CN116861741A

CN116861741A - Stamping data low-cost acquisition method and process energy spectrum construction method

Info

Publication number: CN116861741A
Application number: CN202310819959.7A
Authority: CN
Inventors: 黄海鸿; 甘雷; 李磊; 刘志峰
Original assignee: Hefei University of Technology
Current assignee: Hefei University of Technology
Priority date: 2023-07-05
Filing date: 2023-07-05
Publication date: 2023-10-10

Abstract

The invention belongs to the field of stamping forming production monitoring, and particularly relates to a stamping data low-cost acquisition method and a process energy map construction method. The stamping data low-cost acquisition method comprises the following steps: s1: and obtaining a plurality of discrete sample data in the stamping process through an actual stamping test, wherein the discrete sample data comprise the measured stamping force and the measured process energy. S2: a simulation model for simulating the stamping forming process is constructed, the simulation model comprising a stamping geometry model and a finite element model. S3: and generating simulation punching data including simulation process energy and simulation thickness change by using the simulation model. S4: calculating deviation between the simulated process energy and the measured process energy; and carrying out iterative correction on the simulation model to obtain a required numerical model. S5: and generating simulation stamping data required by the energy spectrum of the construction process by using a numerical model. The invention solves the problem that the processing quality of the stamping workpiece is difficult to be monitored on line accurately with low cost under the existing conditions.

Description

Stamping data low-cost acquisition method and process energy spectrum construction method

Technical Field

The invention belongs to the field of stamping forming production monitoring, and particularly relates to a stamping data low-cost acquisition method and system, corresponding data processing equipment, a process energy map construction method and a stamping workpiece processing quality on-line monitoring method.

Background

In the stamping process, how to monitor the quality of the process has been a problem facing those skilled in the art. Because of the closeness of the stamping die, stamping the workpiece is typically accomplished rapidly in a closed processing environment; it is technically difficult to achieve a direct measurement of the press working quality. In the prior art, the thickness change of a punched workpiece in the machining process is predicted based on the operation parameters or feedback signals of equipment, so that the machining quality of the punched workpiece is estimated.

On the basis, the invention provides a method for monitoring the processing quality of a stamping workpiece based on a process energy spectrum. The method mainly researches the relation between thickness variation, process energy and stamping depth in the process of stamping a workpiece, predicts the thickness variation of the stamping workpiece by combining the process energy and stamping depth variation of the stamping workpiece in the process of machining in the subsequent quality monitoring process, and judges whether the stamping workpiece is wrinkled or broken according to the maximum thickening rate and the maximum thinning rate of the stamping workpiece so as to obtain a machining quality evaluation result of the stamping workpiece.

The online monitoring method for the processing quality of the stamping workpiece based on the process energy spectrum has high monitoring precision, but has a disadvantage that the precision of the quality monitoring scheme is very dependent on the scale of sample data. The scheme needs to carry out a large number of actual tests on each type of stamping workpiece under the specified stamping process condition to obtain enough sample data, then creates a process energy map by using the actual sample data, and guides the actual production. Taking an example of a process energy spectrum of a door-like inner panel with a construction material of AA5052 aluminum alloy and a stamping depth of 28 mm. The technician needs to select stamping depths of 1mm to 28mm at intervals to determine the abscissa of the process energy map, and 16 groups of process conditions are set under each stamping depth, so that 448 inner plate pieces of the vehicle door are required to be processed, and process energy and thickness change data of 448 inner plate pieces of the vehicle door are acquired. In addition, the type of material or shape, structure and dimensions of the workpiece are changed, and the skilled person needs to repeat the above steps. These add significant cost to the construction of the process energy spectrum.

It can be seen from this: under the acquisition strategy of acquiring sample data through actual punching test, one punching process needs to be completed at each punching depth, and after the processing is completed, the punched workpiece needs to be cut to acquire thickness variation data required by the energy spectrum of the construction process, so that great consumption of human resources and material resources is caused, the application cost of the scheme is greatly increased, the test period is overlong, and the test cost is high. Based on this, a need exists for a way to obtain a large amount of reliable stamping data at low cost to avoid the inefficiency and waste of resources caused by the actual test.

Disclosure of Invention

In order to solve the problems of low efficiency and high cost of the traditional method for acquiring punching data through actual punching test, the invention provides a method for acquiring the punching data at low cost and a method for constructing a process energy map.

The invention is realized by adopting the following technical scheme:

the low-cost acquisition method of stamping data is used for acquiring process energy and thickness variation data with mapping relation required by building a process energy map at low cost; the thickness variation data includes a maximum thinning rate and an increased thickening rate. The low-cost acquisition method of the stamping data comprises the following steps:

s1: under the preset stamping process condition, a plurality of discrete sample data are obtained through an actual stamping test, wherein the sample data comprise the measured stamping force and the measured process energy.

S2: a simulation model for simulating the stamping forming process is constructed, the simulation model comprising a stamping geometry model and a finite element model. The stamping geometric model is used for reflecting the shape, structure and size of the stamping forming die and the plate. The finite element model is used for simulating the stamping forming of the plate material, so that the mapping relation between the thickness change of the stamped workpiece and the process energy is obtained.

S3: and generating a simulation maximum thinning rate, a simulation maximum thickening rate and simulation process energy by using the simulation model.

S4: calculating deviation between the simulated process energy and the measured process energy; the following decisions are then made:

A. when the deviation is larger than the correction threshold, actually measuring the model parameters which need to be manually input in the stamping geometric model, and inputting the actually measured value of the model parameters into the stamping geometric model to correct the stamping geometric model; and iteratively correcting the friction coefficient mu in the finite element model.

B. And when the deviation is not greater than the correction threshold value, saving the current simulation model as a numerical model.

S5: and generating process energy and thickness change data with a mapping relation required by building a process energy map by taking the numerical model as a tool.

As a further improvement of the present invention, in step S4, the correction strategy of the friction coefficient is as follows:

s01: initializing friction coefficient population mu (g) and mutation probability F _a Cross probability CR, maximum iteration number ger, individual learning factor b ₁ Population learning factor b ₂ And an inertial weight q.

S02: calculating the current coefficient of frictionEach individual μ of the original coefficient of friction in population μ (g) _i (g) Average absolute percentage error e of corresponding process energy _MA,i (g) And identifying the optimal friction coefficient individual mu with the smallest average absolute percentage error _m (g)。

S03: the original friction coefficient individuals mu are updated based on the following friction coefficient populations mu (g) _i (g) Evolution direction v of (v) _i (g+1)：

In the above, r ₁ And r ₂ A random value between 0 and 1; e is an identity matrix; e, e _MA,m (g) Average absolute percentage error for the individual optimal coefficient of friction; mu (mu) _i (g)、μ _q (g) And mu _r (g) Three original friction coefficient individuals randomly selected from the friction coefficient population mu (g).

S04: according to the direction of evolution v _i (g+1) selecting the original friction coefficient individuals μ in the current friction population μ (g) _i (g) Performing mutation to obtain mutation friction coefficient individual mu _i ' (g), the formula for the mutation operation is as follows:

in the above, F _a (g+1) is the mutation probability of the g+1st iteration.

S05: individuals mu with mutation friction coefficients in current friction population mu (g) _i ' (g) individual μ with original coefficient of friction _i (g) Performing a crossover operation to generate a new individual friction coefficient, the crossover operation strategy being as follows:

in the above, rand _i Is the original coefficient of friction, individual mu _i (g) Individual μ with abrupt coefficient of friction _i ' (g) random numbers between 0 and 1 generated when the interleaving operation is performed.

S06: comparison of the coefficient of abrupt Friction individuals mu _i ' (g) individual μ with original coefficient of friction _i (g) The average absolute percentage error of the process energy of (a) is used for updating the friction coefficient individuals to obtain a friction coefficient population mu (g+1) when the friction coefficient individuals are subjected to g+1 iterations:

in the above, e _MA,i ' (g) is the mutation coefficient of friction individual μ _i ' the corresponding process energy average absolute percent error of (g).

S07: judging whether the current iteration number g is smaller than the maximum iteration number ger or not:

if so, steps S02 to S06 are executed in a loop.

Otherwise, ending the iteration, and outputting the optimal friction coefficient individual mu with the minimum process energy average absolute percentage error in the current friction coefficient population mu (g) _m (g) As an updated coefficient of friction.

The invention also comprises a construction method of the process energy spectrum, which comprises the following steps:

1. under the preset stamping process condition, a plurality of discrete sample data are obtained through an actual stamping test.

2. By adopting the method for acquiring the stamping data at low cost, a numerical model is built by combining the acquired sample data.

The numerical model is used for generating an array composed of a maximum thinning rate and a maximum thickening rate with mapping relations, a stamping depth and stamping force.

3. At a stamping depth h _p And drawing blank process energy maps for the abscissa and the ordinate respectively, and determining a region to be filled with a boundary in the process energy maps according to the constraint of the process on the parameters.

4. And (3) carrying out data filling on the region to be filled in the blank process energy map by using a numerical model, wherein the process is as follows:

4.1: generating simulated stamping depth h from correlation using numerical model _ps Simulated punching force F _s Simulation of maximum thickening ratio delta _MTCs And simulate the maximum thinning rate delta _MTNs First array U1: u1 = { h _ps ，F _s ，Δ _MTCs ，Δ _MTNs }。

4.2: according to the simulation stamping depth h _ps And simulation punching force F _s Calculating corresponding simulation process energy E _s ：

In the above, F _s (x) To simulate the punching force F _s Concerning the simulated stamping depth h _ps Is a fitting function of (a);

4.3: converting the first array U1 into a second array U2: u2= { h _ps ，E _s ，Δ _MTCs ，Δ _MTNs }。

4.4: with simulated stamping depth h in the second array _ps And simulation Process energy E _s Respectively used as the abscissa and the ordinate of the pixel points in the region to be filled to simulate the maximum thickening rate delta _MTCs And simulating the maximum thinning rate delta _MTNs The first attribute value and the second attribute value are respectively corresponding to the pixel points; and (5) completing pixel filling of the blank process energy spectrum.

5. Mirroring the image filled in the steps into two symmetrical parts by taking the longitudinal axis as a symmetry axis; the pucker identification zone and the fracture identification zone, respectively.

6. Coloring each pixel point in the wrinkling recognition area according to a preset color mapping relation and coloring each pixel point in the cracking recognition area according to a second attribute value.

7. And dividing the boundary between the wrinkling region and the safety region in the wrinkling recognition region and the boundary between the cracking region and the safety region in the cracking recognition region according to a preset safety threshold value to obtain a required process energy map.

The invention also comprises an online monitoring method for the processing quality of the stamping workpiece, which comprises the following steps:

step 1: and obtaining a process energy spectrum generated by the construction method of the process energy spectrum of the current stamping workpiece to be processed.

Step 2: collecting real-time stamping depth h of stamping workpiece to be processed in real time in actual processing process _real And a real-time punching force F _real 。

Step 3: according to the real-time stamping depth h in the processing process _real And a real-time punching force F _real Calculating the real-time process energy E _real 。

Step 4: according to the real-time stamping depth h of the stamping workpiece to be processed in the processing process _real And real-time process energy E _real Drawing corresponding state tracks in the wrinkling recognition area and the cracking recognition area of the process energy map.

Step 5: evaluating the processing quality of the processed stamping workpiece according to the state track:

(1) When any point in the state track passes through the rupture zone, judging that the machined stamping workpiece is ruptured.

(2) When the end point of the state track is positioned in the wrinkling zone, judging that the processed stamping workpiece has local wrinkling.

The invention also comprises a punching data low-cost acquisition system which adopts the punching data low-cost acquisition method and combines the measured process energy and the simulation model to obtain a numerical model for acquiring the punching data at low cost. The punching data low-cost acquisition system comprises: the system comprises a measurement process energy acquisition module, a simulation model construction module, an error calculation module, a friction coefficient updating module and a numerical model correction module.

The measuring process energy acquisition module is used for calculating measuring process energy according to a plurality of discrete sample data acquired by an actual punching test.

The simulation model construction module comprises a stamping geometric model and a finite element model; the stamping geometric model is used for reflecting the shape, structure and size of the stamping forming die and the plate. The finite element model is used for simulating the stamping deformation of the plate material, so that the mapping relation between the thickness change of the stamping workpiece and the process energy is obtained. The simulation model is used for generating a simulation maximum thinning rate, a simulation maximum thickening rate and simulation process energy.

The error calculation module is used for calculating deviation between the simulation process energy generated by the constructed simulation model and the measured process energy.

The friction coefficient updating module is used for carrying out iterative correction on the friction coefficient in the finite element model by adopting the correction strategy when the deviation calculated by the error calculating module exceeds the correction threshold value.

The numerical model correction module is used for resetting manually input parameters in the stamping geometric model and updating friction coefficients in the finite element model when the deviation calculated by the error calculation module exceeds a correction threshold value; and when the deviation calculated by the error calculation module does not exceed the correction threshold value, preserving parameters in the corresponding simulation model to obtain a numerical model for generating a mapping relation between the maximum thinning rate, the maximum thickening rate and the process energy of the stamping workpiece.

The invention also includes a data processing apparatus comprising a memory, a processor and a computer program stored on the memory and executable on the processor, when executing the computer program, creating a ram data low cost acquisition system as described above; and then automatically creating a numerical model capable of generating a mapping relation between the maximum thinning rate, the maximum thickening rate and the process energy of the stamping workpiece under the specified process conditions according to a plurality of pieces of discrete sample data in the collected stamping test process.

The technical scheme provided by the invention has the following beneficial effects:

according to the low-cost acquisition method for the stamping data, a large amount of process energy and thickness change data are acquired by constructing a high-precision stamping forming simulation model, and the method is further used for constructing a process energy map with low cost. The processing quality of the stamping workpiece can be monitored on line by using the process energy spectrum.

In the process of constructing and correcting the simulation model, in order to ensure the accuracy and the reliability of the simulation model, the invention designs a new model iteration correction strategy and an improved particle swarm differential algorithm (PSO-DE) for carrying out iteration optimization on friction coefficients in a finite element model. According to the invention, the simulation process energy of the stamping forming simulation model is dynamic response, the friction coefficient is used as an optimization parameter, and the friction coefficient is continuously optimized through an improved algorithm, so that the simulation process energy generated by the constructed simulation model is consistent with the measured process energy obtained by an actual stamping test, and the accuracy of the finally generated process energy and thickness variation data is ensured.

The numerical model is obtained in the low-cost acquisition method of the stamping data, and can replace the measured process energy and measured thickness change data obtained by actual stamping test in the construction process of the energy map of the traditional process, so that the high data acquisition cost generated by actual measurement is greatly reduced under the condition of ensuring similar data precision, and the practical value of the constructed process energy map is improved.

Drawings

Fig. 1 is a flowchart of steps of a low-cost acquisition method of punching data provided in embodiment 1 of the present invention.

FIG. 2 is a flowchart showing the steps for obtaining energy of the measuring process in embodiment 1 of the present invention.

FIG. 3 is a flow chart of the simulation model creation and application process in embodiment 1 of the present invention.

Fig. 4 is a schematic diagram of the accuracy evaluation process of the press forming simulation model in embodiment 1 of the present invention.

Fig. 5 is a flowchart showing the steps of the iterative correction process of friction coefficient in the finite element model according to embodiment 1 of the present invention.

Fig. 6 is a block diagram of a punching data low-cost acquiring system provided in embodiment 2 of the present invention.

Fig. 7 is a flow chart of steps of a method for constructing a process energy spectrum according to embodiment 4 of the present invention.

FIG. 8 is a schematic diagram of the process energy spectrum construction method in embodiment 4 of the present invention.

Fig. 9 is a flowchart of the steps of a method for online monitoring the processing quality of a punched workpiece according to embodiment 5 of the present invention.

Fig. 10 is a graph of simulated process energy and error corresponding to 4 different combinations of stamping control parameters (CB 1, CB2, CB3, CB 4) in a performance test.

FIG. 11 is a graph showing the accuracy versus operating speed of the improved PSO-DE algorithm of the present invention versus a conventional PSO algorithm in finding an individual with an optimal coefficient of friction.

In fig. 11, the part (a) is the average absolute percentage error of the process energy corresponding to the optimal friction coefficient individual obtained by the two algorithms, and the part (b) in fig. 11 is the convergence curve of the two algorithms.

Fig. 12 shows simulated thickness variations and errors corresponding to 4 different press control parameter combinations (CB 1, CB2, CB3, CB 4) in the performance test.

FIG. 13 is a comparison of process energy maps constructed in three different ways.

Wherein MAP1 is a process energy MAP constructed using a large amount of measured stamping data; MAP2 is a process energy MAP constructed by using the stamping data generated by the numerical model in the invention; MAP3 is a process energy MAP constructed using the punching data generated by the unmodified initial simulation model.

FIG. 14 is a spatial distribution diagram of the thickness variation monitoring results of different process energy MAPs in MAP 1-MAP 3.

FIG. 15 is a comparison of the monitoring accuracy of different process energy MAPs in MAP 1-MAP 3.

Wherein part (a) corresponds to the average absolute percentage error of different monitoring results; (b) And part of defects corresponding to different monitoring results are identified with accuracy.

Fig. 16 is a process energy map visualization path of a stamped workpiece thickness variation constructed using the present invention.

Fig. 17 is a graph showing the comparison of the processing quality of the processed real punched workpiece under the different processing paths corresponding to fig. 16.

Detailed Description

The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.

Example 1

The embodiment provides a low-cost acquisition method of punching data, which is used for acquiring process energy and thickness change data with mapping relation required by building a process energy map at low cost; the thickness variation data includes a maximum thinning rate and an increased thickening rate. As shown in fig. 1, the low-cost acquisition method of the punching data comprises the following steps:

under the preset stamping process condition, a plurality of discrete sample data are obtained through an actual stamping test, wherein the sample data comprise the measured stamping force and the measured process energy.

As shown in fig. 2, the collection method of each set of sample data is as follows:

A. l stamping depths are set.

B. And measuring the stamping force of the stamping workpiece corresponding to the L stamping depths through an actual stamping test.

C. Calculating the process energy corresponding to the L stamping depths according to the L stamping depths and the stamping forces corresponding to the L stamping depths:

In the above, E _l ' is the measured process energy at a stamping depth of l; f (x) is the stamping force F with respect to the stamping depth h _p Is a function of the fitting of (a).

A simulation model for simulating the stamping forming process is constructed, the simulation model comprising a stamping geometry model and a finite element model. The stamping geometric model is used for reflecting the shape, structure and size of the stamping forming die and the plate. The finite element model is used for simulating the stamping forming of the plate material, so that the mapping relation between the thickness change of the stamped workpiece and the process energy is obtained.

As shown in fig. 3, the stamping geometric model in this embodiment is constructed by any three-dimensional geometric modeling software including SolidWorks, 3DMax, unity, and the like. The parameters for constructing the stamping geometric model mainly comprise structural parameters of a die such as structural parameters of a male die, a female die and a blank holder, structural parameters of a plate material and the like. The finite element model is built by adopting any finite element analysis software including Dynaform, abaqus and the like. In order to drive the finite element model to operate, stamping control parameters, performance parameters of materials, friction coefficients and the like need to be input into the finite element model, and the parameters need to be reasonably designed by combining with an actual stamping workpiece, and are not described herein.

And generating a simulation maximum thinning rate, a simulation maximum thickening rate and simulation process energy by using the simulation model.

In the solution provided in this embodiment, the process energy is actually the ordinate of the process energy spectrum, and its accuracy directly determines the monitoring range of the spectrum. The simulation model provided in this embodiment essentially needs to replace the actual punching test to generate the relevant punching data, so the simulation process energy and the simulation thickness variation obtained by the simulation model should be consistent with the measured process energy and the measured thickness variation obtained in the actual punching test. In the subsequent work of the embodiment, the main content is to correct the simulation model through iterative correction, so that the simulation process energy and the simulation thickness change generated by the simulation model and the measurement process energy and the measurement thickness change are more converged.

As shown in fig. 4, the accuracy of the stamping forming simulation model is evaluated, and further, the friction coefficient mu in the finite element model is subjected to iterative correction, and in each iteration process, the deviation between the simulation process energy and the measurement process energy is calculated first; the following decisions are then made:

A. when the deviation is larger than the correction threshold, actually measuring the model parameters which need to be manually input in the stamping geometric model, and inputting the actually measured value of the model parameters into the stamping geometric model to correct the stamping geometric model; and correcting the finite element model by using the iterated friction coefficient.

In the scheme provided in this embodiment, the deviation between the simulated process energy and the measured process energy adopts the average absolute percentage error e _MA The specific calculation formula is as follows:

wherein E is _p,i ' is the measured process energy obtained by the actual stamping test when the stamping depth is i; e (E) _p,i Simulation process energy generated by the simulation model when the stamping depth is i is respectively; and N is the number of groups of data of the measured process energy obtained by the actual punching test. In this embodiment, the preset correction threshold is 10%.

After the deviation between the simulation process energy and the measured process energy is calculated, if the deviation is larger than a set correction threshold value, the stamping simulation model needs to be corrected. Because the deviation between the simulation model of the stamping forming and the measured data obtained by the actual stamping test mainly comes from the inconsistency between the modeling parameters and the actual stamping using parameters, the correction of the simulation model of the stamping forming mainly corrects the modeling parameters.

The present embodiment divides correction of the simulation model into two parts: firstly, the stamping geometric model is corrected, and secondly, the finite element model is corrected.

The structural parameters of the stamping geometric model can be ensured to be consistent with the actual parameters according to accurate measurement of the die, the plate material and the like. For control parameters such as stamping speed, blank holder force and the like in the driving parameters of the finite element model, as the control parameters are active input parameters, the input values of the simulation model are always ensured to be consistent with the actual input values, however, fluctuation of the control parameters is unavoidable in the actual forming process, and deviation of a simulation result and a measurement result is caused. In order to avoid misjudgment, a plurality of groups of stamping speed and edge pressing force combinations can be set, and the simulation process energy and the measurement process energy under different control parameter combinations can be compared to eliminate deviation caused by accidental fluctuation. For the material performance parameters of the plates, the material performance of each plate is possibly different, and even if the material performance parameters in a certain stamping forming process are corrected, the deviation from the actual situation can still exist in the next forming process. Therefore, the material performance parameters of the plate materials are preset.

In the finite element model correction process, the technical staff of the embodiment research finds that: for friction coefficients between the die and the plate, although the constructed process energy spectrum can effectively monitor thickness variation of the stamping workpiece under different friction coefficients, the variation of the friction coefficient still affects the quantitative relation between the process energy and the thickness variation, so that the monitoring precision of the constructed process energy spectrum is reduced. Therefore, correction of the friction coefficient in the stamped finite element model is a necessary measure to ensure the data accuracy.

In this embodiment, if a fixed friction coefficient is set only empirically, there is a tendency that there is a deviation from the actual friction coefficient. In addition, because the mold is inevitably worn during operation, human experience is often difficult to take into account the variation in coefficient of friction caused by this factor. Therefore, the present embodiment assists in performing iterative optimization of the friction coefficient by a modified particle swarm differential algorithm (PSO-DE). In the iterative process, the present embodiment uses the process energy of the stamping process as a dynamic response and the friction coefficient as an optimization parameter. As shown in fig. 5, the iterative correction process for the friction coefficient in the finite element model correction process of the present embodiment is as follows:

Wherein the friction coefficient population mu (g) is:

μ(g)＝(μ ₁ (g),μ ₂ (g),…,μ _i (g),…,μ _τ (g))

in the above formula, g is the iteration number; when g=1, the iteration is in an initial state; τ is the number of individuals with the original coefficient of friction in the coefficient of friction population μ (g).

The ith coefficient of friction individual μ in coefficient of friction population μ (g) _i (g) The method comprises the following steps:

μ _i (g)＝(μ _1,i (g),μ _2,i (g))

in the above, mu _1,i (g) Individual μ for the i-th raw coefficient of friction _i (g) Friction coefficient between the middle plate and the blank holder; mu (mu) _2,i (g) Individual μ for the i-th raw coefficient of friction _i (g) Friction coefficient between the middle plate and the female die.

In this example, the i-th original coefficient of friction, individual μ _i (g) Friction coefficient mu between middle plate and blank holder _1,i (g) And the i th original coefficient of friction individual μ _i (g) Friction coefficient mu between middle plate and concave mould _2,i (g) At a preset maximum friction coefficient mu _max With minimum friction coefficient mu _min The formula is as follows:

s02: calculating individual mu of each original friction coefficient in the current friction coefficient population mu (g) _i (g) Average absolute percentage error e of corresponding process energy _MA,i (g) And identifying the optimal friction coefficient individual mu with the smallest average absolute percentage error _m (g)。

In this example, individuals with the original coefficient of friction, μ, in the coefficient of friction population, μ (g) _i (g) Average absolute percentage error e of corresponding process energy _MA,i (g) The calculation formula of (2) is as follows:

in the above, E _l (μ _i (g) If the stamping depth is l, the stamping simulation model adopts the original friction coefficient individual mu _i (g) As process energy at the friction coefficient; h is a _p,max Is the total stamping depth.

in the above, F _a (g+1) is the mutation probability of the g+1st iteration.

if so, steps S02 to S06 are executed in a loop.

Otherwise, ending the iteration, and outputting the optimal friction coefficient individual mu with the minimum process energy average absolute percentage error in the current friction coefficient population mu (g) _m (g) As a coefficient of friction after correction.

And finally, when the maximum iteration times are reached or the deviation between the simulation process energy obtained by the corrected simulation model and the actual measured process energy is smaller than a preset correction threshold value, storing the simulation model as a required numerical model. And then, using the numerical model as a tool to generate process energy and thickness change data with a mapping relation required by building a process energy map.

Example 2

On the basis of the scheme in embodiment 1, this embodiment further provides a punching data low-cost acquisition system, as shown in fig. 6, which adopts the punching data low-cost acquisition method as in embodiment 1, and combines the measured process energy and the simulation model to obtain a numerical model for acquiring punching data at low cost. The punching data low-cost acquisition system in this embodiment includes: the system comprises a measurement process energy acquisition module, a simulation model construction module, an error calculation module, a friction coefficient updating module and a numerical model correction module.

The simulation model construction module comprises a stamping geometric model and a finite element model; the stamping geometric model is used for reflecting the shape, structure and size of the stamping forming die and the plate. The finite element model is used for simulating the stamping forming of the plate material, so that the mapping relation between the thickness change of the stamped workpiece and the process energy is obtained. The simulation model is used for generating a simulation maximum thinning rate, a simulation maximum thickening rate and simulation process energy.

The friction coefficient updating module is used for carrying out iterative correction on the friction coefficient in the finite element model by adopting the correction strategy as in the embodiment 1 when the deviation calculated by the error calculating module exceeds the correction threshold value.

The numerical model correction module is used for resetting manually input parameters in the stamping geometric model and correcting friction coefficients in the finite element model when the deviation calculated by the error calculation module exceeds a correction threshold value; and when the deviation calculated by the error calculation module does not exceed the correction threshold value, preserving parameters in the corresponding simulation model to obtain a numerical model for generating a mapping relation between the maximum thinning rate, the maximum thickening rate and the process energy of the stamping workpiece.

Example 3

The present embodiment provides a data processing apparatus including a memory, a processor, and a computer program stored on the memory and executable on the processor, when executing the computer program, creating a punching data low-cost acquisition system as in embodiment 2; and then automatically constructing a numerical model capable of generating a mapping relation between the maximum thinning rate, the maximum thickening rate and the process energy of the stamping workpiece under the specified process conditions according to a plurality of discrete sample data acquired by the acquired actual stamping test. The numerical model may be used to populate data in a process energy spectrum during the construction of the spectrum.

The data processing apparatus provided in this embodiment is essentially a computer apparatus, which may be an intelligent terminal, a tablet computer, a notebook computer, a desktop computer, a rack-mounted server, a blade server, a tower server, or a rack-mounted server (including an independent server, or a server cluster formed by a plurality of servers) capable of executing a program, or the like.

The computer device of the present embodiment includes at least, but is not limited to: a memory, a processor, and the like, which may be communicatively coupled to each other via a system bus.

In this embodiment, the memory (i.e., readable storage medium) includes flash memory, hard disk, multimedia card, card memory (e.g., SD or DX memory, etc.), random Access Memory (RAM), static Random Access Memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory may be an internal storage unit of a computer device, such as a hard disk or memory of the computer device. In other embodiments, the memory may also be an external storage device of a computer device, such as a plug-in hard disk, smart Media Card (SMC), secure Digital (SD) Card, flash Card (Flash Card) or the like, which are provided on the computer device. Of course, the memory may also include both internal storage units of the computer device and external storage devices. In this embodiment, the memory is typically used to store an operating system and various application software installed on the computer device. In addition, the memory can be used to temporarily store various types of data that have been output or are to be output.

The processor may be a central processing unit (Central Processing Unit, CPU), controller, microcontroller, microprocessor, or other data processing chip in some embodiments. The processor is typically used to control the overall operation of the computer device. In this embodiment, the processor is configured to execute the program code stored in the memory or process the data.

Example 4

On the basis of the scheme of embodiment 1, this embodiment further provides a method for constructing a process energy spectrum, as shown in fig. 7 and 8, which includes the following steps:

2. A numerical model was built using the low cost acquisition method of punching data as in example 1, in combination with the acquired sample data.

3. At a stamping depth h _p And drawing blank process energy maps for the abscissa and the ordinate respectively, and determining a region to be filled with a boundary in the process energy maps according to the constraint of the process on the parameters. The results obtained are shown in FIG. 8 (a).

/>

In the above, F _s (x) To simulate the punching force F _s Concerning the simulated stamping depth h _ps Is a function of the fitting of (a).

4.3: simulating punching force F in first array _s Alternative to simulation Process energy E _s And then to a second plurality U2: u2= { h _ps ，E _s ，Δ _MTCs ，Δ _MTNs }。

4.4: with simulated stamping depths in a second arrayDegree h _ps And simulation Process energy E _s Respectively used as the abscissa and the ordinate of the pixel points in the region to be filled to simulate the maximum thickening rate delta _MTCs And simulating the maximum thinning rate delta _MTNs The first attribute value and the second attribute value are respectively corresponding to the pixel points; and (5) completing pixel filling of the blank process energy spectrum.

5. Mirroring the map filled in the steps into two symmetrical parts by taking the ordinate as a symmetry axis; the pucker identification zone and the fracture identification zone, respectively. The results obtained are shown in part (b) of fig. 8.

6. Coloring each pixel point in the wrinkling recognition area according to a preset color mapping relation and coloring each pixel point in the cracking recognition area according to a second attribute value. The results obtained are shown in part (c) of fig. 8.

Example 5

The embodiment provides an online monitoring method for processing quality of a stamping workpiece, as shown in fig. 9, which comprises the following steps:

step 1: a process energy map of the currently punched workpiece to be processed, which is generated using the process energy map construction method as in example 4, is obtained.

Step 2: collecting real-time stamping depth h of stamping workpiece to be processed in actual processing process in real time _real And a real-time punching force F _real 。

Performance testing

In order to verify the effectiveness of the stamping data low-cost acquisition method provided by the embodiment in the practical application process, the construction cost of the process energy spectrum is reduced. Taking an inner plate of a similar car door as an example, AA5052 aluminum alloy is selected as a material of the inner plate, and a test experiment of an energy spectrum of a construction process is established. Test experiments the profile of the process energy profile was determined with the stamping process conditions combined with 16 sets of stamping speed and edge force. And (3) selecting stamping depths of 1mm to 28mm at intervals to determine the abscissa of the process energy spectrum. And then filling the data required in the process energy map by using the numerical model obtained by the punching data low-cost acquisition method provided by the embodiment.

1. Control parameter setting for simulation model

In the experiment, the initial value of the friction coefficient between the die and the plate is set to 0.14 in the stamping forming simulation model, and the measured process energy and thickness change data are measured data obtained by an actual stamping test under the lubrication of 16ml of lubricating oil.

In the embodiment, 10% of the threshold value is selected as a correction threshold value of the stamping forming simulation model, so that inaccurate process energy spectrum constructed based on the simulation model due to overlarge threshold value selection is avoided, and meanwhile, waste of manpower and material resources caused by frequent correction due to overlarge threshold value selection is prevented. In order to avoid misjudgment on the accuracy evaluation of the stamping forming simulation model caused by fluctuation of control parameters, 4 different control parameter combinations (namely CB1 (v=2 mm/s, F) are also selected in the experiment _h ＝6kN)、CB2(v＝10mm/s,F _h ＝18kN)、CB3(v＝22mm/s,F _h =36 kN) and CB4 (v=30 mm/s, F _h The measured process energy=48 kN) was compared with the simulated process energy obtained by the initial press forming simulation model, and the result and deviation thereof are shown in fig. 10. In the control parameter combination, v represents the punching speed, F _h Representing the binding force. In FIG. 10 e _{After MA-correction} The average absolute percentage error of the simulation process energy representing the corrected press forming simulation model can be calculated as:

in the above, E _l ' is the measured process energy at a stamping depth of l; e (E) _{After l-correction} When the stamping depth is l, the simulation process energy obtained by the corrected simulation model is obtained; h is a _p,max For the total stamping depth, here h _p,max =28 mm; l is the number of sets of measured process energy data. e, e _{MA-before correction} The average absolute percentage error of the simulation process energy representing the initial press forming simulation model before correction can be calculated as:

in the above, E _{l-before correction} And when the stamping depth is l, the simulation process energy obtained by the initial simulation model before correction is obtained.

As can be seen from the data in fig. 10: the simulation process energy of the initial stamping forming simulation model under different control parameter combinations is larger than the measurement process energy, and larger deviation exists. Since the measured process energy was measured under lubrication conditions with a lubricant amount of 16ml, and the friction coefficient set in the initial press forming simulation model was 0.14, the actual corresponding lubrication conditions should be around 8 ml. Analysis shows that the worse the lubrication conditions, the more energy input is required to overcome friction and therefore the more energy is in the process. Therefore, the simulation process energy of the initial press forming simulation model may be greater than the measured process energy. And under each control parameter combination, the average absolute percentage error of the simulation process energy of the initial stamping forming simulation model before correction exceeds 10%, namely the set correction threshold value, so that the initial stamping forming simulation model needs to be corrected.

2. Correction of simulation model

Because the simulation process energy deviation of the initial stamping simulation model under the control parameter combination CB1 is the largest, the experiment continues to correct the friction coefficient of the stamping simulation model by taking the measured process energy under the CB1 as a reference. The friction coefficient correction of the initial stamping forming simulation model adopting the PSO-DE iterative algorithm is used as an experimental group of the scheme of the embodiment, and the parameter setting of the experimental group is shown in table 1.

Table 1: initial parameter setting of PSO-DE algorithm in experimental group

In addition, in this embodiment, a classical Particle Swarm Optimization (PSO) algorithm is further provided to modify the friction coefficient of the initial press forming simulation model, which is used as a control group in the embodiment. The experimental group and the control group are executed for the algorithm 200 times in Matlab, the average absolute percent error of the energy of the simulation process corresponding to the individual optimal friction coefficient executed each time is shown in part (a) of FIG. 11, and the convergence curves of the experimental group and the control group are shown in part (b) of FIG. 11.

As can be seen from fig. 11, the mean absolute percentage error of the energy of the simulation process corresponding to the individual optimal friction coefficient of the PSO-DE algorithm steadily fluctuates between 2% and 5%, while the mean absolute percentage error of the energy of the simulation process corresponding to the individual optimal friction coefficient of the PSO algorithm fluctuates between 4% and 10%, and the precision of the friction coefficient optimized by the PSO-DE algorithm is significantly higher than that of the PSO algorithm. In addition, in the convergence curves of the PSO-DE algorithm and the PSO algorithm, the PSO-DE algorithm converges after about 10 iterations and the PSO algorithm converges after about 55 iterations, and the PSO-DE algorithm is also obviously superior to the PSO algorithm in terms of operation speed.

Simulation process corresponding to optimal friction coefficient individual obtained by 200 times of operation of PSO-DE algorithmIn the energy average absolute percentage error, the minimum value and the optimal friction coefficient corresponding to the minimum value are identified, namely mu ₁ =0.077 and μ ₂ =0.086; wherein mu ₁ Is the friction coefficient between the plate and the blank holder, mu ₂ Is the friction coefficient between the plate and the female die.

After the optimal friction coefficient is obtained, the coefficient is input into a stamping forming simulation model to correct the model, and the simulation process energy of the corrected simulation model under CB1, CB2, CB3 and CB4 is shown in FIG. 12. As can be seen from fig. 12, the corrected simulation model obtains the simulation process energy which is significantly closer to the measured process energy than the initial simulation model under different control parameter combinations. The maximum average absolute percentage error of the energy of the simulation process obtained by the corrected simulation model is not more than 4.51 percent and is less than 10 percent of the correction threshold. In addition, the simulation maximum thinning rate and the simulation maximum thickening rate of the simulation models before and after correction under CB1, CB2, CB3 and CB4 are compared as shown in fig. 12. As seen from fig. 12, under different control parameter combinations, the simulated maximum thinning rate and the simulated maximum thickening rate obtained by the corrected simulation model are obviously closer to the measured maximum thinning rate and the measured maximum thickening rate than the simulated maximum thinning rate and the simulated maximum thickening rate obtained by the initial simulation model. The comparison of the simulation process energy and the simulation thickness change obtained by the simulation models before and after correction proves that the corrected simulation model meets the use requirement, and can be used for further obtaining the simulation process energy and the simulation thickness change data. This demonstrates the effectiveness of the scheme provided by the present embodiment.

3. Building process energy map based on corrected simulation model

After the corrected stamping forming simulation model is obtained, the set stamping process conditions can be input into the simulation model to acquire the simulation process energy and simulation thickness change data. After the data acquisition is completed, a process energy spectrum is constructed, wherein the abscissa is the stamping depth h _p The ordinate is the process energy E; the color of the pixel point of the interval is used for representing each stamping depth and the process energy pair thereofThe thickness change state of the stamping workpiece comprises the maximum thinning rate delta _MTN Maximum thickening ratio delta _MTC Both are represented in two graphs. The process energy spectrum constructed based on the modified press forming simulation model and the process energy spectrum constructed by the other two different data acquisition modes are approximately shown in fig. 13. From this it can be seen that: the image of MAP2 is very close to the image of MAP1 and the difference between the two is large from MAP 3.

4. Precision analysis of process energy spectrum

In order to verify the effectiveness of the process energy spectrum constructed by the corrected stamping forming simulation model provided by the embodiment, the experiment selects the measured thickness change of the plate in the vehicle door under the condition of 27.5mm stamping depth, and the monitoring precision and defect identification accuracy of the process energy spectrum constructed by the corrected stamping forming simulation model are identified.

The monitoring results of the process energy spectrum are shown in fig. 14, and it can be seen from fig. 14: the maximum thinning rate and the maximum thickening rate monitored by the scheme of the embodiment are very similar to the measurement result. Wherein, only 4 unbroken pieces are identified as broken pieces by the spectrogram, and the identification accuracy of the spectrogram on breakage is 87.5%. 5 wrinkling pieces are not correctly identified by the pattern, and the accuracy of the wrinkling identification of the pattern is 84.4%. The defects of the punched workpiece can be effectively identified through the process energy spectrum constructed by the corrected punch forming simulation model.

The average absolute percentage error of the process energy spectrum monitoring result constructed by the scheme of the embodiment is shown as MAP2 in fig. 15 (a), and the maximum thickness variation monitoring error is not more than 8% in the graph, which shows the high monitoring precision of the process energy spectrum constructed by the corrected stamping forming simulation model on the thickness variation of the stamping workpiece. The defect recognition accuracy of the embodiment is more than 80% as shown by MAP2 in fig. 15 (b).

Further, it can be seen from fig. 14 that: the MAP1 is actually closer to the measurement result because it is built up from the actual measurement result. As can be seen from fig. 15, the maximum thinning rate of the punched workpiece by MAP1 (i.e., 3.09%) is 2.02% lower than that by MAP2 (i.e., 5.11%), and the maximum thickening rate by MAP1 (i.e., 3.07%) is 4.74% lower than that by MAP2 (i.e., 7.81%). The crack recognition rate of MAP1 (i.e., 93.75%) was 6.25% higher than that of MAP2 (i.e., 87.5%), and the wrinkle recognition rate of MAP1 (i.e., 93.75%) was 9.35% higher than that of MAP2 (i.e., 84.4%).

The reason why the process energy spectrum monitoring error and defect recognition accuracy constructed by the method of the embodiment are lower than those of MAP1 is that: in the actual stamping forming process, the friction coefficient between the plate and the die and the material performance parameters of the plate are continuously changed. Along with the change of control parameters such as blank holder force and stamping speed, the friction state between the die and the plate can be changed continuously, so that the quantitative relation between the process energy and thickness change of the stamping workpiece is influenced. In the simulation model for stamping forming, once the friction coefficient is set to be a fixed value, the friction coefficient is inconsistent with the actual stamping forming, so that a certain error exists between the simulation process energy obtained by the simulation model and the simulation thickness change, and the monitoring precision of a process energy map constructed by the simulation result is further affected. In addition, the material properties of the plate material are continuously changed in the stamping forming process, the changed material properties can influence the process energy and thickness variation of the stamped workpiece, and the material performance parameters are also constant in the stamping forming simulation model, so that the monitoring precision of a process energy map constructed by simulation results can be reduced. The reduction of the monitoring accuracy also brings about the reduction of the defect recognition accuracy.

However, as a whole, the monitoring error of the process energy spectrum constructed by the low-cost acquisition method combined with the stamping data is still lower than 8%, the defect recognition rate is higher than 80%, the higher level is maintained, and the method is obviously superior to MAP3. In addition, the construction cost of the process energy spectrum constructed by the scheme of the embodiment can be greatly reduced. Specific process energy and sheet consumption pairs are shown in table 2.

Table 2: cost comparison of different process energy spectrum construction modes

Analysis of the above table data shows that: the traditional process energy spectrum construction mode based on test data needs to stamp 448 vehicle door inner plate members, and the formed stamping workpiece needs to be cut and thickness measured after each stamping is finished, so that material resources such as electric energy required by the operation of a press machine, material consumption of the plate materials and human resources are consumed. For the process energy MAP constructed in the scheme, only 4 similar door inner plates are needed to be stamped, and compared with MAP1, the consumption rate of the plates of MAP2 is reduced by 99.11%. Because each time of forming the inner plate of the similar car door consumes a certain amount of process energy, 448 times of forming the inner plate of the similar car door consumes about 111.43kJ of process energy, and 4 times of stamping forming the inner plate of the similar car door, which is used for correcting a stamping forming simulation model, only consumes 3.79kJ of process energy, and 107.64kJ of process energy is saved. Based on the energy efficiency analysis result of the existing centering hydraulic machine, the process energy consumption of the hydraulic machine is about 7% of the total energy consumption of the operation of the hydraulic machine, which means that the method for constructing the process energy spectrum provided by the scheme can save 1537.71kJ of energy when one process energy spectrum is constructed.

In summary, the process energy spectrum constructed by the low-cost acquisition method of the punching data provided by the embodiment can greatly reduce the construction cost of the process energy spectrum and greatly shorten the test period for acquiring the test data on the premise of keeping the monitoring precision similar to that of the conventional process energy spectrum based on the test data, so that the method has good use value and can generate good economic benefit.

5. Application of process energy spectrum constructed in scheme

The process energy spectrum constructed by the embodiment can visualize the thickness change of the stamping workpiece in the processing process, thereby supporting the production decision in the stamping process to avoid the generation of defective parts. When the punching speed is 14mm/s and the blank holder force is 42kN, the thickness variation path 1 of the door-like inner panel is shown as a circle 1 in FIG. 16. It can be seen that the maximum reduction of the punched workpiece is in the safety zone at a punching depth of 26mm and in the fracture zone at a punching to 27 mm. The maximum thinning rate of the punched workpiece further moves inside the fracture zone when punched to 28mm, meaning that the punched workpiece breaks at 27mm and the fracture expands further when punched to 28 mm.

As shown in fig. 17, when the actual processing effect of the punched workpiece is checked, it is found that: when the inner plate of the car door is stamped to 27mm, the stamped workpiece generates obvious small cracks under the thickness change path 1, and when the inner plate is stamped to 28mm, the cracks are further expanded to large cracks, so that the effectiveness and the accuracy of the process energy spectrum constructed in the embodiment on the thickness change and defect monitoring of the stamped workpiece are proved.

In order to suppress defects generated in the stamping process, a technician can perform process control on the stamping process through the constructed process energy map. The blank holder force is reduced by 10kN when the punched workpiece is punched to 26mm, and the thickness variation path of the punched workpiece under this control operation is indicated by a circle 2 in fig. 16, which is referred to as a thickness variation path 2. In the thickness variation path 2, the punched workpiece was punched to 27mm and 28mm, although the risk of wrinkling was increased, i.e., the maximum thickening rate was close to the wrinkling zone, and the maximum thinning rate was in the safety zone.

It can be found in connection with the detection of the actual processing effect in fig. 17 that: the punched workpiece under the thickness variation path 2 is free from cracking and wrinkling when punched to 27mm and 28mm, and the measured maximum thinning rate of the punched workpiece is reduced by 14.56%.

The technical energy map constructed by the scheme provided by the embodiment can realize the visualization of the thickness change of the stamping workpiece, further intuitively observe the defect generation trend of the stamping workpiece, timely control the stamping processing process and effectively avoid the generation of the defect of the stamping workpiece.

The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, and alternatives falling within the spirit and principles of the invention.

Claims

1. The low-cost acquisition method of stamping data is used for acquiring process energy and thickness variation data with mapping relation required by building a process energy map at low cost; the thickness variation data comprises a maximum thinning rate and a maximum thickening rate; the method is characterized by comprising the following steps of:

s1: under the preset stamping process condition, a plurality of discrete sample data are obtained through an actual stamping test, wherein the sample data comprise the measured stamping force and the measured process energy;

s2: constructing a simulation model for simulating a stamping forming process, wherein the simulation model comprises a stamping geometric model and a finite element model; the stamping geometric model is used for reflecting the shape, structure and size of the stamping forming die and the plate; the finite element model is used for simulating the stamping forming of the plate material, so that the mapping relation between the thickness change of the stamped workpiece and the process energy is obtained;

S3: generating a simulation maximum thinning rate, a simulation maximum thickening rate and simulation process energy by using a simulation model;

A. when the deviation is larger than the correction threshold, actually measuring the model parameters which need to be manually input in the stamping geometric model, and inputting the actually measured value of the model parameters into the stamping geometric model to correct the stamping geometric model; and carrying out iterative correction on the friction coefficient mu in the finite element model;

B. when the deviation is not greater than the correction threshold value, the current simulation model is saved as a numerical model;

2. The punching data acquisition method according to claim 1, characterized in that: in step S1, the sample data acquisition method is as follows:

A. setting L stamping depths;

B. measuring the stamping force of the stamping workpiece corresponding to the L stamping depths through an actual stamping test;

3. The punching data acquisition method according to claim 1, characterized in that: in step S4, the correction strategy of the friction coefficient μ is as follows:

s01: initializing friction coefficient population mu (g) and mutation probability F _a Cross probability CR, maximum iteration number ger, individual learning factor b ₁ Population learning factor b ₂ And an inertial weight q;

s02: calculating individual mu of each original friction coefficient in the current friction coefficient population mu (g) _i (g) Average absolute percentage error e of corresponding process energy _MA,i (g) And identifying the optimal friction coefficient individual mu with the smallest average absolute percentage error _m (g)；

In the above, r ₁ And r ₂ A random value between 0 and 1; e is an identity matrix; e, e _MA,m (g) Average absolute percentage error for the individual optimal coefficient of friction; mu (mu) _i (g)、μ _q (g) And mu _r (g) Is in friction coefficient group mu (g)Randomly selected three individuals with original friction coefficients;

s04: according to the direction of evolution v _i (g+1) selecting the original coefficient of friction individuals μ in the current coefficient of friction population μ (g) _i (g) Performing mutation to obtain mutation friction coefficient individual mu _i ' (g), the formula for the mutation operation is as follows:

in the above, F _a (g+1) is the mutation probability of the g+1st iteration;

s05: for individuals mu with abrupt friction coefficient in the current friction coefficient population mu (g) _i ' (g) individual μ with original coefficient of friction _i (g) Performing a crossover operation to generate a new individual friction coefficient, the crossover operation strategy being as follows:

in the above, rand _i Is the original coefficient of friction, individual mu _i (g) Individual μ with abrupt coefficient of friction _i ' (g) a random number between 0 and 1 generated when the interleaving operation is performed;

in the above, e _MA,i ' (g) is the mutation coefficient of friction individual μ _i ' the corresponding process energy average absolute percent error of (g);

s07: judging whether the current iteration number g is smaller than the maximum iteration number ger or not: if yes, executing steps S02-S06 circularly, otherwiseThe iteration is ended, and the optimal friction coefficient individual mu with the minimum process energy average absolute percentage error in the current friction coefficient population mu (g) is output _m (g) As a coefficient of friction after correction.

4. A method of low cost acquisition of stamping data as recited in claim 3, wherein: in step S01, the friction coefficient population μ (g) is:

μ(g)＝(μ ₁ (g),μ ₂ (g),…,μ _i (g),…,μ _τ (g))

in the above formula, g is the iteration number; when g=1, the iteration is in an initial state; τ is the number of individuals with the original coefficient of friction in the coefficient of friction population μ (g);

wherein the ith original friction coefficient individual mu in friction coefficient population mu (g) _i (g) The method comprises the following steps:

μ _i (g)＝(μ _1,i (g),μ _2,i (g))

5. The method for obtaining stamping data at low cost as recited in claim 4, wherein: the i th original coefficient of friction individual μ _i (g) Friction coefficient mu between middle plate and blank holder _1,i (g) And the i th original coefficient of friction individual μ _i (g) Friction coefficient mu between middle plate and concave mould _2,i (g) At a preset maximum friction coefficient mu _max With minimum friction coefficient mu _min The formula is as follows:

6. the method for obtaining stamping data at low cost as recited in claim 5, wherein: in step S02, individuals with the original coefficient of friction μ in the coefficient of friction population μ (g) _i (g) Corresponding process energy mean absolute percentage error e _MA,i (g) The calculation formula of (2) is as follows:

7. The construction method of the process energy spectrum comprises the following steps:

1. under the preset stamping process condition, a plurality of discrete sample data are obtained through an actual stamping test;

2. establishing a numerical model by adopting the stamping data low-cost acquisition method according to any one of claims 1-6 and combining the acquired sample data;

the numerical model is used for generating an array composed of a maximum thinning rate and a maximum thickening rate with mapping relations, stamping depth and stamping force;

3. at a stamping depth h _p Drawing blank process energy maps for the abscissa and the ordinate respectively, and determining a region to be filled with a boundary in the process energy maps according to the constraint of the process on parameters;

4. and (3) carrying out data filling on the region to be filled in the blank process energy map by utilizing the numerical model, wherein the process is as follows:

4.1: generating by means of the numerical model a simulated stamping depth h correlated therewith _ps Simulated punching force F _s Simulation of maximum thickening ratio delta _MTCs And simulate the maximum thinning rate delta _MTNs First array U1: u1 = { h _ps ，F _s ，Δ _MTCs ，Δ _MTNs }；

4.2: according to simulationDepth of press h _ps And simulation punching force F _s Calculating corresponding simulation process energy E _s ：

4.3: converting the first array U1 into a second array U2: u2= { h _ps ，E _s ，Δ _MTCs ，Δ _MTNs }；

4.4: with simulated stamping depth h in the second array _ps And simulation Process energy E _s Respectively used as the abscissa and the ordinate of the pixel points in the region to be filled to simulate the maximum thickening rate delta _MTCs And simulating the maximum thinning rate delta _MTNs The first attribute value and the second attribute value are respectively corresponding to the pixel points; completing the pixel filling of the energy spectrum of the blank process;

5. mirroring the map filled in the steps into two symmetrical parts by taking the ordinate as a symmetry axis; a crinkling identification area and a rupture identification area respectively;

6. coloring each pixel point in the wrinkling recognition area according to a preset color mapping relation and coloring each pixel point in the cracking recognition area according to a second attribute value;

8. The on-line monitoring method for the processing quality of the stamping workpiece is characterized by comprising the following steps of:

step 1: acquiring a process energy spectrum of a current stamping workpiece to be processed, wherein the process energy spectrum is generated by the process energy spectrum construction method according to claim 7;

step 2: real-time miningCollecting real-time stamping depth h of stamping workpiece to be processed in actual processing process _real And a real-time punching force F _real ；

Step 3: according to the real-time stamping depth h in the processing process _real And a real-time punching force F _real Calculating the real-time process energy E _real ；

Step 4: according to the real-time stamping depth h of the stamping workpiece to be processed in the processing process _real And real-time process energy E _real Drawing corresponding state tracks in a wrinkling recognition area and a cracking recognition area of the process energy map;

step 5: and evaluating the processing quality of the processed stamping workpiece according to the state track:

(1) When any point in the state track passes through the fracture zone, judging that the processed workpiece is broken;

(2) And when the end point of the state track is positioned in the wrinkling zone, judging that the processed workpiece has local wrinkling.

9. A punching data low cost acquisition system, characterized by: a method for obtaining punching data at low cost according to any one of claims 1 to 6, which combines the measured process energy and the simulation model to obtain a numerical model for obtaining punching data at low cost, comprising:

The measuring process energy acquisition module is used for calculating measuring process energy according to a plurality of discrete sample data acquired by an actual stamping test;

the simulation model construction module comprises a stamping geometric model and a finite element model; the stamping geometric model is used for reflecting the shape, structure and size of the stamping forming die and the plate; the finite element model is used for simulating the stamping forming of the plate material, so that the mapping relation between the thickness change of the stamped workpiece and the process energy is obtained; the simulation model is used for generating a simulation maximum thinning rate, a simulation maximum thickening rate and simulation process energy;

the error calculation module is used for calculating deviation between the simulation process energy generated by the constructed simulation model and the measured process energy;

a friction coefficient updating module for iteratively correcting the friction coefficient in the finite element model using the correction strategy included in any one of claims 3 to 6 when the deviation calculated by the error calculating module exceeds the correction threshold; the numerical model correction module is used for resetting manually input model parameters in the stamping geometric model and correcting friction coefficients in the finite element model when the deviation calculated by the error calculation module exceeds a correction threshold value; and when the deviation calculated by the error calculation module does not exceed the correction threshold value, preserving parameters in the corresponding simulation model to obtain a numerical model for generating a mapping relation between the maximum thinning rate, the maximum thickening rate and the process energy of the stamping workpiece.

10. A data processing apparatus comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that: creating the ram data low cost acquisition system of claim 9 when the processor executes the computer program; and then automatically constructing a numerical model capable of generating a mapping relation between the maximum thinning rate, the maximum thickening rate and the process energy of the stamping workpiece under the specified process conditions according to a plurality of discrete sample data obtained by an actual stamping test.