Process monitoring device and process monitoring method

Abstract

This invention provides a machining process monitoring device and a machining process monitoring method. The machining process monitoring device includes a data interception unit, a data length adjustment unit, and a judgment processing unit. The data interception unit determines the start and end times of each single cycle of machining, and intercepts first measurement data representing the first physical quantity changing over time during the cyclic machining process from the machining machine, for each single cycle. The data length adjustment unit adjusts the data length of the first measurement data to match the data length of judgment reference data that represents the change of the first physical quantity in a single cycle under normal machining operation, thereby generating second measurement data. The judgment processing unit compares the second measurement data with the judgment reference data to determine whether the machining machine is operating normally.

Description

Technical Field

[0001] This disclosure relates to a device for monitoring processing steps and a method for monitoring processing steps. Background Technology

[0002] In fields such as stamping, injection molding, NC machining, and industrial robotics, processing involves repeated, relatively short, single-cycle machining. The following technique is proposed: using sensors to detect the processing status of such machining machinery, and analyzing the time-series processing sensing data as the detection result to determine whether the machining machinery is operating normally.

[0003] For example, the method for generating judgment data for manufacturing device diagnosis disclosed in Patent Document 1 internally divides the state quantities representing the normal operation of the equipment and the state quantities representing the abnormal operation of the equipment into judgment values, which are then used to judge the processed sensing data, thereby providing accurate judgment results. Patent Document 2 discloses a data parsing device and data parsing method that provides the following method: making the positional relationship of the waveforms of the processed sensing data different to maximize the similarity with the main waveform, and making the positional relationship with the reference waveform appropriate to perform anomaly detection.

[0004] Prior art literature

[0005] Patent documents

[0006] Patent Document 1: Japanese Utility Model Application Publication No. 9-120365

[0007] Patent Document 2: Japanese Patent Application Publication No. 2020-86843 Summary of the Invention

[0008] One aspect of this disclosure provides a processing procedure monitoring device for monitoring a processing machine that performs repetitive single-cycle processing, and for determining whether the processing machine is operating normally. The processing procedure monitoring device comprises:

[0009] The data interception unit determines the start and end times of processing for each single cycle, and intercepts first measurement data representing the first physical quantity that changes over time during the cycle processing in the processing machine, for each single cycle, the first measurement data representing the first physical quantity from the start of processing to the end of processing.

[0010] The data length adjustment unit adjusts the data length of the first measurement data to match the data length of the judgment reference data, which represents the change of the first physical quantity in a single cycle under normal operating conditions of the processing machine, and generates the second measurement data; and

[0011] The judgment processing unit compares the second measurement data with the judgment benchmark data to determine whether the processing machine is operating normally.

[0012] Another aspect of this disclosure provides a method for monitoring a machining process, which monitors a machining machine performing repetitive single-cycle machining to determine whether the machining machine is operating normally. The machining process monitoring method includes:

[0013] The first physical quantity acquisition step involves sequentially acquiring the first physical quantity that changes over time during the cyclic processing in the processing machinery.

[0014] The data extraction step involves determining the start and end times of processing for each single cycle, and extracting first measurement data representing the first physical quantity from the start of processing to the end of processing for each single cycle of the first physical quantity.

[0015] The data length adjustment step involves adjusting the data length of the first measurement data to match the data length of the judgment benchmark data, which uses the change of the first physical quantity in a single cycle under normal operating conditions of the processing machinery as a reference, to generate the second measurement data; and

[0016] The judgment and processing step involves comparing the second measured data with the judgment benchmark data to determine whether the processing machine is operating normally. Attached Figure Description

[0017] Figure 1 This is a block diagram illustrating a structural example of the processing monitoring device 100 according to the first embodiment of this disclosure.

[0018] Figure 2 It is through Figure 1 This is a schematic diagram illustrating an example of data interception processing performed by the data interception unit 8.

[0019] Figure 3 It is used to illustrate the passage Figure 1 Oversampling processing performed by the oversampling filter 9, through Figure 1 A schematic diagram of the resampling process performed by the resampling unit 10.

[0020] Figure 4 It is through Figure 1 A schematic diagram illustrating the captured data before and after the resampling process performed by the resampling unit 10.

[0021] Figure 5 It means through Figure 1 A graph showing the results of the resampling process performed by the resampling unit 10.

[0022] Figure 6 This is a block diagram illustrating a structural example of the processing monitoring device 200 according to the second embodiment of this disclosure.

[0023] Figure 7 It is used for passing through Figure 6 A schematic diagram illustrating the compression process performed by the resampling unit 210.

[0024] Figure 8 This is a schematic diagram used to illustrate a variation of the second embodiment of this disclosure.

[0025] Figure 9 This is a block diagram illustrating a structural example of the processing monitoring device 300 according to the third embodiment of this disclosure.

[0026] Figure 10 It is a graph representing the waveform of sensed data in the blanking process based on an existing stamping press.

[0027] -Symbol Explanation-

[0028] 1. Height sensor

[0029] 2. Load sensor

[0030] 3, 4 controllers

[0031] 5 A / D converters

[0032] Filters 6 and 7

[0033] 8. Data Acquisition Department

[0034] 9. Oversampling filter (data interpolation section)

[0035] 10, 210 Resampling Unit (Data Length Adjustment Unit)

[0036] 11 Judgment and Processing Department

[0037] 12 Output Section

[0038] 13 Storage Department

[0039] 100, 200, 300 Processing Step Monitoring Device

[0040] 131, 231 are the benchmark data for judgment. Detailed Implementation

[0041] (The process of achieving this disclosure)

[0042] In existing technologies, when the processing time of a machining machine fluctuates during cyclic processing, there are concerns that abnormalities in the machine's operation may not be accurately detected. This issue will be addressed below.

[0043] An example of cyclic processing involving repeated identical operations is a cyclic processing method that repeatedly applies load to a workpiece such as metal to process it, for example, stamping, which involves blanking the workpiece. In a stamping press used for stamping, for example, a die is provided with tools such as punches and dies. The press conveys the workpiece by the up-and-down movement of a slide that is almost always in a fixed cycle, and repeatedly blanks the workpiece to form a predetermined shape.

[0044] Presses, for example, include crank-type presses with cranks that are controlled to rotate at a certain cycle. In crank-type presses, where high-speed machining is required to improve productivity, the SPM (Shots Per Minute), which represents the number of machining operations per minute, is set to a high value. The press rotates the crank at high speed to achieve the set SPM, converting this rotational motion into the up-and-down movement of a sliding member (slider), thus machining the material at high speed.

[0045] However, even when attempting to control the processing at a constant speed, it is difficult to maintain a consistently high processing speed with high precision, resulting in fluctuations in processing speed. These fluctuations in processing speed also lead to fluctuations in processing time. As the press becomes larger, the energy required for control increases, making it difficult to maintain speed with high precision. Therefore, the larger the press, the more significant the fluctuations in processing speed. Firstly, immediately after the press begins operation, the slide and the relatively heavy die mounted on the slide need to move from a stopped state, resulting in a slide speed that is lower than the set SPM (Speed ​​Per Minute).

[0046] Figure 10 It is a graph representing the waveform of sensed data in the blanking process based on an existing stamping press. Figure 10 (a) is a graph showing the change over time of the height of the template (e.g., the height from the template) as measured by a height sensor. Figure 10 (b) is a graph showing the change in the time of the load [N] applied to the punch as measured by the load sensor. Figure 10 of (a), Figure 10 The data set in (b) is obtained by overlaying multiple data sets acquired by sensors installed on the die of the stamping press.

[0047] Figure 10 of (a), Figure 10 The acquisition of data from (b) Figure 10 (a) The process begins when the demolding height is below the start threshold and ends when the process is above the end threshold. The start threshold and end threshold are... Figure 10In (a), the same threshold is set, but they may also be set to different values. Furthermore, Figure 10 The graph is obtained by overlaying multiple measurement data to match the positions at the start of the process.

[0048] exist Figure 10 The data represents measurements taken under normal processing conditions, i.e., when the machining machinery is performing a normal cyclic processing cycle. According to... Figure 10 It is known that even under normal processing conditions, the end time of processing is not fixed. This is because the processing time fluctuates, resulting in variations in the measured data, i.e., the length of the processing time. Furthermore, for... Figure 10 The punch load data shown in (b) indicates a short punching period for the workpiece, with steep variations in load during this period. Therefore, even under normal processing conditions, the waveform shape characterizing the punching process exhibits a temporal positional shift between multiple load data points. This positional shift becomes even more pronounced in the physical direction immediately after the press begins operation, during the period before the SPM stabilizes.

[0049] Therefore, in the prior art proposed in Patent Document 1, which uses whether or not an anomaly is detected by whether or not it falls within the range of a reference value, the reference value needs to be set to a wide range that takes into account the deviation in the waveform of a normal processing state, and thus cannot detect minor anomalies.

[0050] Furthermore, in the prior art proposed in Patent Document 2, which shifts the measured waveform along the time axis so that the peak position is the same as the reference waveform, even if the reference waveform and the measured waveform can be aligned locally, they cannot be consistent overall. Moreover, in the case of physical quantities where the peak appears large, the positional relationship can be consistent even locally, but in the case of physical quantities where the peak does not appear large, such as in the case of various physical quantities based on various sensors, the positional relationship cannot be consistent.

[0051] Furthermore, in existing technologies, to accurately capture steep changes in load data during the punching process, a higher sampling frequency is required in the A / D conversion process that converts analog signals to digital signals to refine the sampling. However, if the sampling frequency is high, the sampling of load data in time regions outside the slowly changing punching period, as well as the generally slowly changing height data of the die release, also becomes more refined, increasing the amount of data. Therefore, there is a problem of increased data processing volume and processing time required to detect anomalies.

[0052] The purpose of this disclosure is to provide a machining process monitoring device and method that can more accurately determine whether a machining machine is operating normally, even when the machining time of each individual cycle in a cyclic machining process fluctuates. The machining process monitoring device according to the embodiments described below solves the above-mentioned problems.

[0053] The embodiments will now be described in detail with appropriate reference to the accompanying drawings. However, unnecessary details may be omitted. For example, detailed descriptions of known matters and repetitive descriptions of substantially the same structures may be omitted. This is to avoid unnecessary redundancy in the following description and to facilitate understanding by those skilled in the art. Furthermore, the applicant has provided the drawings and the following description to enable those skilled in the art to fully understand this disclosure, but does not intend to limit the subject matter of the claims by these.

[0054] (First Embodiment)

[0055] [structure]

[0056] Figure 1 This is a block diagram illustrating a structural example of the processing monitoring device 100 according to the first embodiment of this disclosure. The processing monitoring device 100 includes: a height sensor 1, a load sensor 2, controllers 3 and 4, an A / D converter 5, and filters 6 and 7. The height sensor 1 and the load sensor 2 are mounted on a processing machine performing cyclic processing. The processing machine is, for example, a crank-type stamping press equipped with a mold containing tools such as punches and dies.

[0057] The height sensor 1 is, for example, an optical, radio wave, or ultrasonic distance sensor. In this embodiment, the height sensor 1 detects the height of the die-off platen of the stamping machine, such as the distance between the die and the die-off platen. The controller 3 outputs a voltage signal corresponding to the height detected by the height sensor 1. The load sensor 2 is, for example, a piezoelectric force sensor or an electrical force sensor such as a strain gauge, and detects the load applied to the punch of the stamping machine when the workpiece is being punched. The controller 4 outputs a voltage signal corresponding to the load detected by the load sensor 2. The height detected by the height sensor 1 and the load detected by the load sensor 2 are examples of the "first physical quantity" of this disclosure. The height detected by the height sensor 1 is an example of the "second physical quantity" of this disclosure. The second physical quantity is information representing the position of a component of the processing machine, such as the die-off platen, that moves during a single cycle of the processing machine to apply a load to the workpiece. The controllers 3 and 4 are examples of the "physical quantity acquisition unit" of this disclosure.

[0058] A / D converter 5 converts the input voltage signal into a digital signal. A / D converter 5 operates at a sampling frequency Fs (samples per second). That is, A / D converter 5 converts the voltage signals received from controllers 3 and 4 into digital signals and outputs them at sampling intervals of 1 / Fs (seconds).

[0059] Filters 6 and 7 filter the digital signals corresponding to height sensor 1 and load sensor 2, respectively. Filters 6 and 7 include low-pass filters (LPFs), median filters, etc., and have the function of removing noise from the input data. Filters 6 and 7 remove noise from height sensor 1 and load sensor 2 themselves, as well as noise mixed in the path from the output of height sensor 1 and load sensor 2 to the input of filters 6 and 7. If the noise is small enough to be negligible, at least one of filters 6 and 7 can be omitted. In the case of omission, the phase of the signals is made consistent by using a delay unit, etc., so that only one signal is delayed.

[0060] The processing monitoring device 100 also includes: a data acquisition unit 8, an oversampling filter 9, a resampling unit 10, a judgment processing unit 11, an output unit 12, and a storage unit 13.

[0061] The data acquisition unit 8 performs alignment processing on multiple signals s1 from the height sensor 1 and multiple signals s2 from the load sensor 2 to ensure that the conditions for the start and end of processing are consistent, thus setting this as the first captured data (Shot data). Details of the data acquisition processing performed by the data acquisition unit 8 will be described later. The data acquisition unit 8 outputs the first captured data to the oversampling filter 9. The first captured data is an example of the "first measurement data" of this disclosure.

[0062] The oversampling filter 9 oversamples the digital signal groups contained in the first captured data at a sampling frequency of 8Fs. The digital signal groups contained in the first captured data include signals from the height sensor 1 and signals from the load sensor 2. The details of the oversampling (sometimes referred to as upsampling) performed in the oversampling filter 9 are described later. The oversampling filter 9 is an example of the "data interpolation unit" of this disclosure. The oversampling filter 9 is, for example, a low-pass filter, such as an FIR (Finite Impulse Response) filter.

[0063] The storage unit 13 is a recording medium that records various information, including programs and data required to realize the functions of the processing monitoring device 100. The storage unit 13 can be, for example, any of a hard disk drive (HDD), an optical drive, or a solid-state drive (SSD). The storage unit 13 can be internal, external, or a NAS (network-attached storage) type, and can also be implemented via cloud computing.

[0064] Storage unit 13 stores, for example, judgment reference data 131. Judgment reference data 131 is information that represents changes in measured data during a single cycle of operation of the machining machine under normal conditions. Judgment reference data 131 may be waveform data such as height data and load data under normal machining conditions, serving as a reference for judgments such as anomaly detection. Judgment reference data 131 may also include auxiliary data such as upper and lower limits of this waveform data. Judgment reference data 131 includes a reference data length representing the data length of judgment reference data 131. "Data length" refers, for example, to the length of data as a number of samples. For example, the more samples, the longer the data length.

[0065] The resampling unit 10 receives first captured data that has passed through the oversampling filter 9 and a reference data length based on the judgment reference data 131. The resampling unit 10 performs resampling processing on the first captured data that has passed through the oversampling filter 9 to make its data length consistent with the reference data length, generating second captured data. The resampling unit 10 outputs the second captured data to the judgment processing unit 11. The resampling unit 10 is an example of the "data length adjustment unit" of this disclosure, and the second captured data is an example of the "second measurement data" of this disclosure.

[0066] The judgment processing unit 11 performs judgment processing such as anomaly detection based on the second captured data and outputs a judgment result. For example, the judgment processing unit 11 compares the second captured data with the judgment reference data 131, and judges that an anomaly exists if the comparison result is not between the upper limit and the lower limit.

[0067] Output unit 12 is an output interface for outputting judgment results. Output unit 12 can be, for example, a light-emitting device such as an LED that emits light to report certain information to the user. Output unit 12 can also be a display device such as a liquid crystal display or an organic EL display capable of displaying information. Output unit 12 can also be a sound output device such as a speaker that reports information via sound. Output unit 12 can also be a communication interface for enabling communication between the processing monitoring device 100 and external devices. This communication interface communicates according to existing wired or wireless communication standards.

[0068] Each structural element of the processing monitoring device 100 can be implemented by a circuit corresponding to each structural element, or by an arithmetic circuit. Such an arithmetic circuit includes a CPU (Central Processing Unit), RAM (Random Access Memory), ROM (Read Only Memory), etc., and controls the operation of the processing monitoring device 100, such as the data capture unit 8, oversampling filter 9, resampling unit 10, and judgment processing unit 11, based on information processing. This information processing is achieved by executing a program through the arithmetic circuit. The arithmetic circuit can also be implemented by one or more dedicated processors. Furthermore, regarding the structural elements of the arithmetic circuit, functions may be appropriately omitted, replaced, or added depending on the implementation method.

[0069] Figure 1 The document describes various structural elements that the machining process monitoring device 100 can have, but having all of them is not mandatory. When operating, the machining process monitoring device 100 only needs to have at least a data acquisition unit 8, an oversampling filter 9, a resampling unit 10, a judgment processing unit 11, and judgment reference data 131. In other words, the machining process monitoring device 100 may not have sensors 1 and 2, controllers 3 and 4, an A / D converter 5, filters 6 and 7, and part or all of the output unit 12. When the machining process monitoring device 100 operates, it only needs to be able to acquire signals from sensors 1 and 2, controllers 3 and 4, an A / D converter 5, and filters 6 and 7.

[0070] [action]

[0071] Figure 2 It is through Figure 1 This is a schematic diagram illustrating an example of data interception processing performed by the data interception unit 8. Figure 2 (a) represents the signal s1, which originates from the height sensor 1, input to the data interception unit 8. Signal s1 represents the signal of the state before being intercepted by the data interception unit 8. Figure 2 (b) represents the signal output from the data capture unit 8 corresponding to signal s1. Figure 2 The signal shown in (b) is generated as a result of data interception processing based on the data interception unit 8 performed on signal s1. Figure 2 The signal shown has passed through A / D converter 5, therefore it is a discrete signal, but... Figure 2 In the middle, it is used to represent continuously in a schematic way.

[0072] Specifically, the data interception department has 8 pairs of... Figure 2The height sensor 1 shown in (a) monitors the source signal s1 and determines when the strength of signal s1 is below the start threshold (when processing begins) and above the end threshold (when processing ends). Next, the data capture unit 8 captures the waveform of signal s1 from the start of processing to the end of processing. The start threshold and end threshold are... Figure 2 The same threshold is set in (a), but they can also be set to different values.

[0073] exist Figure 2 In this paper, only the data interception processing for signal s1 originating from height sensor 1 is described. The data interception unit 8 also performs data interception processing for signal s2 originating from load sensor 2. That is, for signal s2, the data interception unit 8 executes the waveform from the same processing start time to the processing end time as signal s1. The signals s1 and s2 intercepted by the data interception unit 8 constitute the first captured data. As described above, signals s1 and s2 are aligned to data from the same processing start time to the processing end time, so the data length of all waveforms constituting the first captured data is the same.

[0074] Figure 3 It is used for passing through Figure 1 Oversampling processing performed by the oversampling filter 9, through Figure 1 A schematic diagram illustrating the resampling process performed by the resampling unit 10. Figure 3 (a) represents an example of captured data output from oversampling filter 9. Figure 3 (b) represents an example of the second captured data output from the resampling unit 10.

[0075] exist Figure 3 In (a), black circles represent sampling points sampled at sampling frequency Fs, and white circles represent oversampled sampling points. Correspondingly, in Figure 3 In (a), the solid line represents the measurement sampling timing based on the sampling frequency Fs, and the dashed line represents the oversampling timing based on 8 times the sampling frequency 8Fs.

[0076] As described, the oversampling filter 9 is, for example, an FIR-based low-pass filter. Preferably, the oversampling filter 9 does not produce a phase lag relative to the input signal in the output signal. Furthermore, the filter constant of the oversampling filter 9 is preferably set such that the cutoff frequency is almost the same as half of the sampling frequency Fs (the Nyquist frequency), which is obtained, for example, as the reciprocal of the reference data length.

[0077] like Figure 3As shown in (a), between the sampling points (black circle symbols) sampled at the sampling frequency Fs, the high-frequency components above the cutoff frequency are interpolated by the oversampling filter 9 to form a smooth curve of 7 points (white circle symbols).

[0078] Next, the resampling unit 10 targets the data that has passed through the oversampling filter 9. Figure 3 The first captured data, as shown in (a), is subjected to resampling processing to make its data length consistent with the reference data length, generating... Figure 3 The second captured data is shown in (b). The resampling unit 10 will... Figure 3 The period 1 / Fs shown by the solid line in (a) is the sampling time resampling time. Figure 3 The solid line in (b) indicates the timing of the sampling. Figure 3 In (b), the resampling frequency is represented by Fr. Figure 3 The sampling timing of (b) is determined as the reciprocal of the reference data length.

[0079] in addition, Figure 3 The illustration shows a case where the processing time in the first captured data (i.e., the data length of the first captured data) is slightly shorter than the processing time in the judgment reference data 131 (i.e., the reference data length of the judgment reference data 131). Regarding the reference data length and the data length of the first captured data, even if the sampling period is the same, they may not be consistent due to fluctuations in the processing time. Figure 3 As shown in the example, the length of the first captured data may be shorter than the length of the reference data.

[0080] Therefore, the resampling unit 10 performs resampling processing to make the data lengths of both data identical. Specifically, the sampling period of the first captured data is adjusted so that the start and end times of processing the first captured data are identical to the start and end times of processing the judgment reference data 131, respectively. As a result, the data length of the first captured data is the same as the reference data length of the judgment reference data 131. Figure 3 In the example shown, the start and end times of processing the data adjusted to be the first captured data are consistent with the start and end times of processing the judgment reference data 131, respectively. As a result, in this embodiment, Figure 3 In (a), the measurement sampling time interval shown by the solid line is converted to a slightly larger one.

[0081] Furthermore, from the sampling points that were oversampled by 8 times, that is, from Figure 3 Among the black and white circle symbols in (a), choose the one closest in time. Figure 3 In (b), the sampling points used to determine the sampling timing, indicated by the solid line, are denoted as […]. Figure 3The second captured data, resampled and indicated by the black circle symbol in (b), is thus obtained. Figure 3 In (b), the interpolation value with the smaller error corresponding to the sampling timing is indicated by the solid line.

[0082] In this way, by performing oversampling processing, it is not necessary to set the sampling frequency Fs of the A / D converter 5 to be extremely high than the rate of change of the physical quantity for the sake of resampling processing. Therefore, the sampling frequency Fs of the A / D converter 5 can be suppressed to a low level, which simplifies the hardware structure, reduces the amount of data to be processed, and speeds up the decision processing. In addition, even when the data length is not constant due to fluctuations in processing time, the resampling unit 10 can convert the first captured data into a second captured data with the same length as the reference data, so that it is always consistent at the beginning and end of processing.

[0083] Figure 4 It is through Figure 1 A schematic diagram illustrating the captured data before and after the resampling process performed by the resampling unit 10. Figure 4 (a) represents the captured data before resampling based on the resampling unit 10. Figure 4 The captured data in (a) is signal information from load sensor 2. Figure 4 (b) indicates by means of Figure 4 The captured data of (a) is the second captured data obtained by performing resampling processing based on the resampling unit 10. Figure 4 (c) represents an example of the benchmark data 131. Figure 4 The data shown in (a), (b), and (c) are, for example, time series data arranged at equal intervals along the time axis.

[0084] In making Figure 4 The processing of (a) captured data begins with Figure 4 If the processing starts in the judgment benchmark data 131 of (c) are consistent, then... Figure 4 (a) The processing of Ta at the end of the captured data is... Figure 4 The processing time difference ΔT is caused by fluctuations in processing time between the processing end Tc and the judgment benchmark data 131 in (c). Furthermore, if for Figure 4 The captured data of (a) and Figure 4 If the judgment benchmark data 131 of (c) is compared, the number of samples from the start of processing to the end of processing will also be different.

[0085] In contrast, such as Figure 4 As shown in (b), the second captured data obtained by performing resampling processing based on the resampling unit 10 has the same characteristics as... Figure 4 The second captured data (c) has the same data length as the judgment reference data 131, and the number of samples taken from the start to the end of processing is also the same. Thus, the resampling unit 10 makes the data length of the second captured data, which is the object of judgment, consistent with the data length of the judgment reference data 131, which is the basis for judgment. Therefore, the resampling unit 10 can make the peak position appearing in the second captured data consistent with the peak position in the judgment reference data 131. Here, the peak position, for example, refers to the time when the peak appears in the data, and refers to the time based on the start of processing. Therefore, even if the processing time fluctuates, the judgment processing unit 11 can perform judgments based on the second captured data compared with the judgment reference data 131 with high accuracy.

[0086] Figure 5 This is a graph representing the result of resampling, which reduces the deviation of the waveform representing the punching process in the second captured data. Figure 5 (a) is obtained in each cycle of the cyclic processing, and will Figure 4 The graph shown in (b) is an overlay of the second captured data. Figure 5 In (a), only the peak position and its surrounding area are enlarged in the second captured data. Figure 5 (b) is for Figure 5 (a) shows a graph of multiple second-capture data, representing the frequency of the load at sampling time t1. Figure 5 In (b), the horizontal axis represents frequency, and the vertical axis represents the load or the signal level corresponding to the load.

[0087] Figure 5 The sampling time t1 shown in (a) is the rising portion of the waveform. In the prior art, the rising portion of the waveform is the part of the waveform that shows significant changes when abnormalities occur in the cyclic processing or when the processing time of the cyclic processing fluctuates.

[0088] As described above, the resampling unit 10 makes the data length of the second captured data, which is the object of judgment, consistent with the data length of the judgment reference data 131, which is the basis for judgment. Therefore, even if the processing time of the cyclic processing fluctuates, the resampling unit 10 can ensure that the waveforms, such as peak positions, which are characteristics of the cyclic processing, are consistent across the multiple waveform data corresponding to multiple cycles. Thus, as... Figure 5 As shown in (b), the frequency of the load at sampling time t1 between multiple waveform data has a steep peak. That is, the resampling unit 10 can reduce the standard deviation of the load at sampling time t1 between multiple waveform data.

[0089] Therefore, when the machining process monitoring device 100 makes judgments based on the waveform data of the judgment reference data 131, such as setting an upper limit and a lower limit, and performing anomaly detection on the second captured data, it can make judgments with high accuracy even when the difference between the upper limit and the lower limit is set relatively narrow. In addition to judgments using the upper and lower limits, for example, when making judgments based on statistical data such as cosine similarity and Mahalanobis distance, the machining process monitoring device 100 can reduce the impact of fluctuations in machining time and make judgments with high accuracy.

[0090] The output unit 12 outputs the judgment result based on the judgment processing unit 11. For example, the output unit 12 is a display showing the judgment result. Thus, the user operating the processing machine can know the judgment result. Alternatively, the output unit 12 may also output the judgment result to the processing machine. For example, if a judgment result indicating an abnormality is input, the processing machine can stop operating, preventing the production of abnormal processed products, malfunctions of the processing machine, accidents caused by abnormal operation, etc.

[0091] [Effects, etc.]

[0092] As described above, the machining process monitoring device 100 according to this embodiment monitors a machining machine that performs repeated single-cycle machining and determines whether the machining machine is operating normally. The machining process monitoring device 100 includes: a data capture unit 8, a resampling unit 10 (an example of a data length adjustment unit), and a judgment processing unit 11. The data capture unit 8 determines the start and end times of each single-cycle machining process and captures first captured data from the start to the end of each single-cycle machining process, based on a first physical quantity that changes over time during the cyclic machining process. The resampling unit 10 adjusts the data length of the first captured data to match the data length of the judgment reference data 131, thereby generating second captured data. The judgment reference data 131 is information represented by the change of the first physical quantity in a single cycle under the condition that the machining machine is operating normally. The judgment processing unit 11 compares the second captured data with the judgment reference data 131 to determine whether the machining machine is operating normally. The first physical quantity is, for example, load.

[0093] With the above structure, even if the processing time of each single cycle in the cyclic processing fluctuates, since the resampling unit 10 makes the data length of the second captured data consistent with the data length of the judgment reference data 131, it is possible to determine whether the processing machine is operating normally more accurately than the prior art.

[0094] The data capture unit 8 can also determine the start and end times of processing based on a second physical quantity that changes at the start and end of processing in the processing machine. The second physical quantity is information representing the position of a component of the processing machine that moves to apply a load to the material being processed in a single cycle, such as the height of the die-off plate of a stamping press.

[0095] With the above structure, the start and end times of each single cycle in the cyclic processing can be aligned with good accuracy, and the resampling unit 10 can make the data length of the second captured data consistent with the data length of the judgment reference data 131 with better accuracy.

[0096] The processing monitoring device 100 may also include an oversampling filter 9. The oversampling filter 9 interpolates the data between multiple data points representing the time change of the first physical quantity contained in the first captured data. The resampling unit 10 adjusts the data length of the first captured data to be consistent with the data length of the judgment reference data 131 by using the multiple data points and the data interpolated by the oversampling filter 9, thereby generating the second captured data.

[0097] With the above structure, the amount of data for the first physical quantity can be reduced due to the processing of interpolation data. Therefore, the structure of the processing monitoring device 100 can be simplified and the cost reduced, the amount of data reduced and the processing workload decreased, and the judgment processing speed increased.

[0098] (Second Implementation)

[0099] Figure 6 This is a block diagram illustrating a structural example of the processing monitoring device 200 according to the second embodiment of this disclosure. Figure 6 Compared to the processing monitoring device 200 Figure 1 The processing monitoring device 100 replaces the resampling unit 10 and includes a resampling unit 210. Furthermore, Figure 6 In the storage unit 13 of the processing monitoring device 200, replacing Figure 1 The judgment benchmark data 131 is used to save the judgment benchmark data 231.

[0100] The resampling unit 210 compresses at least a portion of the data output from the oversampling filter 9. For example, the resampling unit 210 does not compress the waveform portion representing the processing characteristics in the first captured data, but only compresses other waveform portions.

[0101] Figure 7 It is used for the purpose of Figure 6 A schematic diagram illustrating the compression process performed by the resampling unit 210. Figure 7 (a) is a graph representing the data input to the resampling unit 210.

[0102] In this embodiment, the first captured data is divided into a complex number of n intervals (where n is an integer greater than 2). Figure 7 In the example shown in (a), the first captured data is divided into: from the start of processing t a0 At time t a1 The first interval S a1 Data D1 from time t a1 At time t a2 The second interval S a2 Data D2 from time t a2 t at the end of processing a3 The third interval S a3 Data D3 in the data is segmented by either the data truncation unit 8 or the resampling unit 210. Figure 7 In the example shown in (a), the first interval S a1 This represents the interval from the start of processing until the punch descends and reaches the material being processed. Interval S (2nd interval) a2 This corresponds to the interval where the punch cuts the material. Interval 3, S a3 This corresponds to the period from when the material is punched by the punch until the end of the processing.

[0103] Each interval S a1 S a2 S a3 The length of time t a1 t a2 t a3 The value is preset based on the type of machining based on the machining machinery. For example, in a stamping-based cyclic machining process, the timing of the punch cutting the material in each machining cycle is determined according to the die structure, assuming the machining starts at t. a0 Using this as a benchmark, it can be defined as a certain range. Therefore, the interval S can be defined as... a1 S a2 S a3 The ratio of their lengths is set to a fixed value. Figure 7 In the example shown in (a), the interval S a1 S a2 S a3 The length ratio is set to 2:1:7.

[0104] The first captured data is preferably segmented into portions with large time variations (in... Figure 7 In the example shown in (a), the second interval S is... a2 ), parts with small changes (in Figure 7 In the example shown in (a), the first interval S is... a1 and the third interval S a3 ).

[0105] Figure 7 (b) indicates that by oversampling the Figure 7 The first captured data in (a) is compressed using compression processing based on the resampling unit 210, resulting in a graph of the compressed data. The compressed data has: from the start of processing t b0 At time t b1 The first interval S b1 From time t b1 At time t b2 The second interval S b2 From time t b2 t at the end of processing b3 The third interval S b3 The compressed interval S b1 S b2 S b3 Corresponding to Figure 7 The uncompressed interval S of (a) a1 S a2 S a3 .

[0106] The resampling unit 210 compresses the portions of the oversampled first captured data that have small temporal variations, while not compressing the portions with large variations, or compressing them at a smaller compression ratio than the portions with small variations. For example, the resampling unit 210 compresses the oversampled data... Figure 7 The interval S of (a) a1 S a2 S a3 The waveform data in the image are compressed using compression ratios C1, C2, and C3 respectively, and thus set as follows: Figure 7 The interval S of (b) b1 S b2 S b3 The compressed data. Here, the compression ratio is the ratio of the amount of data before compression to the amount of data after compression.

[0107] For example, the resampling unit 210 allocates a number of samples of different time lengths to the data in each interval, thereby compressing the data in each interval with different compression ratios.

[0108] exist Figure 7 In the example, C1 = 4, C2 = 1, and C3 = 3. That is, Figure 7 The first interval S of (a) a1 The waveform data in the image is compressed to 1 / 4, thus setting it as... Figure 7 The first interval S of (b) b1 Compressed data, Figure 7 The third interval S of (a) a3 The waveform data in the image is compressed to 1 / 3, thus setting it as... Figure 7 The third interval S of (b) b3 The compressed data. Within these intervals, the number of samples is reduced. In contrast, Figure 7 The second interval S of (a) a2 The waveform data in the video is not compressed.

[0109] Figure 6 The judgment benchmark data 231 is divided into... Figure 7 Each interval S b1 S b2 S b3 Each corresponds to a base interval. That is, the base data 231 is also divided into a complex number of n base intervals. Furthermore, Figure 6 The judgment benchmark data 231 and each benchmark interval Figure 1 Compared with the judgment benchmark data 131, in order to target Figure 7 The interval S of (a) a1 S a2 S a3 The compression ratios C1, C2, and C3 of the executed compression are compressed respectively. It is determined that each reference interval of the reference data 231 has a reference data length, which are referred to as the first reference data length, the second reference data length, and the third reference data length, respectively.

[0110] The lengths of the first to third reference data of the judgment reference data 231 are input to the resampling unit 210. The resampling unit 210 performs a process on the compressed data, adjusting the lengths of each interval S... b1 S b2 S b3 The data length is resampled to be consistent with the lengths of the first to third reference data to generate the second captured data.

[0111] In this embodiment, the resampling unit 210 can perform resampling processing even when the sampling periods before and after resampling are set to be significantly different.

[0112] As described above, in the processing monitoring device 200 according to this embodiment, the judgment reference data 131 may also have a plurality of n reference intervals divided in time at a predetermined ratio. The resampling unit 210 divides the first captured data into a plurality of n intervals in time at a predetermined ratio. Furthermore, the resampling unit 210 adjusts the data length of the first captured data in each of the plurality of n intervals so that it matches the data length of the judgment reference data 131 in the corresponding plurality of n reference intervals, thereby generating the second measurement data.

[0113] The aforementioned processing step monitoring device 200 is more effective when the time for characterizing a processing point (e.g., during blanking) is shorter than the overall processing time of each individual cycle in a cyclic process. Specifically, by measuring physical quantities such as the blanking load applied to tools like punches during blanking operations based on a stamping press, the processing step monitoring device 200 can not only monitor the tool's state in detail but also grasp the overall state throughout the processing time, specifically the actions of pressing the workpiece based on the die and pressing it into the scrap die. Furthermore, as in this embodiment, by compressing data from an appropriate portion of the overall processing time of each individual cycle, both detailed partial state monitoring and overall state monitoring can be achieved. Moreover, data compression reduces the amount of data acquired and processed, thus speeding up the decision-making process.

[0114] Therefore, the processing procedure monitoring device 200 can perform judgment processing by using data that does not compress the large changes in the processing and compresses less data for the slow changes, thus obtaining judgment results more quickly without compromising judgment accuracy. Furthermore, the processing procedure monitoring device 200 can perform overall anomaly detection from the start to the end of processing in one go. By using a judgment method based on the correlation of sampled values ​​from two different times—that is, a judgment method using Mahalanobis distance—the processing procedure monitoring device 200 can achieve even higher performance.

[0115] The resampling unit 210 can also adjust the data length of the first measurement data in at least one interval by compressing the first measurement data in at least one interval among the complex n intervals, so that it is consistent with the data length of the judgment reference data 131 in the corresponding reference interval.

[0116] Through the above structure and data compression, the amount of data acquired and processed can be reduced, and the decision-making process can be accelerated.

[0117] (A variation of the second embodiment)

[0118] The following is for reference Figure 8 A variation of the second embodiment of this disclosure will now be described. In this variation, Figure 6 Height sensor 1 measures the height of the slide block in a crank-type press. For example, height sensor 1 measures the die height. The die height is, for example, the distance from the lower surface of the slide block to the upper surface of the pad plate. The height of the slide block is an example of the "second physical quantity" of this disclosure.

[0119] Figure 8 (a) is a graph showing the change in the height of the slider over time as measured by height sensor 1. Figure 8(b) is a graph showing the change in the time of the load applied to the punch as measured by load sensor 2.

[0120] like Figure 8 As shown in (a), the time variation of the slide height in the crank-type press approximates a sine wave. The blanking of the workpiece occurs between the start of processing and the moment when the slide descends to the bottom dead center.

[0121] Data capture unit 8 or resampling unit 210 is determined Figure 8 The moment t when the slider height of (a) is below the first intermediate threshold. c1 The time t is lower than the second intermediate threshold which is smaller than the first intermediate threshold. c2 The first intermediate threshold and the second intermediate threshold are, for example, pre-stored in the storage unit 13. Furthermore, as described in the first embodiment, when the intensity of the signal s1 from the height sensor 1 is lower than the start threshold at the start of processing t... c0 When processing ends at a point t higher than the termination threshold c3 It was determined by the data interception section 8.

[0122] Therefore, the first captured data is divided into: from the start of processing t c0 At time t c1 The first interval S c1 Data D1 from time t c1 At time t c2 The second interval S c2 Data D2 from time t c2 t at the end of processing c3 The third interval S c3 Data D3 in the data. Figure 8 In the example shown, the first interval S c1 This represents the interval from the start of processing until the punch descends and reaches the material being processed. Interval S (2nd interval) c2 This corresponds to the interval where the punch cuts the material. Interval 3, S c3 This corresponds to the period from when the material is punched by the punch until the end of the processing.

[0123] As described above, in this modified example, the resampling unit 210 segments the first captured data at multiple times at multiple thresholds, such as a first intermediate threshold and a second intermediate threshold, that are different from each other below the height of the slider being measured.

[0124] In the second embodiment, Figure 7 The intervals S of (a) a1 S a2 S a3 The length of time t a1 t a2 ta3 The value is preset; in contrast, in this variant, it is determined using the detection result of the slider height based on height sensor 1. Each interval S a1 S a2 S a3 The segmentation is synchronized with the detection waveform based on height sensor 1, so even if the processing time of each individual cycle in the cyclic processing fluctuates, each segmented region is accurately determined. Therefore, by segmenting in sync with the detection waveform based on height sensor 1, it is possible to obtain the effect of adjusting for fluctuations in each interval according to each correction time, thus enabling the... Figure 5 The frequency of the load at the specific sampling time shown in (b) is set to a more unbiased, steeper frequency, i.e., a frequency with a smaller standard deviation. Therefore, with this variation, anomaly detection can be performed with greater accuracy.

[0125] (Third Implementation)

[0126] Figure 9 This is a block diagram illustrating a structural example of the processing step monitoring device 300 according to the third embodiment of this disclosure. (Compared to...) Figure 6 The processing monitoring device 200 is equipped with an oversampling filter 9. Figure 9 The machining process monitoring device 300 has multiple oversampling filters. In the illustrated example, the machining process monitoring device 300 has three oversampling filters 9a, 9b, and 9c.

[0127] Oversampling filters 9a, 9b, and 9c are, for example, low-pass filters with independently set cutoff frequencies. That is, the cutoff frequencies of oversampling filters 9a, 9b, and 9c can be different from each other or the same.

[0128] Inputting the oversampling filter 9a from the start of processing t a0 At time t a1 The first interval S up to now a1 Data D1 (refer to) Figure 7 (a)). Input from time t to the oversampling filter 9b. a1 At time t a2 The second interval S up to now a2 Data D2 in the data. Input from time t to the oversampling filter 9c. a2 t at the end of processing a3 The third interval S up to now a3 Data D3 in the middle.

[0129] The cutoff frequencies of the oversampling filters 9a, 9b, and 9c are adjusted according to the compression ratios C1, C2, and C3 of the input data D1, D2, and D3. For example, in the first interval S... a1When the waveform data is compressed to 1 / 4 (C1 = 4), the cutoff frequency of the oversampling filter 9a is set to 1 / 4 of the Nyquist frequency. In the first interval S... b1 When the waveform data is not compressed (C2 = 1), the cutoff frequency of the oversampling filter 9b is set to the Nyquist frequency. In the first interval S... c1 When the waveform data is compressed to 1 / 3 (when C3 = 3), the cutoff frequency of the oversampling filter 9c is set to 1 / 3 of the Nyquist frequency.

[0130] When there is only one oversampling filter, the resampling point may be obtained when the timing value of the sampling point closest to the sampling point before resampling is obtained, and the higher frequency components based on the sampling frequency before compression are obtained, and the resampling point is calculated. In contrast, with this embodiment having multiple oversampling filters 9a, 9b, and 9c, it is easy to prepare oversampling filters 9a, 9b, and 9c that meet the sampling definition, and accurate values ​​with high-frequency components removed can be obtained in resampling.

[0131] Therefore, the processing monitoring device 300 can obtain measurement data that meets the sampling definition and compresses the data volume to include the correct frequency components, and can more accurately perform anomaly detection based on the judgment processing unit 11.

[0132] In this embodiment, an example is described where the type of oversampling filter is set to be the same as the number of intervals, with a one-to-one correspondence between the two. However, this embodiment is not limited to this. For example, if the compression ratio applied to multiple intervals is almost the same, then one oversampling filter can also be set for these multiple intervals.

[0133] Furthermore, when a high compression ratio is applied to a certain range, such as when a compression ratio of 8 times or more is applied, the compression itself has the function of oversampling. Therefore, it is not necessary to oversample the data in that range to a sampling frequency of Fs or higher.

[0134] As described above, the processing monitoring device 300 according to this embodiment may also include a plurality of n oversampling filters corresponding to a plurality of n intervals. The plurality of n oversampling filters perform oversampling processing on multiple data points representing the time variation of a first physical quantity contained in the first measurement data within each of the plurality of n intervals. The plurality of n oversampling filters have mutually independent cutoff frequencies.

[0135] With the above structure, even when the input data is compressed, the resampling unit 210 can still satisfy the sampling definition and resample the input data. Therefore, the processing monitoring device 300 can easily obtain measurement data that satisfies the sampling definition and compresses the data volume to include the correct frequency components, and can more accurately perform anomaly detection based on the judgment processing unit 11.

[0136] The resampling unit 210 can also compress the first measurement data in at least one of the complex n intervals at a predetermined compression ratio, adjusting the data length of the first measurement data in at least one interval to match the data length of the judgment reference data 131 in the corresponding reference interval. The cutoff frequencies of the complex n oversampling filters are set based on the compression ratio.

[0137] Through the above structure, in addition to the effects of the processing procedure monitoring device 300, it is also possible to perform both detailed local status monitoring and overall status monitoring by compressing appropriate portions of the overall processing time for each single cycle. Furthermore, by compressing the data, the amount of data acquired and processed can be reduced, thereby speeding up the judgment and processing.

[0138] (Other implementation methods)

[0139] As described above, the embodiments disclosed in this application have been illustrated as technical examples. However, the technology in this disclosure is not limited to this and can be applied to embodiments with appropriate modifications, substitutions, additions, omissions, etc. Furthermore, the structural elements described in the above embodiments can be combined to create new embodiments. Other embodiments are illustrated below.

[0140] In the above embodiment, a structure for processing two signals detected by height sensor 1 and load sensor 2 has been described. However, this disclosure is not limited to this, and it is possible to detect the timing of the start and end of processing based on at least one signal. Therefore, the number of sensors for detecting physical quantities is not limited to two, but may be one or more. The at least one signal is not limited to those obtained by height sensor 1 or load sensor 2, and may be, for example, a signal from a controller such as a PLC (Programmable Logic Controller), or a signal indicating the rotation angle or position of the press.

[0141] In the above implementation, an example of the first captured data being oversampled by a factor of 8 was described. However, the oversampling factor can be greater than 1 and is not limited to 8. If the oversampling factor is large, the amount of data and computation increases, but it can obtain interpolated values ​​with small errors. Considering this trade-off, an oversampling factor of around 8 is preferred.

[0142] In the above embodiments, an oversampling filter was described as an example of a data interpolation unit. However, the data interpolation unit is not limited to this; it is sufficient to obtain sampled values ​​between multiple data points. For example, the data interpolation unit may also perform linear interpolation between two sampled points or curve interpolation based on sampled values ​​of three or more points.

[0143] In the above embodiment, the judgment reference data 131 was described as waveform data used as a reference in the judgment process. However, the judgment reference data 131 is not limited to this; it can have a reference data length. The judgment reference data 131 can also be a learned model generated by learning multiple time series data of the reference data length through machine learning. The judgment processing unit 11 can also use such a learned model to output a judgment result to the input of time series data with the same data length as the reference data, i.e., the second captured data.

[0144] As in the second and third embodiments, the effect of using compressed and reduced data for judgment processing is significant when a fully learned model is used in the judgment processing. This is because in machine learning of time series data, as the length of the time series data increases, the size of the generated fully learned model tends to increase, and consequently, the amount of inference processing based on the fully learned model increases. Furthermore, as the data length increases, there is a problem of difficulty in performing learning suitable for the model. By using waveform data that is characteristically compressed without loss of processing, the above-mentioned problems can be solved, and judgment can be performed at high speed.

[0145] The processing monitoring device and method disclosed herein can determine whether the processing machinery is operating normally more accurately than the prior art, even when the processing time of each individual cycle in a cyclic processing fluctuates.

[0146] As described above, embodiments have been illustrated as technical examples in this disclosure. Therefore, accompanying drawings and detailed descriptions are provided.

[0147] Therefore, the structural elements described in the accompanying drawings and detailed descriptions include not only those necessary for solving the problem, but also those essential for illustrating the above-described technology rather than for solving the problem. Thus, these non-essential structural elements are described in the drawings and detailed descriptions, but should not be directly considered as essential.

[0148] Furthermore, the above-described embodiments are used to illustrate the technology in this disclosure, and various changes, substitutions, additions, omissions, etc., can be made within the scope of the claims or their equivalents.

[0149] Industrial availability

[0150] This disclosure relates to a machining process monitoring device and a machining process monitoring method that can be applied to monitor machining machines that perform repetitive single-cycle processing to determine whether the machining machine is operating normally.

Claims

1. A processing procedure monitoring device for monitoring a processing machine that performs repetitive single-cycle processing and determining whether the processing machine is operating normally, the processing procedure monitoring device comprising: The data interception unit determines the start and end times of processing for each single cycle, and intercepts first measurement data representing the first physical quantity that changes over time during the cyclic processing in the processing machine, for each single cycle, the first measurement data representing the first physical quantity from the start of processing to the end of processing. The data length adjustment unit adjusts the data length of the first measurement data to match the data length of the judgment reference data, which represents the change of the first physical quantity in a single cycle under normal operating conditions of the processing machine, thereby generating the second measurement data; and The judgment processing unit compares the second measured data with the judgment benchmark data to determine whether the processing machine is operating normally. The judgment benchmark data has a complex number of n benchmark intervals that are divided in time at a predetermined ratio. The data length adjustment unit divides the first measurement data into a plurality of n intervals over time according to the prescribed ratio. The data length adjustment section adjusts the data length of the first measurement data in each of the complex n intervals so that it is consistent with the data length of the judgment benchmark data in the corresponding complex n benchmark intervals, thereby generating the second measurement data.

2. The processing procedure monitoring device according to claim 1, wherein, The data interception unit determines the start and end times of processing based on a second physical quantity that changes at the start and end times of processing in the processing machine.

3. The processing procedure monitoring device according to claim 2, wherein, The processing machinery applies a load to the material being processed in the single cycle to perform the processing. The second physical quantity is information representing the position of a component of the processing machinery that moves during the single cycle in order to apply a load to the material being processed.

4. The processing procedure monitoring device according to claim 1 or 2, wherein, The processing monitoring device further includes a data interpolation unit, which interpolates data between multiple data points representing the time variation of the first physical quantity included in the first measurement data. The data length adjustment unit adjusts the data length of the first measurement data by using the plurality of data points and the data interpolated by the data interpolation unit so that it is consistent with the data length of the judgment reference data, thereby generating the second measurement data.

5. The processing procedure monitoring device according to claim 4, wherein, The data interpolation unit interpolates the data between the multiple data points by oversampling the multiple data points.

6. The processing procedure monitoring device according to claim 1, wherein, The data length adjustment unit divides the first measurement data at multiple times in the processing machine when the value of the second physical quantity that changes at the start and end of processing is lower than multiple different thresholds.

7. The processing procedure monitoring device according to any one of claims 1 to 3, wherein, The data length adjustment unit adjusts the data length of the first measurement data in at least one of the at least n intervals by compressing the first measurement data in the at least one interval, so that it is consistent with the data length of the judgment reference data in the corresponding reference interval.

8. The processing procedure monitoring device according to any one of claims 1 to 3, wherein, The processing step monitoring device further comprises: a complex number of n oversampling filters, which perform oversampling processing on multiple data points representing the time change of the first physical quantity contained in the first measurement data in each of the complex number of n intervals, and correspond to the complex number of n intervals respectively. The complex number n oversampling filters have mutually independent cutoff frequencies.

9. The processing procedure monitoring device according to claim 8, wherein, The data length adjustment unit adjusts the data length of the first measurement data in at least one of the n intervals by compressing the first measurement data in at least one interval at a predetermined compression rate, so that the data length is consistent with the data length of the judgment reference data in the corresponding reference interval. The cutoff frequencies of the n complex oversampling filters are set based on the compression ratio.

10. The processing procedure monitoring device according to any one of claims 1 to 3, wherein, The data length adjustment unit adjusts the data length of the first measurement data by resampling the first measurement data so that it is consistent with the data length of the judgment benchmark data.

11. The processing procedure monitoring device according to any one of claims 1 to 3, wherein, The processing monitoring device further includes an output unit that outputs the judgment result based on the judgment processing unit.

12. The processing procedure monitoring device according to any one of claims 1 to 3, wherein, The processing machinery is a stamping machine.

13. The processing procedure monitoring device according to any one of claims 1 to 3, wherein, The first physical quantity is the load detected by the load sensor.

14. A method for monitoring a machining process, wherein the method monitors a machining machine performing a repetitive single-cycle machining process, and determines whether the machining machine is operating normally, the method comprising: The first physical quantity acquisition step involves sequentially acquiring the first physical quantity that changes over time during the cyclic processing in the processing machinery. The data extraction step involves determining the start and end times of processing for each single cycle, and extracting first measurement data representing the first physical quantity from the start of processing to the end of processing for each single cycle of the first physical quantity. The data length adjustment step involves adjusting the data length of the first measurement data to match the data length of the judgment reference data, which represents the change of the first physical quantity in a single cycle under normal operating conditions of the processing machinery, thereby generating the second measurement data; and The judgment and processing step involves comparing the second measured data with the judgment benchmark data to determine whether the processing machinery is operating normally. The judgment benchmark data has a complex number of n benchmark intervals that are divided in time at a predetermined ratio. The data length adjustment step divides the first measurement data into a complex number of n intervals over time according to the prescribed ratio. The data length adjustment step adjusts the data length of the first measurement data in each of the complex n intervals so that it is consistent with the data length of the judgment benchmark data in the corresponding complex n benchmark intervals, thereby generating the second measurement data.

Images