Method for diagnosing malfunctions in vacuum valve devices

Abstract

To provide a failure diagnosis method which can acquire proper abnormality diagnosis according to chronological variation change and initial variation of an individual vacuum valve device.SOLUTION: A failure diagnosis method of a vacuum valve device is a method for diagnosing failure on the basis of vibration acceleration measured by an acceleration sensor provided on vibration acceleration the vacuum valve device, and comprises: calculating one evaluation value using vibration data on vibration acceleration obtained for each operation; calculating, as a predicted value of a next evaluation value, a value being an average from initial to current operations of the vacuum valve device and obtained by multiplying the average value from initial to previous operations by the previous number of times, adding the current evaluation value, and then dividing the value by the current number of times; calculating a reference value obtained by adding variation for each operation of the device individual body, a length of stop period, and a state change caused by handling of the device therebetween to the predicted value; and outputting a fact that there is a failure sign if the next evaluation value exceeds the reference value.SELECTED DRAWING: Figure 5

Description

Technical Field

[0006] , ,

[0007] ,

[0001] The present invention relates to a method for diagnosing a failure of a vacuum valve device that performs an opening / closing operation of an inlet / outlet of a vacuum chamber such as a semiconductor manufacturing apparatus.

Background Art

[0002] The vacuum valve device was mainly pneumatically driven to generate a certain driving force. However, in order to realize a high-speed opening / closing operation while suppressing vibration by changing the operation speed flexibly and performing smooth opening / closing, motor drive with speed control that can control the driving force and vary the speed is being adopted.

[0003] In Patent Document 1, it is proposed to provide a vibration sensor (acceleration sensor) in a vacuum valve device that performs motor drive, measure the vibration, decelerate to suppress the vibration when it exceeds a set magnitude, and apply control to the motor to recover to the original speed when it becomes smaller than the set magnitude.

[0004] Furthermore, in Patent Document 2, a method is shown for identifying a failure from the position of a different part by comparing a measured vibration waveform with a recorded normal vibration waveform.

[0005] On the other hand, in rotating equipment using a motor, various methods for judging abnormalities based on the magnitude of vibration have been proposed and put into practical use.

[0006] For example, in Patent Document 3, while conventionally the safety region, caution region, and abnormal region were determined by the amplitude of the vibration generated during the rotation of the motor, it is now divided for each rotation frequency component of the motor, and the safety region, caution region, and abnormal region are determined by the magnitude of each vibration.

[0007] Furthermore, Patent Document 4 describes a semiconductor manufacturing apparatus in which multiple sensors detect multiple physical quantities to measure and understand feature quantities. A fault prediction method is shown in which feature vectors are learned in advance from the feature quantities of both the fault state and the normal state, and an anomaly diagnosis is performed based on the difference between these and the feature vectors during actual operation. [Prior art documents] [Patent Documents]

[0008] [Patent Document 1] Patent No. 6392479 [Patent Document 2] Patent No. 6491390 [Patent Document 3] Special Publication No. 61-025884 [Patent Document 4] Patent No. 6860406 [Overview of the project] [Problems that the invention aims to solve]

[0009] As described in Patent Document 1, conventional vacuum valve devices use vibration sensors (accelerometers) to detect vibrations and utilize them for control. Furthermore, as described in Patent Document 2, the detected vibrations are compared with vibrations under normal conditions, and the fault is identified from the location information of the differences.

[0010] On the other hand, apart from vacuum valve devices, methods for determining abnormalities include setting safe, caution, and abnormal ranges for motor rotational speed vibrations, as described in Patent Document 3, and as described in Patent Document 4, identifying the characteristic quantities of the fault state and normal state of physical quantities detected in semiconductor manufacturing equipment and comparing them to diagnose abnormalities.

[0011] Since vacuum valve devices are composed of a combination of numerous mechanical parts, the vibrations generated during operation (the amplitude and frequency of each of the three-dimensional vector components) change subtly depending on the machining precision of each part and the assembly precision of the device as a whole.

[0012] Therefore, if a uniform abnormality standard is set for vibration and an abnormality is judged based on a uniform abnormal state, the initial vibration will differ for each vacuum valve device, and the change in vibration over time will also differ. For example, in the case of a good product with a good combination of component precision and assembly precision, the vibration may not reach a uniform abnormality, but the cumulative number of operations and cumulative operating time may exceed the allowable number of operations and time.

[0013] In this case, the system may continue operating without detecting the abnormality, potentially leading to a sudden failure (which could be described as extremely excessive vibration).

[0014] Therefore, the problem that the present invention aims to solve is to obtain appropriate abnormality diagnosis in response to the initial vibration and changes in vibration over time of individual vacuum valve devices. [Means for solving the problem]

[0015] To solve the above problems, the present invention provides a fault diagnosis method for a vacuum valve device, which diagnoses a fault based on vibration acceleration measured by an acceleration sensor installed on the vacuum valve device, and is characterized in that it calculates one evaluation value using vibration data acquired for each operation of the vibration acceleration, calculates the average of the evaluation values ​​from the first operation of the vacuum valve device to the current operation, which is calculated by multiplying the average value up to the previous operation by the number of previous operations, adds the evaluation value of the current operation, and divides by the current number of operations to calculate a predicted value for the next evaluation value, calculates a reference value by adding the variation for each operation for each individual device, the length of the downtime, and the allowable value due to changes in the state due to handling of the device during that time to the predicted value, and outputs that there is a fault indication when the next evaluation value exceeds the reference value.

[0016] The method for diagnosing a failure of the vacuum valve device is characterized in that the vibration data is time-series data consisting of three axis components (X, Y, and Z) acquired separately for the Close operation in which the vacuum valve device is sealed and the Open operation in which the vacuum valve device is opened, and the evaluation value is obtained by adding the absolute values ​​of the three axis components or by vector synthesis of the three axis components and summing them up from the start to the end of one operation.

[0017] In the method for diagnosing a failure of the vacuum valve device, the vibration data is time-series data composed of three-axis components of the X-axis, Y-axis, and Z-axis separately acquired during the Close operation in which the vacuum valve device is closed and the Open operation in which the vacuum valve device is opened, and the evaluation value is obtained by adding or vector-synthesizing the maximum values of each of the three-axis components from the start to the end of one operation.

[0018] In the method for diagnosing a failure of the vacuum valve device, the vibration data is time-series data composed of three-axis components of the X-axis, Y-axis, and Z-axis separately acquired during the Close operation in which the vacuum valve device is closed and the Open operation in which the vacuum valve device is opened, and the evaluation value is obtained by extracting the vibration data at arbitrary intervals, adding the absolute values of the three-axis components or vector-synthesizing the three-axis components, and summing from the start to the end of one operation.

[0019] In the method for diagnosing a failure of the vacuum valve device, when the vacuum valve device has not operated for a predetermined time or more, it is regarded as the first operation.

[0020] In the method for diagnosing a failure of the vacuum valve device, the predicted value is obtained by increasing the weight of the evaluation value of the current operation compared to the product of the average value up to the previous operation and the previous number of times, and then adding it and dividing by the current number of times.

[0021] In the method for diagnosing a failure of the vacuum valve device, the acceleration sensor is installed on the upper part, side surface, or lower surface of the housing covering the valve body driven by the motor.

[0022] In the method for diagnosing a failure of the vacuum valve device, the acceleration sensor is installed on the drive circuit board inside the housing covering the valve body driven by the motor.

[0023] In the method for diagnosing a failure of the vacuum valve device, the vibration data is time-series data composed of three-axis components of the X-axis, Y-axis, and Z-axis, and the degree of abnormality is diagnosed from the failure signs of the evaluation values of each of the three-axis components. This is the feature.

Advantages of the Invention

[0024] According to the present invention, appropriate abnormality diagnosis can be obtained according to the initial vibration and the vibration change over time of each vacuum valve, and there is no delay in abnormality diagnosis due to individual differences in vacuum valves. Therefore, there is an effect that highly reliable abnormality diagnosis can be provided.

Brief Description of the Drawings

[0025] [Figure 1] It is a diagram showing an overview of the vacuum valve device used in the method for diagnosing a failure of the vacuum valve device of the present invention. [Figure 2] It is a diagram showing the sealed state by the valve body of the vacuum valve device used in the method for diagnosing a failure of the vacuum valve device of the present invention. [Figure 3] It is a diagram showing a state where the valve body of the vacuum valve device used in the method for diagnosing a failure of the vacuum valve device of the present invention is at the uppermost part of the vertical movement. [Figure 4] It is a diagram showing a state where the valve body of the vacuum valve device used in the method for diagnosing a failure of the vacuum valve device of the present invention is in the middle part of the vertical movement. [Figure 5] It is a flowchart showing the flow of the method for diagnosing a failure of the vacuum valve device of the present invention. [Figure 6] It is a graph showing an example of the waveform of the vibration acceleration (initial Close operation) in the method for diagnosing a failure of the vacuum valve device of the present invention. [Figure 7] It is a graph showing an example of the waveform of the vibration acceleration (initial Open operation) in the method for diagnosing a failure of the vacuum valve device of the present invention. [Figure 8] It is a graph showing an example of the waveform of the vibration acceleration (later Close operation) in the method for diagnosing a failure of the vacuum valve device of the present invention. [Figure 9]This graph shows an example of the vibration acceleration waveform (late open operation) in the fault diagnosis method for a vacuum valve device according to the present invention. [Figure 10] This graph shows an example of the vibration acceleration waveform (during abnormal late-stage Close operation) in the fault diagnosis method for vacuum valve devices according to the present invention. [Figure 11] This graph shows an example of the vibration acceleration waveform (during abnormal late-open operation) in the fault diagnosis method for vacuum valve devices according to the present invention. [Figure 12] This graph shows an example of vibration acceleration evaluation (close operation) in the fault diagnosis method for vacuum valve devices according to the present invention. [Figure 13] This graph shows an example of vibration acceleration evaluation (Open operation) in the fault diagnosis method for vacuum valve devices according to the present invention. [Modes for carrying out the invention]

[0026] Embodiments of the present invention will be described in detail below with reference to the drawings. Components having the same function will be denoted by the same reference numerals, and repeated descriptions of such components may be omitted. [Examples]

[0027] Figure 1 shows an overview of the vacuum valve device used in the fault diagnosis method for vacuum valve devices according to the present invention.

[0028] The vacuum valve device 100 is installed between chambers 110 and 120 and consists of a valve body 200 that opens and closes the opening of chamber 110, a stem 500 that supports the valve body 200, a drive unit 400 that moves the stem 500 vertically and horizontally, and a housing 600 that covers the valve body 200. The opening of the housing 600 and the openings of chambers 110 and 120 are connected via a flange 130.

[0029] The valve body 200 can seal the opening of the housing 600 by contacting the opening edge, and a sealing material 300 is attached in an annular shape along the contact portion. By pressing the valve body 200 against the opening edge in such a way as to crush the sealing material 300, the opening of the chamber 110 is sealed.

[0030] The drive unit 400 has a mechanical configuration for moving the stem 500 vertically or horizontally, and a drive source for that configuration. For example, the mechanical configuration may include a vertical movement mechanism having a cam. The vertical movement mechanism converts the rotational force from the motor, which is the drive source, into linear motion via a ball screw.

[0031] A pin is passed horizontally through the stem 500 and connected to the cam. When the valve body 200, supported by the stem 500 by the vertical movement mechanism, reaches the top of its vertical movement, it comes into contact with a stopper, preventing further upward movement, and only the cam rotates.

[0032] The cam has a roughly elliptical shape, and as it rotates from an inclined state towards a horizontal state, it moves the stem 500 horizontally together with the pin. The valve body 200, supported by the stem 500, also moves horizontally and contacts the opening edge of the housing 600 via the sealing material 300.

[0033] By moving the stem 500 vertically and horizontally, the valve body 200 can seal or open the opening of the chamber 110. The cam and pin structure is designed with four pins in total, two on each side of the stem 500 (upper and lower), which suppresses rolling in all directions and allows the valve body 200 to move almost horizontally.

[0034] An acceleration sensor 700 is installed on the top of the housing 600. The acceleration sensor 700 should be capable of detecting acceleration in three axes: X-axis acceleration component 710, Y-axis acceleration component 720, and Z-axis acceleration component 730. The vibration acceleration measured by the acceleration sensor 700 is used to estimate whether or not the vacuum valve device 100 has malfunctioned.

[0035] Figure 2 shows the sealing state of the vacuum valve device used in the fault diagnosis method for the vacuum valve device. When the valve body 200 is pushed in horizontally, the cross-section of the sealing material 300 sandwiched between it and the opening edge of the housing 600 deforms into a roughly elliptical shape, resulting in a high degree of sealing of the chamber 110.

[0036] Figure 3 shows the state in which the valve body of a vacuum valve device, used in the fault diagnosis method for vacuum valve devices, is at the top of its vertical movement. The pin of the stem 500 is not shifted horizontally. If the vertical movement mechanism moves upward from this position, the pin will move the stem 500 horizontally. Conversely, if the vertical movement mechanism moves downward, it will move the stem 500 vertically downward.

[0037] Figure 4 shows the state in which the valve body of a vacuum valve device used in a fault diagnosis method for a vacuum valve device is in the middle of its vertical movement. The chamber 110 is open, and the stem 500 moves vertically according to the upward or downward movement of the vertical movement mechanism.

[0038] For each Close operation, which moves the valve body 200 from the open state to the closed state, and for each Open operation, which moves the valve body 200 from the closed state to the open state, the vibration acceleration is measured by the acceleration sensor 700, and the state of the vacuum valve device 100 (normal, abnormal, or faulty) is estimated from the accumulated data.

[0039] Figure 5 is a flowchart illustrating the procedure for diagnosing a fault in a vacuum valve device. In Step 1, after the i-th Close or Open operation is completed, the absolute values ​​of the three-axis vibration acceleration data Ax, Ay, and Az are added together from the start time 0 to the end time te, and an evaluation value α(i) of the vibration data acquired for each operation is calculated. Here, K is a coefficient for adjusting the magnitude of α.

[0040]

number

[0041] In Step 2, the current (i-1) evaluation value α(i-1) and the average evaluation value up to the previous (i-2) evaluation value αav(i-2) are used to calculate the average evaluation value up to the present αav(i-1). This formula allows us to calculate the average evaluation value up to the present from the current evaluation value and the average evaluation value up to the previous evaluation value, so it is not necessary to retain all past evaluation values ​​in order to calculate the average evaluation value.

[0042]

number

[0043] In Step 3, the current average value αav(i-1) is used as the predicted value for the next cycle, and the tolerance value e0 is added to this to determine the next reference value E(i). The tolerance value is set considering factors such as the variability of each individual device operation, the length of the downtime, and the changes in the device's state due to handling during that time.

[0044]

number

[0045] In step 4, it is determined whether the evaluation value α(i) is greater than the reference value E(i). If the evaluation value α(i) exceeds the reference value E(i), it is determined that the vacuum valve device 100 is in some kind of abnormal state, and an abnormality alarm is triggered. Then, the process continues to the next evaluation value (i+1). This process is repeated to detect any abnormal state in the vacuum valve device 100.

[0046] Figures 6-11 are graphs showing examples of vibration acceleration waveforms in the fault diagnosis method for vacuum valve devices. The horizontal axis represents time, and the vertical axis represents acceleration. (1) is the X-axis component when vibration acceleration is large, (2) is the Y-axis component when vibration acceleration is large, (3) is the Z-axis component when vibration acceleration is large, (4) is the X-axis component when vibration acceleration is small, (5) is the Y-axis component when vibration acceleration is small, and (6) is the Z-axis component when vibration acceleration is small.

[0047] Figure 6 shows the vibration acceleration during the initial Close operation with a small number of movement steps. In the region where the vibration is large immediately after the start of operation, the valve body 200 is moving vertically and at high speed. In the region where the vibration is small in the latter half, the valve body 200 is moving horizontally and at low speed. Near where the sealing material 300 of the valve body 200 contacts the opening edge of the housing 600, the operation is performed at a low speed to prevent excessive vibration, and therefore the vibration is not noticeable. Thus, even under normal conditions, there is a difference in the magnitude of vibration acceleration between (1)(2)(3) and (4)(5)(6), indicating that some degree of variability should be expected.

[0048] Figure 7 shows the vibration acceleration during the initial Open operation. The region with low vibration acceleration in the first half is generally the region where the valve body 200 moves horizontally, and the seal material 300 is operating at a low speed to suppress vibration during delamination. The region with high vibration acceleration in the second half is where the valve body 200 moves vertically and at high speed.

[0049] Figure 8 shows the vibration acceleration during the Close operation in the later stages after a certain number of operation cycles. Figure 9 shows the vibration acceleration during the Open operation in the later stages. Clearly, compared to the initial stages shown in Figures 6 and 7, an overall increase in vibration due to time progression can be observed. Furthermore, when comparing the cases with high and low vibration acceleration, a variation similar to or greater than that in the initial stages can be observed.

[0050] Figure 10 shows the vibration acceleration during the Close operation when additional abnormalities occur in the later stages. Figure 11 shows the vibration acceleration during the Open operation when additional abnormalities occur. From this, it can be seen that the vibration acceleration is increasing more than the increase in vibration acceleration over time shown in Figures 8 and 9, but in the case of small changes, it is difficult to distinguish from the case of large changes over time.

[0051] Thus, while it may be possible for a person to visually inspect the vibration acceleration waveform itself and determine increases over time or increases during abnormal situations, automatically judging the waveform itself would require storing all the data and judging the differences each time, which would necessitate a huge amount of recording capacity as the number of operations increases. Therefore, the vibration acceleration waveforms of the three axes are expressed as an overall evaluation value.

[0052] Table 1 shows the evaluation values ​​calculated using the aforementioned equation 1 for the example vibration acceleration shown in Figures 6-11, and Table 2 shows the evaluation values ​​for the Close operation.

[0053] [Table 1]

[0054] [Table 2]

[0055] From this, it seems possible to determine the differences between the early and late stages (changes over time) and the differences in abnormalities between the later stages (abnormality detection). However, it is clear that it is difficult to uniformly define abnormality detection criteria for the early and late stages. In other words, it is necessary to define abnormality detection criteria based on the evaluation values ​​of the early stages and the evaluation values ​​of the changes over time.

[0056] Figure 12 is a graph showing an example of vibration acceleration evaluation in a fault diagnosis method for a vacuum valve device (close operation). Figure 13 is a graph showing an example of vibration acceleration evaluation in a fault diagnosis method for a vacuum valve device (open operation). The horizontal axis represents the number of measurements, and the vertical axis represents the evaluation value α(i). The predicted value αAV(i-1) and the reference value (sum of the predicted value and the allowable value) E(i) are shown, respectively.

[0057] In this example, an abnormality is detected when the evaluated value exceeds the baseline value (actual value). The diagram shows that an abnormality occurs at trial A, and at that time the evaluated value exceeds the baseline value, allowing for an abnormality detection. This abnormality is mainly caused by an increase in load due to deterioration of the grease in the ball screw of the drive unit.

[0058] Occasionally, before an abnormality occurs, the evaluation value may change suddenly, albeit at a small rate. This is due to external factors such as the device being shut down for a period of time for adjustments or being removed and reinstalled, and does not indicate an essential change in the valve device's state. Therefore, by establishing a baseline value based on the predicted value, misjudgments due to these factors can be prevented.

[0059] Furthermore, while an abnormality can be determined from the results of a Close operation by setting a certain threshold value, in the results of an Open operation, a certain threshold value cannot distinguish between fluctuations in the evaluation value during the process and abnormal conditions. Therefore, predicted values ​​are used to correctly identify abnormal states.

[0060] In this example, the variation in vibration acceleration in a single vacuum valve device was shown. However, if multiple vacuum valve devices are compared, the vibration acceleration itself will change even under the same initial conditions, and the degree of variation will also change. However, as shown in the embodiment of the present invention, the predicted value is successively corrected from the initial evaluation value of vibration acceleration. This allows for the acquisition of a predicted value that reflects the variation and magnitude of vibration acceleration in each individual device. By using this as a reference value with a certain tolerance added, it becomes possible to determine abnormalities in each device.

[0061] Furthermore, the closing and opening speeds of this valve device can be arbitrarily set according to the application of the customer's equipment. In that case, if the speed is increased to prioritize throughput, the vibration acceleration will increase, and if the speed is decreased to prioritize low vibration, the vibration acceleration will decrease. According to the embodiment of the present invention, since the predicted values ​​are set appropriately in response to such changes in conditions, there is also the effect that the detection of abnormal conditions is less affected by the usage conditions.

[0062] Furthermore, as shown here, the predicted value αAV(i-1) is a historical average, so it includes some degree of change over time and shows a slight increasing trend. Therefore, it is possible to predict maintenance timing by judging the increase in the predicted value αAV(i-1) itself and using this value alone.

[0063] Note that in equation 1, for the sake of simplicity, the absolute values ​​of each component of vibration acceleration are added together, but it is also acceptable to use the vector sum of each component. This allows for an evaluation that is closer to the magnitude of the actual acceleration.

[0064]

number

[0065] Furthermore, while the above evaluation values ​​for vibration acceleration are combined into a single value by adding time-series data, it is also possible to represent each vibration acceleration component with its maximum value. In this case, since not all vibration accelerations are evaluated, the accuracy will be reduced, but the calculation process can be simplified.

[0066]

number

[0067] In the formula in Math 2, the average of past evaluation values ​​and the previous evaluation value are treated equally, and the average of the previous evaluation value is calculated. However, it is also possible to assign weights to past and previous evaluation values. If A > B, and this ratio is large, then more weight is given to evaluation values ​​that are closer to the present than to past values, which has the effect of obtaining predicted values ​​that reflect the current situation.

[0068]

number

[0069] The abnormality assessment described so far assumes that it is performed on all data sequences. However, the vibrations that appear will differ between Close and Open operations because the way the load is applied to the drive source and the load state over time are different. Therefore, it is desirable to perform abnormality assessments for Close operations and Open operations separately.

[0070] Furthermore, judging every operation would result in hundreds of thousands of data points in post-processing. Also, since signs of anomalies are unlikely to appear as significant changes in a single operation, but rather over dozens or hundreds of operations, it is desirable to thin out the vibration acceleration data stream and perform anomaly detection using vibration data extracted at arbitrary intervals. The number of thinned-out points is (J-1), and vibration acceleration is obtained every J points, which are then used as the data stream for anomaly detection. Typically, J is set to around 100 to 200 points, and obtaining 1,000 to 500 anomaly scores over 100,000 operations makes post-processing easier.

[0071] However, if the vacuum valve device 100 is idle for a certain period of time, there may be changes in its state due to aging or maintenance of the related equipment itself, which could cause a sudden change in vibration. Therefore, even if J is set but the number of operations has not yet reached J, it is desirable to acquire the vibration acceleration and perform abnormality detection processing. In this case, the next operation will be the first.

[0072] Up to this point, evaluation values ​​have been obtained from the three axes of vibration acceleration, but it is also possible to perform similar anomaly detection by focusing on just one axis. If this is done with an inexpensive single-axis acceleration sensor, data processing will only require processing for one axis, making it simpler.

[0073] Alternatively, each of the three axes can be evaluated individually. Although the number of data points triples, this allows for more detailed anomaly detection tailored to the characteristics of each of the three axes. Specifically, the X-axis component, which inherently has no movable components, is sensitive to anomalies caused by play in the device. The Y-axis component, which is greatly affected by the closing and opening movements of the valve body 200, is sensitive to anomalies in the closed state of the valve body 200. The Z-axis component, which is greatly affected by the vertical movement of the stem 500, is sensitive to anomalies in the drive source such as the motor or ball screw, and the vertical drive mechanism.

[0074] Furthermore, although the acceleration sensor 700 is installed on the top of the housing 600, it may be installed elsewhere. If there is no space to install the acceleration sensor 700 on the top, it can be installed on the side or bottom of the housing 600, and similar abnormality detection will be possible. In this case, depending on the installation location, it may be close to the ball screw or the motor, which will make it more sensitive to abnormalities in those components, resulting in more detailed abnormality detection.

[0075] Furthermore, if drive circuit boards are provided inside the housing 600 of the vacuum valve device 100, the acceleration sensor 700 may be installed in one corner of the circuit board. In this case, there is the advantage that external wiring between the acceleration sensor 700 and the circuit is not required.

[0076] In the embodiments of the present invention, deterioration of the ball screw sliding surface (grease deterioration) was shown as a cause of abnormality, but other factors may also be present. Due to the deterioration of the sealing material 300, the suction force may increase, which may increase the vibration when the sealing material 300 peels off during the Open operation. This is particularly noticeable during the Open operation, so it is expected that abnormality detection during the Open operation will precede any other abnormalities.

[0077] Furthermore, the cam converts the vertical movement of the ball screw into the horizontal movement of the valve body 200, but deterioration of this cam, such as wear, can increase vibration. Since the operating range of the ball screw in which these increases in vibration acceleration occur differs, the time of increase differs in the vibration acceleration waveforms shown in Figures 6 to 11 of the embodiment of the present invention. However, the evaluation value is reflected as an increase in the evaluation value regardless of the factor.

[0078] In the embodiments described above, the tolerance value e0 is set to a constant value, but in order to reflect the individuality of each device, it may also be determined by the ratio of each predicted value.

[0079]

number

[0080] Here, K1 is a constant, and it is best to set n to 5 or greater, where n is n times the ratio σ / μ of the standard deviation σ and mean μ of the predicted value αav(m). This exceeds the variability of the predicted value, allowing for appropriate abnormality determination that takes into account the variability of each individual device. Note that it is also possible to set an arbitrary tolerance value by considering the actual variability of the devices and individual devices, rather than determining it using such a ratio.

[0081] According to the present invention, it is possible to obtain appropriate abnormality diagnoses in response to the initial vibrations of individual vacuum valves and changes in vibrations over time. Since there is no delay in abnormality diagnosis due to individual differences in vacuum valves, it has the effect of providing highly reliable abnormality diagnoses.

[0082] The embodiments of the present invention have been described above, but the invention is not limited to these embodiments. [Explanation of symbols]

[0083] 100: Vacuum valve device 110: Chamber 120: Chamber 130: Flange 200: Valve body 300: Sealant 400: Drive mechanism 500: Stem 600: Cabinet 700: Accelerometer 710:X-axis acceleration component 720: Y-axis acceleration component 730: Z-axis acceleration component

Claims

1. A method for diagnosing a malfunction based on vibration acceleration measured by an acceleration sensor installed in a vacuum valve device, Using the vibration data acquired for each operation, a single evaluation value is calculated. The average of the evaluation values ​​from the first operation of the vacuum valve device to the current operation is calculated by multiplying the average value up to the previous operation by the number of previous operations, adding the evaluation value of the current operation, and dividing by the current number of operations. This is then used as the predicted value for the next evaluation value. A reference value is calculated by adding to the aforementioned predicted value the variability for each individual device operation, the length of the downtime, and the allowable value due to changes in the device's state caused by handling during that time. If the next evaluation value exceeds the aforementioned threshold value, it will output that there is a potential failure. A method for diagnosing failures in a vacuum valve device, characterized by the features described herein.

2. The vibration data is time-series data consisting of three axis components (X, Y, and Z) acquired separately during the Close operation in which the vacuum valve device is sealed and the Open operation in which the vacuum valve device is opened. The aforementioned evaluation value is obtained by adding the absolute values ​​of the three axis components or by vector synthesis of the three axis components and summing them up from the start to the end of one operation. The method for diagnosing a failure in a vacuum valve device according to feature 1.

3. The vibration data is time-series data consisting of three axis components (X, Y, and Z) acquired separately during the Close operation in which the vacuum valve device is sealed and the Open operation in which the vacuum valve device is opened. The aforementioned evaluation value is obtained by adding or vector-combining the maximum values ​​of each of the three axis components from the start to the end of one movement. The method for diagnosing a failure in a vacuum valve device according to feature 1.

4. The vibration data is time-series data consisting of three axis components (X, Y, and Z) acquired separately during the Close operation in which the vacuum valve device is sealed and the Open operation in which the vacuum valve device is opened. The aforementioned evaluation value is obtained by extracting the vibration data at any number of intervals, adding the absolute values ​​of the three-axis components or performing vector synthesis of the three-axis components, and summing the values ​​from the start to the end of one operation. The method for diagnosing a failure in a vacuum valve device according to feature 1.

5. If the vacuum valve device does not operate for a predetermined period of time or longer, it will be considered to have performed its first operation. The method for diagnosing a failure in a vacuum valve device according to feature 1.

6. The aforementioned predicted value is calculated by adding the evaluation value of the current action (with a greater weight than the average value up to the previous action multiplied by the number of previous actions) and then dividing by the number of current actions. The method for diagnosing a failure in a vacuum valve device according to feature 1.

7. The acceleration sensor is installed on the top, side, or bottom surface of the housing that covers the motor-driven valve body. The method for diagnosing a failure in a vacuum valve device according to feature 1.

8. The acceleration sensor is installed on a drive circuit board inside the housing that covers the motor-driven valve body. The method for diagnosing a failure in a vacuum valve device according to feature 1.

9. The vibration data is time-series data consisting of three axial components: X, Y, and Z. The degree of abnormality is diagnosed from the fault prediction indicators of the evaluation values ​​of each of the three axial components. The method for diagnosing a failure in a vacuum valve device according to feature 1.

Images