WO2021145156A1 - 異常診断システム及び異常診断方法 - Google Patents
異常診断システム及び異常診断方法 Download PDFInfo
- Publication number
- WO2021145156A1 WO2021145156A1 PCT/JP2020/047726 JP2020047726W WO2021145156A1 WO 2021145156 A1 WO2021145156 A1 WO 2021145156A1 JP 2020047726 W JP2020047726 W JP 2020047726W WO 2021145156 A1 WO2021145156 A1 WO 2021145156A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- abnormality
- principal component
- component analysis
- series data
- equipment
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0218—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults
- G05B23/0221—Preprocessing measurements, e.g. data collection rate adjustment; Standardization of measurements; Time series or signal analysis, e.g. frequency analysis or wavelets; Trustworthiness of measurements; Indexes therefor; Measurements using easily measured parameters to estimate parameters difficult to measure; Virtual sensor creation; De-noising; Sensor fusion; Unconventional preprocessing inherently present in specific fault detection methods like PCA-based methods
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0218—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults
- G05B23/0224—Process history based detection method, e.g. whereby history implies the availability of large amounts of data
- G05B23/024—Quantitative history assessment, e.g. mathematical relationships between available data; Functions therefor; Principal component analysis [PCA]; Partial least square [PLS]; Statistical classifiers, e.g. Bayesian networks, linear regression or correlation analysis; Neural networks
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M99/00—Subject matter not provided for in other groups of this subclass
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B2223/00—Indexing scheme associated with group G05B23/00
- G05B2223/02—Indirect monitoring, e.g. monitoring production to detect faults of a system
Definitions
- the present invention relates to an abnormality diagnosis system and an abnormality diagnosis method for diagnosing an abnormality in equipment having an equipment drive unit.
- dies for width pressing are arranged on both sides of the slab to be conveyed, and power is transmitted from the motor to the dies via a crank to press the slab in a fixed cycle.
- the current waveform or load waveform that repeatedly appears at this time is collected for each material, and the current waveform or load waveform for one crank cycle is analyzed by PCA for each material to extract an abnormal waveform.
- an object of the present invention is to provide an abnormality diagnosis system and an abnormality diagnosis method capable of suppressing a misdiagnosis that an equipment abnormality has occurred when an equipment abnormality has not actually occurred.
- the present invention provides the following (1) to (10).
- An abnormality diagnosis system that diagnoses abnormalities in equipment that repeats a certain pattern.
- a sampling unit that repeats a pattern corresponding to the fixed pattern and samples time-series data indicating the state of the equipment.
- a calculation unit that diagnoses an abnormality in the equipment by performing principal component analysis based on the time-series data sampled by the sampling unit. Have, The calculation unit The first principal component analysis is performed on one pattern in the time series data indicating the state, and the first abnormality presence / absence judgment is performed based on the result.
- a plurality of temporally continuous patterns including the pattern obtained by performing the first principal component analysis in the time series data indicating the state are collectively the first.
- the principal component analysis of 2 is performed, and the presence or absence of the second abnormality is determined based on the result.
- An abnormality diagnosis system characterized in that an equipment abnormality is diagnosed when the second abnormality determination is abnormal.
- the first abnormality presence / absence determination and the second abnormality presence / absence determination are characterized in that they are based on the Q statistic obtained from the results of the first principal component analysis and the second principal component analysis.
- the number of sampling points of the time series data used for the first principal component analysis in the calculation unit is N1
- the dimension compression by the first principal component analysis is k1 dimension
- the second principal component analysis is performed.
- the number of sampling points of the time series data used for is N2 points
- the dimension compression by the second principal component analysis is k2 dimensions
- the number of the plurality of patterns is M (M ⁇ 2), N2 ⁇ N1 ⁇ M
- the abnormality diagnosis system according to (1) or (2) above, wherein k2 ⁇ k1 ⁇ M is established.
- the equipment is a sizing press equipment in which a mold is driven by a drive motor to reduce the width of the slab, and the time-series data indicating the state of the equipment is the current waveform of the drive motor or the load waveform of the mold.
- the abnormality diagnosis system according to any one of (1) to (3) above, wherein the abnormality diagnosis system is provided.
- the data pre-processing unit further includes a data pre-processing unit that performs pre-processing including a process of normalizing the height of the waveform and a process of classifying the time-series data sampled by the sampling unit according to the width reduction amount.
- the abnormality diagnosis system according to (4) above.
- An abnormality diagnosis method for diagnosing an abnormality in equipment that repeats a certain pattern A step of sampling time-series data indicating the state of the equipment, which repeats a pattern corresponding to the fixed pattern, and The step of performing the first principal component analysis on one pattern in the time series data showing the above state, The step of determining the presence or absence of the first abnormality from the result of the first principal component analysis, When the first abnormality presence / absence determination is abnormal, a plurality of temporally continuous patterns including the pattern obtained by performing the first principal component analysis in the time series data indicating the state are collectively the first.
- the first abnormality presence / absence determination and the second abnormality presence / absence determination are characterized in that they are based on the Q statistic obtained from the results of the first principal component analysis and the second principal component analysis.
- the dimension compression by the first principal component analysis is k1 dimension
- the time series data is used for the second principal component analysis.
- the dimension compression by the second principal component analysis is k2 dimensions
- the number of the plurality of patterns is M (M ⁇ 2), N2 ⁇ N1 ⁇ M or k2 ⁇ k1
- the equipment is a sizing press equipment in which a mold is driven by a drive motor to reduce the width of the slab, and the time-series data indicating the state of the equipment is the current waveform of the drive motor or the load waveform of the mold.
- the abnormality diagnosis method according to any one of (6) to (8) above, wherein the abnormality is present.
- the time-series data sampled above is further provided with a step of performing preprocessing including a process of normalizing the height of the waveform and a process of classifying by the amount of width reduction (9). ) Described in the abnormality diagnosis method.
- an abnormality diagnosis system and an abnormality diagnosis method capable of suppressing a misdiagnosis that an equipment abnormality has occurred when an equipment abnormality has not actually occurred.
- FIG. 1 is a schematic diagram and a block diagram showing an example in which an abnormality diagnosis system according to an embodiment is used in a sizing press facility.
- FIG. 2 is a diagram showing an example of a sample waveform when extracting an abnormal waveform.
- FIG. 3 is a diagram for explaining a PCA performed using the sample waveform of FIG.
- FIG. 4 is a flowchart showing an abnormality diagnosis method executed in the abnormality diagnosis system according to the embodiment.
- FIG. 5 is a diagram showing a comparison between the case where the load waveform, which is time-series data indicating the load, is a normal waveform and the case where the load waveform is an equipment abnormality waveform.
- FIG. 1 is a schematic diagram and a block diagram showing an example in which an abnormality diagnosis system according to an embodiment is used in a sizing press facility.
- FIG. 2 is a diagram showing an example of a sample waveform when extracting an abnormal waveform.
- FIG. 3 is a diagram for explaining a PCA
- FIG. 6 is a diagram showing a comparison between the case where all the load waveforms, which are time-series data indicating the load, are normal waveforms and the case where the slip generation waveform exists.
- FIG. 7 is a diagram showing the relationship between the number of cycles and the Q statistic (mean square error) when the slip generation waveform is included for one cycle (one waveform).
- FIG. 1 is a schematic diagram and a block diagram showing an example in which an abnormality diagnosis system according to an embodiment is used in a sizing press facility.
- the sizing press facility 10 performs width reduction on the slab 11 in the hot rolling line after continuous casting.
- the sizing press facility 10 is provided corresponding to a pair of dies 2, an inlet side pinch roll 3 and an outlet side pinch roll 4, and a pair of dies 2, and a pair of drive motors for driving the dies 2. It has 5 and a pair of cranks 6 that connect the mold 2 and the drive motor 5, respectively.
- the slab 11 is carried between the pair of dies 2 facing each other by the entry side pinch roll 3, and is swung by the power transmitted from the drive motor 5 via the crank 6.
- the width is reduced by the mold 2 and carried out by the exit side pinch roll 4. While the slab 11 is being conveyed intermittently, the slab 11 is repeatedly reduced in width by the mold 2 at regular intervals to become a slab having a predetermined width.
- the abnormality diagnosis system 20 diagnoses an equipment abnormality of the sizing press equipment 10, and has a sampling unit 21, a data preprocessing unit 22, a storage unit 23, a calculation unit 24, and an output unit 25.
- the sampling unit 21 samples time-series data indicating the state of the equipment.
- the time-series data shows the load such as the current value of the drive motor 5 and the load of the mold 2 when the slab 11 is reduced in width by the mold 2.
- Crank angle information can be used as the time information for obtaining time series data.
- the time-series data indicating the load such as the current value and the load becomes one waveform (pattern) in which one cycle of the crank 6 has a peak, and such a waveform appears repeatedly. Specifically, it is a current waveform and a load waveform.
- the data preprocessing unit 22 preprocesses the data sampled by the sampling unit 21. As the preprocessing, a process of normalizing the height of the waveform corresponding to one cycle of the crank 6 and a process of classifying the data according to the width killing amount (width reduction amount) at the time of width reduction are performed.
- the storage unit 23 stores the preprocessed data.
- the storage unit 23 stores time-series data (normal waveform), which is a reference waveform and indicates a load in a normal state, which is obtained in advance.
- the calculation unit 24 performs an abnormality diagnosis using principal component analysis (PCA) on the waveform of the time-series data (current value and load) of the load after preprocessing stored in the storage unit 23.
- PCA principal component analysis
- the analysis method using PCA is a method of extracting a waveform having a shape different from that of a normal chart waveform (abnormality) from a large number of chart waveforms.
- a multivariate statistical process management method is used with PCA. To use.
- FIG. 2 is a diagram showing an example of a sample waveform when an abnormal waveform is extracted using PCA.
- the horizontal axis is time
- the vertical axis is a load such as a power value and a load.
- N samples are collected at predetermined intervals in the waveform collection time zone (number of samples: N). For example, the number of samples N is 50 if the sampling time is 5 seconds and the sampling time is 100 ms.
- the image diagram when this is analyzed by PCA is shown in FIG. 3, and for the sake of simplicity, the diagram in the case of a two-dimensional space where the number of data points is two points is shown, and the first element axis (X axis) and the first element axis (X axis) of the data.
- the axis with the largest dispersion is the first principal component with respect to the points in space represented by the two-element axis (Y axis).
- the axis orthogonal to the first principal component is shown as the second principal component as shown in the figure.
- the second principal component axis that maximizes the variance is determined.
- the kth principal component (k ⁇ N) is the axis orthogonal to the first to (k-1) principal components.
- the one with the maximum dispersion is the k-th principal component.
- the Q statistic is evaluated.
- the Q statistic is the sum of squares (sum of squared errors) of the distance between each sample and the principal component axis.
- the mean square error obtained by dividing this value by the number of samples is the Q statistic. It is preferable to use as.
- the Q statistic is evaluated only by the distance from the first principal component. Since the abnormal sample has a large distance from the principal component axis, the Q statistic increases when there is an abnormal sample, and it can be determined that the Q statistic is abnormal, and the possibility of equipment abnormality can be grasped. can.
- a threshold value for the Q statistic can be set, and if the Q statistic is equal to or greater than the threshold value, it can be determined that there is a possibility of equipment abnormality.
- the Q statistic becomes large, and it is determined that the Q statistic has an abnormality.
- the calculation unit 24 extracts one waveform for one crank cycle from the time-series data of the load after preprocessing stored in the storage unit 23, and performs PCA on the waveform to obtain a Q statistic. Ask. Then, the presence or absence of abnormality in the Q statistical value is determined. For example, when the Q statistic is larger than the threshold value, it is determined that the Q statistic is abnormal. If there is no abnormality in the Q statistic, a signal without abnormality is sent to the output unit 25. On the other hand, if there is an abnormality in the Q statistic, there is a possibility of equipment abnormality, but there may be a case where the load is removed, which will be described later.
- FIG. 4 is a flowchart showing an abnormality diagnosis method executed in the abnormality diagnosis system 20.
- sizing press facility 10 power is transmitted from the drive motor 5 to the die 2 via the crank 6 while intermittently conveying the slab 11, and the die 2 reduces the width of the slab 11 at regular intervals.
- a sizing press is performed. Such a sizing press is continuously performed on a plurality of slabs 11.
- the abnormality of the sizing press equipment 10 is diagnosed by the following method from the time series data showing the load such as the current value of the drive motor 5 and the load of the die 2 during the sizing press. do.
- the abnormality diagnosis system 20 samples the time-series data of the load in the sizing press of the slab 11 of the predetermined material by the sampling unit 21 (step 1).
- time-series data indicating the load becomes one waveform having a peak for one cycle of the crank 6, and such a waveform appears repeatedly.
- time-series data is a current waveform of the drive motor 5 and a load waveform of the mold 2.
- the abnormality diagnosis system 20 preprocesses the data sampled by the sampling unit 21 by the data preprocessing unit 22 (step 2).
- the process of normalizing the height of the waveform corresponding to one cycle of the crank 6 and the process of classifying the data according to the width killing amount (width reduction amount) at the time of width reduction are performed. Will be done.
- the reason why the data is classified according to the width reduction amount (width reduction amount) when the width reduction is applied is that the crank angle at which the load is generated differs depending on the width reduction amount, and the waveform varies. Specifically, the smaller the width killing amount, the later the load generation timing.
- the data after such preprocessing is stored in the storage unit 23.
- the abnormality diagnosis system 20 extracts one waveform corresponding to one cycle of the crank 6 from the time-series data after the pretreatment of the slab of the material stored in the storage unit 23, and the one waveform.
- PCA is performed for (step 3).
- step 3 PCA is performed using the waveform of the time series data at the normal time (normal waveform) as the reference waveform, and the Q statistic is obtained from the result by the multivariate statistical process management method as described above.
- step 4 determine whether the Q statistic is abnormal. For example, when the Q statistic is equal to or greater than the threshold value, it is determined that the Q statistic is abnormal.
- the load waveform which is time-series data indicating the load, becomes an equipment abnormality waveform different from the normal waveform in all cycles. Therefore, if there is no abnormality in one waveform corresponding to one cycle, it can be assumed that there is no equipment abnormality.
- the horizontal axis is the crank angle corresponding to time.
- the vertical axis is the load of the mold. The same applies to FIG. 6 described below.
- the waveform of the cycle in which the slip occurs is a slip generation waveform similar to the equipment abnormality waveform.
- the load waveform which is time-series data indicating the load, is in a state where slip generation waveforms are sporadically inserted into the normal waveform. Therefore, when the slip generation waveform is accidentally extracted and PCA is performed, the Q statistic becomes large as in the case of equipment abnormality.
- Such a phenomenon in which no load is applied occurs in about 1 to 2 cycles out of about 10 cycles, but cannot be predicted.
- one waveform in step 3 is obtained from the time-series data after the pretreatment of the slab of the material stored in the storage unit 23.
- a plurality of continuous waveforms for a plurality of cycles including the waveforms are extracted, and PCA is performed on the plurality of waveforms (step 5).
- step 5 the waveform of the time-series data at the normal time (normal waveform) is used as the reference waveform, and PCA is collectively (simultaneously) performed on a plurality of waveforms (multiple patterns that are continuous in time) for a plurality of cycles. From the result, the Q statistic is obtained by the multivariate statistical process management method as described above.
- step 6 it is determined whether or not the Q statistic obtained in step 5 is abnormal (step 6).
- the load waveform is different from the normal waveform in all cycles. Therefore, even when the PCA having a plurality of waveforms is executed in step 5, Q The statistic (mean squared error) is the same value as when PCA is executed for one waveform.
- the abnormality due to the PCA of one waveform determined in step 4 is a load relief due to pinch roll slip that occurs sporadically in about 1 to 2 cycles out of 10 cycles, a plurality of waveforms for a plurality of cycles Of these, the waveforms other than the slip occurrence waveform are almost normal waveforms as shown in FIG. 6, so the Q statistic is smaller than that in the case of equipment abnormality.
- step 6 it is possible to determine whether the equipment is abnormal or not based on the Q statistic. If the Q statistic is abnormal, for example, if the Q statistic is equal to or greater than the threshold value, it is diagnosed that there is an equipment abnormality. If the Q statistic is not abnormal, for example, if the Q statistic is less than the threshold value, it is diagnosed that there is no equipment abnormality.
- the number of sampling points of the time series data of one pattern used in the first principal component analysis is N1
- the dimension after dimension compression by the first principal component analysis is k1 dimension
- the number of sampling points of the time series data of the plurality of patterns used in the second principal component analysis is N2 points
- the dimension after dimension compression by the second principal component analysis is k2 dimension
- the number of patterns of the plurality of patterns is M (M ⁇ ). 2), that is, the waveform corresponding to one cycle used in the first principal component analysis is defined as the M waveform.
- N2 ⁇ N1 ⁇ M or k2 ⁇ k1 ⁇ M since the characteristics of the time-series signals for the M waveform are more important than the disturbance of the waveform in one cycle, N2 ⁇ N1 ⁇ M or k2 ⁇ k1 ⁇ M. Data may be acquired so that the relationship is established.
- the abnormality diagnosis using the detailed data of one pattern is performed, and in the case of the abnormality diagnosis using a plurality of patterns, the characteristics of one waveform are saved to some extent.
- the shape of the data as an M waveform can be saved.
- the first abnormality determination can detect minute disturbance of the waveform of one pattern
- the second abnormality determination determines the repeatability of the waveform disturbance and the deviation of the appearance of the waveform (time delay). Focused detection will be possible, and highly accurate and highly reliable abnormality detection will be possible.
- FIG. 7 is a diagram showing the relationship between the number of cycles and the Q statistic (mean squared error) when the slip generation waveform is included for one cycle (one waveform).
- the Q statistic mean squared error
- FIG. 7 it can be seen that as the number of cycles (that is, the number of waveforms) used in the calculation of PCA increases, the value of the Q statistic (mean squared error) becomes smaller, making it easier to distinguish between equipment abnormality and load relief. For example, if the Q statistic (mean square error) is 10,000 or more, it is possible to determine that the equipment is abnormal and if it is less than 10,000, the load is released due to slipping. Further, from FIG. 7, it can be seen that when the number of cycles is 3 or more (the number of waveforms is 3 or more), it becomes easier to distinguish between equipment abnormality and load release due to slip.
- the abnormality diagnosis system 20 for diagnosing the abnormality of the sizing press equipment 10 it is possible to suppress the misdiagnosis that the equipment abnormality has occurred when the equipment abnormality has not actually occurred. can do. Therefore, it is possible to reduce equipment stoppages and the like due to misdiagnosis, and it is possible to suppress a decrease in production efficiency.
- the preprocessing unit 22 normalizes the height of the waveform corresponding to one cycle of the crank 6, and classifies the data according to the width killing amount (width reduction amount) at the time of width reduction. Since the processing is performed, it is possible to suppress the variation of the waveform due to the width killing amount (width reduction amount).
- the present invention is not limited to the above embodiment and can be variously modified within the scope of the idea of the present invention.
- the present invention is applied to the abnormality detection of the sizing press equipment, but it can be applied as long as the operation of the equipment drive unit repeats a certain pattern.
- an example of a waveform is shown as a repeating pattern of time-series data indicating a load, but the present invention is not limited to this.
- the time series data indicating the load current or load is exemplified, but the present invention is not limited to this.
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Automation & Control Theory (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
Description
前記一定のパターンに対応するパターンを繰り返す、前記設備の状態を示す時系列データをサンプリングするサンプリング部と、
前記サンプリング部でサンプリングした前記時系列データに基づいて主成分分析を行って前記設備の異常を診断する演算部と、
を有し、
前記演算部は、
前記状態を示す時系列データの中の一つのパターンについて第1の主成分分析を行って、その結果に基づいて第1の異常有無判定を行い、
前記第1の異常有無判定が異常ありの場合に、前記状態を示す時系列データの中の前記第1の主成分分析を行ったパターンを含む時間的に連続する複数のパターンについて一括して第2の主成分分析を行って、その結果に基づいて第2の異常有無判定を行い、
前記第2の異常有無判定が異常ありの場合に設備異常と診断することを特徴とする異常診断システム。
前記一定のパターンに対応するパターンを繰り返す、前記設備の状態を示す時系列データをサンプリングするステップと、
前記状態を示す時系列データの中の一つのパターンについて第1の主成分分析を行うステップと、
前記第1の主成分分析の結果から第1の異常有無判定を行うステップと、
前記第1の異常有無判定が異常ありの場合に、前記状態を示す時系列データの中の前記第1の主成分分析を行ったパターンを含む時間的に連続する複数のパターンについて一括して第2の主成分分析を行うステップと、
前記第2の主成分分析の結果から第2の異常有無判定を行うステップと、
を有し、
前記第2の異常有無判定が異常ありの場合に設備異常であると診断することを特徴とする異常診断方法。
本実施形態では、本発明をサイジングプレス設備の異常診断に適用した場合について説明する。
図1は、サイジングプレス設備に一実施形態に係る異常診断システムを用いた例を示す模式図及びブロック図である。
図4は、異常診断システム20において実行される異常診断方法を示すフローチャートである。
3 入側ピンチロール
4 出側ピンチロール
5 駆動モータ
6 クランク
10 サイジングプレス設備
11 スラブ
20 異常診断システム
21 サンプリング部
22 データ前処理部
23 記憶部
24 演算部
25 出力部
Claims (10)
- 一定のパターンを繰り返す動作を行う設備の異常診断を行う異常診断システムであって、
前記一定のパターンに対応するパターンを繰り返す、前記設備の状態を示す時系列データをサンプリングするサンプリング部と、
前記サンプリング部でサンプリングした前記時系列データに基づいて主成分分析を行って前記設備の異常を診断する演算部と、
を有し、
前記演算部は、
前記状態を示す時系列データの中の一つのパターンについて第1の主成分分析を行って、その結果に基づいて第1の異常有無判定を行い、
前記第1の異常有無判定が異常ありの場合に、前記状態を示す時系列データの中の前記第1の主成分分析を行ったパターンを含む時間的に連続する複数のパターンについて一括して第2の主成分分析を行って、その結果に基づいて第2の異常有無判定を行い、
前記第2の異常有無判定が異常ありの場合に設備異常と診断することを特徴とする異常診断システム。 - 前記第1の異常有無判定及び前記第2の異常有無判定は、前記第1の主成分分析及び前記第2の主成分分析の結果から求められたQ統計量に基づくことを特徴とする請求項1に記載の異常診断システム。
- 前記演算部における前記第1の主成分分析に使用する時系列データのサンプリング点数をN1点、前記第1の主成分分析による次元圧縮をk1次元、とし、前記第2の主成分分析に使用する時系列データのサンプリング点数をN2点、前記第2の主成分分析による次元圧縮をk2次元、前記複数のパターンの数をM個(M≧2)とするとき、N2<N1×M又はk2<k1×Mが成立することを特徴とする請求項1又は請求項2に記載の異常診断システム。
- 前記設備は駆動モータにより金型を駆動してスラブを幅圧下するサイジングプレス設備であり、前記設備の状態を示す時系列データは、駆動モータの電流波形又は前記金型の荷重波形であることを特徴とする請求項1から請求項3のいずれか一項に記載の異常診断システム。
- 前記サンプリング部でサンプリングした前記時系列データについて、波形の高さを正規化する処理、及び、幅圧下量により分類する処理を含む前処理を行うデータ前処理部をさらに有することを特徴とする請求項4に記載の異常診断システム。
- 一定のパターンを繰り返す動作を行う設備の異常診断を行う異常診断方法であって、
前記一定のパターンに対応するパターンを繰り返す、前記設備の状態を示す時系列データをサンプリングするステップと、
前記状態を示す時系列データの中の一つのパターンについて第1の主成分分析を行うステップと、
前記第1の主成分分析の結果から第1の異常有無判定を行うステップと、
前記第1の異常有無判定が異常ありの場合に、前記状態を示す時系列データの中の前記第1の主成分分析を行ったパターンを含む時間的に連続する複数のパターンについて一括して第2の主成分分析を行うステップと、
前記第2の主成分分析の結果から第2の異常有無判定を行うステップと、
を有し、
前記第2の異常有無判定が異常ありの場合に設備異常であると診断することを特徴とする異常診断方法。 - 前記第1の異常有無判定及び前記第2の異常有無判定は、前記第1の主成分分析及び前記第2の主成分分析の結果から求められたQ統計量に基づくことを特徴とする請求項6に記載の異常診断方法。
- 前記第1の主成分分析に使用する時系列データのサンプリング点数をN1点、前記第1の主成分分析による次元圧縮をk1次元、とし、前記第2の主成分分析に使用する時系列データのサンプリング点数をN2点、前記第2の主成分分析による次元圧縮をk2次元、前記複数のパターンの数をM個(M≧2)とするとき、N2<N1×M又はk2<k1×Mが成立することを特徴とする請求項6又は請求項7に記載の異常診断方法。
- 前記設備は駆動モータにより金型を駆動してスラブを幅圧下するサイジングプレス設備であり、前記設備の状態を示す時系列データは、駆動モータの電流波形又は前記金型の荷重波形であることを特徴とする請求項6から請求項8のいずれか一項に記載の異常診断方法。
- 前記でサンプリングした前記時系列データについて、波形の高さを正規化する処理、及び、幅圧下量により分類する処理を含む前処理を行うステップをさらに有することを特徴とする請求項9に記載の異常診断方法。
Priority Applications (6)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202080089721.8A CN114846315A (zh) | 2020-01-14 | 2020-12-21 | 异常诊断系统以及异常诊断方法 |
KR1020227021529A KR20220104042A (ko) | 2020-01-14 | 2020-12-21 | 이상 진단 시스템 및 이상 진단 방법 |
US17/758,358 US20230024947A1 (en) | 2020-01-14 | 2020-12-21 | Abnormality diagnosis system and abnormality diagnosis method |
BR112022013365A BR112022013365A2 (pt) | 2020-01-14 | 2020-12-21 | Sistema para diagnóstico de anormalidade e método para diagnóstico de anormalidade |
EP20913851.0A EP4092501A4 (en) | 2020-01-14 | 2020-12-21 | FAULT DIAGNOSTIC SYSTEM AND FAULT DIAGNOSTIC METHOD |
JP2021514439A JP6915763B1 (ja) | 2020-01-14 | 2020-12-21 | 異常診断システム及び異常診断方法 |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2020-003769 | 2020-01-14 | ||
JP2020003769 | 2020-01-14 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2021145156A1 true WO2021145156A1 (ja) | 2021-07-22 |
Family
ID=76863703
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2020/047726 WO2021145156A1 (ja) | 2020-01-14 | 2020-12-21 | 異常診断システム及び異常診断方法 |
Country Status (8)
Country | Link |
---|---|
US (1) | US20230024947A1 (ja) |
EP (1) | EP4092501A4 (ja) |
JP (1) | JP6915763B1 (ja) |
KR (1) | KR20220104042A (ja) |
CN (1) | CN114846315A (ja) |
BR (1) | BR112022013365A2 (ja) |
TW (1) | TWI760050B (ja) |
WO (1) | WO2021145156A1 (ja) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2023113056A (ja) * | 2022-02-02 | 2023-08-15 | 株式会社日立製作所 | タクト分割データ収集システム、及び、異常検知システム |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2002181038A (ja) * | 2000-12-18 | 2002-06-26 | Mitsubishi Heavy Ind Ltd | 異常診断装置 |
JP2008224519A (ja) * | 2007-03-14 | 2008-09-25 | Omron Corp | 検査装置及びその検査装置における見直し時期の判定方法 |
WO2016039088A1 (ja) * | 2014-09-12 | 2016-03-17 | 株式会社神戸製鋼所 | 回転機械異常検出装置および該方法ならびに回転機 |
JP5991042B2 (ja) | 2011-06-29 | 2016-09-14 | Jfeスチール株式会社 | 異常監視システムおよび異常監視方法 |
JP2017032567A (ja) | 2015-08-05 | 2017-02-09 | Jfeスチール株式会社 | モータの監視装置および方法 |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP4762088B2 (ja) * | 2006-08-31 | 2011-08-31 | 株式会社東芝 | プロセス異常診断装置 |
CN103191931B (zh) * | 2012-01-10 | 2015-03-04 | 宝山钢铁股份有限公司 | 热连轧机零调后二侧偏差控制方法 |
TWI571329B (zh) * | 2016-03-02 | 2017-02-21 | 中國鋼鐵股份有限公司 | 軋延追蹤方法 |
WO2017176401A1 (en) * | 2016-04-05 | 2017-10-12 | Siemens Energy, Inc. | Shaft ground monitoring with diagnostic waveform analysis |
CN109996615B (zh) * | 2016-11-28 | 2020-07-07 | 东芝三菱电机产业系统株式会社 | 轧制设备的异常诊断方法及装置 |
-
2020
- 2020-12-21 WO PCT/JP2020/047726 patent/WO2021145156A1/ja unknown
- 2020-12-21 CN CN202080089721.8A patent/CN114846315A/zh active Pending
- 2020-12-21 BR BR112022013365A patent/BR112022013365A2/pt unknown
- 2020-12-21 US US17/758,358 patent/US20230024947A1/en active Pending
- 2020-12-21 KR KR1020227021529A patent/KR20220104042A/ko unknown
- 2020-12-21 EP EP20913851.0A patent/EP4092501A4/en active Pending
- 2020-12-21 JP JP2021514439A patent/JP6915763B1/ja active Active
-
2021
- 2021-01-07 TW TW110100529A patent/TWI760050B/zh active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2002181038A (ja) * | 2000-12-18 | 2002-06-26 | Mitsubishi Heavy Ind Ltd | 異常診断装置 |
JP2008224519A (ja) * | 2007-03-14 | 2008-09-25 | Omron Corp | 検査装置及びその検査装置における見直し時期の判定方法 |
JP5991042B2 (ja) | 2011-06-29 | 2016-09-14 | Jfeスチール株式会社 | 異常監視システムおよび異常監視方法 |
WO2016039088A1 (ja) * | 2014-09-12 | 2016-03-17 | 株式会社神戸製鋼所 | 回転機械異常検出装置および該方法ならびに回転機 |
JP2017032567A (ja) | 2015-08-05 | 2017-02-09 | Jfeスチール株式会社 | モータの監視装置および方法 |
Non-Patent Citations (1)
Title |
---|
See also references of EP4092501A4 |
Also Published As
Publication number | Publication date |
---|---|
EP4092501A1 (en) | 2022-11-23 |
JPWO2021145156A1 (ja) | 2021-07-22 |
JP6915763B1 (ja) | 2021-08-04 |
TWI760050B (zh) | 2022-04-01 |
BR112022013365A2 (pt) | 2022-09-13 |
US20230024947A1 (en) | 2023-01-26 |
TW202138076A (zh) | 2021-10-16 |
KR20220104042A (ko) | 2022-07-25 |
CN114846315A (zh) | 2022-08-02 |
EP4092501A4 (en) | 2023-07-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP3454289B1 (en) | Plant abnormality detection method and system | |
US11392114B2 (en) | Abnormality determination support apparatus | |
CN108620950B (zh) | 一种车削刀具加工状态监测方法及系统 | |
KR102417415B1 (ko) | 딥러닝 이상 감지 기반 생산 제품 불량 검출 방법 및 시스템 | |
CN115239711B (zh) | 一种缝制设备的在线运行异常识别系统 | |
WO2021145156A1 (ja) | 異常診断システム及び異常診断方法 | |
CN109141625B (zh) | 一种滚珠丝杠副的在线状态监测方法 | |
CN113924207B (zh) | 异常探测装置、异常探测方法以及存储介质 | |
KR102026069B1 (ko) | 반도체 설비의 센서 데이터 분할 시스템 및 그 방법 | |
KR102356222B1 (ko) | 압력변화를 이용한 금형의 지능형 모니터링 시스템 | |
CN110108474A (zh) | 一种旋转机械运行稳定性在线监测与评估方法及系统 | |
CN111678699A (zh) | 一种面向滚动轴承早期故障监测与诊断方法及系统 | |
CN117314890B (zh) | 打扣加工的安全控制方法、装置、设备及存储介质 | |
Brophy et al. | AI-based condition monitoring of the drilling process | |
CN111538755A (zh) | 一种基于归一化互相关与单位根检验的设备运行状态异常检测方法 | |
CN117258932A (zh) | 智能砂磨机的温度监测系统及方法 | |
CN115034094B (zh) | 一种金属加工机床运行状态预测方法及系统 | |
KR102356220B1 (ko) | 음파변화를 이용한 금형의 지능형 모니터링 시스템 | |
CN115329942A (zh) | 一种基于人工智能的螺栓紧固装置及其方法 | |
RU2809934C1 (ru) | Система диагностики аномалий и способ диагностики аномалий | |
CN115655764B (zh) | 振动趋势分析方法、装置、电子设备及存储介质 | |
EP3951530B1 (en) | Process state monitoring device and process state monitoring method | |
CN110929425B (zh) | 一种基于Lasso算法的纸张纵向伸长率的软测量方法 | |
CN112567306A (zh) | 制造过程监视装置 | |
CN115112853A (zh) | 螺栓红冲加工过程在线质量检测方法 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
ENP | Entry into the national phase |
Ref document number: 2021514439 Country of ref document: JP Kind code of ref document: A |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 20913851 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 20227021529 Country of ref document: KR Kind code of ref document: A |
|
REG | Reference to national code |
Ref country code: BR Ref legal event code: B01A Ref document number: 112022013365 Country of ref document: BR |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
ENP | Entry into the national phase |
Ref document number: 2020913851 Country of ref document: EP Effective date: 20220816 |
|
ENP | Entry into the national phase |
Ref document number: 112022013365 Country of ref document: BR Kind code of ref document: A2 Effective date: 20220704 |