Automatic lubrication control method and system for plate drawing forming and storage medium

Abstract

The invention relates to the technical field of plate drawing forming control, and discloses an automatic lubrication control method and system for plate drawing forming and a storage medium. The method comprises the steps that the surface of a deep drawing die is subjected to zoning treatment through a lubricating area dividing module, and parameters of all areas are initialized; carrying out real-time acquisition and pretreatment on process parameters; lubrication demand parameters are obtained through fuzzy control calculation; a proportional servo valve is used for pressure regulation; the lubrication pressure curve is corrected on line; and the lubrication parameters are analyzed and stored in a database. According to the method, the lubricating state of each area is intelligently adjusted by monitoring and analyzing the technological parameters in real time, so that the stability of the drawing forming quality is guaranteed.

Description

Technical Field

[0001] The present application relates to the field of sheet metal deep drawing control, and in particular to an automatic lubrication control method, system and storage medium for sheet metal deep drawing. Background Art

[0002] In the sheet metal drawing process, the lubrication state directly affects the forming quality and die life. Traditional lubrication control methods mainly use constant pressure oil supply or simple closed-loop control to maintain constant lubrication pressure and flow during the drawing process. With the improvement of industrial automation level, some adaptive lubrication control methods have been developed, such as zoned oil supply control based on pressure sensor feedback, real-time control of oil supply adjustment according to displacement sensor signal, etc. These methods have improved the lubrication effect and reduced product defects to a certain extent.

[0003] However, the existing lubrication control methods have the following shortcomings: Due to the dynamic changes in stress state and friction conditions during sheet deformation, a single closed-loop control cannot accurately reflect the real-time lubrication needs of each area. At the same time, traditional control methods lack in-depth analysis and mining of process parameters, and do not establish a mapping relationship between lubrication parameters and forming quality. In addition, the optimization process of control parameters often relies on empirical settings and lacks scientific data support, resulting in unstable lubrication effects and difficulty in adapting to the needs of different working conditions. Summary of the invention

[0004] The present application provides an automatic lubrication control method, system and storage medium for sheet metal deep drawing, which are used to intelligently adjust the lubrication status of each area through real-time monitoring and analysis of process parameters, thereby ensuring the stability of deep drawing quality.

[0005] In a first aspect, the present application provides an automatic lubrication control method for sheet metal deep drawing, the automatic lubrication control method for sheet metal deep drawing comprising: partitioning the surface of the drawing die through a lubrication area partitioning module to obtain a plurality of independent lubrication control areas including a flange area and a fillet area, and initializing parameters of each lubrication control area according to physical parameters in a workpiece material library to obtain a reference lubrication pressure value of each area; real-time acquisition of process parameters according to a sensor network distributed in each lubrication control area, filtering and standardization processing through a data preprocessing module to obtain a process state digital model; based on the process state digital model, analyzing and calculating the stress state of each lubrication control area through a fuzzy control algorithm to obtain real-time lubrication demand parameters of each area; based on the real-time lubrication demand parameters, pulsatingly adjusting the lubrication pressure of each lubrication control area using a proportional servo valve control system to form a dynamic lubrication pressure control curve; online evaluation of the sheet metal thinning rate and surface roughness, and correction of the dynamic lubrication pressure control curve according to the evaluation results to obtain optimized lubrication parameters; correlation analysis of the optimized lubrication parameters, establishment of a lubrication parameter optimization model, and storage of the lubrication parameter optimization model in a process parameter database.

[0006] In a second aspect, the present application provides an automatic lubrication control system for sheet metal deep drawing, the automatic lubrication control system for sheet metal deep drawing comprising: The partitioning module is used to partition the surface of the drawing die through the lubrication area partitioning module to obtain multiple independent lubrication control areas including the flange area and the fillet area, and initialize the parameters of each lubrication control area according to the physical parameters in the workpiece material library to obtain the reference lubrication pressure value of each area; The acquisition module is used to collect process parameters in real time according to the sensor network distributed in each lubrication control area, and obtain a digital model of the process state by filtering and standardizing the data through the data preprocessing module; A calculation module, used to analyze and calculate the stress state of each lubrication control area according to the process state digital model through a fuzzy control algorithm to obtain the real-time lubrication demand parameters of each area; A regulating module, for pulsatingly regulating the lubrication pressure of each lubrication control area by using a proportional servo valve control system based on the real-time lubrication demand parameter to form a dynamic lubrication pressure control curve; A correction module, used for online evaluation of the plate thinning rate and surface roughness, and correcting the dynamic lubrication pressure control curve according to the evaluation results to obtain optimized lubrication parameters; The analysis module is used to perform correlation analysis on the optimized lubrication parameters, establish a lubrication parameter optimization model, and store the lubrication parameter optimization model in a process parameter database.

[0007] A third aspect of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the above-mentioned automatic lubrication control method for sheet metal deep drawing.

[0008] In the technical solution provided by the present application, the surface of the drawing die is partitioned by the lubrication area division module, and the accurate lubrication control of key parts such as the flange area and the fillet area is realized, and the problem of uneven lubrication caused by the traditional single oil supply method is avoided. At the same time, the parameters of each lubrication control area are initialized according to the physical parameters in the workpiece material library, ensuring the scientificity and rationality of the reference lubrication pressure value. The sensor network arranged in each lubrication control area collects the process parameters in real time, and the data preprocessing module performs filtering and standardization processing to establish an accurate process state digital model, which provides a reliable data basis for the dynamic adjustment of the lubrication parameters. The fuzzy control algorithm is used to analyze and calculate the stress state of each lubrication control area, and the real-time lubrication demand parameters of each area are accurately obtained, which reflects the intelligence and adaptability of the control strategy. The proportional servo valve control system is used to pulsate the lubrication pressure of each lubrication control area to form a dynamic lubrication pressure control curve, and the precise regulation of the lubrication pressure is realized. By online evaluation of the sheet thinning rate and surface roughness, the dynamic lubrication pressure control curve is corrected according to the evaluation results, and the optimized lubrication parameters are obtained to ensure the stability of the forming quality. Finally, the correlation analysis of the optimized lubrication parameters was carried out, and the lubrication parameter optimization model was established and stored in the process parameter database, which achieved continuous optimization of control parameters and experience accumulation, and provided data support for lubrication control under different working conditions. The entire control process realizes full process automation from parameter collection, data processing to control execution, which significantly improves the processing quality and production efficiency of sheet metal deep drawing. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0010] Figure 1 A schematic diagram of an embodiment of an automatic lubrication control method for deep drawing of a sheet material in an embodiment of the present application; Figure 2 A schematic diagram of the effect of forced lubrication on forming force in an embodiment of the present application; Figure 3 This is a schematic diagram of an embodiment of an automatic lubrication control system for sheet metal deep drawing in an embodiment of the present application. DETAILED DESCRIPTION

[0011] The embodiments of the present application provide an automatic lubrication control method, system and storage medium for sheet metal deep drawing. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described here can be implemented in an order other than that illustrated or described here. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0012] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 In the embodiment of the present application, an embodiment of the automatic lubrication control method for sheet metal deep drawing forming includes: Step S101, partitioning the surface of the drawing die through the lubrication area partitioning module to obtain multiple independent lubrication control areas including the flange area and the fillet area, and initializing the parameters of each lubrication control area according to the physical parameters in the workpiece material library to obtain the reference lubrication pressure value of each area; Step S102, collecting process parameters in real time according to the sensor network distributed in each lubrication control area, filtering and standardizing the process parameters through a data preprocessing module, and obtaining a process state digital model; Step S103, analyzing and calculating the stress state of each lubrication control area through a fuzzy control algorithm according to the process state digital model, and obtaining the real-time lubrication demand parameters of each area; Step S104: Based on the real-time lubrication demand parameter, the lubrication pressure of each lubrication control area is pulsatingly adjusted by using a proportional servo valve control system to form a dynamic lubrication pressure control curve; Step S105, online evaluation of the plate thinning rate and surface roughness, and correction of the dynamic lubrication pressure control curve according to the evaluation results to obtain optimized lubrication parameters; Step S106: performing correlation analysis on the optimized lubrication parameters, establishing a lubrication parameter optimization model, and storing the lubrication parameter optimization model in a process parameter database.

[0013] It is understandable that the execution subject of the present application may be an automatic lubrication control system for sheet metal deep drawing, or a terminal or a server, which is not specifically limited here. The present application embodiment is described by taking a server as the execution subject as an example.

[0014] Specifically, the surface of the drawing die is partitioned by the lubrication area partitioning module, and the surface of the drawing die is regionally identified according to the geometric features, and the surface of the drawing die is divided into a flange area and a fillet area. The flange area is located at the edge of the plate and bears the maximum deformation, and the fillet area is located in the transition area of ​​the mold cavity. The oil circuit network is divided in each area according to the lubrication requirements, and each oil circuit is equipped with an independent pressure sensor and flow sensor. The friction coefficient, yield strength and hardening index of the plate to be processed are extracted from the workpiece material library to generate a material property data set. According to the geometric characteristics and material property data sets of each area, the stress state of the plate in each area is numerically analyzed, and the stress distribution is calculated by finite element analysis. According to the stress distribution, the weighted allocation method is used to allocate the initial pressure coefficient to each oil circuit, and the reference lubrication pressure value of each area is calculated. The calculation of the reference lubrication pressure value needs to take into account the deformation degree, friction state and material properties of each area, and the reference lubrication pressure value is written into the control parameter table corresponding to each area through the PLC control unit. When the sensor network collects process parameters in real time, the pressure sensor collects real-time pressure data of each area, the flow sensor monitors the flow state of the lubricant, the displacement sensor tracks the deformation degree of the sheet, and the force sensor measures the forming force during the drawing process. Figure 2 The figure is a schematic diagram of the effect of forced lubrication on forming force in the embodiment of the present application; where, under the same process parameters such as mold and blank shape, blank holder force, and when the punch displacement is the same, a smaller forming load indicates better lubrication conditions, while poor friction conditions often lead to a larger forming load. Figure 2 It can be seen from the middle curve that in the initial stage of deep drawing, the deformation of the flange area is small, and the difference in the effect of different lubrication conditions on the forming load is not obvious. As the flange area sheet flows into the center of the die, the tangential compressive stress becomes larger and larger, the wrinkling and thickening become more and more serious, the friction environment becomes more and more severe, and the difference in the effect of different lubrication conditions on the forming load begins to become obvious. In addition, from Figure 2It can be seen from the figure that the maximum load under each lubrication condition is about 90kN, because this is the bearing limit of the material selected for the test in the forming of this geometric parameter. If this load limit is exceeded, it will break. Therefore, it is unscientific to judge the quality of lubrication performance only from the maximum load value, but it can be evaluated by comparing the load values ​​at the same punch stroke moment. When the blank holder force is 20kN and the punch stroke is 30mm, the forming force after applying 5MPa and 9MPa forced lubrication is reduced by 8.0% and 8.6% respectively compared with no forced lubrication; when the blank holder force is 35kN and the punch stroke is 30mm, the forming force after applying 5MPa and 9MPa forced lubrication is reduced by 2.4% and 3.5% respectively compared with no forced lubrication; when the blank holder force is 50kN and the punch stroke is 30mm, the forming force after applying 5MPa and 9MPa forced lubrication is reduced by 3.4% and 8.9% respectively compared with no forced lubrication. After forced lubrication under the conditions of blank holder force of 20kN and 50kN, the forming load is reduced by a large margin, reaching a maximum of 8.9%. The test results show that applying forced lubrication during the deep drawing process helps to reduce the forming force. The greater the forced lubrication pressure within the test range, the greater the reduction in forming force.

[0015] The raw data collected by the sensor is de-noised by a Butterworth filter to obtain filtered process parameter data. The Butterworth filter has the maximum flat amplitude-frequency characteristic and effectively suppresses high-frequency noise by setting the cutoff frequency. The filtered process parameter data is classified according to pressure data, flow data, displacement data and force data to form a classified data group. Each type of data in the classified data group is normalized to obtain standardized process parameters. The standardized process parameters are data-associated according to each lubrication control area to construct a process parameter association matrix. The process parameter association matrix is ​​reduced in dimension by principal component analysis to obtain a digital model of the process state.

[0016] According to the process state digital model, the pressure data, flow data and force data of each lubrication control area are extracted as fuzzy control input variables. The pressure data, flow data and force data are fuzzified according to three levels: low, medium and high to obtain a fuzzy set. The fuzzy set is defuzzified by the weighted average method to obtain the stress state value of each lubrication control area. The stress state value is ratioed with the reference lubrication pressure value to form the lubrication pressure deviation data. The lubrication pressure deviation data is incrementally calculated to generate the lubrication pressure compensation value. The lubrication pressure compensation value is superimposed with the reference lubrication pressure value to obtain the real-time lubrication demand parameters of each area. Based on the real-time lubrication demand parameters, the proportional servo valve opening of each lubrication control area is calculated to obtain the valve opening control data. The valve opening control data is converted into a PWM control signal to generate the valve pulsation frequency and duty cycle data. The PWM control signal adjusts the output power by changing the pulse width to achieve precise control of the valve opening. The valve pulsation frequency and duty cycle data are converted from digital to analog to output the valve control voltage signal. The actual output pressure of the proportional servo valve in each lubrication control area is collected to form pressure feedback data. The pressure feedback data and the real-time lubrication demand parameters are used for deviation calculation to generate a pressure compensation signal. The pressure compensation signal and the pressure feedback data are arranged in time series to form a dynamic lubrication pressure control curve.

[0017] During the sheet drawing process, the thickness of each measuring point of the sheet is collected by an ultrasonic thickness sensor to obtain the sheet thickness distribution data. The difference between the sheet thickness distribution data and the initial thickness of the sheet is calculated to obtain the thinning data of each measuring point. The thinning data is normalized to form the sheet thinning rate data. The surface roughness measurement value of the sheet is collected to generate the surface quality data. The sheet thinning rate data and the surface quality data are weighted and superimposed to generate the quality evaluation parameters. The dynamic lubrication pressure control curve is proportionally corrected by the quality evaluation parameters to obtain the optimized lubrication parameters. When performing correlation analysis on the optimized lubrication parameters, the pressure value, flow value and time series data are first extracted from the optimized lubrication parameters, and the data are grouped according to the lubrication control area. The correlation between the pressure value and the flow value in the grouped data is calculated by the Pearson correlation coefficient to generate the parameter correlation matrix. The parameter correlation matrix is ​​subjected to eigenvalue decomposition to obtain the lubrication parameter eigenvector. The lubrication parameter eigenvector and the baseline lubrication pressure value of each lubrication control area are subjected to deviation analysis to form the pressure correction coefficient. The pressure correction coefficient is numerically fitted by the weighted average method to obtain the lubrication pressure control equation. Based on the lubrication pressure control equation, the least square method is used to establish the lubrication parameter optimization model. The lubrication parameter optimization model and the corresponding process parameters are encoded and stored in the process parameter database.

[0018] For example: In practical applications, take the deep drawing of a steel plate as an example. First, the deep drawing die is divided into zones, 8 oil circuits are set in the flange area, and 6 oil circuits are set in the fillet area. The material parameters of the steel plate are extracted: friction coefficient 0.15, yield strength 235MPa, hardening index 0.21. After calculating the stress distribution of each area, the initial pressure coefficient is obtained: 1.2 in the flange area and 1.0 in the fillet area. Then, according to the real-time data collected by the sensor, the high-frequency noise is removed by Butterworth filtering, and the data is normalized. The stress state of each area is calculated by the fuzzy control algorithm, and the pressure compensation value is generated after comparison with the reference pressure. The proportional servo valve adjusts the opening according to the compensation value, and adopts PWM signal control to achieve precise pressure regulation. After the sheet thinning rate and surface roughness data are processed, the dynamic pressure curve is corrected. The lubrication parameter optimization model is established through correlation analysis to achieve precise control of the deep drawing process.

[0019] In the embodiment of the present application, the surface of the drawing die is partitioned by the lubrication area division module, and the accurate lubrication control of key parts such as the flange area and the fillet area is realized, and the problem of uneven lubrication caused by the traditional single oil supply method is avoided. At the same time, the parameters of each lubrication control area are initialized according to the physical parameters in the workpiece material library, ensuring the scientificity and rationality of the reference lubrication pressure value. The sensor network arranged in each lubrication control area collects the process parameters in real time, and the data preprocessing module performs filtering and standardization processing to establish an accurate process state digital model, which provides a reliable data basis for the dynamic adjustment of the lubrication parameters. The fuzzy control algorithm is used to analyze and calculate the stress state of each lubrication control area, and the real-time lubrication demand parameters of each area are accurately obtained, which reflects the intelligence and adaptability of the control strategy. The lubrication pressure of each lubrication control area is pulsated by using the proportional servo valve control system to form a dynamic lubrication pressure control curve, which realizes the precise regulation of the lubrication pressure. By online evaluation of the sheet thinning rate and surface roughness, the dynamic lubrication pressure control curve is corrected according to the evaluation results, and the optimized lubrication parameters are obtained to ensure the stability of the forming quality. Finally, the correlation analysis of the optimized lubrication parameters was carried out, and the lubrication parameter optimization model was established and stored in the process parameter database, which achieved continuous optimization of control parameters and experience accumulation, and provided data support for lubrication control under different working conditions. The entire control process realizes full process automation from parameter collection, data processing to control execution, which significantly improves the processing quality and production efficiency of sheet metal deep drawing.

[0020] In a specific embodiment, the process of executing step S101 may specifically include the following steps: (1) The surface of the drawing die is identified according to its geometric features and the surface of the drawing die is divided into the flange area and the fillet area; (2) Divide the oil circuit network in each area according to lubrication requirements, and set up independent pressure sensors and flow sensors for each oil circuit; (3) Extract the friction coefficient, yield strength, and hardening index of the plate to be processed from the workpiece material library to generate a material property data set; (4) Calculate the stress distribution of each region according to the geometric characteristics and material property data sets of each region; (5) Based on the stress distribution, the initial pressure coefficient is assigned to each oil circuit, and the reference lubrication pressure value of each area is calculated; (6) The reference lubrication pressure value is written into the control parameter table corresponding to each area through the PLC control unit.

[0021] Specifically, a three-dimensional digital model is established when the surface of the drawing die is identified, and the die surface is divided into a flange area and a fillet area by geometric feature analysis. The flange area is located at the outer edge of the sheet and mainly controls the flow of the sheet; the fillet area is the key transition area for sheet forming; the area division adopts the curvature analysis method, and the boundary is determined by calculating the main curvature value of each point on the die surface. A lubricating oil circuit network is set in each area, and the oil circuit spacing is determined according to the deformation gradient. The oil circuits are densely arranged in the area with large deformation, and the spacing is appropriately increased in the area with small deformation. Each oil circuit is equipped with an independent pressure sensor and flow sensor. The pressure sensor has a measurement range of 0-20MPa and an accuracy of 0.1%. The flow sensor has a range of 0-15L / min. When extracting material parameters from the workpiece material library, the three key parameters are mainly friction coefficient, yield strength and hardening index. The friction coefficient reflects the friction state between the material and the die, which is usually obtained through a slider friction test; the yield strength represents the stress value when the material begins to plastically deform, which is determined by a tensile test; the hardening index describes the work hardening characteristics of the material, which is calculated from the true stress-true strain curve. These parameters constitute the material property data set and provide basic data for subsequent stress calculations.

[0022] The stress distribution calculation of each region is based on the finite element analysis method. The contact analysis model between the plate and the mold is established, and the parameters in the material property data set are imported into the model. The calculation process includes: establishing a geometric model, dividing mesh units, setting boundary conditions, defining contact pairs, applying loads, etc. The stress distribution cloud map of each region is obtained through iterative calculation, focusing on the radial stress in the flange area and the equivalent stress in the fillet area.

[0023] The process of calculating the reference lubrication pressure value based on the stress distribution can be expressed as: in: is the base lubrication pressure value; is the lubrication pressure adjustment coefficient; is the regional deformation coefficient; is the material correction factor; is the weight coefficient of the i-th stress component; is the value of the i-th stress component; is the local strain rate coefficient; n is the number of stress components involved in the calculation.

[0024] After the PLC control unit receives the reference lubrication pressure value data, it first converts the data format to convert the floating point number into 16-bit integer data. Then the data is written into the data register corresponding to each area according to the preset storage address. The data is written in batch transmission mode, and the communication between the PLC and the host computer is realized through the Modbus-TCP protocol. In addition to the reference lubrication pressure value, the control parameter table also contains parameters such as the upper and lower pressure limits and alarm thresholds.

[0025] For example: for the deep drawing process of a certain automobile cover. First, a three-dimensional model of the mold is established, and the mold surface is divided into three main areas through curvature analysis. 12 annular oil circuits are set in the flange area and 8 in the fillet area. The oil circuit spacing is 30mm in the high stress area and 50mm in the low stress area. The plate parameters are extracted from the material library: friction coefficient 0.12 (measured by slider friction test), yield strength 320MPa (obtained by tensile test), hardening index 0.24 (stress-strain curve fitting). Finite element analysis uses quadrilateral shell units to divide the grid, with a unit size of 3mm, to obtain the stress distribution cloud map of each area. Based on the stress distribution data, the baseline lubrication pressure value of the flange area is calculated to be 12MPa, and the fillet area is 15MPa. After the format conversion, the data is written into the D100-D115 data register group of the PLC.

[0026] In a specific embodiment, the process of executing step S102 may specifically include the following steps: (1) Collect pressure data, flow data, displacement data, and force data from the sensor network to generate raw process parameter data streams; (2) Using a Butterworth filter to remove noise from the original process parameter data stream to obtain filtered process parameter data; (3) Classifying the filtered process parameter data according to pressure data, flow data, displacement data, and force data to form a classified data group; (4) Normalizing the pressure data, flow data, displacement data, and force data in the classified data group to obtain standardized process parameters; (5) Data association of standardized process parameters according to each lubrication control area is performed to construct a process parameter association matrix; (6) The process parameter correlation matrix is ​​reduced in dimension through principal component analysis to obtain a digital model of the process status.

[0027] Specifically, during the process of collecting process parameters by the sensor network, the pressure sensor collects the pressure signals of each oil circuit in real time, with a sampling frequency of 1000Hz and a measurement range of 0-20MPa; the flow sensor collects the flow state data of the lubricating oil, with a sampling frequency of 500Hz and a range of 0-15L / min; the displacement sensor tracks the deformation degree of the plate, with a sampling frequency of 1000Hz and a measurement range of 0-300mm; the force sensor measures the forming force during the drawing process, with a sampling frequency of 1000Hz and a range of 0-1000kN. The data collected by each sensor is converted into a digital signal through high-speed A / D conversion to form the original process parameter data stream. The Butterworth filter is used to remove noise from the original data, which has the maximum flat amplitude-frequency characteristic. The cutoff frequency of the filter is set according to the characteristics of each type of data: 200Hz for pressure data, 100Hz for flow data, 200Hz for displacement data, and 200Hz for force data. The order of the filter is set to 4, and the original data is filtered by calculating the response function. For pressure data, the sampled signal is filtered through a high-pass filter to remove the DC component, and then filtered through a low-pass filter to suppress high-frequency noise. Flow data mainly has power frequency interference, which is processed using a band-stop filter. Displacement data and force data need to retain the rapidly changing effective signal while removing high-frequency noise interference.

[0028] The filtered process parameter data are classified and sorted according to different types. The pressure data contains the real-time pressure value of each oil circuit, forming a pressure data matrix; the flow data records the flow changes of each oil circuit, forming a flow data matrix; the displacement data describes the deformation of each measuring point of the plate, and establishes a displacement data matrix; the force data characterizes the load changes during the forming process, and generates a force data matrix. The data points in each data matrix are aligned according to the timestamp to ensure the time sequence of the data. When normalizing the classified data, the maximum and minimum value normalization method is adopted. First, the maximum and minimum values ​​of each type of data are calculated, and then the data is mapped to the [0, 1] interval according to the normalization formula. The normalization of pressure data considers the design pressure upper limit of each oil circuit, the normalization of flow data is based on the maximum allowable flow, the normalization of displacement data refers to the maximum deformation, and the normalization of force data is based on the upper limit of the forming force. The normalized data eliminates the dimension effect, which is convenient for the comprehensive analysis of different types of data.

[0029] When the standardized process parameters are analyzed for association according to each lubrication control area, the corresponding relationship of the data is first determined. The pressure data, flow data, displacement data and force data in the same area are combined to form the eigenvector of the area. The eigenvectors of each area constitute the process parameter association matrix, the rows of the matrix represent different moments, and the columns represent different parameter types. The degree of association between parameters is evaluated by calculating the correlation coefficient, and the mapping relationship between parameters is established. When the principal component analysis is used to reduce the dimension of the process parameter association matrix, the covariance matrix of the data is first calculated, and then the eigenvalues ​​and eigenvectors are solved. The eigenvectors are sorted according to the size of the eigenvalues, and the principal components with a cumulative contribution rate of more than 85% are selected. The original high-dimensional data is projected into the principal component space through the principal component transformation to obtain a digital model of the process state after dimensionality reduction. This model retains the main features of the original data while reducing data redundancy.

[0030] For example, during the deep drawing process of a certain automobile chassis part, the sensor network collected 50,000 sets of raw data points within 10 seconds. These data include pressure data (sampling frequency 1000Hz), flow data (sampling frequency 500Hz) of 12 oil circuits, displacement data (sampling frequency 1000Hz) of 16 displacement measurement points, and force data (sampling frequency 1000Hz) of 4 force sensors. The raw data is processed by a 4th-order Butterworth filter, and the high-frequency noise amplitude of the oil pressure data is reduced from ±0.5MPa to ±0.05MPa, and the fluctuation of the flow data is reduced from ±5L / min to ±0.5L / min. The filtered data is grouped by type to form 5 data matrices. Through maximum and minimum value normalization, the pressure data (0-20MPa), flow data (0-15L / min), displacement data (0-200mm) and force data (0-1000kN) are mapped to a unified interval [0, 1]. A process parameter association matrix (50000×44 dimensions) was established, and a digital model of process status containing 7 principal components was obtained after dimensionality reduction using principal component analysis, with a cumulative contribution rate of 87%.

[0031] In a specific embodiment, the process of executing step S103 may specifically include the following steps: (1) Extract the pressure data, flow data and force data of each lubrication control area from the process state digital model as fuzzy control input variables; (2) Fuzzy processing is performed on the pressure data, flow data, and force data according to three levels: low, medium, and high to obtain a fuzzy set; (3) Defuzzification of the fuzzy set is performed by weighted average method to obtain the stress state value of each lubrication control area; (4) Perform a ratio calculation on the stress state value and the reference lubrication pressure value to form lubrication pressure deviation data; (5) Incrementally calculate the lubrication pressure deviation data to generate a lubrication pressure compensation value; (6) The lubrication pressure compensation value is superimposed on the reference lubrication pressure value to obtain the real-time lubrication demand parameters of each area.

[0032] Specifically, when extracting data from the digital model of the process state, for each lubrication control area, the pressure data, flow data and force data are organized into a feature array as the input variable of the fuzzy control. The pressure data contains the time series data of the real-time pressure value of each oil circuit, the flow data records the change process of the lubricating oil flow, and the force data represents the load state during the deep drawing process. These data are aligned according to the sampling time to form a data sequence. When the extracted data is fuzzy processed, the membership function of the three language variables of pressure, flow and force is established. The pressure data is divided into three levels according to the actual measurement value: low pressure (0-8MPa), medium pressure (6-14MPa), and high pressure (12-20MPa), and the fuzzy relationship is described by the triangular membership function; the flow data is divided into three levels: low flow (0-5L / min), medium flow (4-10L / min), and high flow (8-15L / min), and the trapezoidal membership function is used; the force data is divided into three levels: low load (0-200kN), medium load (150-750kN), and high load (700-1000kN), and the Gaussian membership function is used. The membership value of each data point at different levels is calculated by the membership function to form a fuzzy set.

[0033] The defuzzification operation of fuzzy sets adopts the weighted average method, and its calculation formula is: in: is the stress state value; is the i-th level membership value of pressure data; is the j-th level membership value of the traffic data; is the k-th level membership value of the force data; is the pressure weight coefficient; is the flow weight coefficient; is the force weight coefficient.

[0034] The expression for the ratio of the stress state value to the reference lubrication pressure value is: in: is the lubrication pressure deviation; is the stress state value of the rth region; is the reference lubrication pressure value of the rth area; is the stress compensation coefficient; is the regional deformation coefficient; is the pressure correction factor; is the global adjustment coefficient; n is the number of lubrication control areas.

[0035] When calculating the incremental lubrication pressure deviation data, the dynamic step method is used to calculate the pressure compensation increment. The calculation of the compensation value takes into account the size, rate of change and cumulative amount of the deviation, and the compensation step is dynamically adjusted according to different working conditions. The calculated lubrication pressure compensation value is superimposed on the reference lubrication pressure value to obtain the real-time lubrication demand parameters of each area.

[0036] For example, during the deep drawing process of the inner panel of a car door, the real-time data of the flange area is extracted from the digital model of the process state, including the oil circuit pressure of 10MPa, the lubricating oil flow of 8L / min, and the forming force of 650kN. These data are substituted into the fuzzy rule base for calculation, and the membership of the pressure data in the medium pressure level is 0.7 and the membership of the high pressure level is 0.3; the membership of the flow data in the medium flow level is 0.8 and the membership of the high flow level is 0.2; the membership of the force data in the medium load level is 0.6 and the membership of the high load level is 0.4. The stress state value of the flange area is calculated by the weighted average method, and the ratio operation is performed with the reference lubrication pressure value of 12MPa to form the lubrication pressure deviation data. The compensation value of 1.5MPa is calculated based on the deviation data, and the real-time lubrication demand pressure of 13.5MPa is obtained after superposition with the reference pressure. The entire calculation process reflects the logical association between the data and realizes the precise regulation of the lubrication parameters.

[0037] In a specific embodiment, the process of executing step S104 may specifically include the following steps: (1) Calculate the proportional servo valve opening of each lubrication control area based on the real-time lubrication demand parameters to obtain valve opening control data; (2) Convert valve opening control data into PWM control signal to generate valve pulsation frequency and duty cycle data; (3) Perform digital-to-analog conversion on valve pulsation frequency and duty cycle data, and output valve control voltage signal; (4) Collect the actual output pressure of the proportional servo valve in each lubrication control area to form pressure feedback data; (5) Calculate the deviation between the pressure feedback data and the real-time lubrication demand parameter to generate a pressure compensation signal; (6) Arrange the pressure compensation signal and pressure feedback data in time series to form a dynamic lubrication pressure control curve.

[0038] Specifically, the proportional servo valve opening calculation first uses the real-time lubrication demand parameters to establish a mathematical model of the valve opening. The valve opening calculation formula is: in: is the valve opening control quantity; is the mth lubrication demand component; is the pressure response coefficient; is the time decay factor; is the nth flow component; is the flow weight coefficient; is the pulsation angular frequency; M, N are the number of pressure and flow components; t is the time variable.

[0039] The pressure deviation calculation expression is: in: is the pressure compensation signal; is the i-th feedback pressure value; is the feedback gain coefficient; is the time weight factor; is the pressure feedback reference value; is the global compensation coefficient; I is the number of feedback data points.

[0040] In the actual sheet metal deep drawing lubrication control process, when calculating the proportional servo valve opening of each lubrication control area according to the real-time lubrication demand parameters, the corresponding relationship between valve opening and flow is established. The proportional servo valve opening range is 0-100%, and the corresponding flow range is 0-15L / min. The lubrication demand parameter is converted into the target flow value through the flow characteristic curve, and then the required opening value is calculated according to the valve characteristic curve. For the case of multiple oil circuits in parallel, it is necessary to consider the pressure balance and coordinate the opening of each oil circuit. When converting the valve opening control data into a PWM control signal, a PWM control method based on frequency modulation is adopted. The base frequency of the PWM signal is set to 1kHz, and the duty cycle range is 0-100%. The valve opening value is converted into a duty cycle through linear mapping, and an opening of 0% corresponds to a duty cycle of 0%, and an opening of 100% corresponds to a duty cycle of 100%. The frequency of the PWM signal can be adjusted according to the dynamic characteristics of the valve. A higher frequency is used for valves with faster response, and a lower frequency is used for valves with slower response.

[0041] The digital-to-analog conversion of valve pulsation frequency and duty cycle data uses a 16-bit DAC converter to convert digital quantities into 0-10V analog voltage signals. During the conversion process, voltage compensation is performed according to the control characteristics of the valve to ensure that the control signal is linearly related to the valve opening. The update frequency of the DAC is set to 10kHz to meet the requirements of rapid valve response. The valve control voltage signal drives the proportional servo valve coil after current amplification to achieve precise control of the valve opening. The pressure feedback data is collected through pressure sensors installed on each oil circuit. The pressure sensor has a range of 0-20MPa, a sampling frequency of 1kHz, and a resolution of 0.01MPa. The collected pressure signal is converted into a digital quantity through A / D conversion to form a pressure feedback data sequence. The pressure feedback data contains the real-time pressure value, acquisition time and channel identification of each oil circuit, which is used to monitor the changes in lubrication pressure in real time.

[0042] The deviation calculation between the pressure feedback data and the real-time lubrication demand parameter adopts a dynamic comparison method. The pressure feedback data is aligned according to the acquisition time, and the difference with the lubrication demand parameter is calculated. The difference data is subjected to sliding average filtering to eliminate the influence of instantaneous fluctuations. The required pressure compensation amount is calculated according to the size and change trend of the deviation. The generation of the pressure compensation signal takes into account the dynamic response characteristics of the pressure, and the proportional-integral-differential (PID) control algorithm is used for calculation. The pressure compensation signal is time-series associated with the pressure feedback data to establish a dynamic lubrication pressure control curve. The control curve records the change process of the pressure setting value, the actual pressure value and the compensation amount, reflecting the dynamic adjustment process of the lubrication pressure. The control curve is analyzed to extract key characteristic parameters as the basis for optimizing the lubrication control strategy.

[0043] For example: In the deep drawing of a certain automobile chassis part, the real-time lubrication demand pressure set in the flange area is 15MPa. According to the pressure value, the proportional servo valve opening is calculated to be 75%, and the corresponding PWM control signal is generated with a frequency of 1kHz and a duty cycle of 75%. After D / A conversion, a 7.5V control voltage is output. The actual pressure collected by the pressure sensor is 14MPa, which deviates from the set value by 1MPa. The compensation amount of 0.8MPa is calculated through PID control and superimposed on the original control signal. The data before and after pressure compensation form a dynamic control curve, which records the process of gradually adjusting the pressure from 14MPa to the target value of 15MPa. The entire control process reflects the working principle of closed-loop feedback and realizes the precise adjustment of lubrication pressure.

[0044] In a specific embodiment, the process of executing step S105 may specifically include the following steps: (1) The thickness of the plate is collected at each measuring point by an ultrasonic thickness sensor to obtain the plate thickness distribution data; (2) Calculate the difference between the plate thickness distribution data and the plate initial thickness to obtain the thinning data of each measuring point; (3) Normalizing the thinning data to form plate thinning rate data; (4) Collect the surface roughness measurement values ​​of the plate and generate surface quality data; (5) Perform weighted superposition processing on the sheet metal thinning rate data and the surface quality data to generate quality assessment parameters; (6) The dynamic lubrication pressure control curve is proportionally corrected using the quality assessment parameters to obtain the optimized lubrication parameters.

[0045] Specifically, when the ultrasonic thickness sensor collects the thickness of the plate, the measurement points are arranged according to the preset measurement grid. The measurement grid adopts polar coordinate distribution mode, and the measurement area is divided radially and circumferentially from the center of the plate to the outside. The radial direction is divided into 8 measurement circles, and 12 measurement points are arranged on each measurement circle, forming a total of 96 measurement points. The ultrasonic thickness sensor adopts a dual crystal probe with an operating frequency of 10MHz, a measurement accuracy of 0.01mm, and a measurement range of 0.4-5mm. During measurement, the probe is kept perpendicular to the surface of the plate, and the coupling agent is used to achieve good transmission of ultrasonic waves. Multiple sampling is performed on each measurement point to obtain the average value to obtain the actual thickness value of the point. The thickness data of all measurement points are stored according to the position coordinates to form the plate thickness distribution data. When the plate thickness distribution data is calculated with the initial thickness, the nominal initial thickness value of the plate is first read and used as the reference thickness. The actual thickness value of each measurement point is subtracted from the reference thickness to obtain the absolute thinning amount of the point. During the calculation process, the position information of the measurement point is recorded at the same time to establish the corresponding relationship between the thinning amount and the position. For points with abnormal measured values, interpolation correction is performed through the data of surrounding points to ensure the continuity of the thinned data.

[0046] When normalizing the thinning data, the maximum thinning amount is used as the normalization benchmark. The maximum and minimum thinning amounts of all measuring points are calculated, and the thinning amount of each measuring point is mapped to the interval [0, 1]. The normalization calculation takes into account the distribution characteristics of thinning, and a piecewise linear mapping is used for the area where the thinning is concentrated to improve the resolution of the data. The normalized data reflects the proportional relationship of each measuring point relative to the maximum thinning amount, forming the plate thinning rate data. The surface roughness of the plate is measured using a portable roughness meter with a measurement range of Ra 0.05-10μm, an evaluation length of 4mm, and a sampling length of 0.8mm. Representative areas are selected on the surface of the plate for measurement, including areas with large deformation and smaller areas. Multi-point sampling is performed in each measuring area to obtain the average roughness value of the area. After the roughness data is filtered to remove outliers, a surface quality data matrix is ​​formed according to the measurement area.

[0047] The weighted superposition of the plate thinning rate data and the surface quality data adopts the zoning weight method. Different weight coefficients are assigned to different areas according to the process requirements. The thinning rate weight of the area with large deformation is higher, and the roughness weight of the area with high surface quality requirements is higher. The comprehensive quality evaluation parameters of each area are obtained by weighted summation. The quality evaluation parameters comprehensively reflect the thinning state and surface quality of the plate.

[0048] The calculation formula for correcting the dynamic lubrication pressure control curve by the quality evaluation parameter is: in: is the optimized lubrication parameter; is the evaluation parameter of the kth region; is the regional hardening coefficient; is the strain rate factor; is the material yield coefficient; is the friction compensation factor; is the lth quality characteristic value; is the pressure correction factor; is the dynamic response frequency; K, L are the number of regions and the number of features; t is the time variable.

[0049] For example: During the deep drawing process of the front fender of a car, the workpiece is measured by ultrasonic thickness measurement. Eight measuring circles are arranged in the radial direction, with 12 points in each circle, for a total of 96 measuring points. The difference between the measured thickness data and the initial thickness of 2.0 mm is calculated to obtain the thinning amount at each point. The thinning rate distribution is obtained after normalization, and the surface roughness value is measured at the same time. According to the process requirements, a weight of 0.7 is assigned to the thinning rate of the fillet area, and a weight of 0.6 is assigned to the surface quality of the flange area, and the quality assessment parameters of each area are calculated. The quality assessment parameters are substituted into the correction formula, the original lubrication pressure curve is corrected, and the optimized lubrication parameters are obtained. The entire calculation process reflects the correlation between the sheet forming quality and the lubrication parameters.

[0050] In a specific embodiment, the process of executing step S106 may specifically include the following steps: (1) Extract pressure values, flow values, and time series data from the optimized lubrication parameters and group the data according to the lubrication control area; (2) The correlation between the pressure value and the flow value in the grouped data is calculated using the Pearson correlation coefficient to generate a parameter correlation matrix; (3) Perform eigenvalue decomposition on the parameter correlation matrix to obtain the lubrication parameter eigenvector; (4) Perform deviation analysis on the lubrication parameter characteristic vector and the reference lubrication pressure value of each lubrication control area to form a pressure correction coefficient; (5) The pressure correction coefficient is numerically fitted by the weighted average method to obtain the lubrication pressure control equation; (6) Based on the lubrication pressure control equation, the lubrication parameter optimization model is established using the least squares method; (7) The lubrication parameter optimization model and the corresponding process parameters are encoded and stored in the process parameter database.

[0051] Specifically, when extracting data from the optimized lubrication parameters, the pressure value, flow value and corresponding time series data of each lubrication control area are organized into a standard format. The pressure value contains the real-time pressure measurement results of each oil circuit, the flow value records the flow state of the lubricating oil, and the time series data marks the collection time of each group of data. The data grouping is carried out according to the flange area and the fillet area, and the data in each area maintains a time series correspondence. When calculating the correlation between the pressure value and the flow value by the Pearson correlation coefficient, their covariance and standard deviation are calculated for the data pairs in the same area. The value range of the Pearson correlation coefficient is [-1, 1], 1 represents a complete positive correlation, -1 represents a complete negative correlation, and 0 represents no correlation. The calculated correlation coefficient is organized into a matrix form, the rows and columns of the matrix correspond to different measurement parameters, and each element represents the degree of correlation between the corresponding parameters.

[0052] When performing eigenvalue decomposition on the parameter correlation matrix, the matrix is ​​represented as a combination of eigenvalues ​​and eigenvectors. The eigenvalues ​​are obtained by solving the characteristic equation, and the size of the eigenvalues ​​reflects the importance of the corresponding eigenvector. The eigenvector represents the main direction of data change and contains information about the regularity of parameter changes. The eigenvalues ​​are sorted from large to small, and the main eigenvectors are selected to form the lubrication parameter eigenvector set. When performing deviation analysis on the lubrication parameter eigenvector and the reference lubrication pressure value, the difference between the pressure distribution represented by the eigenvector and the reference pressure is calculated. The difference calculation takes into account the absolute deviation and relative deviation of the pressure value, and determines the allowable deviation range in combination with the process requirements. According to the deviation analysis results, a correction coefficient is assigned to each area, and the correction coefficient reflects the direction and amplitude of the pressure adjustment.

[0053] The numerical fitting of the pressure correction coefficient adopts the weighted average method to comprehensively calculate the correction coefficient of each area. The weight distribution is based on the regional importance and process sensitivity, and the fitting process uses piecewise functions to handle different working conditions. The fitted lubrication pressure control equation describes the mathematical relationship of pressure adjustment, and the equation contains the action terms of various influencing factors. When the least squares method is used to establish the lubrication parameter optimization model, the objective function and constraints are established based on the control equation. The objective function reflects the optimization goal of pressure adjustment, and the constraints ensure that the adjustment results meet the process requirements. The objective function is minimized by iterative calculation to obtain the optimized parameter value. The optimization model takes into account the synergistic effect of multiple lubrication areas to ensure the overall lubrication effect.

[0054] Finally, when the lubrication parameter optimization model is stored in the database, data encoding and format conversion are required. The encoding rules of process parameters include area identification, parameter type and timestamp information to ensure the uniqueness and traceability of data. In addition to storing the optimization model itself, the database also contains related process parameters and control strategies.

[0055] For example, in the processing of a certain automobile side cover drawing part, the pressure and flow values ​​of each area are extracted from the optimized lubrication parameter data. In the fillet area, the Pearson correlation coefficient of the pressure and flow values ​​is 0.85, indicating that the pressure and flow have a strong positive correlation. The eigenvector obtained after eigenvalue decomposition shows that the pressure change is mainly concentrated in the early stage of drawing. The eigenvector is compared with the reference pressure of 12MPa in the area, and the correction coefficient is calculated. The pressure control equation is established by the weighted average method, and the least squares method is used to optimize the final control parameters. The optimization results are encoded and stored in the database in the format of "area number_parameter type_timestamp". The whole process reflects the logic of data processing and the systematic nature of parameter optimization.

[0056] In a specific embodiment, the process of performing the step of performing deviation analysis on the lubrication parameter characteristic vector and the reference lubrication pressure value of each lubrication control area may specifically include the following steps: (1) The lubrication parameter feature vector is reduced in dimension through matrix operation to obtain reduced-dimensional feature data; (2) Using discrete Fourier transform to perform spectrum analysis on the reference lubrication pressure value to obtain frequency domain component data; (3) Numerical comparison of the reduced dimension feature data and the frequency domain component data to generate the difference coefficient; (4) Perform linear transformation on the reduced dimension feature data based on the difference coefficient to obtain the normalized feature value; (5) Compare the normalized characteristic value with the reference lubrication pressure value point by point and calculate the relative deviation value; (6) Construct a correction factor matrix based on the relative deviation value and extract the main components through singular value decomposition; (7) The main components are reconstructed through bilinear interpolation to form the pressure correction coefficient.

[0057] Specifically, when the eigenvector of the lubrication parameter is reduced in dimension, the eigenvector is organized into a matrix form, the rows of the matrix represent different sampling moments, and the columns represent different eigencomponents. The matrix is ​​reduced in dimension by the principal component analysis method, the covariance matrix of the eigenvector is calculated, the eigenvalue decomposition is performed on the covariance matrix, and the main components are selected by sorting the eigenvalues. The principal component analysis retains the main variation characteristics of the data while reducing the data dimension. The eigendata after dimensionality reduction retains the main information of the original data. When the discrete Fourier transform is used to perform spectrum analysis on the reference lubrication pressure value, the pressure data sequence in the time domain is converted to the frequency domain. The Fourier transform decomposes the time domain signal into a superposition of sinusoidal components of different frequencies, and the calculation is realized by the fast Fourier transform algorithm. The frequency domain analysis can reflect the periodic characteristics of the pressure change and identify the main frequency components. The frequency domain component data contains the amplitude spectrum and the phase spectrum, which records the frequency characteristics of the pressure change.

[0058] The numerical comparison between the reduced-dimensional feature data and the frequency domain component data uses the correlation analysis method to calculate the correlation between the two sets of data in different frequency bands. The correlation analysis takes into account the matching relationship between amplitude and phase, and generates a difference coefficient that reflects the degree of data difference. The difference coefficient quantifies the deviation between the feature data and the benchmark data in the frequency domain, providing a basis for subsequent data processing. When the reduced-dimensional feature data is linearly transformed based on the difference coefficient, a transformation matrix is ​​constructed to map the feature data to a new feature space. The transformation process takes into account the weight effect of the difference coefficient and performs weighted processing on data in different frequency bands. The linearly transformed data is normalized to obtain standardized eigenvalues. The normalized eigenvalues ​​eliminate the dimension effect and are easy to compare with the benchmark values.

[0059] When the normalized eigenvalue is compared point-to-point with the reference lubrication pressure value, the data at the corresponding moment is calculated for difference. The relative deviation value reflects the degree of change of the characteristic data relative to the reference value, and the calculation result forms a deviation sequence. The deviation sequence records the deviation change trend during the pressure regulation process and provides a basis for pressure correction. The correction factor matrix is ​​constructed based on the relative deviation value, and the deviation data is organized into a matrix form. The correction factor matrix is ​​decomposed by the singular value decomposition method to obtain the left singular matrix, singular value matrix and right singular matrix. The size of the singular value reflects the importance of the corresponding component, and the main singular value and its corresponding singular vector are selected for reconstruction. The data reconstruction of the main component adopts the bilinear interpolation method to perform interpolation calculations between discrete data points. Bilinear interpolation takes into account the data changes in both the horizontal and vertical directions, ensuring the smoothness of the reconstructed data. The reconstructed data forms a pressure correction coefficient, which is used to adjust the control parameters of the lubrication pressure.

[0060] For example, during the deep drawing process of a certain automobile roof, the lubrication parameter eigenvector was subjected to matrix dimensionality reduction. The original data contained 32 eigencomponents, and 8 main components were retained through principal component analysis. The benchmark lubrication pressure sequence was subjected to fast Fourier transform to obtain frequency domain component data, and the main frequency components were concentrated in the range of 0-10Hz. The reduced eigendata were compared with the frequency domain components to calculate the difference coefficient matrix. Linear transformation and normalization were performed based on the difference coefficient to obtain the standardized eigenvalue. After comparison with the benchmark pressure value, the relative deviation sequence was obtained, and the correction factor matrix was constructed and singular value decomposition was performed. The components corresponding to the first three main singular values ​​were selected for bilinear interpolation reconstruction to obtain the final pressure correction coefficient.

[0061] The above describes the automatic lubrication control method for deep drawing of sheet metal in the embodiment of the present application. The following describes the automatic lubrication control system for deep drawing of sheet metal in the embodiment of the present application. Figure 3 In the embodiment of the present application, an embodiment of the automatic lubrication control system for sheet metal deep drawing forming includes: A partitioning module 201 is used to partition the surface of the drawing die through the lubrication area partitioning module to obtain multiple independent lubrication control areas including the flange area and the fillet area, and initialize the parameters of each lubrication control area according to the physical parameters in the workpiece material library to obtain the reference lubrication pressure value of each area; The acquisition module 202 is used to collect the process parameters in real time according to the sensor network distributed in each lubrication control area, and perform filtering and standardization processing through the data preprocessing module to obtain a digital model of the process state; The calculation module 203 is used to analyze and calculate the stress state of each lubrication control area according to the process state digital model through a fuzzy control algorithm to obtain the real-time lubrication demand parameters of each area; The regulating module 204 is used to perform pulsation regulation on the lubrication pressure of each lubrication control area by using a proportional servo valve control system based on the real-time lubrication demand parameter to form a dynamic lubrication pressure control curve; A correction module 205 is used to perform online evaluation on the plate thinning rate and surface roughness, and to correct the dynamic lubrication pressure control curve according to the evaluation result to obtain optimized lubrication parameters; The analysis module 206 is used to perform correlation analysis on the optimized lubrication parameters, establish a lubrication parameter optimization model, and store the lubrication parameter optimization model in a process parameter database.

[0062] Through the cooperation of the above components, the surface of the drawing die is partitioned by the lubrication area division module, and the precise lubrication control of key parts such as the flange area and the fillet area is realized, avoiding the problem of uneven lubrication caused by the traditional single oil supply method. At the same time, the parameters of each lubrication control area are initialized according to the physical parameters in the workpiece material library, ensuring the scientificity and rationality of the reference lubrication pressure value. The sensor network arranged in each lubrication control area collects the process parameters in real time, and the data preprocessing module performs filtering and standardization processing to establish an accurate process state digital model, which provides a reliable data basis for the dynamic adjustment of lubrication parameters. The fuzzy control algorithm is used to analyze and calculate the stress state of each lubrication control area, and the real-time lubrication demand parameters of each area are accurately obtained, which reflects the intelligence and adaptability of the control strategy. The proportional servo valve control system is used to pulsate the lubrication pressure of each lubrication control area to form a dynamic lubrication pressure control curve, realizing the precise regulation of the lubrication pressure. By online evaluation of the sheet thinning rate and surface roughness, the dynamic lubrication pressure control curve is corrected according to the evaluation results, and the optimized lubrication parameters are obtained to ensure the stability of the forming quality. Finally, the correlation analysis of the optimized lubrication parameters was carried out, and the lubrication parameter optimization model was established and stored in the process parameter database, which achieved continuous optimization of control parameters and experience accumulation, and provided data support for lubrication control under different working conditions. The entire control process realizes full process automation from parameter collection, data processing to control execution, which significantly improves the processing quality and production efficiency of sheet metal deep drawing.

[0063] The present application also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein instructions are stored in the computer, and when the instructions are executed on a computer, the computer executes the steps of the automatic lubrication control method for sheet metal deep drawing.

[0064] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0065] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program codes.

[0066] As described above, 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 make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An automatic lubrication control method for sheet metal deep drawing, characterized in that: The automatic lubrication control method for sheet metal deep drawing comprises: The surface of the drawing die is partitioned through the lubrication area division module to obtain multiple independent lubrication control areas including the flange area and the fillet area. The parameters of each lubrication control area are initialized according to the physical parameters in the workpiece material library to obtain the reference lubrication pressure value of each area. The process parameters are collected in real time by the sensor network distributed in each lubrication control area, and filtered and standardized by the data preprocessing module to obtain a digital model of the process status; According to the process state digital model, the stress state of each lubrication control area is analyzed and calculated by a fuzzy control algorithm to obtain the real-time lubrication demand parameters of each area; Based on the real-time lubrication demand parameter, a proportional servo valve control system is used to pulsate-regulate the lubrication pressure of each lubrication control area to form a dynamic lubrication pressure control curve; The thinning rate and surface roughness of the plate are evaluated online, and the dynamic lubrication pressure control curve is corrected according to the evaluation results to obtain optimized lubrication parameters; A correlation analysis is performed on the optimized lubrication parameters, a lubrication parameter optimization model is established, and the lubrication parameter optimization model is stored in a process parameter database.

2. The automatic lubrication control method for sheet metal deep drawing according to claim 1, characterized in that: The surface of the drawing die is partitioned by the lubrication area partitioning module to obtain multiple independent lubrication control areas including the flange area and the fillet area, and the parameters of each lubrication control area are initialized according to the physical parameters in the workpiece material library to obtain the reference lubrication pressure value of each area, including: Performing regional recognition on the surface of the drawing die according to geometric features, and dividing the surface of the drawing die into a flange area and a fillet area; Divide the oil circuit network in each area according to lubrication requirements, and set independent pressure sensors and flow sensors for each oil circuit; Extract the friction coefficient, yield strength and hardening index of the plate to be processed from the workpiece material library to generate a material property data set; Calculating stress distribution in each region according to the geometric features of each region and the material property data set; According to the stress distribution, an initial pressure coefficient is allocated to each oil circuit, and a reference lubrication pressure value of each area is calculated; The reference lubrication pressure value is written into the control parameter table corresponding to each area through the PLC control unit.

3. The automatic lubrication control method for sheet metal deep drawing according to claim 1, characterized in that: The process parameters are collected in real time according to the sensor network distributed in each lubrication control area, and filtered and standardized by the data preprocessing module to obtain a digital model of the process state, including: Collecting pressure data, flow data, displacement data and force data from the sensor network to generate a raw process parameter data stream; Denoising the original process parameter data stream using a Butterworth filter to obtain filtered process parameter data; Classifying the filtered process parameter data according to pressure data, flow data, displacement data and force data to form a classified data group; Normalizing the pressure data, flow data, displacement data and force data in the classified data group respectively to obtain standardized process parameters; The standardized process parameters are data-associated according to each lubrication control area to construct a process parameter association matrix; The process parameter association matrix is ​​subjected to dimensionality reduction processing through principal component analysis to obtain a process state digital model.

4. The automatic lubrication control method for sheet metal deep drawing according to claim 1, characterized in that: According to the process state digital model, the stress state of each lubrication control area is analyzed and calculated by a fuzzy control algorithm to obtain the real-time lubrication demand parameters of each area, including: Extracting pressure data, flow data and force data of each lubrication control area from the process state digital model as fuzzy control input variables; The pressure data, flow data and force data are fuzzified according to three levels: low, medium and high to obtain a fuzzy set; Defuzzification operation is performed on the fuzzy set by weighted average method to obtain the stress state value of each lubrication control area; Performing a ratio operation on the stress state value and a reference lubrication pressure value to form lubrication pressure deviation data; performing incremental calculation on the lubrication pressure deviation data to generate a lubrication pressure compensation value; The lubrication pressure compensation value is superimposed on the reference lubrication pressure value to obtain the real-time lubrication demand parameter of each area.

5. The automatic lubrication control method for sheet metal deep drawing according to claim 1, characterized in that: Based on the real-time lubrication demand parameter, the lubrication pressure of each lubrication control area is pulsatingly adjusted by using a proportional servo valve control system to form a dynamic lubrication pressure control curve, including: Calculating the proportional servo valve opening of each lubrication control area according to the real-time lubrication demand parameter to obtain valve opening control data; Convert the valve opening control data into a PWM control signal to generate valve pulsation frequency and duty cycle data; Performing digital-to-analog conversion on the valve pulsation frequency and duty cycle data, and outputting a valve control voltage signal; Collect the actual output pressure of the proportional servo valve in each lubrication control area to form pressure feedback data; Calculate the deviation between the pressure feedback data and the real-time lubrication demand parameter to generate a pressure compensation signal; The pressure compensation signal and the pressure feedback data are arranged in time series to form a dynamic lubrication pressure control curve.

6. The automatic lubrication control method for sheet metal deep drawing according to claim 1, characterized in that: The online evaluation of the plate thinning rate and surface roughness and the correction of the dynamic lubrication pressure control curve according to the evaluation results to obtain the optimized lubrication parameters include: The thickness of each measuring point of the plate is collected by ultrasonic thickness measuring sensor to obtain the plate thickness distribution data; Calculate the difference between the plate thickness distribution data and the plate initial thickness to obtain the thinning data of each measuring point; Performing normalization operation on the thinning data to form plate thinning rate data; Collect the surface roughness measurement values ​​of the plate and generate surface quality data; Performing weighted superposition processing on the sheet metal thinning rate data and the surface quality data to generate quality assessment parameters; The dynamic lubrication pressure control curve is proportionally corrected according to the quality evaluation parameter to obtain optimized lubrication parameters.

7. The automatic lubrication control method for sheet metal deep drawing according to claim 1, characterized in that: The step of performing correlation analysis on the optimized lubrication parameters, establishing a lubrication parameter optimization model, and storing the lubrication parameter optimization model in a process parameter database includes: Extracting pressure values, flow values ​​and time series data from the optimized lubrication parameters, and grouping the data according to lubrication control areas; The correlation between the pressure value and the flow value in the grouped data is calculated by the Pearson correlation coefficient to generate a parameter correlation matrix; Performing eigenvalue decomposition on the parameter correlation matrix to obtain a lubrication parameter eigenvector; Performing deviation analysis on the lubrication parameter characteristic vector and the reference lubrication pressure value of each lubrication control area to form a pressure correction coefficient; The pressure correction coefficient is numerically fitted by a weighted average method to obtain a lubrication pressure control equation; Based on the lubrication pressure control equation, a lubrication parameter optimization model is established by using the least square method; The lubrication parameter optimization model and the corresponding process parameters are encoded and stored in a process parameter database.

8. The automatic lubrication control method for sheet metal deep drawing according to claim 7, characterized in that: The step of performing deviation analysis on the lubrication parameter characteristic vector and the reference lubrication pressure value of each lubrication control area to form a pressure correction coefficient includes: Performing dimension reduction processing on the lubrication parameter feature vector through matrix operation to obtain dimension reduction feature data; Performing spectrum analysis on the reference lubrication pressure value by using discrete Fourier transform to obtain frequency domain component data; Performing numerical comparison between the dimension reduction feature data and the frequency domain component data to generate a difference coefficient; Performing a linear transformation on the dimension-reduced feature data based on the difference coefficient to obtain a normalized feature value; Comparing the normalized characteristic value with the reference lubrication pressure value point-to-point to calculate the relative deviation value; Constructing a correction factor matrix according to the relative deviation value, and extracting main components by singular value decomposition; The data of the main components are reconstructed by bilinear interpolation to form a pressure correction coefficient.

9. An automatic lubrication control system for sheet metal deep drawing, used to implement the automatic lubrication control method for sheet metal deep drawing as claimed in any one of claims 1 to 8, characterized in that: The automatic lubrication control system for sheet metal deep drawing includes: The partitioning module is used to partition the surface of the drawing die through the lubrication area partitioning module to obtain multiple independent lubrication control areas including the flange area and the fillet area, and initialize the parameters of each lubrication control area according to the physical parameters in the workpiece material library to obtain the reference lubrication pressure value of each area; The acquisition module is used to collect process parameters in real time according to the sensor network distributed in each lubrication control area, and obtain a digital model of the process state by filtering and standardizing the data through the data preprocessing module; A calculation module, used to analyze and calculate the stress state of each lubrication control area according to the process state digital model through a fuzzy control algorithm to obtain the real-time lubrication demand parameters of each area; A regulating module, for pulsatingly regulating the lubrication pressure of each lubrication control area by using a proportional servo valve control system based on the real-time lubrication demand parameter to form a dynamic lubrication pressure control curve; A correction module, used for online evaluation of the plate thinning rate and surface roughness, and correcting the dynamic lubrication pressure control curve according to the evaluation results to obtain optimized lubrication parameters; The analysis module is used to perform correlation analysis on the optimized lubrication parameters, establish a lubrication parameter optimization model, and store the lubrication parameter optimization model in a process parameter database.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the automatic lubrication control method for sheet metal deep drawing as described in any one of claims 1 to 8 is implemented.

Images