Pressing device, method for detecting abnormalities in a pressing device, and abnormality detection program

Abstract

To suitably detect a clutch or brake abnormality.SOLUTION: A press device 1 includes: an eccentric shaft 16 which advances / retracts a slide 18 in the press direction following rotation; a clutch brake 13 which rotationally drives or brakes the eccentric shaft 16; a pressure gauge 453 which detects hydraulic oil pressure of the clutch brake 13; a storage part 52 which stores data 522 for abnormality determination being normal pressure data when the slide 18 advances / retracts in advance; and a control part 53 which determines the abnormality in the clutch brake 13 on the basis of the measurement value of the pressure detected by the pressure gauge 453 and the normal pressure data stored in the storage part 52.SELECTED DRAWING: Figure 4

Description

Technical Field

[0001] The present invention relates to a press apparatus, an abnormality detection method for a press apparatus, and an abnormality detection program.

Background Art

[0002] Conventionally, as a press apparatus, one provided with a wet clutch brake that performs rotational driving and braking of an eccentric shaft is known (see, for example, Patent Document 1).

[0003] Generally, since a wet clutch brake operates a hydraulic cylinder using a servo valve, the responsiveness of the cylinder fluctuates depending on various operating conditions such as oil temperature, which affects the pressing operation. As described above, since the sound operation of the clutch brake is important also for the mechanical safety of the entire apparatus, when an abnormality is found in the cylinder operation of the clutch brake, it is necessary to detect this abnormality promptly. Therefore, conventionally, when an actually measured pressure value does not reach a pressure command value for hydraulic oil of the clutch brake within a set time, it has been determined that an abnormality has occurred.

Prior Art Literature

Patent Literature

[0004]

Patent Document 1

Summary of Invention

Problem to be Solved by the Invention

[0005] However, in the above-mentioned conventional abnormality detection method, since an abnormality is detected based on the delay time of the response with respect to the pressure command value, a time lag corresponding to at least the delay time inevitably occurs when detecting an abnormality. Therefore, when an abnormality is detected, there is a possibility that a situation that may adversely affect the press operation, such as excessive overrun or clutch slippage, has already occurred.

[0006] This invention has been made in view of the above circumstances, and aims to suitably detect abnormalities in the clutch or brake. [Means for solving the problem]

[0007] The press apparatus according to the present invention is A rotating shaft that moves the slide forward and backward in the pressing direction as it rotates, A clutch or brake that rotates or brakes the aforementioned rotating shaft, A detection unit for detecting parameters related to the pressure of the clutch or brake working fluid, A storage unit that stores in advance normal data of the parameters when the slide moves forward and backward, A determination unit that determines an abnormality in the clutch or brake based on the measured values ​​of the parameters detected by the detection unit and the normal data stored in the storage unit, Equipped with, The determination unit determines the command value of the parameter The operation of the clutch or brake based on the above In each of the multiple time domains, The normal data and the measured values ​​associated with the variation in the time width of the time domain. The system compares these values ​​to determine if there is an abnormality in the clutch or brake.

[0008] The method for detecting abnormalities in a press device according to the present invention is: A rotating shaft that moves the slide forward and backward in the pressing direction as it rotates, A clutch or brake that rotates or brakes the aforementioned rotating shaft, A detection unit for detecting parameters related to the pressure of the clutch or brake working fluid, A storage unit that stores in advance normal data of the parameters when the slide moves forward and backward, A method for detecting abnormalities in a press device, comprising: The control unit of the press device performs a determination step to determine if there is an abnormality in the clutch or brake based on the measured value of the parameter detected by the detection unit and the normal data stored in the storage unit. In the determination step, the command value of the parameter The operation of the clutch or brake based on the above In each of the multiple time domains, The normal data and the measured values ​​associated with the variation in the time width of the time domain.The system compares these values ​​to determine if there is an abnormality in the clutch or brake.

[0009] The abnormality detection program for a press device according to the present invention is A rotating shaft that moves the slide forward and backward in the pressing direction as it rotates, A clutch or brake that rotates or brakes the aforementioned rotating shaft, A detection unit for detecting parameters related to the pressure of the clutch or brake working fluid, A storage unit that stores in advance normal data of the parameters when the slide moves forward and backward, An abnormality detection program for a press device, comprising: The control unit of the press device is configured to function as a determination unit that determines whether the clutch or brake is abnormal based on the measured values ​​of the parameters detected by the detection unit and the normal data stored in the storage unit. The determination unit determines the command value of the parameter The operation of the clutch or brake based on the above In each of the multiple time domains, The normal data and the measured values ​​associated with the variation in the time width of the time domain. The system compares these values ​​to determine if there is an abnormality in the clutch or brake. [Effects of the Invention]

[0010] According to the present invention, abnormalities in the clutch or brake can be suitably detected. [Brief explanation of the drawing]

[0011] [Figure 1] This figure shows the main body of the press apparatus according to this embodiment. [Figure 2] Figure 1 is a cross-sectional view of the clutch brake provided in the main body of the device. [Figure 3] This block diagram shows a schematic control configuration of a press apparatus according to this embodiment. [Figure 4] This flowchart shows the flow of the anomaly detection process in this embodiment. [Figure 5] This graph shows an example of the pressure change of the clutch brake's hydraulic fluid corresponding to the entire slide motion. [Figure 6] It is a graph showing an example of actually measured values of hydraulic oil pressure and normal pressure data in abnormality determination by the k-nearest neighbor method over the entire slide motion, where (a) is an example of a waveform when the state is normal, and (b) is an example of a waveform when the state is abnormal. [Figure 7] It is a graph showing an example of actually measured values of hydraulic oil pressure and normal pressure data in abnormality determination by the k-nearest neighbor method in the first region of the entire slide motion, where (a) is an example of a waveform when the state is normal, and (b) is an example of a waveform when the state is abnormal. [Figure 8] It is a graph showing an example of actually measured values of hydraulic oil pressure and normal pressure data in abnormality determination by the k-nearest neighbor method in the second region of the entire slide motion, where (a) is an example of a waveform when the state is normal, and (b) is an example of a waveform when the state is nearly abnormal. [Figure 9] It is a graph showing an example of actually measured values of hydraulic oil pressure and normal pressure data in abnormality determination by the k-nearest neighbor method in the third region of the entire slide motion, where (a) is an example of a waveform when the state is normal, and (b) is an example of a waveform when the state is nearly abnormal. [Figure 10] It is a graph showing an example of actually measured values of hydraulic oil pressure and normal pressure data in abnormality determination by the sum of squared error method, where (a) is an example of a waveform when the state is normal, and (b) is an example of a waveform when the state is abnormal. [Figure 11] It is a graph showing an example of a regression line and a normal range in abnormality determination by Gaussian process regression. [Figure 12] It is a graph showing an example of actually measured values and normal data when the parameter is the opening degree of a servo valve, where (a) is an example of a waveform when the state is normal, and (b) is an example of a waveform when the state is abnormal. MODE FOR CARRYING OUT THE INVENTION

[0012] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

[0013] [Configuration of Press Apparatus] Figure 1 shows the main body 100 of the press apparatus 1 according to this embodiment, and Figure 2 is a cross-sectional view of the clutch brake 13 provided in the main body 100. As shown in Figure 1, the press apparatus 1 according to this embodiment is a forging press apparatus that performs forging and comprises an apparatus body 100. The apparatus body 100 comprises a bed 23, an upright 22, a crown 21, a bolster 24, a slide 18, a rotation amount measuring instrument 35, and a drive unit 10.

[0014] The bed 23, upright 22, and crown 21 constitute the frame of the press machine 1. These bed 23, upright 22, and crown 21 are fastened to each other by inserting tie rods 25a into their interiors and tightening them with tie rod nuts 25b.

[0015] The bolster 24 is fixed on the bed 23, and multiple lower molds 32 are fixed to its upper part. The slide 18 is supported by a guide 19 provided on the upright 22 so that it can move up and down. Multiple upper dies 31 are fixed to the lower part of the slide 18. The multiple upper dies 31 and multiple lower dies 32 are arranged in corresponding pairs and are positioned along the left-right direction of the apparatus, with each pair facing the other in the upper and lower directions. As the slide 18 descends, the upper dies 31 and lower dies 32 move closer together, and the workpiece is forged between them. The direction in which the slide 18 moves up and down is not particularly limited, but in this embodiment, it will be described as moving up and down. A conveying device 40 is provided near the upper mold 31 and the lower mold 32. When the upper mold 31 and the lower mold 32 are separated, this conveying device 40 sequentially conveys the molded material to multiple sets of upper molds 31 and lower molds 32 arranged in a line.

[0016] The rotation amount measuring instrument 35 measures the amount of rotation (rotation angle) of the eccentric shaft 16 that moves the slide 18 forward and backward, and outputs the measured amount of rotation of the eccentric shaft 16 to the control device 50 described later (see Figure 3).

[0017] The drive unit 10 is configured to move the slide 18 forward and backward, and comprises a motor 11, a flywheel 12, a clutch brake 13, a transmission shaft 14, a reduction gear 15, an eccentric shaft 16, and a connecting rod 17. Of these, the transmission shaft 14 and the eccentric shaft 16 correspond to examples of rotating shafts according to the present invention.

[0018] The motor 11 is fixed to a frame part such as the crown 21. Power from the motor 11 is transmitted to the flywheel 12 via the belt 11a, causing the flywheel 12 to rotate. The flywheel 12 is rotatably supported and stores rotational energy. The transmission shaft 14 transmits the rotational motion of the flywheel 12 to the reduction gear 15. The reduction gear 15 reduces the rotational motion of the transmission shaft 14 and transmits it to the eccentric shaft 16. The eccentric shaft 16 has a main shaft portion 16a that is rotatably supported by a frame portion such as a crown 21 or upright 22 via a bearing 41. The eccentric shaft 16 has a hollow portion that runs through along the rotational axis, and the transmission shaft 14 is arranged within this hollow portion so as to be rotatable relative to the eccentric shaft 16. The connecting rod 17 connects the eccentric shaft 16 and the slide 18, converting the rotational motion of the eccentric shaft 16 into linear motion and transmitting it to the slide 18.

[0019] The clutch brake 13 is positioned on one axial side of the transmission shaft 14 and performs rotational driving and braking of the transmission shaft 14 (i.e., the eccentric shaft 16). Specifically, as shown in Figure 2, the clutch brake 13 is a separate type comprising a clutch 131 and a brake 132 that operate independently.

[0020] The clutch 131 can disconnect and reconnect the input shaft 133, which is fixed to the transmission shaft 14, and the flywheel 12, and has the function of switching between connecting and disconnecting power from the flywheel 12 to the transmission shaft 14 (eccentric shaft 16). This clutch 131 is a so-called wet clutch and comprises an outer hub 131a connected to the flywheel 12 and rotating in conjunction with it, an inner hub 131b fixed to the input shaft 133 and rotating in conjunction with it, a group of discs 131c in which multiple clutch discs Dc1 linked to the outer hub 131a and clutch discs Dc2 linked to the inner hub 131b are arranged alternately on top of each other, and a piston 131d.

[0021] The inner hub 131b has a hydraulic path formed therein that supplies hydraulic fluid to the piston 131d side and the disc group 131c side via the input shaft 133. A rotary joint 134 is provided at the tip of the input shaft 133 for circulating hydraulic fluid to the clutch 131 and brake 132. The rotary joint 134 is rotatably supported at the tip of the input shaft 133 and is connected to the hydraulic fluid piping of the hydraulic system 45 (see Figure 3). Even when the input shaft 133 is rotating relative to the rotary joint 134, hydraulic fluid can still flow between the piping of the hydraulic system 45 and the hydraulic path provided on the input shaft 133. Normally, the piston 131d is pressurized by a spring in a direction that separates it from the disc group 131c. When hydraulic fluid is supplied to the piston 131d side through the inner hub 131b, it moves in a direction that compresses the disc group 131c against the spring pressure. When the disc group 131c is compressed by the piston 131d, the clutch discs Dc1 and Dc2 come into frictional contact with each other, and the outer hub 131a and the inner hub 131b are connected, creating a connection that allows power transmission from the flywheel 12 to the input shaft 133 (transmission shaft 14). In this connection state, the rotational motion of the flywheel 12 is transmitted in the order of the transmission shaft 14, the reduction gear 15, and the eccentric shaft 16, and then converted into translational motion of the slide 18 via the connecting rod 17, causing the slide 18 to move up and down (see Figure 1). Furthermore, when the pressure-receiving oil of piston 131d is released, the spring pushes piston 131d back, causing clutch discs Dc1 and Dc2 to separate from each other. This then disengages the connection between outer hub 131a and inner hub 131b, cutting off power transmission from flywheel 12 to input shaft 133.

[0022] The brake 132 has the function of braking the rotation of the transmission shaft 14 (eccentric shaft 16). This brake 132 is a so-called wet brake and comprises an outer hub 132a fixed to a unit cover 135, an inner hub 132b fixed to an input shaft 133 and rotating in conjunction with it, a group of discs 132c in which multiple brake discs Db1 linked to the outer hub 132a and brake discs Db2 linked to the inner hub 132b are arranged alternately on top of each other, and a piston 132d.

[0023] The inner hub 132b has a hydraulic path formed therein that supplies hydraulic fluid to the piston 132d side via the input shaft 133. Normally, the piston 132d is pressurized by a spring (not shown) in a direction that presses it against the disc group 132c. As a result, the disc group 132c is compressed by the piston 132d, causing the brake discs Db1 and Db2 to come into frictional contact with each other, and the outer hub 132a and inner hub 132b become connected, resulting in a braking state for the input shaft 133. Then, when hydraulic fluid is supplied to the piston 132d through the inner hub 132b, it moves in a direction away from the disc group 132c against the spring pressure. As a result, the piston 132d is pushed back, and the brake discs Db1 and Db2 become separated from each other. Then, the connection between the outer hub 132a and the inner hub 132b is released, and the braking state of the input shaft 133 is released.

[0024] By applying a wet configuration to the clutch 131 and brake 132, high cooling performance for the disc group can be obtained. Due to this high cooling performance, a higher torque capacity can be obtained in both the clutch and brake functions compared to a dry configuration of the same size. Therefore, the wet configuration makes it possible to achieve the required torque capacity while reducing the size and weight of the clutch 131 and brake 132. Furthermore, because the clutch 131 and brake 132 are placed in a sealed space due to the wet configuration, they are not affected by dust generated during forging, resulting in high cleanliness, and in addition, vibration and noise can be reduced.

[0025] Figure 3 is a block diagram showing the schematic control configuration of the press machine 1. As shown in this figure, the press device 1 includes the main body 100, as well as a hydraulic device 45 and a control device 50.

[0026] As described above, the hydraulic system 45 is for supplying hydraulic fluid to the clutch brake 13. The hydraulic system 45 includes a hydraulic pump 451 that pressurizes the hydraulic fluid, at least one servo valve (control valve) 452 located in the fluid passage between the hydraulic pump 451 and the clutch brake 13 to control the supply of hydraulic fluid, and a pressure gauge 453. The hydraulic system 45 supplies high-pressure hydraulic fluid to the clutch brake 13 and discharges hydraulic fluid from the clutch brake 13 by driving a hydraulic pump 451 and a servo valve 452. The hydraulic piping connecting the hydraulic system 45 and the clutch brake 13 includes a fluid passage for supplying hydraulic fluid and a fluid passage for discharging hydraulic fluid. The pressure gauge 453 detects (measures) the pressure of the hydraulic fluid supplied to the clutch brake 13 and outputs it to the control device 50. The pressure gauge 453 may also be located inside the main body 100 of the device.

[0027] The control device 50 includes an alarm unit 51, a storage unit 52, and a control unit 53. Of these, the alarm unit 51 outputs an alarm to notify of an abnormality in the main unit 100 of the device. The manner of the alarm is not particularly limited; for example, an alarm display may be output to a display (not shown), or an alarm sound may be output from a speaker (not shown).

[0028] The memory unit 52 is a memory composed of, for example, RAM (Random Access Memory) or ROM (Read Only Memory), and stores various programs and data, as well as functioning as a work area for the control unit 53. In this embodiment, the memory unit 52 pre-stores an anomaly detection program 521 and anomaly determination data 522. The anomaly detection program 521 is a program for executing the anomaly detection process described later (see Figure 4).

[0029] The abnormality detection data 522 is data relating to the pressure of the hydraulic fluid supplied to the clutch brake 13, and is data for determining whether the measured value of said pressure is abnormal (determining whether it is abnormal or not). This abnormality detection data 522 may be data of the hydraulic fluid pressure when the slide 18 moves forward and backward, and data when the clutch brake 13 is operating normally (hereinafter referred to as "normal pressure data (normal data)"), or it may include an algorithm that performs abnormality detection using machine learning or the like with the normal pressure data. Alternatively, the abnormality detection data 522 may be a learning model constructed by machine learning using, for example, previously measured normal pressure data as training data. Learning techniques that can be used for the abnormality detection data 522 include, for example, support vector regression, Gaussian process regression, k-nearest neighbors, sum of squared errors method, neural networks, etc. The specific details of the anomaly detection process using anomaly detection data 522 will be described later.

[0030] The control unit 53 is composed of, for example, a CPU (Central Processing Unit) and provides integrated control of each part of the press machine 1. Specifically, the control unit 53 controls the operation of the motor 11 and the conveying equipment 40, drives the hydraulic system 45 to operate the clutch brake 13, and detects abnormalities in the main body 100 based on the output from the pressure gauge 453, etc.

[0031] [Anomaly detection processing] Next, we will explain the abnormality detection process that determines and detects abnormalities in the clutch brake 13. Figure 4 is a flowchart showing the flow of the anomaly detection process, and Figure 5 is a graph showing an example of the pressure change of the hydraulic fluid of the clutch brake 13 corresponding to the entire slide motion. Figures 6 to 9 are graphs showing an example of the measured value of the hydraulic fluid pressure and normal pressure data in the anomaly determination using the k-nearest neighbor method described later, of which Figure 6 is for the entire slide motion, and Figures 7 to 9 are for the first region R1 to the third region R3 of the whole, described later. Figure 10 is a graph showing an example of the measured value of the hydraulic fluid pressure and normal pressure data in the anomaly determination using the sum of squared errors method described later. Figure 11 is a graph showing an example of the regression line and normal range in the anomaly determination using Gaussian process regression described later. In Figures 6 to 10, (a) shows an example of the waveform in the normal case, and (b) shows an example of the waveform in the abnormal (or near abnormal) case.

[0032] The abnormality detection process is a process that detects abnormalities in the clutch brake 13 during the operation of the main unit 100 of the device. This abnormality detection process is performed, for example, when the press device 1 starts operation, by the control unit 53 reading and executing the abnormality detection program 521 from the storage unit 52.

[0033] As shown in Figure 4, first the control unit 53 starts operating the main unit 100 of the device (step S1). Specifically, in this operation, the control unit 53 rotates the flywheel 12 with the clutch brake 13 disengaging the flywheel 12 from the reduction gear 15 before the press working process. When the press working process starts in this state, the control unit 53 drives the clutch brake 13 according to the position of the slide 18 so that a preset slide motion is achieved. After the slide 18 completes one reciprocating motion with a predetermined slide motion, the material is pressed between the upper die 31 and the lower die 32, and then the slide 18 rises to separate the upper die 31 and the lower die 32, completing one press working process.

[0034] In the press working process described above, the control unit 53 drives the clutch brake 13 in four stages by appropriately switching the servo valve 452 of the hydraulic system 45. As shown in Figure 5, the four stages of driving include driving the soft clutch, full clutch, soft brake, and full brake in the order in which they are executed. Full clutch operation is the operation in which the pressure of pistons 131d and 132d of the clutch brake 13 (i.e., the hydraulic fluid pressure) is at its highest value (1st value). Full clutch operation allows a large torque to be transmitted from the flywheel 12 to the reduction gear 15, enabling the maximum load of the device body 100 to be applied to the slide 18. The soft clutch is driven in such a way that the hydraulic fluid pressure of the clutch brake 13 is lower than the first value, resulting in a second value, which allows the slide 18 to be accelerated when no large load is applied to the slide 18. The soft brake is driven in a third-value manner, where the pressure of the hydraulic fluid in the clutch brake 13 is less than the spring force, allowing the slide 18 to be decelerated with minimal impact. Full braking is achieved by releasing the pressure of the hydraulic fluid in the clutch brake 13, which allows the maximum braking force to be applied to the slide 18.

[0035] At this time, the control unit 53 outputs four pressure command values ​​corresponding to each drive state to the drive circuit of the hydraulic system 45, causing the drive circuit to drive the servo valve 452 by performing feedback control or feedforward control based on the measured value of the hydraulic fluid. This enables four stages of drive of the clutch brake 13. However, the change in the pressure command value and the change in the hydraulic fluid pressure (measured value) do not perfectly coincide, and the hydraulic fluid pressure follows the pressure command value with a predetermined time lag or predetermined pressure change.

[0036] Next, as shown in Figure 4, the control unit 53 measures (detects) the pressure of the hydraulic fluid supplied from the hydraulic system 45 to the clutch brake 13 when operation starts (step S2). Specifically, the control unit 53 acquires the measured value output from the pressure gauge 453 of the hydraulic system 45 and stores it in the memory unit 52. As described above, this measured value is also used to drive (control the opening degree of) the servo valve 452. In the following explanation, unless otherwise specified, the "measured value" of the hydraulic fluid pressure refers to the pressure value measured in step S2, and does not include the measured data used when creating normal pressure data or the measured data that constitutes the normal pressure data itself.

[0037] Next, the control unit 53 performs an abnormality determination of the clutch brake 13 (measured value) based on the measured pressure detected in step S2 and the abnormality determination data 522 stored in the storage unit 52 (step S3). In this embodiment, for example, normal pressure data is used as the abnormality detection data 522, and abnormality detection is performed using the k-nearest neighbor method. Specifically, in the k-nearest neighbor method, the distance d is first calculated using the following equation (1).

number

[0038] Furthermore, in each of the multiple regions (time intervals) corresponding to the pressure command value, the clutch brake 13 may be judged to be abnormal by comparing the measured value corresponding to that region with the normal pressure data. In this case, the normal pressure data is set in association with multiple regions. Specifically, for example, as shown in Figures 6(a) and (b), three regions may be set: a first region R1 from the start of driving the clutch brake 13 to the start of full clutching, a second region R2 from the start of full clutching to the start of soft braking, and a third region R3 from the start of soft braking to a predetermined time after the start of full braking. Then, as shown in Figures 7 to 9, in each of the first region R1, the second region R2, and the third region R3, the distance d may be calculated using equation (1), and the measured value may be judged to be abnormal by comparing the distance d with a threshold. If the time width of each region R1 to R3 changes (if the timing of the change in the pressure command value changes), for example, the normal pressure data may be extended or shortened to match the change in the time width. In this way, by performing abnormality detection of the measured value in the region corresponding to the pressure command value, abnormality detection of the clutch brake 13 can be suitably performed regardless of the operating state of the device body 100. That is, the time width of each region R1 to R3 may fluctuate according to the operating state of the device body 100, such as the temperature of each part that affects the rotation state of the slide 18. Therefore, if the measured value and the normal pressure data are simply compared at the same time position, a discrepancy (difference in the operating state being compared) will occur between the measured value and the normal pressure data in accordance with the change in the operating state of the device body 100, making comparison difficult. In this regard, by performing abnormality detection based on the comparison between the measured value and the normal pressure data in correspondence with each region R1 to R3 corresponding to the pressure command value, the occurrence of such discrepancies between the measured value and the normal pressure data can be suppressed, and abnormality detection of the clutch brake 13 can be suitably performed. The interval in which the above calculation is performed may be set to any quantity and time range. Furthermore, the threshold value may be changed according to the operating state of the device body 100 (for example, the oil temperature of the hydraulic fluid). In addition, the threshold value may be set relatively high at points where the measured value is likely to deviate from the normal pressure data (such as immediately after a change in the pressure command value), and relatively low at points where the measured value is likely to match the normal pressure data.

[0039] Alternatively, instead of the k-nearest neighbor method given by equation (1), the sum of squared errors method may be used as the machine learning algorithm for anomaly detection. In the sum of squared errors method, the sum of squared errors φ is calculated using the following equation (2).

number

[0040] Alternatively, the normal range set by, for example, Gaussian process regression may be used as the abnormality detection data 522. Specifically, as shown in Figure 11, for example, a regression equation can be obtained based on normal pressure data, and a Gaussian distribution representing the normal range can be set for it. In the example in Figure 11, a 95% confidence interval of ±1.96σ is shown as the normal range, but this range can be set arbitrarily. Then, an abnormality of the clutch brake 13 can be determined based on whether the measured value falls within the normal range or not. Furthermore, in this method, as with the k-nearest neighbor method, the measured values ​​may be compared with the normal range by associating them with multiple regions, or the interval in which the calculation is performed may be set to an arbitrary quantity and time range.

[0041] Furthermore, the various anomaly detection methods described above are preferably executed on a device capable of high-speed calculations, such as a PC (personal computer). However, in actual control devices 50, execution may be performed on devices with lower processing power than a PC, such as a PLC (Programmable Logic Controller). With a PLC, complex calculations are difficult, and securing memory space is also difficult because it also performs machine control simultaneously, making it difficult to compare (detect anomalies) across all measurement points. Therefore, in order to realize application to PLCs, a method of dividing the calculation time into small segments, commonly called search windows, can be considered. In the case of the clutch brake 13, the meaning of the search window differs depending on the section. The pressure rise section up to the soft clutch pressure is effective for detecting the stroke speed of the clutch brake 13 cylinder. After reaching the soft clutch pressure and full clutch pressure, it is effective for detecting slippage of the clutch 131. Depending on the location of the search window, it is also possible to detect the type of abnormality. Alternatively, the calculations performed by the machine learning algorithm (processing in step S3) may be performed by a high-speed computing device such as a PC, independent of the control device 50.

[0042] Next, as shown in Figure 4, the control unit 53 determines whether the measured pressure value is normal or not based on the result of step S3 (step S4). If it determines that it is normal (step S4; Yes), the control unit 53 proceeds to step S2 described above.

[0043] On the other hand, if in step S4 the measured pressure value is determined to be abnormal (step S4; No), the control unit 53 determines that an abnormality has occurred in the clutch brake 13 and activates the alarm unit 51 to notify the system of the abnormality (step S5). The notification method at this time is not particularly limited, and the notification method may be changed according to the calculation result of step S3, for example, by changing the intensity of the alarm according to the degree of deviation between the measured value and the normal value. In addition, the device body 100 may be stopped in conjunction with the notification of the occurrence of an abnormality.

[0044] Next, the control unit 53 determines whether or not to terminate the abnormality detection process (step S6). If it determines not to terminate the process (step S6; No), it proceeds to step S2 described above and continues detecting abnormalities. Then, if it is determined that the abnormality detection process should be terminated, for example, due to the termination of operation of the main unit 100 (step S6; Yes), the control unit 53 terminates the abnormality detection process.

[0045] [Technical effects of this embodiment] As described above, according to this embodiment, with respect to the hydraulic fluid pressure of the clutch brake 13, an abnormality determination of the clutch brake 13 is made based on the measured value detected by the pressure gauge 453 and the abnormality determination data 522 (normal pressure data) stored in the memory unit 52. In this way, since abnormality is determined by comparing the measured value with the normal value, the response delay from the pressure command value can be suppressed compared to the conventional method, which detected abnormalities based on the response delay time to the pressure command value. Therefore, abnormalities can be detected quickly, and consequently, abnormalities in the clutch brake 13, including the hydraulic device 45, can be suitably detected.

[0046] Furthermore, according to this embodiment, abnormality detection of the clutch brake 13 is performed using machine learning algorithms such as the k-nearest neighbor method, the sum of squared errors method, or Gaussian process regression. Therefore, a machine learning algorithm that utilizes past normal pressure data, which represents normal performance, can effectively detect abnormalities in the clutch brake 13.

[0047] Furthermore, according to this embodiment, the normal pressure data (abnormality determination data 522) is set in association with multiple regions R1 to R3 with different pressure command values, and in each of the multiple regions R1 to R3, the clutch brake 13 is determined to be abnormal by comparing the measured value corresponding to that region with the normal pressure data. This allows for optimal detection of abnormalities in the clutch brake 13, regardless of the operating state of the device body 100. In other words, while the time intervals of each region R1 to R3 may vary depending on the operating state of the device body 100, the influence of fluctuations in the time intervals of each region R1 to R3 can be suppressed by comparing the measured values ​​associated with each region R1 to R3 with normal pressure data.

[0048] [others] Although embodiments of the present invention have been described above, the present invention is not limited to the embodiments described above. For example, in the above embodiment, the pressure of the hydraulic fluid of the clutch brake 13 is detected, and an abnormality of the clutch brake 13 is determined based on that value. However, the parameter used for this abnormality determination can be any parameter related to the pressure of the hydraulic fluid of the clutch brake 13, and may be, for example, the opening degree of the servo valve 452 (servo opening degree), as shown in Figures 12(a) and (b). Alternatively, both the hydraulic fluid pressure and the servo opening degree may be used as parameters. In this case, for example, if the hydraulic fluid pressure is normal but the third opening degree is abnormal, a malfunction of the servo valve 452 may be suspected, and a judgment may be made according to the difference in the determination results of the two parameters.

[0049] Furthermore, in the above embodiment, an abnormality detection method is used by comparing measured values ​​with normal values. However, in addition to this, a conventional abnormality detection method based on the response delay time to the pressure command value may also be used in combination. In this case, if an abnormality is detected by either abnormality detection method, an alarm (or further device shutdown) may be issued. Alternatively, if an abnormality is detected by the abnormality detection method of the above embodiment, only an alarm may be issued, and if an abnormality is detected by the conventional abnormality detection method, the device may also be shut down. This is because the conventional abnormality detection method has a response delay to the command value compared to the abnormality detection method of this embodiment, raising concerns about more severe abnormalities.

[0050] Furthermore, if the measured pressure value detected in step S2 of the abnormality detection process is determined to be a normal value, it may be stored (added) to the abnormality determination data 522 as normal pressure data.

[0051] Furthermore, in the above embodiment, the clutch brake 13 is a separate type comprising a clutch 131 and a brake 132 that operate individually, but it may also be a combination type in which the clutch and brake are linked to each other. Furthermore, although the above embodiment involves determining an abnormality in the clutch brake 13, it is also possible to determine an abnormality in either the clutch or the brake. Moreover, the press device according to the present invention only needs to be equipped with at least one of the clutch or the brake.

[0052] Furthermore, details shown in the above embodiments can be modified as appropriate without departing from the spirit of the invention. [Explanation of Symbols]

[0053] 1 Pressing device 13. Clutch Brake 14. Transmission shaft (rotating shaft) 16. Eccentric axis (rotation axis) 18 slides 45 Hydraulic System 50 Control device 52 Storage section 53 Control Unit (Determination Unit) 100 Main unit of the device 131 Clutch 132 Brake 451 Hydraulic pump 452 Servo valve (control valve) 453 Pressure gauge (detection unit) 521 Anomaly Detection Program 522 Data for anomaly detection d distance R1 1st area R2 2nd area R3 3rd area φ sum of squares error

Claims

1. A rotating shaft that moves the slide forward and backward in the pressing direction as it rotates, A clutch or brake that rotates or brakes the aforementioned rotating shaft, A detection unit for detecting parameters related to the pressure of the clutch or brake working fluid, A storage unit that stores in advance normal data of the parameters when the slide moves forward and backward, A determination unit that determines an abnormality in the clutch or brake based on the measured values ​​of the parameters detected by the detection unit and the normal data stored in the storage unit, Equipped with, The determination unit compares the normal data and the measured value, which are associated with the variation in the time width of each of the multiple time domains corresponding to the operation of the clutch or brake based on the command value of the parameter, to determine if the clutch or brake is abnormal. Pressing device.

2. The determination unit determines whether the clutch or brake is abnormal based on the measured value and the normal data, using the k-nearest neighbor method, the sum of squared errors method, or Gaussian process regression. The press apparatus according to claim 1.

3. When the determination unit uses the k-nearest neighbor method or the sum of squared errors method, it calculates the distance or sum of squared errors using the k-nearest neighbor method, compares the distance or sum of squared errors with a preset threshold, and determines whether the clutch or brake is abnormal. The press apparatus according to claim 2.

4. When using Gaussian process regression, The storage unit has, as normal data, a regression equation for the parameters and a Gaussian distribution set as a normal range for the regression equation. The determination unit determines whether the clutch or brake is abnormal based on whether the measured value falls within the normal range. The press apparatus according to claim 2.

5. The system includes a control unit that controls the operation of the clutch or brake, The control unit outputs command values ​​for the parameters, which are divided into multiple time domains, when the slide moves forward and backward. The storage unit has the normal data set in association with the plurality of time domains, The determination unit compares the measured value corresponding to each of the plurality of time domains with the normal data to determine if there is an abnormality in the clutch or brake. The press apparatus according to claim 1.

6. The clutch or brake is a clutch brake that includes both a clutch and a brake. The aforementioned plurality of time domains include three domains: a first domain from the start of clutch brake operation to the start of full clutch operation; a second domain from the start of full clutch operation to the start of soft braking; and a third domain from the start of soft braking to the elapsed time after the start of full braking. The press apparatus according to claim 5.

7. The parameter includes at least one of the pressure of the hydraulic fluid supplied to the clutch or brake and the opening degree of a control valve that controls the supply of hydraulic fluid to the clutch or brake. The press apparatus according to claim 1.

8. The clutch or brake is of the wet type. The press apparatus according to claim 1.

9. A rotating shaft that moves the slide forward and backward in the pressing direction as it rotates, A clutch or brake that rotates or brakes the aforementioned rotating shaft, A detection unit for detecting parameters related to the pressure of the clutch or brake working fluid, A storage unit that stores in advance normal data of the parameters when the slide moves forward and backward, A method for detecting abnormalities in a press device, comprising: The control unit of the press device performs a determination step to determine if there is an abnormality in the clutch or brake based on the measured value of the parameter detected by the detection unit and the normal data stored in the storage unit. In the determination step, in each of a plurality of time domains corresponding to the operation of the clutch or brake based on the command value of the parameter, the normal data associated with the variation in the time width of the time domain and the measured value are compared to determine an abnormality in the clutch or brake. Method for detecting abnormalities in a press machine.

10. A rotating shaft that moves the slide forward and backward in the pressing direction as it rotates, A clutch or brake that rotates or brakes the aforementioned rotating shaft, A detection unit for detecting parameters related to the pressure of the clutch or brake working fluid, A storage unit that stores in advance normal data of the parameters when the slide moves forward and backward, An abnormality detection program for a press device, comprising: The control unit of the press device is configured to function as a determination unit that determines whether the clutch or brake is abnormal based on the measured values ​​of the parameters detected by the detection unit and the normal data stored in the storage unit. The determination unit compares the normal data and the measured value, which are associated with the variation in the time width of each of the multiple time domains corresponding to the operation of the clutch or brake based on the command value of the parameter, to determine if the clutch or brake is abnormal. Anomaly detection program for press machines.

Images