Structure abnormality prediction method and device

Abstract

To predict a structure abnormality in a processing machine or device that has periodicity like a press device in advance with high accuracy.SOLUTION: An AE detection signal from an AE sensor provided in a part of a structure is converted into AE image information as visual information in accordance with the strength of the AE detection signal in each one cycle. The AE image information is arranged in an image map area having vertical and horizontal axes with one axis representing the number of cycles and the other axis representing a fixed time for one cycle. Cumulative mapping image information corresponding to the number of cycles is acquired, and abnormality prediction is performed on the basis of image determination with the mapping image information and either one or both of an image pattern based on prestored mapping image information for a normal state and prestored mapping image information for an abnormal sign.EFFECT: Occurrences of defective products can be almost eliminated because occurrence of an abnormality is not detected but a sign of an abnormality occurrence is detected.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a method and apparatus for accurately predicting structural abnormalities in periodic processing machines and equipment. [Background technology]

[0002] Forging is highly productive and suitable for mass production, but on the other hand, if there is an abnormality in the die, a large number of defective products will be produced. For this reason, it is important to understand the condition of the die in order to continue stable production.

[0003] Sensor technology is used to detect deterioration and abnormalities in industrial machinery and structures. Sensors are adopted that are appropriate for what needs to be detected. For example, the following Patent Documents 1 to 4 disclose techniques for detecting and predicting abnormalities using the AE sensor (Acoustic Emission Sensor: acoustic reflection detection sensor) employed in the present invention.

[0004] Patent Document 1 (Japanese Patent Application Laid-Open No. 150090 / 1977) proposes a configuration in which, when examining the fatigue strength of materials in material mechanics, in order to accurately grasp the timing of crack initiation, its progression and behavior, etc., multiple sets of preamplifiers and AE sensors are connected to an AE positioning monitoring device with a zone selection function, one end of a pinpoint contact type waveguide is connected to each of the AE sensors, the waveguides are supported by a waveguide fixture made of acoustic insulating material, and the other end of the waveguide is acoustically connected to the specimen.

[0005] Furthermore, Patent Document 2 (JP Patent Publication No. 4-310857) proposes that in order to monitor the progression of minute cracks in bridge structures that do not affect their strength, multiple AE sensors are arranged to form spatial filters near the crack tip position in the bridge structural member and in the predicted crack progression direction, thereby creating limited monitoring areas, and extracting only AE signals from vibrations caused by running wheels that originate within each monitoring area, and estimating the progression of the crack from the changes in these signals over time.

[0006] Furthermore, Patent Document 3 (Japanese Patent Laid-Open Publication No. 8-159151) proposes that in order to accurately estimate the remaining life of a rolling bearing, vibration information, temperature information, and load information of the rolling bearing unit while it is in operation are monitored simultaneously, and the results are compared with fluctuations in vibration information and temperature information associated with bearing damage corresponding to each load condition measured in advance, thereby predicting the degree of bearing damage and remaining life.

[0007] Furthermore, Patent Document 4 (JP 2004-170397 A) proposes a system in which one AE sensor is attached to each measurement location, corresponding to each predicted location of crack initiation or propagation in the structure, in order to detect and monitor the level of damage to the entire structure; the AE sensor performs primary processing of the AE signal detected by the composite probe, stores the primary processed data, and communicates it to the outside; the AE sensor is connected to a data processing unit that requests and receives output of the primary processed data from the AE sensor and performs secondary processing of the received primary processed data; and the safety of the structure is evaluated based on the secondary processed data; the primary processing of the AE signal in the AE sensor involves counting the number of times a threshold value is exceeded per unit time that is input in advance, and creating primary processed data that indicates an evaluation rank according to this count; and the secondary processing in the data processing unit involves noise removal by comparing the AE signal with supplementary data that is stored in the data processing unit in advance or transmitted from an external source.

[0008] However, Patent Documents 1 and 2 are intended to detect cracks that have already "occurred" while they are still minute, Patent Document 3 is intended to "estimate the remaining life," and Patent Document 4 is intended to grasp the extent of damage by attaching an AE sensor to a "predicted location of crack occurrence" or a "location where cracks have already occurred," and none of these are intended to detect cracks or abnormalities before they occur.

[0009] For example, forging is highly productive and can be mass-produced, but if an abnormality occurs in the die, a large number of defective products will be produced immediately. Moreover, since it is not possible to directly check the die during production, where and how the abnormality occurred can only be clearly understood from the large number of defective products that are produced.

[0010] In other words, if Patent Documents 1 to 4 are applied to a forging processing site, it is too late to stop operations after an abnormality occurs in the die, the location of the abnormality is unpredictable, and even if the lifespan can be estimated, the lifespan may be shortened due to high operation from that point onwards, or there may be abnormalities that do not affect the lifespan of the die but affect the product, so ultimately none of them are capable of "predicting abnormalities that will occur in the near future in advance." [Prior art documents] [Patent documents]

[0011] [Patent Document 1] Japanese Patent Publication No. 52-150090 [Patent Document 2] Japanese Patent Application Publication No. 4-310857 [Patent Document 3] Japanese Patent Application Publication No. 8-159151 [Patent Document 4] Japanese Patent Application Laid-Open No. 2004-170397 Summary of the Invention [Problem to be solved by the invention]

[0012] The problem that the present invention aims to solve is that it is not possible to accurately predict in advance mold abnormalities that will occur in the near future. [Means for solving the problem]

[0013] In order to solve the above problems, the structural abnormality prediction method of the present invention is a method for predicting structural abnormalities in cyclical processing machines such as presses, in which an AE detection signal from an AE sensor provided in a part of the structure is converted into AE image information as visual information according to the strength of the AE detection signal for each cycle, the AE image information is arranged in an image map area with vertical and horizontal axes, one axis representing the number of cycles and the other axis representing a fixed time for one cycle, cumulative mapping image information according to the number of cycles is obtained, and an abnormality is predicted by image judgment of the mapping image information together with pre-stored mapping image information for normal conditions and / or mapping image information registered as a sign of an abnormality. [Effects of the Invention]

[0014] The present invention can predict the near-future occurrence (manifestation) of structural abnormalities in periodically moving parts of a periodic processing mechanism. Therefore, for example, in forging processing, measures can be taken before die abnormalities occur, thereby preventing the production of a large number of defective products. [Brief explanation of the drawings]

[0015] [Figure 1] 1A is a diagram showing the configuration of a cold forging device as a processing machine, and FIG. 1B is a diagram showing the overall configuration of a structural abnormality prediction device of the present invention. [Figure 2] FIG. 10 is a diagram for explaining the effectiveness of the AE sensor. [Figure 3] FIG. 10 is a diagram showing the behavior of AE waves in one cycle for each cumulative number of shots. [Figure 4] 10(a) to 10(d) are diagrams for explaining color graph generation. [Figure 5] 10(a) to 10(d) are diagrams for explaining abnormality prediction judgment using image patterns. DETAILED DESCRIPTION OF THE INVENTION

[0016] The present invention is a method for predicting structural abnormalities in periodic processing machines such as presses, in order to accurately predict mold abnormalities that will occur in the near future. The method converts an AE detection signal from an AE sensor installed in a part of the structure into AE image information as visual information according to the strength of the AE detection signal for each cycle, arranges the AE image information in an image map area with vertical and horizontal axes, one axis representing the number of cycles and the other axis representing a fixed time for one cycle, obtains cumulative mapping image information according to the number of cycles, and performs image judgment of the mapping image information against either or both of pre-stored mapping image information for normal conditions and mapping image information registered as an abnormality precursor.

[0017] In this example, the structural abnormality prediction device of the present invention, which employs the method of the present invention, is implemented in a cold forging machine (for bevel gears) as a processing machine with periodicity. To implement the present invention, the structural abnormality prediction device 1 of the present invention includes an AE sensor S1 provided in a part of the cold forging machine shown in Fig. 1(a), an image data processing unit 1A, a storage unit 1B, and an image determination unit 1C, as shown in Fig. 1(b).

[0018] The image data processor 1A converts the AE detection signal from the AE sensor S1 into AE image information as visual information according to the strength of the AE detection signal for each cycle. The image data processor 1A also arranges the AE image information in an image map area with vertical and horizontal axes, one axis representing the number of cycles and the other axis representing a fixed time for one cycle, to obtain cumulative mapping image information according to the number of cycles.

[0019] The memory unit 1B stores mapping image information that is cumulatively collected at any time, and in this example, stores, for example, an image pattern based on mapping image information in a normal state (referred to as a "normal image pattern") and an image pattern based on mapping image information registered as an abnormality precursor (referred to as an "abnormal image pattern"). Note that in this example, the memory unit 1B stores all of the mapping image information that is cumulatively collected at any time, but does not store it as an image pattern for a certain fixed cycle unit.

[0020] The image determination unit 1C performs image determination between the image pattern that appears after accumulating the mapping image information for a certain number of cycles and the normal image pattern in the memory unit 1B or the image data registered in the memory unit 1B as an abnormality precursor. Furthermore, when the image determination unit 1C predicts an abnormality, that is, when it determines that an abnormality will occur in the near future, it activates a notification means (described later) or stops the operation of the cold forging device. In this example, the operation is stopped after the notification is issued.

[0021] Furthermore, the structural abnormality prediction device 1 is equipped with the above-mentioned notification means for generating an alarm or the like to indicate that an abnormality has been predicted, and display means for displaying the alarm and various other displays, as well as the hardware configuration provided in a general personal computer, such as input means for inputting commands, etc., storage means for storing the method of the present invention as software, RAM, CPU, etc. Note that the various sensors including the AE sensor S1 in the structural abnormality prediction device 1 shown in Figure 1(b) may be configured remotely from one another as long as they are capable of sending and receiving signals.

[0022] The cold forging device as a periodic processing machine has the basic components of a forging device, which are a slide 2, a bolster 3, an upper die 4, a lower die 5, a die height 6, an inner punch 7, a punch 8, a die 9 (metal mold), a die cushion 10, a cushion plate 11, and a knockout 12.

[0023] In this example, the structural abnormality prediction device 1 of the present invention has the following sensors. S1 is an AE sensor. The AE sensor S1 is installed on the cushion plate 11. S2 is a load sensor. The load sensor S2 is placed in the pressure-receiving section directly below the die 9 to measure the load applied to the main parts such as the die 9, knockout 12, and inner punch 7 during molding. S3 is a displacement sensor that measures the press stroke and the gap between the dies to clarify the positional relationship of the dies.

[0024] Furthermore, in addition to the above-mentioned sensors S1 to S3, the structural abnormality prediction device 1 of the present invention may additionally employ sensors such as temperature, sound, vibration (acceleration) and various strain sensors, and may additionally employ an image inspection method using a camera.

[0025] Furthermore, the structural abnormality prediction device 1 of the present invention can also make evaluations in relation to equipment and peripheral equipment by collecting data such as control signals for equipment and peripheral equipment such as presses, actuator currents, hydraulic and pneumatic pressures, and proximity sensors.

[0026] In this example, the material is molded as follows: Before molding, the cushion plate 11 with the die 9 attached is positioned above the lower die 5 by the die cushion 10. At the start of molding, the punch 8 and die 9 come into contact as the slide 2 descends. Thereafter, the punch 8 pushes the die 9, causing the cushion plate 11 to descend, and the material is pushed out into the molding section P by the fixed knockout 12, completing molding at the bottom dead center.

[0027] In forging, the contact state with the die changes in complex ways due to the plastic deformation of the material, and the stress state also changes accordingly. Also, in actual production sites, there are many uncertain factors such as contact and sliding between other dies, peripheral equipment, and environmental disturbances, and until now, it has been extremely difficult to identify the relationship between these uncertain factors and die abnormalities, so the only way to reduce the number of defective products was to discover abnormalities as early as possible. In other words, it has not been possible to "know beforehand that an abnormality has occurred."

[0028] In this invention, the AE sensor is the main sensor for "predicting the occurrence of abnormalities before they occur." In this example, other sensors are also introduced to separate data caused by molding from data caused by the above uncertain factors (details omitted).

[0029] Here, we will explain why the AE sensor is used as the main sensor in this invention. The AE sensor detects the change in elastic energy within a solid body as a wave that accompanies deformation or destruction, and the data detected by the AE sensor provides clues to understanding the dynamic structural changes within the material that will be used to make the product in the mold.

[0030] What is important in this application is not to look for large changes that accompany an abnormality, nor to look for changes that accompany the progression of the abnormality, but to find "slight changes that appear before an abnormality occurs," and this is because it is ideal for capturing these slight (weak, minute) changes. The present invention makes it possible to reliably capture these slight changes that may be overlooked, thereby predicting abnormalities.

[0031] Next, we will explain the process of "predicting abnormalities before they occur" using the AE sensor as the main sensor. To confirm the effectiveness of the data detected by the AE sensor, Figure 2 shows a comparison of AE waves when a crack occurred in the die (punch) and the die was replaced with a new one (a), and when the die was used and was deemed unusable (b). In Figure 2, the horizontal axis is time, and the vertical axis is the AE wave (maximum value) on the left and the knockout load, which is the main load in forming, on the right.

[0032] In Figure 2(a)(b), almost no change was observed in the knockout load (right vertical axis), but AE waves were hardly generated between 1.0 and 1.5 seconds when a new mold was used, but were generated significantly when the NG mold was used. This shows that there is a large change in AE waves before and after mold replacement, confirming the effectiveness of the AE sensor detection data.

[0033] Next, we analyzed the detection signals from the AE sensor collected during press forming. Figure 3 shows the relationship between stroke and AE waves, with the horizontal axis representing the distance to the bottom dead center of the die height (stroke) and the vertical axis representing the maximum value of AE waves per unit time, plotted logarithmically to make it easier to capture small changes. In addition, to distinguish between load application and unloading, the stroke on the horizontal axis is plotted as a negative value for the distance before the bottom dead center and a positive value for the distance after the bottom dead center.

[0034] In Figure 3, AE waves change significantly over time as the mold is used, as shown in Figure 2, so AE waves from the same mold after 100 shots, 31,000 shots, and 54,000 shots are shown. As a result, it can be seen that AE waves change significantly in specific parts over time as the mold is used.

[0035] A large amount of data is collected during production, and when observing the transition of waveforms from this large amount of data, it is common to use feature quantities such as peak and average values ​​of the time series data as trend data and focus only on the characteristic parts. However, focusing on only a part of the data, such as feature quantities, makes it difficult to capture small changes in the time series.

[0036] That is, when an operator visually compares behavioral changes from a large amount of collected data from a cyclical processing machine, the data displays (for example, graphs) of the same time unit are generally overlapped or arranged vertically. However, when overlapping, even if an abnormal waveform is observed, it is difficult to grasp, for example, which cycle it is, and when arranging the data vertically, there is a limit to the number of cycles that can be viewed in an arrangement, so even if an abnormal waveform is found within the arranged range, it is possible to find an abnormal waveform outside the arranged range, but it is difficult to find an abnormal waveform outside the arranged range.

[0037] Therefore, in the present invention, in order to obtain an overview of the large amount of collected time-series detection data from the AE sensor, the data is visualized as shown in Fig. 4. Hereinafter, the visualized graph will be referred to as a color graph (mapping image information). The color graph is generated in the image data processing unit 1A as follows. Note that image data for a certain range (a certain number of shots range) in the color graph is referred to as an "image pattern."

[0038] The processing in the image data processing unit 1A will be described with reference to Fig. 4. The image data processing unit 1A obtains an AE detection signal for one shot of AE waves (AE detection signals) as a time-series graph with time on the horizontal axis and effective value on the vertical axis, as shown in Fig. 4(a). Next, in this example, for example, the image data processing unit 1A converted the signal into a graph with a logarithmic vertical axis, as shown in Fig. 4(b), in order to make it easier to grasp the changes in the AE waves.

[0039] The image data processing unit 1A converts one shot of AE waves (AE detection signals) shown in Fig. 4(b) into a single color bar (AE image information). The color bar shows the vertical axis shown in Fig. 4(b), i.e., the effective value of the AE waves, using shades of color. Note that although the illustration shows shades of black and white, the actual color is used.

[0040] Figure 4(c) shows five color bars (five shots) arranged vertically (axis). In this way, the horizontal axis is fixed to the displacement of only one shot, the vertical axis is the number of shots, and the strength of the detection data from the AE sensor is expressed in color (shade of color, etc.). In other words, the AE detection signal during a molding operation in which one shot is repeated is converted into a color bar, and this color bar is cumulatively arranged in map information, and image data processing is performed on the mapping image information to obtain a color graph.

[0041] The image determination unit 1C determines whether an abnormality is predicted as follows. In this example, the storage unit 1B stores both an image pattern based on mapping image information of a normal state as shown in Fig. 5(a) (hereinafter referred to as a normal image pattern) and an image pattern based on mapping image information of an abnormality precursor as shown in Fig. 5(b) (hereinafter referred to as an abnormal image pattern). Note that a plurality of abnormal image patterns are stored, the number of which corresponds to the number of types of abnormality that can be identified.

[0042] The image determination unit 1C receives a color graph, which is an accumulation of color bars for each shot, as needed, and compares the image pattern (hereinafter referred to as the real image pattern) generated by accumulating the one shot just input with the normal image pattern and the abnormal image pattern. The image determination unit 1C uses not only the position of the AE image information in the map area but also the color (intensity) as information for determination.

[0043] An example of a judgment made by the image judgment unit 1C will be described below. For example, the normal image pattern in Fig. 5(a) means that no color bar indicating an expected abnormality appears within a range of a set fixed number of shots. Therefore, when a real image pattern such as that shown in Fig. 5(c)(d) appears over a set range of a set fixed number of shots, the image judgment unit 1C judges that the real image pattern does not match any image pattern other than the normal image pattern, i.e., the image pattern in Fig. 5(a), and predicts the occurrence of an abnormality, so in this example, it issues a warning signal (sound and display) and stops operation.

[0044] On the other hand, the abnormal image pattern registered in advance as a sign of an abnormality in Fig. 5(b) means the appearance of color bars indicating all possible abnormalities, obtained empirically or through testing. Therefore, the image judgment unit 1C judges whether the image pattern matches or does not match in a brute-force manner for each shot, and when a real image pattern such as that shown in Fig. 5(c) appears over a set range of a certain number of shots, it does not yet match the abnormal image pattern in Fig. 5(b), so it postpones the prediction of an abnormality. When a real image pattern such as that shown in Fig. 5(d) appears, it is judged to be a mismatch with the abnormal image pattern in Fig. 5(b), and as it predicts the occurrence of an abnormality, it issues a warning signal (sound and display) and stops operation in this example.

[0045] In this example, a method is shown in which both a normal image pattern and an abnormal image pattern are used, and if an abnormality is determined based on the normal image pattern but there is a grace period for the abnormal image pattern, an abnormality is not determined until the color bars have accumulated and the real image pattern completely matches the abnormal image pattern, in which case an abnormality is determined if both the normal image pattern and the abnormal image pattern are determined to be abnormal.However, a determination method that uses either one of them may also be used depending on the processing machine or device.

[0046] In this way, the present invention makes it possible to predict structural abnormalities in machinery and other devices that have a periodicity that is difficult to detect using a normal graph display viewed on a continuous time axis by fixing the period (cycle) time on one axis and accumulating a color bar display for each number of cycles on the other axis, making it easier for humans to detect slight signs of an abnormality before it occurs.In addition, as software processing, image pattern judgment is made using the cumulative display (mapping image information) of color bars over a certain range, making it easy to grasp intermittent, sudden, temporal, and periodic changes before an abnormality occurs, and enabling abnormalities to be predicted reliably.

[0047] In short, the present invention does not predict abnormalities based on the numerical values ​​and waveforms of each piece of detection data from the AE sensor, as in the past, but rather on the image pattern that appears in a color graph where the color bars for each shot are accumulated.In other words, the present invention is characterized by the ability to instantly determine even slight changes over a long period of time.

[0048] For example, if the horizontal axis is a continuous, non-fixed time axis, previous data will disappear from the display as time passes, making it difficult to compare with previous data. Also, software processing cannot detect changes by comparing all the data recorded up to that point, so for example, thresholds and waveforms must be set in advance, and when these set values ​​are detected, an abnormality is determined, that is, an abnormality is detected.

[0049] With the present invention, in this example, the display moves vertically (from the bottom to the top of the display surface) not with the passage of time but with the number of shots, but since the horizontal axis is fixed as the time axis for one shot, it is easy to constantly compare the changes in the detection data of the AE sensor for each shot cycle.

[0050] Furthermore, instead of detecting preset thresholds or waveforms that indicate anomalies, software processing can detect them by matching or mismatching the cumulative image patterns that appear on a shot-by-shot basis, which makes it possible to reliably find signs of anomalies and prevent them from occurring in the first place. [Explanation of symbols]

[0051] 1. Structural abnormality prediction device 1A Image data processing unit 1B Storage section 1C Image judgment section S1 AE sensor 9 Die (mold)

Claims

1. A method for predicting structural abnormalities in periodic processing machinery such as a press, the method comprising: converting an AE detection signal from an AE sensor provided in a part of the structure into AE image information as visual information according to the strength of the AE detection signal for each cycle; arranging the AE image information in an image map area with vertical and horizontal axes, one axis representing the number of cycles and the other axis representing a fixed time for one cycle; obtaining cumulative mapping image information according to the number of cycles; and predicting abnormalities by image judgment of the mapping image information and either or both of an image pattern based on pre-stored mapping image information for a normal state and an image pattern based on pre-stored mapping image information for an abnormality precursor.

2. an image data processing unit that converts an AE detection signal from the AE sensor into AE image information as visual information according to the strength of the AE detection signal for each cycle, arranges the AE image information in an image map area having vertical and horizontal axes, one axis representing the number of cycles and the other axis representing a fixed time for one cycle, and obtains cumulative mapping image information according to the number of cycles; a memory unit that stores the mapping image information and either or both of an image pattern based on mapping image information in a normal state and an image pattern based on mapping image information of an abnormality precursor; and an image judgment unit that performs image judgment of the mapping image information and either or both of the image pattern based on mapping image information in the normal state and the image pattern based on mapping image information of an abnormality precursor stored in the memory unit.

Images