a difference between the second position data and the first position data as an actual displacement difference; The control parameters of the controller are adjusted according to the absolute value of the error between the actual displacement difference and the target compensation value.

7. The method according to claim 1, wherein Before inputting the target feature data into the trained target network model to determine the predicted compensation value through processing by the target network model, the method further includes: Acquire a training data set, wherein the training data set includes third temperature data of the punching machine worktable, fourth temperature data of the slider, historical wear data of the sensor head, historical bottom dead center position data of the slider, and label data; Processing the third temperature data of the punching machine worktable, the fourth temperature data of the slider, and the historical wear data of the sensor head to determine training feature data; Inputting the training feature data into an initial network model to determine a training compensation value after being processed by the initial network model; Determining a loss value based on a difference between the training compensation value and the label data; The initial network model is trained according to the loss value to generate a trained target network model.

8. A compensation device for the bottom dead point position of a punch press slide, characterized in that: include: a first acquisition module, configured to acquire first temperature data of the punching machine worktable, second temperature data of the slider, wear data of the sensor head, and first position data of the slider at a current moment, wherein the first position data is bottom dead center position data of the slider at a current moment, and the sensor is disposed on a guide plate of the punching machine worktable; a first processing module, configured to process the first temperature data of the punching machine worktable, the second temperature data of the slider, and the wear data of the sensor head to determine target characteristic data; a first determination module, configured to input the target feature data into a trained target network model, so as to determine a predicted compensation value after processing by the target network model; A control module is used to control the slider according to the predicted compensation value and the first position data to achieve compensation for the bottom dead point position of the slider.

9. The device according to claim 8, wherein The first processing module is specifically configured to: Performing feature extraction on the first temperature data to obtain first feature data of the punching machine worktable; Performing feature extraction on the second temperature data of the slider to obtain second feature data of the slider; performing feature extraction on the wear data of the sensor head to obtain third feature data of the sensor head; The first feature data, the second feature data and the third feature data are fused to determine target feature data.

10. An electronic device, characterized in that: The electronic device includes: a processor and a memory storing computer program instructions; When the processor executes the computer program instructions, the method for compensating the bottom dead point position of the punch slide as described in any one of claims 1 to 7 is implemented.

Images