Machining state estimation apparatus, and machining state estimation method

Abstract

This machining state estimation apparatus comprises a processor that executes first and second prediction processes in a first period. A storage device stores first and second parameters that define the machining state.

Description

Machining state estimation device and machining state estimation method

[0001] The present disclosure relates to a machining state estimating device and a machining state estimating method.

[0002] Patent Document 1 discloses a technique for obtaining a judgment value by combining a state quantity of a normal piece of equipment and a state quantity of an abnormal piece of equipment in a machine that repeats the same work in a relatively short cycle, such as a press machine. The judgment device in Patent Document 1 issues an alarm when the state quantity of the target machine exceeds or falls below the judgment value.

[0003] Japanese Patent Application Publication No. 9-120365

[0004] The present disclosure provides a machining state estimating device and a machining state estimating method that facilitate accurate estimation of the machining state.

[0005] A machining state estimation device according to one aspect of the present disclosure is a machining state estimation device that estimates the machining state of a press machine that repeatedly performs press working, and includes: a processor that executes a first prediction process and a second prediction process related to estimation of the machining state in a first cycle; and a storage device, wherein the storage device stores first and second parameters that define the machining state of the press machine, and a plurality of reference data that respectively indicate changes in the machining load of the press machine corresponding to each combination of the first and second parameters, the processor acquires measurement data that indicates measurement results of the machining load of the press machine, and acquires first and second current state parameters that correspond respectively to the first and second parameters and indicate estimation results of the machining state of the press machine at a predetermined reference time before measurement of the measurement data, and The first parameter is varied within a predetermined first range from a first current state parameter, and the second parameter is fixed to a second current state parameter, and reference data corresponding to each combination of the varying first parameter and the fixed second parameter is searched for that reference data having a maximum similarity, which is an index of the degree of similarity to the measurement data; the first parameter corresponding to the searched reference data and the fixed second parameter are determined as first and second estimated parameters representing the processing state at the time of measuring the measurement data, respectively, in the first prediction process; and the processor, in the second prediction process, obtains a progress state parameter generated by varying the second current state parameter within a predetermined second range, varies the first parameter within a first range from the first estimated parameter, and fixes the second parameter to the progress state parameter, and searches for reference data corresponding to each combination of the varying first parameter and the fixed second parameter, which has a maximum similarity to the measurement data; a first parameter corresponding to the searched reference data and a fixed second parameter are determined as first and second estimated parameters representing the processing state at the time of measurement of the measurement data in a second prediction process, respectively; and the processor outputs the first estimated parameter determined in either the first or second prediction process as an estimation result for the first parameter in a first period.When a predetermined condition is satisfied, the processor outputs the second estimated parameter determined in either the first or second prediction process as an estimation result for the second parameter.

[0006] A machining state estimation method according to one aspect of the present disclosure is a machining state estimation method for estimating the machining state of a press machine that repeatedly performs press working, the method comprising: a step of a processor acquiring measurement data indicating a measurement result of a machining load by the press machine; a step of the processor acquiring first and second current state parameters corresponding to first and second parameters that define the machining state of the press machine, respectively, and indicating an estimation result of the machining state of the press machine at a predetermined reference time before the measurement of the measurement data; a step of the processor executing a first prediction process and a second prediction process related to estimation of the machining state in a first cycle; and a step of the processor outputting estimation results related to the first and second parameters, wherein the first prediction process comprises: a step of varying the first parameter within a predetermined first range from the first current state parameter and fixing the second parameter to the second current state parameter, and searching for reference data that has a maximum similarity, which is an index of the degree of similarity to the measurement data, from a plurality of reference data indicating changes in the machining load corresponding to each combination of the varying first parameter and the fixed second parameter; The second prediction process includes a step of determining, in the first prediction process, a first parameter corresponding to the searched reference data and a fixed second parameter as first and second estimated parameters, respectively, representing the processing state at the time of measurement of the measurement data; the second prediction process includes a step of acquiring a progress state parameter generated by varying the second current state parameter within a predetermined second range; a step of varying the first parameter from the first estimated parameter within a first range and fixing the second parameter to the progress state parameter, and searching for reference data that has the greatest similarity to the measurement data from a plurality of reference data corresponding to each combination of the varying first parameter and the fixed second parameter; and a step of determining, in the second prediction process, the first parameter corresponding to the searched reference data and the fixed second parameter as first and second estimated parameters, respectively, representing the processing state at the time of measurement of the measurement data; and the step of outputting the estimation results for the first and second parameters includesThe method includes a step of outputting a first estimated parameter determined in either the first or second prediction process as an estimation result for the first parameter at a first period, and a step of outputting a second estimated parameter determined in either the first or second prediction process as an estimation result for the second parameter when a predetermined condition is satisfied.

[0007] According to the present disclosure, it is possible to easily estimate the machining state with high accuracy.

[0008] 1 is a block diagram illustrating an example of the configuration of a machining state estimation device according to an embodiment of the present disclosure. FIG. 1 is a schematic cross-sectional view showing a press machine to which the load sensor shown in FIG. 1 is attached. FIG. 2 is a schematic graph showing an example of a waveform measured by the load sensor shown in FIG. 1. FIG. 3 is a schematic cross-sectional view for explaining an area of ​​the punched contour of the press machine of FIG. 2. FIG. 4 is a table showing an example of state data. FIG. 5 is a flowchart illustrating a procedure of the machining state estimation process executed by the CPU of the machining state estimation device of FIG. 1. FIG. 7 is a flowchart illustrating the initial state estimation process shown in FIG. 8. FIG. 9 is a flowchart illustrating the clearance and workpiece thickness estimation process shown in FIG. 9. FIG. 10 is a flowchart illustrating the reference waveform generation process corresponding to the state data shown in FIG. 9. FIG. 11 is a flowchart illustrating the progress state estimation process shown in FIG. 7. FIG. 12 is a flowchart illustrating the workpiece thickness estimation process shown in FIG. 13. FIG. 14 is a graph illustrating the results of the machining state estimation process according to the present embodiment.

[0009] (Findings that Form the Basis of the Present Disclosure) The present inventors have conducted extensive research to accurately estimate the processing state of a press machine in press processing, particularly punching, and have found the following: Here, the "processing state" refers to at least one of the amount of tool wear, clearance, and thickness of a workpiece.

[0010] The load applied to the punch or workpiece during punching depends on the amount of wear on the punch, the amount of wear on the die, the clearance, the thickness of the workpiece, and other factors.

[0011] The punch wear amount and the die wear amount are examples of a punch wear parameter, which is an index indicating the degree of wear of the punch, and a die wear parameter, which is an index indicating the degree of wear of the die, respectively. Tool wear amounts such as punch wear amount and die wear amount are expressed, for example, as dimensional changes from the design values ​​of the tool. Tool wear amounts may also be expressed as changes in shape, volume, mass, etc. Furthermore, tool wear amounts may also be expressed as the radius of an arc when wear is approximated as an arc.

[0012] The clearance is the gap between the die and the punch. For example, the clearance is the gap between the die and the punch when a punched hole is made in a workpiece. The clearance may be expressed as the ratio of the gap between the die and the punch to the thickness of the workpiece.

[0013] Since the load depends on these parameters, it is conceivable to estimate these parameters from the load waveform obtained during machining. For example, if the amount of tool wear, such as the amount of punch wear and the amount of die wear, can be estimated, it becomes possible to know the desired timing for grinding or regrinding (hereinafter simply referred to as "grinding") the tool in a processing machine that performs cyclic machining. Grinding the tool at the desired timing can prevent situations such as machining workpieces with worn tools and producing a large number of defective products, thereby increasing productivity.

[0014] The inventors have found that, in a processing machine that performs cycle processing, there is an advantage in estimating the processing state by using a history of estimation results of the processing state for punching.

[0015] In machining that is performed many times, for example, tens of thousands of cycles, fluctuations in values ​​such as clearance, punch wear amount, and die wear amount are gradual compared to fluctuations in workpiece thickness, etc. For example, values ​​such as clearance, punch wear amount, and die wear amount usually do not change significantly from the values ​​in the immediately preceding punching, so in the history of estimation results that show fluctuations in these values ​​(changes over time or with respect to the number of machining cycles), the fluctuations in these values ​​are gradual and the amount of change is small.

[0016] On the other hand, variations in the thickness of the workpiece may occur even within a short period of time. For example, the thickness of the workpiece may change significantly from the value in the immediately preceding punching.

[0017] The inventors discovered that values ​​such as clearance, punch wear, and die wear fluctuate over long periods, while values ​​such as workpiece thickness fluctuate over short periods, and based on this knowledge, they discovered that the machining state can be estimated with greater accuracy, leading to the present invention.

[0018] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the accompanying drawings as appropriate. However, more detailed description than necessary may be omitted. For example, detailed description of well-known matters or redundant description of substantially identical configurations may be omitted. This is to avoid unnecessary redundancy in the following description and to facilitate understanding by those skilled in the art. Note that the inventors provide the accompanying drawings and the following description to enable those skilled in the art to fully understand the present disclosure, and do not intend for them to limit the subject matter described in the claims.

[0019] 1 is a block diagram showing an example of the configuration of a machining state estimating device 100 according to an embodiment of the present disclosure. The machining state estimating device 100 includes a CPU 1, a storage device 2, an input interface (I / F) 3, and an output interface (I / F) 4.

[0020] The CPU 1 performs information processing to realize the functions of the machining state estimating device 100, which will be described later. Such information processing is realized, for example, by the CPU 1 operating in accordance with instructions of a program 21 stored in the storage device 2. The CPU 1 is an example of a processor of the present disclosure. The processor is not limited to a CPU as long as it includes an arithmetic circuit that performs calculations for information processing. For example, the processor may be configured by a circuit such as an MPU or FPGA.

[0021] The storage device 2 is a recording medium for recording various information including data such as a waveform library 23 and state data 22, which will be described later, and a program 21 required to implement the functions of the machining state estimating device 100. The storage device 2 is implemented, for example, by a semiconductor storage device such as a flash memory or a solid state drive (SSD), a magnetic storage device such as a hard disk drive (HDD), or other recording media, either singly or in combination. The storage device 2 may include a volatile memory such as an SRAM or a DRAM.

[0022] The input interface 3 is an interface circuit that connects the machining state estimating device 100 to an external device in order to input information such as the detection result by the load sensor 11 to the machining state estimating device 100. Such external devices are, for example, the load sensor 11, other information processing terminals, etc. The input interface 3 may be a communication circuit that performs data communication in accordance with an existing wired communication standard or wireless communication standard.

[0023] The output interface 4 is an interface circuit that connects the machining state estimating device 100 to an external output device in order to output information from the machining state estimating device 100. Such an output device is, for example, a display or another information processing terminal. The output interface 4 may be a communication circuit that performs data communication in accordance with an existing wired communication standard or wireless communication standard. The input interface 3 and the output interface 4 may be realized by similar hardware.

[0024] Fig. 2 is a schematic cross-sectional view showing a press 50 to which the load sensor 11 shown in Fig. 1 is attached. For ease of explanation, Fig. 2 shows an X-axis, a Y-axis, and a Z-axis that are perpendicular to each other. The Z-axis indicates the vertical direction.

[0025] The press 50 is an example of a processing machine that performs cyclic processing, repeating the same processing. The press 50 includes a bolster 51 and a slide 52 that repeatedly performs a cyclic movement up and down relative to the bolster 51 from the top dead center to the bottom dead center. A die backing plate 61 is attached to the top of the bolster 51, and a die plate 62 is attached to the top of the die backing plate 61. The die plate 62 holds a die 63.

[0026] A punch backing plate 71 is attached to the bottom of the slide 52, and a punch plate 72 is attached to the bottom of the punch backing plate 71. The punch plate 72 grips the punch 73. The press 50 further includes a stripper plate 74. The stripper plate 74 is attached to fasteners such as bolts and the punch plate 72 or punch backing plate 71 via positioning guides such as posts (not shown), for example. The stripper plate 74 is biased downward by, for example, a compression spring, and has the function of guiding the punch 73 to keep the position of the punch 73 constant, the function of removing material adhering to the punch 73 after punching the workpiece 80, and / or the function of fixing the workpiece 80 during punching.

[0027] The load sensor 11 is installed, for example, between the punch 73 and the punch backing plate 71. The load sensor 11 is, for example, a piezoelectric force sensor or an electric force sensor such as a strain gauge type, and measures the load applied to the punch 73 when the punch 73 punches out the workpiece 80.

[0028] 3 is a schematic graph showing an example of a waveform measured by the load sensor 11. The horizontal axis of the graph in FIG. 3 represents time, and the vertical axis represents load. The time on the horizontal axis is based (0) on the time when the punch 73 is at the top dead center, which is its initial position. The punch 73 moves in the negative direction of the Z axis from the top dead center toward the bottom dead center.

[0029] The graph in Figure 3 shows a mountain-shaped waveform in which, during punching, a load begins to be applied to the workpiece 80, and therefore to the punch 73 and the load sensor 11, from the moment the punch 73 descends and contacts the workpiece 80, and the load suddenly decreases to almost zero after the workpiece 80 is punched. The punching start timing of the punching can be measured, for example, based on the time when the load exceeds a rising threshold in the measured waveform. Such a rising threshold may be defined as an absolute value or as a percentage of the peak value of the load.

[0030] Instead of the example of FIG. 3, the horizontal axis of the graph may represent the distance that the punch 73 has advanced in the negative direction of the Z axis (the distance advanced is taken as a positive value).

[0031] [2. Operation] [2-1. Overview of Operation] An overview of the machining state estimation process will be described with reference to Figures 4 to 6. Figure 4 is a schematic diagram illustrating an example of the outline of the machining state estimation process executed by the machining state estimation device 100 of Figure 1.

[0032] The CPU 1 acquires unit waveforms (hereinafter referred to as "standard reference data") per unit length of the punched contour of the press 50 from the waveform library 23, and generates area waveforms (area data) corresponding to the two areas A1 and A2, respectively. The CPU 1 synthesizes all the area waveforms to generate a reference waveform (hereinafter referred to as "total reference data"), and compares the measured waveform with the reference waveform.

[0033] The unit waveform, the area waveform, and the reference waveform are waveform data that indicate the change over time in the processing load of the press machine 50, and can be compared with the measured waveform.

[0034] Since the unit waveform is associated with parameters indicating at least one of the tool wear amount, clearance, or workpiece thickness, the parameters of each of areas A1 and A2 can be estimated by searching for a reference waveform that closely matches the measured waveform.

[0035] Fig. 5 is a schematic cross-sectional view for explaining areas A1 and A2 of the punching contour of the press 50. The cross-sectional view of Fig. 5 shows only the punch 73 and the die 63 for ease of understanding.

[0036] The punching contour is the contour of the portion of the workpiece 80 that is punched by the punching process using the press machine 50. The shapes of the punch 73 and the die 63 are designed to achieve a desired punching contour. The punching contour may be the design value of the contour of the punch 73 as viewed from the punching direction, or the design value of the contour of the opening of the die 63 as viewed from the punching direction.

[0037] The areas A1 and A2 of the punched outline are obtained by dividing the punched outline. Where the punched outline is divided is determined in advance depending on the shape of the punched outline. In the example of FIG. 5 , the punched outline is a rectangle with rounded corners. The areas A1 and A2 of the punched outline are obtained by dividing the punched outline. In the example of FIG. 5 , the punched outline is divided into two, and one of the divided punched outlines is area A1 and the other is area A2.

[0038] FIG. 6 is a table showing an example of the state data 22. The state data 22 includes a contour parameter that defines information about the punching contour, a tool state parameter that defines the state of the tool, and a workpiece state parameter that defines the state of the workpiece. In the example of FIG. 6, the contour parameter is the area length along the punching contour. In the example of FIG. 6, the tool state parameters are the punch wear amount, the die wear amount, and the clearance. In the example of FIG. 6, the workpiece state parameter is the workpiece thickness.

[0039] 6 is an example of "area shape information" in the present disclosure. Regarding the contour parameters, the area lengths of the first area A1 and the second area A2 are represented as W1 and W2, respectively.

[0040] Regarding the tool state parameters shown in FIG. 6, the punch wear amount, die wear amount, and clearance in the first area A1 are represented as P1, D1, and C1, respectively, and the punch wear amount, die wear amount, and clearance in the second area A2 are represented as P2, D2, and C2, respectively.

[0041] 6, the workpiece thickness T is constant across all zones. However, this embodiment is not limited to this, and the workpiece thickness may also take on different values ​​for each zone, similar to the amount of wear on the punch, the amount of wear on the die, and the clearance.

[0042] For example, the punch wear amounts P1 and P2 can be set to any of the following candidate values: 0 μm, 2 μm, 4 μm, 6 μm, 8 μm, 10 μm, and 12 μm. For example, the die wear amounts D1 and D2 can be set to any of the following candidate values: 0 μm, 2 μm, 4 μm, 6 μm, 8 μm, 10 μm, and 12 μm. For example, the clearances C1 and C2 can be set to any of the following candidate values: 3 μm, 4 μm, 5 μm, 6 μm, and 7 μm. For example, the workpiece thickness T can be set to any of the following candidate values: 46 μm, 48 μm, 50 μm, 52 μm, and 54 μm. Note that the candidate values ​​for the punch wear amount, die wear amount, clearance, and workpiece thickness are not limited to these, and the number of candidate values ​​is not limited to the above.

[0043] As in the above example, when there are seven candidate values ​​for punch wear amount, seven candidate values ​​for die wear amount, five candidate values ​​for clearance, and five candidate values ​​for workpiece thickness, 1,225 different unit waveforms are pre-registered in the waveform library 23. In this way, the waveform library 23 is a four-dimensional table in which unit waveforms corresponding to arrays of punch wear amount, die wear amount, clearance, and workpiece thickness are registered.

[0044] The waveform library 23 stores in advance unit waveforms per unit length of the punched contour corresponding to all combinations of punch wear amount, die wear amount, clearance, and workpiece thickness. The unit length is a predetermined unit length, for example, 1 mm. In this embodiment, the unit waveform is a waveform that represents the relationship between time and load, similar to the measured waveform in FIG. 3.

[0045] The unit waveform can be obtained, for example, by actually measuring the punching load or by multiplying the waveform obtained by simulation by the ratio of the unit length to the total length of the punched profile. For example, if the unit length is 1 [mm] and the total length of the punched profile is L [mm], the unit waveform can be obtained by actually measuring the punching load or by multiplying the waveform obtained by simulation by 1 / L.

[0046] 4 and 6, the CPU 1 acquires unit waveforms corresponding to the combination of punch wear amount, die wear amount, clearance, and workpiece thickness for each zone from the waveform library 23. Next, the CPU 1 generates a zone waveform for each zone by multiplying each unit waveform by the zone length. As shown in FIG. 6, the CPU 1 generates a reference waveform that indicates the load over the entire length of the punched profile by combining two zone waveforms.

[0047] As shown in FIG. 6, the CPU 1 searches for a reference waveform that matches the measured waveform to the greatest extent, and estimates a combination of parameters corresponding to the unit waveform of each section that is the basis of the searched reference waveform as an estimated parameter set that represents the machining state of that section.

[0048] [2-2. Flowchart] [2-2-1. Overall flow] Fig. 7 is a flowchart illustrating the procedure of the machining state estimation process executed by the CPU 1 of the machining state estimation device 100 of Fig. 1. The process of Fig. 7 is performed every time the CPU 1 acquires a measurement waveform from the load sensor 11. For example, the execution interval of the process of Fig. 7 is approximately equal to the machining period of cycle machining.

[0049] First, the CPU 1 acquires from the load sensor 11 a measurement waveform that indicates the measurement result of the load applied to the load sensor 11 during press working by the press machine 50 (S1).

[0050] Next, the CPU 1 determines whether a predetermined period of time has elapsed since the tool change was performed (S2). For example, the CPU 1 determines whether the predetermined period of time has elapsed since receiving a tool change signal indicating that the tool change was performed. The CPU 1 may determine that the predetermined period of time has elapsed if press working has been performed a predetermined number of times or more since receiving the tool change signal. Such a tool change signal is transmitted to the CPU 1, for example, by the user pressing a tool change completion button provided on the press machine 50, the user interface of the machining state estimation device 100, or the like.

[0051] If it is determined that the predetermined period has not elapsed since the tool was replaced (No in S2), the CPU 1 executes a first state estimation process (hereinafter referred to as "initial state estimation process") S3. The initial state estimation process S3 will be described in detail later.

[0052] In step S2, if it is determined that a predetermined period of time has elapsed since the tool was replaced (Yes in S2), the CPU 1 executes a second state estimation process (hereinafter referred to as a "progress state estimation process") S4. The progress state estimation process S4 will be described in detail later.

[0053] [2-2-2. Initial State Estimation Processing S3] FIG. 8 is a flowchart illustrating the initial state estimation processing S3 shown in FIG.

[0054] In the initial state estimation process S3, the CPU 1 acquires the state data 22 estimated in the previous processing state estimation process (S31). If there is no previously estimated state data 22 in step S31, the CPU 1 creates the state data 22 and sets parameters to initial values. For example, the initial value of the punch wear amount is 0 μm, the initial value of the die wear amount is 0 μm, the initial value of the clearance is 5 μm, and the initial value of the workpiece thickness is 50 μm.

[0055] Next, the CPU 1 initializes the tool state parameters (S32). In step S32, if the punch 73 has been replaced or re-ground, the punch wear amount is set to an initial value of 0 μm. Similarly, if the die 63 has been replaced or re-ground, the die wear amount is set to an initial value of 0 μm.

[0056] Next, the CPU 1 executes a clearance and workpiece thickness estimation process (S33). In the clearance and workpiece thickness estimation process S33, the clearance and the workpiece thickness are estimated.

[0057] Fig. 9 is a flowchart illustrating the clearance and workpiece thickness estimation process S33 shown in Fig. 8. In the clearance and workpiece thickness estimation process S33, the CPU 1 first executes a reference waveform generation process S331 corresponding to the status data.

[0058] Fig. 10 is a flowchart illustrating the reference waveform generation process S331 corresponding to the status data shown in Fig. 9. In the reference waveform generation process S331 corresponding to the status data, the CPU 1 first obtains, for each zone, a unit waveform corresponding to the parameter value of the status data 22 from the waveform library 23 (S3310).

[0059] Next, the CPU 1 generates a segment waveform for each segment by multiplying each unit waveform by the segment length (S3311).

[0060] Next, the CPU 1 generates a reference waveform that indicates the load over the entire length of the punched contour by synthesizing all the area waveforms (S3312). Synthesizing a plurality of waveforms means, for example, taking the sum of a plurality of waveforms.

[0061] Returning to FIG. 9, the CPU 1 calculates the degree of coincidence between the reference waveform corresponding to the status data 22 generated in step S331 and the measured waveform acquired in step S1 (S332).

[0062] Here, the degree of match is an index indicating the degree of match between two waveforms. The degree of match is, for example, the cosine similarity, Euclidean distance, or Manhattan distance between two waveforms during the punching period. Instead of the degree of match, the CPU 1 may calculate a loss, which is an index indicating the degree of mismatch between two waveforms. Both the degree of match and the degree of mismatch are examples of "similarity," which is an index indicating the degree of similarity between two waveforms.

[0063] Next, the CPU 1 determines whether the loop processing in the clearance and workpiece thickness estimation processing S33 has converged (completed) (S333). Convergence means that all candidate values ​​that can be selected based on a predetermined selection rule have been set in all areas of the provisional state data. In step S333, the CPU 1 determines, as a convergence determination, whether all candidate values ​​for combinations of clearance and workpiece thickness have been set as clearance and workpiece thickness in areas A1 and A2 of the provisional state data.

[0064] In this embodiment, in step S333, the CPU 1 determines whether the loop processing has converged based on whether all candidate values ​​have been set, but the present disclosure is not limited to this. For example, the CPU 1 may determine whether the loop processing has converged based on the provisional state data in step S334 and the change in the degree of match in step S336.

[0065] If the CPU 1 determines in step S333 that the loop processing in the clearance and work thickness estimation processing S33 has not converged (No in S333), it executes step S334, and if it determines that the loop processing has converged (Yes in S333), it terminates the clearance and work thickness estimation processing S33.

[0066] In step S334, the CPU 1 prepares provisional state data by changing the state data 22 for each zone so that the clearance and the workpiece thickness are set to either of the candidate values ​​for the clearance and the workpiece thickness (S334). Note that in step S334, the other parameters of the provisional state data, the punch wear amount and the die wear amount, are fixed to the previously estimated punch wear amount and die wear amount, respectively.

[0067] Next, the CPU 1 executes a reference waveform generation process S335 corresponding to the provisional state data.

[0068] Fig. 11 is a flowchart illustrating the reference waveform generation process S335 corresponding to the provisional state data shown in Fig. 9. Compared with the reference waveform generation process S331 corresponding to the state data in Fig. 10, the reference waveform generation process S335 corresponding to the provisional state data includes step S3350 instead of step S3310.

[0069] 11, the CPU 1 first acquires unit waveforms corresponding to the parameter values ​​of the provisional state data for each section from the waveform library 23 (S3350). The subsequent steps S3351 and S3352 are similar to steps S3311 and S3312, respectively, of the reference waveform generation process S331 corresponding to the state data in FIG.

[0070] Returning to FIG. 9, the CPU 1 calculates the degree of coincidence between the reference waveform corresponding to the tentative state data generated in step S335 and the measured waveform acquired in step S1 (S336).

[0071] Next, the CPU 1 determines whether the degree of match calculated in step S336 has increased compared to the degree of match calculated in the most recent step S332 (S337). If the CPU 1 determines that the degree of match has increased (Yes in S337), the process proceeds to step S338, but if the CPU 1 determines that the degree of match has not increased (No in S337), the process returns to step S333.

[0072] In step S338, the CPU 1 updates the state data 22 so that the provisional state data prepared in step S334 becomes the state data 22 (S338). After completing step S338, the CPU 1 returns to step S331.

[0073] As described above, if the CPU 1 determines in step S333 that the loop processing in the clearance and workpiece thickness estimation processing S33 has converged (Yes in S333), it ends the clearance and workpiece thickness estimation processing S33.

[0074] In the above example, the CPU 1 may set the clearances in the areas A1 and A2 of the provisional state data to 3 μm, 4 μm, 5 μm, 6 μm, and 7 μm, and the workpiece thicknesses to 46 μm, 48 μm, 50 μm, 52 μm, and 54 μm. Furthermore, the set values ​​in the areas A1 and A2 may be different from each other. When the CPU 1 has completed all loops corresponding to all possible combinations of the clearance set value and the workpiece thickness set value, it ends the clearance and workpiece thickness estimation process S33.

[0075] Returning to Fig. 8, the CPU 1 updates the current state data and the progress state data (S34). The current state data is an example of the first state data (see Fig. 12). The progress state data is an example of the second to seventh state data (see Fig. 12). In step S34, the CPU 1 prepares state data (first to seventh state data) to be used in prediction flows F1 to F7 of the progress state estimation process S4, which will be described later.

[0076] Generally, tool condition parameters change as wear progresses, but the degree of change is gradual compared to the workpiece thickness. For example, the amount of punch wear and the amount of die wear gradually increase as the processing cycle progresses. In addition, it is known that in press processing, the side surface is gradually worn away due to friction generated between the punch 73 and the workpiece 80, and the clearance at the tip of the punch 73 where it contacts the workpiece 80 also gradually increases as the processing cycle progresses.

[0077] Based on this knowledge, in step S34, the CPU 1 prepares multiple pieces of progress status data (second to seventh status data) in addition to the status data 22 (current status data, first status data) predicted in the clearance and work thickness estimation process S33.

[0078] Each progress state data is prepared by performing a process to change the punch wear amount, die wear amount, and / or clearance in the current state data to a value one step larger than the current state data. For example, the state data 22 predicted in the clearance and workpiece thickness estimation process S33 is used as the first state data corresponding to prediction flow F1. The second to seventh state data corresponding to prediction flows F2 to F7, respectively, are state data in which at least one parameter is different from the first state data.

[0079] Specifically, if the punch wear amount P1 in the first area A1 is 0 μm in the first status data, then the punch wear amount P1 in the first area A1 is set to 2 μm, which is one step larger than 0 μm, in the second status data. Similarly, if the punch wear amount P1 in the second area A2 is 0 μm in the first status data, then the punch wear amount P1 in the second area A2 is set to 2 μm in the third status data.

[0080] Similarly, in the fourth state data, the die wear amount D1 in area A1 is set to a value one level larger than that of the first state data, in the fifth state data, the die wear amount D2 in area A2 is set to a value one level larger than that of the first state data, in the sixth state data, the clearance C1 in area A1 is set to a value one level larger than that of the first state data, and in the seventh state data, the clearance C2 in area A2 is set to a value one level larger than that of the first state data.

[0081] [2-2-3. Progression State Estimation Processing S4] FIG. 12 is a flowchart illustrating the progress state estimation processing S4 shown in FIG.

[0082] First, in the progress state estimation process S4, the CPU 1 executes the prediction flows F1 to F7. The prediction flows F1 to F7 are executed, for example, sequentially by the CPU 1. Alternatively, the prediction flows F1 to F7 may be executed in parallel and asynchronously by multiple processing cores constituting the CPU 1. When the prediction flows F1 to F7 are executed in parallel and asynchronously, the CPU 1 may wait until all of the prediction flows F1 to F7 have been executed, and then synchronize the execution to combine the prediction flows.

[0083] In the prediction flow F1, the CPU 1 acquires the first state data (S41a).

[0084] In step S2 in Figure 7, immediately after the judgment changes from No to Yes, that is, immediately after it is determined that a predetermined period of time has passed since the tool was replaced, the progress status estimation process S4 is executed, and the value set in step S34 in Figure 8 is obtained.

[0085] 7, in the progress state estimation process S4 that is executed after one or more determinations of "Yes" have been made in step S2, the CPU 1 acquires, as the first state data, state data corresponding to the same prediction flow, i.e., state data previously estimated in the prediction flow F1. For example, the acquired state data is state data updated in step S428 (described later) in the prediction flow F1.

[0086] In the prediction flow F2, the CPU 1 acquires the second state data corresponding to the prediction flow F2 (S41b). The same applies to the prediction flows F3 to F7. For example, in the prediction flow F7, the CPU 1 acquires the seventh state data corresponding to the prediction flow F7 (S41g).

[0087] In the prediction flow F1, the CPU 1 executes the workpiece thickness estimation process S42 after step S41a. Similarly, in the prediction flows F2 to F7, the CPU 1 executes the workpiece thickness estimation process S42 after the step of acquiring the state data.

[0088] Fig. 13 is a flowchart illustrating the workpiece thickness estimation process S42 shown in Fig. 12. The workpiece thickness estimation process S42 differs from the clearance and workpiece thickness estimation process S33 shown in Fig. 9 in the estimated parameters and the step (S429) of calculating the average value of the degree of coincidence. That is, the clearance and workpiece thickness estimation process S33 is a process for estimating the clearance and workpiece thickness, whereas the workpiece thickness estimation process S42 is a process for estimating the workpiece thickness. The workpiece thickness changes in a shorter cycle than the tool state parameters.

[0089] In the workpiece thickness estimation process S42 in the prediction flow F1, the CPU 1 first executes the reference waveform generation process S331 corresponding to the status data shown in Fig. 10. The status data used in the workpiece thickness estimation process S42 in the prediction flow F1 is the first status data acquired in step S41a.

[0090] Next, CPU 1 calculates the degree of match between the reference waveform corresponding to the status data generated in step S331 and the measured waveform acquired in step S1 (S422), and determines whether the loop processing in the work thickness estimation processing S42 has converged (S423).

[0091] If the CPU 1 determines in step S423 that the loop processing in the work thickness estimation processing S42 has not converged (No in S423), it executes step S424, and if it determines that the loop processing has converged (Yes in S423), it executes step S429.

[0092] In step S424, the CPU 1 prepares provisional state data by changing the state data so as to set the workpiece thickness to one of the candidate values ​​(S424). Note that in step S424, the other parameters of the provisional state data, namely, the punch wear amount, the die wear amount, and the clearance, are fixed to the previously estimated punch wear amount, the die wear amount, and the clearance, respectively.

[0093] Next, the CPU 1 executes the reference waveform generation process S335 corresponding to the provisional state data shown in FIG. 11, and calculates the degree of match between the reference waveform corresponding to the provisional state data generated in step S335 and the measurement waveform acquired in step S1 (S426).

[0094] Next, the CPU 1 determines whether the degree of match calculated in step S426 has increased compared to the degree of match calculated in the most recent step S422 (S427). If the CPU 1 determines that the degree of match has increased (Yes in S427), the process proceeds to step S428, but if the CPU 1 determines that the degree of match has not increased (No in S427), the process returns to step S423.

[0095] In step S428, the CPU 1 updates the state data so that the provisional state data prepared in step S424 becomes the state data (S428). After completing step S428, the CPU 1 returns to step S331.

[0096] If it is determined in step S423 that the loop processing in the workpiece thickness estimation processing S42 has converged (Yes in S423), the CPU 1 calculates the average value of the accumulated degrees of agreement (S429). The accumulated degrees of agreement are a set of degrees of agreement calculated after the tool state parameters have been updated. That is, in step S429, the CPU 1 accumulates the degrees of agreement calculated in step S422 of the workpiece thickness estimation processing S42, and calculates the average value of the accumulated degrees of agreement. Note that if the tool state parameters in the state data are updated, the accumulated degrees of agreement in step S429 are reset to zero.

[0097] After step S429, the CPU 1 ends the workpiece thickness estimation process S42.

[0098] Returning to FIG. 12, after the workpiece thickness estimation process S42, the CPU 1 combines the execution of each prediction flow F1 to F7.

[0099] Next, the CPU 1 determines whether a predetermined period has elapsed since the tool state parameters were last updated (S43). Alternatively, the CPU 1 may determine whether the number of machining cycles since the tool state parameters in the state data were last updated exceeds a preset threshold. In this embodiment, the predetermined period in step S43 is set to be longer than the machining period of the cycle machining. Therefore, when determining whether the number of machining cycles has exceeded the threshold in step S43, the threshold is set to an integer of 2 or greater, for example.

[0100] If the CPU 1 determines that a predetermined period has elapsed since the tool state parameters were last updated (Yes in S43), it executes step S44, and if it determines that the predetermined period has not elapsed (No in S43), it terminates the progress state estimation process S4.

[0101] In step S44, the CPU 1 selects the prediction flow with the highest average degree of coincidence from among the prediction flows F1 to F7. The average degree of coincidence is the value calculated in step S429 of the workpiece thickness estimation process S42.

[0102] The CPU 1 acquires the state data corresponding to the prediction flow selected in step S44 as the latest state data, and updates the first to seventh state data to the latest state data (S45).

[0103] Step S45 corresponds to step S34 of the initial state estimation process S3, and in addition to the state data predicted immediately before, state data for a state in which wear has progressed by one stage is set.

[0104] That is, in step S45, the state data itself corresponding to the prediction flow selected in step S44 is set as the first state data. Also, in step S45, state data having at least one parameter different from the first state data are used as the second to seventh state data corresponding to the prediction flows F2 to F7, respectively.

[0105] Specifically, if the punch wear amount P1 in the first area A1 is 4 μm in the first status data, then the punch wear amount P1 in the first area A1 is set to 6 μm, which is one step larger than 4 μm, in the second status data. Similarly, if the punch wear amount P1 in the second area A2 is 6 μm in the first status data, then the punch wear amount P1 in the second area A2 is set to 8 μm in the third status data.

[0106] Similarly, in the fourth state data, the die wear amount D1 in area A1 is set to a value one level larger than that of the first state data, in the fifth state data, the die wear amount D2 in area A2 is set to a value one level larger than that of the first state data, in the sixth state data, the clearance C1 in area A1 is set to a value one level larger than that of the first state data, and in the seventh state data, the clearance C2 in area A2 is set to a value one level larger than that of the first state data.

[0107] When step S45 is executed, the tool state parameters are updated. In particular, the updated first state data indicates the latest results of the estimation process according to this embodiment. As described above, the predetermined period of step S43 is set to be longer than the machining cycle of cyclic machining, which is approximately equal to the execution interval of the process shown in FIG. 7. Therefore, when the CPU 1 reaches step S43 for the first time after completing step S45, it determines that the predetermined period has not elapsed since the tool state parameters were last updated (No in S43).

[0108] [2-3. Prediction of tool state change] Fig. 14 is a graph illustrating the results of the machining state estimation process according to this embodiment. The graph in Fig. 14 shows changes in the results of the machining state estimation process executed by the CPU 1 of the machining state estimation device 100 in Fig. 1 while cycle machining is repeatedly performed by the press machine 50 in Fig. 2.

[0109] 14, the horizontal axis of graphs (a) to (d) represents the number of machining cycles (number of shots) since the punch 73 and die 63 were replaced. The vertical axis of graph (a) represents the amount of punch wear, the vertical axis of graph (b) represents the amount of die wear, the vertical axis of graph (c) represents the clearance, and the vertical axis of graph (d) represents the workpiece thickness. In graphs (a) to (d), the actual values ​​of the parameters are shown by dashed lines, and the estimated values ​​(predicted values) calculated in the estimation process are shown by solid lines.

[0110] In graphs (a) to (d) of Fig. 14, symbol B1 represents the timing at which the determination in step S2 shown in Fig. 7 switches from No to Yes, and symbol B2 represents the timing at which the determination in step S43 shown in Fig. 12 switches from No to Yes.

[0111] The time corresponding to the number of shots indicated by B1 is, for example, T1. The time T1 between time 0 when the punch 73 and the die 63 are replaced and time T1 corresponds to the predetermined period of step S2 shown in Fig. 7. The time interval T2 between the time corresponding to the number of shots indicated by B1 and the time corresponding to the number of shots indicated by B2, and the time interval T2 between the times corresponding to the number of shots indicated by B2, correspond to the predetermined period of step S43 shown in Fig. 12.

[0112] In this embodiment, for example, the process of Fig. 7 is performed every time the CPU 1 acquires a measurement waveform from the load sensor 11. As a result, the execution interval of the process of Fig. 7 becomes approximately equal to the machining cycle of the cycle machining. In contrast, the predetermined period T2 of step S43 is set to a period longer than the machining cycle of the cycle machining.

[0113] 14, the CPU 1 can predict the value of the workpiece thickness, which may change in a shorter cycle than the tool state parameters, in a short cycle. On the other hand, for the punch wear amount, die wear amount, and clearance, which change more slowly than the workpiece state parameters, the CPU 1 predicts the values ​​in a cycle T2 longer than the prediction cycle of the workpiece thickness.

[0114] Generally, the waveform measured by the load sensor 11 as shown in Fig. 3 contains disturbance factors such as noise. Therefore, if all the state values ​​of the tool state parameters and the workpiece state parameters are predicted at once based on the waveform measured by the load sensor 11, errors in the prediction are likely to occur due to disturbance factors such as noise. In addition, to predict a combination of multiple parameters at once, the CPU 1 is required to perform calculations assuming all possible combinations of the parameters, i.e., 1225 possible states in this embodiment, resulting in an enormous amount of processing.

[0115] According to this embodiment, by efficiently varying the frequency of prediction and status data updates between tool state parameters that change over a long period and workpiece state parameters that change over a short period, the prediction accuracy can be improved while reducing the amount of calculations performed by the CPU 1. That is, in the prediction flow F1 of FIG. 12, the CPU 1 predicts only the workpiece thickness. In the prediction flows F2 to F7, the CPU 1 calculates the average degree of coincidence based on knowledge of workpiece thickness fluctuations, with only one of the tool state parameters differing from the first status data of the prediction flow F1. This calculation method reduces the prediction accuracy of the tool state parameters to a low degree of susceptibility to disturbance factors such as noise. Furthermore, the number of parameter combinations to be predicted in each of the prediction flows F1 to F7 can be reduced, thereby reducing the amount of calculations required of the CPU 1.

[0116] The machining state estimation device 100 may output the state data, which is the estimation result in step S33, as the latest state estimation result in cycle machining, by displaying it on a display or the like. Furthermore, if the clearance and / or workpiece thickness in the estimated state data is not within a predetermined range, a notification may be issued to the user. This allows the user to know whether tool maintenance, workpiece setting, etc. have been performed correctly. Such notification may be issued by, for example, lighting or flashing an LED in red, emitting a warning sound from a speaker, or the like.

[0117] Similarly, the machining state estimating device 100 may output, for example, the state data that is the estimation result in step S42 of the prediction flow F1 as the latest state estimation in cyclic machining, by displaying it on a display.

[0118] [3. Effects, etc.] As described above, the machining state estimation device 100 according to this embodiment estimates the machining state of the press machine 50 that repeatedly performs press working. The machining state estimation device 100 includes a CPU 1 that executes a first prediction process (F1) and a second prediction process (F2 to F7) related to estimation of the machining state in a first cycle, and a storage device 2. The storage device 2 stores workpiece state parameters and tool state parameters that define the machining state of the press machine 50. The workpiece state parameters are an example of first parameters, and the tool state parameters are an example of second parameters. The storage device 2 stores multiple pieces of reference data that respectively indicate changes in the machining load applied by the press machine 50 corresponding to each combination of the workpiece state parameters and the tool state parameters.

[0119] The CPU 1 acquires (S1) measurement data indicating the measurement results of the processing load by the press machine 50. The CPU 1 acquires (S31) first state data, which is an example of first and second current state parameters indicating the estimated results of the processing state of the press machine 50 at a predetermined reference time before the current measurement data was measured, for example, at the time of the previous measurement data. The first current state parameter corresponds to the workpiece state parameter, and the second current state parameter corresponds to the tool state parameter.

[0120] In a first prediction process (F1), the CPU 1 varies the workpiece state parameter within a predetermined first range from a first current state parameter, and fixes the tool state parameter to a second current state parameter, and searches for reference data that maximizes the similarity, which is an index of the degree of similarity with the measurement data, from a plurality of reference data corresponding to each combination of the varying workpiece state parameter and the fixed tool state parameter (S42). The CPU 1 determines the workpiece state parameter corresponding to the searched reference data and the fixed tool state parameter as first and second estimated parameters, respectively, that represent the machining state at the time of measuring the measurement data in the first prediction process (F1).

[0121] In the second prediction process (F2 to F7), the CPU 1 acquires progress state data (second to seventh state data), which are examples of progress state parameters generated by varying the second current state parameter within a predetermined second range. The CPU 1 varies the workpiece state parameter within a first range from the first estimated parameter, and fixes the tool state parameter to the progress state parameter, and searches for reference data that has the highest similarity to the measurement data from multiple reference data corresponding to each combination of the varying workpiece state parameter and the fixed tool state parameter (S42). The CPU 1 determines the workpiece state parameter corresponding to the searched reference data and the fixed tool state parameter as first and second estimated parameters, respectively, that represent the machining state at the time of measuring the measurement data in the second prediction process.

[0122] The CPU 1 outputs the first estimated parameter determined in either the first or second prediction process as an estimation result for the workpiece state parameter in a first cycle, and when a predetermined condition is satisfied, the CPU 1 outputs the second estimated parameter determined in either the first or second prediction process as an estimation result for the tool state parameter (S45).

[0123] According to this configuration, by fixing the tool state parameters in the first and second prediction processes, the amount of calculation by the CPU 1 can be reduced, and it becomes easier to estimate the machining state with high accuracy.

[0124] The CPU 1 may output the estimation result of the tool state parameter in a second cycle corresponding to the predetermined period when a predetermined period longer than the first cycle has elapsed since the previous estimation result of the tool state parameter was output (Yes in S43). With this configuration, the amount of calculation by the CPU 1 can be further reduced.

[0125] The CPU 1 may output the estimation result for the tool state parameter in a second cycle longer than the first cycle when press working has been performed a predetermined number of times or more since the previous output of the estimation result for the tool state parameter. With this configuration, the amount of calculation by the CPU 1 can be further reduced.

[0126] The tool condition parameters may include wear parameters that define the degree of wear on the tools of the press 50 .

[0127] The CPU 1 may change the second current state parameter related to the wear parameter within the second range by setting the second current state parameter related to the wear parameter to a value greater than the previous estimation result of the wear parameter (S34). This configuration can further reduce the amount of calculation by the CPU 1.

[0128] The CPU 1 may change the second current state parameter related to the wear parameter within the second range by setting the second current state parameter related to the wear parameter to the smallest value among one or more candidate values ​​that are greater than the previous estimation result for the wear parameter. This configuration can further reduce the amount of calculation performed by the CPU 1.

[0129] The CPU 1 may compare the average value of the similarity calculated in the first prediction process with the average value of the similarity calculated in the second prediction process, and output the second estimated parameter determined in the prediction process with the larger average value as the estimation result for the second parameter (S44, S45). This configuration makes it easier to estimate the machining state with high accuracy.

[0130] (Other Embodiments) As described above, the embodiments have been described as examples of the technology in the present disclosure. However, the technology in the present disclosure is not limited to these, and can be applied to embodiments in which modifications, substitutions, additions, omissions, etc. are made as appropriate. Furthermore, it is also possible to combine the components described in the above embodiments to create new embodiments. Therefore, other embodiments will be described below as examples.

[0131] (Variation 1) In the above embodiment, an example has been described in which the initial state estimation process S3 and the progress state estimation process S4 are switched depending on whether a predetermined period of time has elapsed since the tool was replaced (see FIG. 7 ). However, the present disclosure is not limited to this. For example, the CPU 1 may switch between the initial state estimation process S3 and the progress state estimation process S4 depending on whether the degree of match calculated in step S332 of the initial state estimation process S3 executed immediately before is within a predetermined range.

[0132] For example, if the degree of coincidence is within a predetermined range in step S332 (for example, if the degree of coincidence is equal to or greater than a predetermined value), the CPU 1 determines that a predetermined period of time has elapsed since the tool change (Yes in S2) and executes the progress state estimation process S4. On the other hand, if the degree of coincidence is not within the predetermined range, the CPU 1 determines that a predetermined period of time has not elapsed since the tool change (No in S2) and executes the initial state estimation process S3.

[0133] In this modification, if the state estimation is performed accurately by the initial state estimation process S3, it is possible to quickly proceed to the progress state estimation process S4. Therefore, in this modification, the estimation process can be performed accurately even if wear progresses quickly.

[0134] (Variation 2) In the above embodiment, an example has been described in which whether or not to execute steps S44 and S45 is switched depending on whether a predetermined period of time has elapsed since the tool state parameters were last updated (S43) (see FIG. 12). However, the present disclosure is not limited to this. For example, the CPU 1 may switch whether or not to execute steps S44 and S45 depending on whether or not the degree of coincidence calculated in step S422 of the workpiece thickness estimation process S42 executed immediately before is within a predetermined range.

[0135] For example, if the degree of match calculated in step S422 is within a predetermined range (e.g., if the degree of match is equal to or greater than a predetermined value), the CPU 1 determines that the predetermined period has not elapsed (No in S43) and ends step S4. On the other hand, if the degree of match is not within the predetermined range, the CPU 1 determines that the predetermined period has elapsed (Yes in S43) and executes steps S44 and S45.

[0136] In this modification, when the accuracy of the condition estimation by the prediction flow F1 decreases, the first to seventh condition data can be quickly updated to the condition data with the highest degree of agreement among the prediction flows F1 to F7. Therefore, even if wear progresses rapidly, the estimation process can be performed accurately to keep up with the progress.

[0137] (Variation 3) In the above embodiment, an example was described in which the determination of whether the loop processing in the workpiece thickness estimation processing S42 in FIG. 13 has converged (completed) (S423) is made based on whether all selectable candidate values ​​have been set in the provisional state data. However, the present disclosure is not limited to this. For example, the determination of whether the loop processing has converged may be made based on whether three candidate values ​​have been set in the provisional state data. The three candidate values ​​are, for example, the workpiece thickness T in the first state data (state data 22) acquired in step 41a and two candidate values ​​closest to the workpiece thickness T.

[0138] For example, if the work thickness T in the status data 22 acquired in step 42 is 50 μm, the determination of whether the loop processing has converged may be made based on whether the work thickness is set to 48 μm, 50 μm, and 52 μm in the provisional status data.

[0139] Generally, the variation in workpiece thickness between adjacent machining cycles in cycle machining may be small. Therefore, even if the variation range of the estimated workpiece thickness is small, high prediction accuracy can be maintained.

[0140] According to this modification, the number of loop processes required for the workpiece thickness estimation process S42 can be reduced, and the amount of calculation required of the CPU 1 can be further reduced.

[0141] (Variation 4) In the above embodiment, an example in which the punched outline is divided into areas A1 and A2 has been described. However, the present disclosure is not limited to this and can also be applied to cases in which the punched outline is not divided. When the punched outline is not divided, the processes S3311 and S3352 for generating area waveforms can be omitted, and the number of loops in the estimation process can be reduced. Therefore, the amount of calculation required of the CPU 1 can be reduced.

[0142] (Variation 5) In the above embodiment, the state data 22 includes the punch wear amount, die wear amount, clearance, and workpiece thickness as parameters (see FIG. 6 ), and the CPU 1 estimates these four parameters. However, the present disclosure is not limited to this, and may be configured to estimate at least one of the parameters of punch wear amount, die wear amount, and clearance in addition to the workpiece thickness. For example, even a machining state estimation device configured to estimate the workpiece thickness and punch wear amount can estimate the workpiece thickness and punch wear amount, and can estimate the workpiece thickness and punch wear amount more accurately than conventional methods.

[0143] (Variation 6) In the above embodiment, an example has been described in which the CPU 1 executes the reference waveform generation process S331 corresponding to the status data, but the present disclosure is not limited to this. For example, reference waveforms corresponding to all combinations of the areas A1 and A2 and the parameters may be calculated in advance by the CPU 1 or an external calculation device, and all calculated reference waveforms may be linked to the areas A1 and A2 and the combinations of the parameters and stored in advance in the storage device 2.

[0144] 9 and step S422 in Fig. 13, the CPU 1 calculates the degree of match between the reference waveform stored in the storage device 2 and the measured waveform acquired in step S1, and identifies the reference waveform with the highest degree of match. The identified reference waveform is linked to the areas A1 and A2 and the combination of parameters, so that the wear amount, clearance, and other parameters of each area can be estimated from the identified reference waveform.

[0145] According to this configuration, the CPU 1 does not need to generate a plurality of reference waveforms in real time, and therefore the processing load and processing time of the CPU 1 can be reduced.

[0146] (Variation 7) In the above embodiment, an example was described in which one of the punch wear amount, die wear amount, and clearance in each of the second to seventh state data differs from that in the first state data. However, the present disclosure is not limited to this. For example, the CPU 1 may execute a further prediction flow F8. In prediction flow F8, for example, the punch wear amount in both areas A1 and A2 is set to a value one level larger than the punch wear amount in the first state data. Similarly, a new prediction flow F9 may be further provided, in which one or more tool state parameters are set to differ from those in the first state data in one or more areas.

[0147] (Examples of Aspects) Examples of aspects of the present disclosure will be described below.

[0148] <Aspect 1> A processing state estimation device that estimates a processing state of a press machine that repeatedly performs press processing, comprising: a processor that executes a first prediction process and a second prediction process related to estimation of the processing state in a first cycle; and a storage device, wherein the storage device stores first and second parameters that define the processing state of the press machine, and a plurality of reference data that respectively indicate changes in processing load by the press machine corresponding to each combination of the first and second parameters, wherein the processor acquires measurement data that indicates a measurement result of the processing load by the press machine, and acquires first and second current state parameters that correspond respectively to the first and second parameters and indicate an estimation result of the processing state of the press machine at a predetermined reference time before the measurement of the measurement data, and wherein the processor performs the following in the first prediction process: the first parameter is varied within a predetermined first range from the first current state parameter, and the second parameter is fixed to the second current state parameter, and reference data corresponding to each combination of the varying first parameter and the fixed second parameter is searched for that reference data having a maximum similarity, which is an index of the degree of similarity with the measurement data; the first parameter corresponding to the searched reference data and the fixed second parameter are determined as first and second estimated parameters representing the processing state at the time of measuring the measurement data, respectively, in the first prediction process; and the processor, in the second prediction process, obtains a progress state parameter generated by varying the second current state parameter within a predetermined second range; the first parameter is varied within the first range from the first estimated parameter, and the second parameter is fixed to the progress state parameter, and reference data corresponding to each combination of the varying first parameter and the fixed second parameter is searched for that reference data having a maximum similarity with the measurement data; determining the first parameter corresponding to the searched reference data and the fixed second parameter as first and second estimated parameters representing the processing state at the time of measurement of the measurement data, respectively, in the second prediction process;the processor outputs the first estimated parameter determined in either the first or second prediction processing as an estimation result related to the first parameter in the first period, and the processor outputs the second estimated parameter determined in either the first or second prediction processing as an estimation result related to the second parameter when a predetermined condition is satisfied.

[0149] <Aspect 2> In the machining state estimating device according to Aspect 1, the processor outputs the estimation result for the second parameter when a predetermined period longer than the first period has elapsed since the processor output the previous estimation result for the second parameter, thereby outputting the estimation result for the second parameter in a second period corresponding to the period.

[0150] <Aspect 3> The machining state estimation device according to Aspect 1 or 2, wherein the processor outputs the estimation result for the second parameter in a second cycle longer than the first cycle by outputting the estimation result for the second parameter when the press working has been performed two or more predetermined times since outputting the previous estimation result for the second parameter.

[0151] <Aspect 4> The machining state estimating device according to any one of Aspects 1 to 3, wherein the processor outputs the second estimated parameter determined in either the first or second prediction process as an estimation result related to the second parameter when the similarity calculated in either the first or second prediction process is less than a predetermined threshold.

[0152] <Aspect 5> The machining state estimating device according to any one of Aspects 1 to 4, wherein the first parameter is a workpiece thickness parameter that defines a thickness of a workpiece machined by the press machine, and the second parameter is a tool state parameter that defines a state of a tool of the press machine.

[0153] <Aspect 6> The machining state estimating device according to aspect 5, wherein the tool state parameters include a wear parameter that defines a degree of wear of a tool of the press machine.

[0154] <Aspect 7> The machining state estimating device according to Aspect 6, wherein the processor changes the second current state parameter related to the wear parameter within the second range by setting the second current state parameter related to the wear parameter to a value greater than a previous estimation result related to the wear parameter.

[0155] <Aspect 8> The machining state estimating device according to Aspect 6, wherein the processor changes the second current state parameter related to the wear parameter within the second range by setting the second current state parameter related to the wear parameter to a minimum value among one or more candidate values ​​that are greater than a previous estimation result related to the wear parameter.

[0156] <Aspect 9> The machining state estimating device according to any one of Aspects 1 to 8, wherein the processor compares an average value of the similarities calculated in the first prediction process with an average value of the similarities calculated in the second prediction process, and outputs the second estimated parameter determined in the prediction process with the larger average value as an estimation result for the second parameter.

[0157] <Aspect 10> A processing state estimation method for estimating a processing state of a press machine that repeatedly performs press processing, comprising: a step by a processor acquiring measurement data indicating a measurement result of a processing load by the press machine; a step by the processor acquiring first and second current state parameters corresponding to first and second parameters that define the processing state of the press machine, respectively, and indicating an estimation result of the processing state of the press machine at a predetermined reference time before the measurement of the measurement data; a step by the processor executing a first prediction process and a second prediction process related to estimation of the processing state in a first cycle; and a step by the processor outputting estimation results related to the first and second parameters, wherein the first prediction process comprises: a step of varying the first parameter within a predetermined first range from the first current state parameter and fixing the second parameter to the second current state parameter, and searching for reference data that has a maximum similarity, which is an index of the degree of similarity to the measurement data, from a plurality of reference data indicating a change in the processing load corresponding to each combination of the varying first parameter and the fixed second parameter; the second prediction process includes a step of determining, in the first prediction process, the first parameter corresponding to the searched reference data and the fixed second parameter as first and second estimated parameters representing the processing state at the time of measurement of the measurement data, respectively; the second prediction process includes a step of acquiring a progress state parameter generated by changing the second current state parameter within a predetermined second range; a step of varying the first parameter from the first estimated parameter within the first range and fixing the second parameter to the progress state parameter, and searching for reference data that maximizes the similarity with the measurement data from the plurality of reference data corresponding to each combination of the varying first parameter and the fixed second parameter; and a step of determining, in the second prediction process, the first parameter corresponding to the searched reference data and the fixed second parameter as first and second estimated parameters representing the processing state at the time of measurement of the measurement data, respectively.The step of outputting the estimation results for the first and second parameters includes: a process of outputting the first estimated parameter determined in either the first or second prediction processing as the estimation result for the first parameter in the first period; and a process of outputting the second estimated parameter determined in either the first or second prediction processing as the estimation result for the second parameter when a predetermined condition is satisfied.

[0158] The present disclosure is applicable to presses.

[0159] REFERENCE SIGNS LIST 1 CPU (processor) 2 Storage device 3 Input interface 4 Output interface 11 Load sensor 21 Program 22 Status data 23 Waveform library 50 Press machine 51 Bolster 52 Slide 61 Die backing plate 62 Die plate 63 Die 71 Punch backing plate 72 Punch plate 73 Punch 74 Stripper plate 80 Workpiece 100 Machining status estimation device

Claims

1. A processing state estimation device for estimating a processing state of a press machine which repeatedly performs press processing, comprising: a processor which executes a first prediction process and a second prediction process relating to estimation of the processing state in a first cycle; and a storage device, wherein the storage device stores first and second parameters which define the processing state of the press machine, and a plurality of reference data which respectively indicate changes in processing load by the press machine corresponding to each combination of the first and second parameters, wherein the processor acquires measurement data which indicates a measurement result of the processing load, and acquires first and second current state parameters which correspond respectively to the first and second parameters and which indicate an estimation result of the processing state of the press machine at a predetermined reference time before the measurement of the measurement data, wherein the processor in the first prediction process varies the first parameter within a predetermined first range from the first current state parameter and fixes the second parameter to the second current state parameter, and searches for reference data which maximizes a similarity which is an index of the degree of similarity with the measurement data from the plurality of reference data which correspond to each combination of the varying first parameter and the fixed second parameter, the first parameter corresponding to the searched reference data and the fixed second parameter are respectively determined as first and second estimated parameters representing the processing state at the time of measurement of the measurement data in the first prediction process, and the processor, in the second prediction process, obtains a progress state parameter generated by changing the second current state parameter within a predetermined second range, varies the first parameter from the first estimated parameter within the first range and fixes the second parameter to the progress state parameter, and searches for reference data that maximizes the similarity to the measurement data from the multiple reference data corresponding to each combination of the varying first parameter and the fixed second parameter, and determines the first parameter corresponding to the searched reference data and the fixed second parameter as first and second estimated parameters representing the processing state at the time of measurement of the measurement data in the second prediction process,the processor outputs the first estimated parameter determined in either the first or second prediction processing as an estimation result related to the first parameter in the first period, and the processor outputs the second estimated parameter determined in either the first or second prediction processing as the estimation result related to the second parameter when a predetermined condition is satisfied.

2. The machining state estimation device according to claim 1, wherein the processor outputs an estimation result for the second parameter in a second cycle corresponding to the period by outputting the estimation result for the second parameter when a predetermined period longer than the first cycle has elapsed since the processor output the previous estimation result for the second parameter.

3. The processing state estimation device according to claim 1, wherein the processor outputs the estimation result for the second parameter in a second period longer than the first period by outputting the estimation result for the second parameter when the press processing has been performed two or more predetermined times since the output of the previous estimation result for the second parameter.

4. The machining state estimation device according to claim 1, wherein the processor outputs the second estimated parameter determined in either the first or second prediction process as an estimation result for the second parameter when the similarity calculated in either the first or second prediction process is less than a predetermined threshold value.

5. The machining state estimation device according to claim 1, wherein the first parameter is a work thickness parameter that specifies the thickness of the workpiece machined by the press machine, and the second parameter is a tool state parameter that specifies the state of a tool of the press machine.

6. The machining state estimation device according to claim 5, wherein the tool state parameters include a wear parameter that defines the degree of wear of a tool of the press machine.

7. The machining state estimation device according to claim 6, wherein the processor varies the second current state parameter for the wear parameter within the second range by setting the second current state parameter for the wear parameter to a value greater than a previous estimation result for the wear parameter.

8. The machining state estimation device according to claim 6, wherein the processor varies the second current state parameter for the wear parameter within the second range by setting the second current state parameter for the wear parameter to the smallest value among one or more candidate values ​​that are greater than a previous estimation result for the wear parameter.

9. The machining state estimating device according to claim 1, wherein the processor compares the average value of the similarity calculated in the first prediction process with the average value of the similarity calculated in the second prediction process, and outputs the second estimated parameter determined in the prediction process having the larger average value as the estimation result for the second parameter.

10. A processing state estimation method for estimating a processing state of a press machine which repeatedly performs press processing, comprising: a step in which a processor acquires measurement data indicating a measurement result of a processing load by the press machine; a step in which the processor acquires first and second current state parameters corresponding to first and second parameters defining the processing state of the press machine, respectively, which indicate an estimation result of the processing state of the press machine at a predetermined reference time before the measurement of the measurement data; a step in which the processor executes a first prediction process and a second prediction process relating to the estimation of the processing state in a first cycle; and a step in which the processor outputs the estimation results relating to the first and second parameters, wherein the first prediction process comprises: a step of varying the first parameter within a predetermined first range from the first current state parameter and fixing the second parameter to the second current state parameter, and searching for reference data which maximizes a similarity, which is an index of the degree of similarity to the measurement data, from a plurality of reference data indicating a change in the processing load corresponding to each combination of the varying first parameter and the fixed second parameter; the second prediction process includes a step of determining the first parameter corresponding to the searched reference data and the fixed second parameter as first and second estimated parameters representing the processing state at the time of measurement of the measurement data, respectively, in the first prediction process, wherein the second prediction process includes a step of acquiring a progress state parameter generated by changing the second current state parameter within a predetermined second range, a step of varying the first parameter from the first estimated parameter within the first range and fixing the second parameter to the progress state parameter, and searching for reference data having a maximum similarity to the measurement data from the multiple reference data corresponding to each combination of the varying first parameter and the fixed second parameter, and a step of determining the first parameter corresponding to the searched reference data and the fixed second parameter as first and second estimated parameters representing the processing state at the time of measurement of the measurement data, respectively, in the second prediction process,The machining state estimating method, wherein the step of outputting the estimation results for the first and second parameters includes: a process of outputting the first estimated parameter determined in either the first or second prediction processing as an estimation result for the first parameter in the first period; and a process of outputting the second estimated parameter determined in either the first or second prediction processing as an estimation result for the second parameter when a predetermined condition is satisfied.