Abnormality detector, method, and program

Abstract

To flexibly detect an abnormality of a manufacturing apparatus.SOLUTION: An abnormality detector in accordance with an embodiment includes an acquisition unit, designation unit, and detection unit. The acquisition unit acquires analysis data including at least operation data that is one of operation data on a manufacturing apparatus which intermittently machines at least one product and history data on a manufacture history of the product machined by the manufacturing apparatus. The designation unit designates an operating condition of the manufacturing apparatus in relation to the operation data on the basis of the acquired analysis data. The detection unit detects an abnormality of the manufacturing apparatus on the basis of the operation data and designated operating condition.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] An embodiment of the present invention relates to an anomaly detection device, method, and program. [Background technology]

[0002] In the manufacturing industry, abnormalities in manufacturing equipment are detected based on the operation data of the manufacturing equipment. In many cases, manufacturing equipment prepares for processing the next product (e.g., cleanup, adjustment, cleaning, maintenance) every time it processes a product (e.g., manufacturing, processing, assembly, inspection) (intermittent processing). Manufacturing equipment may temporarily halt the processing it is currently doing to make preparations.

[0003] In particular, there is a demand for flexible detection of abnormalities in manufacturing equipment according to the "operational state" of the manufacturing equipment. For example, it may be desirable to detect abnormalities more strictly in operation data in a state where the manufacturing equipment is processing products (processing state) than in operation data in a state where the manufacturing equipment is not processing products (non-processing state). However, operation data that does not include the operating state of the manufacturing equipment cannot satisfy the above demand. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] International Publication No. 2021 / 241580 Summary of the Invention [Problem to be solved by the invention]

[0005] The problem to be solved by the present invention is to flexibly detect anomalies in manufacturing equipment. [Means for solving the problem]

[0006] The anomaly detection device according to the embodiment includes an acquisition unit, a setting unit, and a detection unit. The acquisition unit acquires analysis data including at least operation data of a manufacturing device that intermittently processes at least one product and history data related to a manufacturing history of the product processed by the manufacturing device. The setting unit sets an operation state of the manufacturing device for the operation data based on the acquired analysis data. The detection unit detects an anomaly of the manufacturing device based on the operation data and the set operation state. [Brief description of the drawings]

[0007] [Figure 1] FIG. 1 is a block diagram showing an example of the functional configuration of an anomaly detection device according to a first embodiment. [Diagram 2] 4 is a flowchart showing an example of the operation of the anomaly detection device according to the first embodiment. [Diagram 3] FIG. 4 is a diagram showing an example of action data according to the first embodiment. [Figure 4] FIG. 4 is a diagram showing an example of history data according to the first embodiment. [Diagram 5] FIG. 4 is a diagram showing a first example of an operating state according to the first embodiment. [Figure 6] FIG. 4 is a diagram showing a second example of an operating state according to the first embodiment. [Figure 7] FIG. 4 is a diagram showing an example of setting an operation state according to the first embodiment. [Figure 8] 5A to 5C are diagrams showing examples of abnormality detection for each operating state according to the first embodiment. [Figure 9] FIG. 4 is a diagram showing a first example of anomaly detection according to the first embodiment. [Figure 10] FIG. 11 is a diagram showing a second example of anomaly detection according to the first embodiment. [Figure 11] FIG. 11 is a view showing a third example of anomaly detection according to the first embodiment. [Figure 12] FIG. 11 is a view showing a fourth example of anomaly detection according to the first embodiment. [Figure 13] FIG. 11 is a block diagram showing an example of the functional configuration of an anomaly detection device according to a second embodiment. [Figure 14] 10 is a flowchart showing an example of the operation of the anomaly detection device according to the second embodiment. [Figure 15] FIG. 13 is a block diagram showing an example of the functional configuration of an anomaly detection device according to a third embodiment. [Figure 16] 13 is a flowchart showing an example of the operation of the anomaly detection device according to the third embodiment. [Figure 17] FIG. 2 is a block diagram showing an example of a hardware configuration of the anomaly detection device according to each embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0008] Hereinafter, each embodiment will be described with reference to the drawings. In the following embodiments, parts with the same reference numerals perform the same operations, and duplicated descriptions will be omitted as appropriate.

[0009] (First embodiment) 1 is a block diagram showing an example of a functional configuration of an anomaly detection device 1 according to a first embodiment. The anomaly detection device 1 is a device that detects an anomaly in a manufacturing device M that intermittently processes at least one product (material, work-in-progress). The anomaly detection device 1 is communicatively connected to a manufacturing device DB 2 and a manufacturing history DB 3. The anomaly detection device 1 includes an acquisition unit 11, a setting unit 12, and a detection unit 13.

[0010] The acquisition unit 11 acquires various data. For example, the acquisition unit 11 acquires analysis data 100 including at least the operation data 200 out of the operation data 200 and the history data 300. Specifically, the acquisition unit 11 acquires the operation data 200 from the manufacturing equipment DB 2 and acquires the history data 300 from the manufacturing history DB 3. The acquisition unit 11 transfers the acquired analysis data 100 to the setting unit 12.

[0011] The operation data 200 is data obtained by measuring numerical values ​​(hereinafter also referred to as "operation numerical values") related to the operation of the manufacturing equipment M over time. The operation data 200 may be acquired over time by various sensors attached to the manufacturing equipment M. The acquired operation data 200 is stored in the manufacturing equipment DB2.

[0012] The various sensors may be pre-installed in the manufacturing equipment M or may be newly installed. The various sensors may or may not be connected to the manufacturing equipment M. For example, a sensor such as a camera may measure the operating values ​​of the manufacturing equipment M from a location remote from the manufacturing equipment M.

[0013] The operating values ​​of the manufacturing equipment M may be input values ​​required for the operation of the manufacturing equipment M, output values ​​generated in conjunction with the operation of the manufacturing equipment M, or values ​​measuring the product manufactured in conjunction with the operation of the manufacturing equipment M. First, the operating values ​​may be the temperature, pressure, current, voltage, etc. of the manufacturing equipment M. Second, the operating values ​​may be data of the location (module) where the manufacturing equipment M processes the product. The operating values ​​may be the temperature of the location where the manufacturing equipment M processes the product at high or low temperature, or the pressure of the location where the manufacturing equipment M processes the product at high or low pressure. The operating values ​​may be energy, parameters, etc. for controlling the temperature, pressure, etc. of the location. Third, the operating values ​​may be data measuring the state of the product while it is being processed by the manufacturing equipment M. The operating values ​​may be data measuring the transition or change in the temperature, color, shape, etc. of the product.

[0014] Unless there is a special reason, the operational data 200 is constantly acquired and recorded over time. For example, since the manufacturing equipment M processes products intermittently, the sensor continues to measure the operational values ​​of the manufacturing equipment M even during the time period when the manufacturing equipment M is not processing products.

[0015] For example, when a camera measures the color of a product manufactured by manufacturing equipment M, operational data 200 may be recorded that measures the background color of a location where the manufacturing equipment M would normally process a product during a time period when the manufacturing equipment M is not manufacturing the product.

[0016] The history data 300 is data related to the manufacturing history of a product processed by the manufacturing equipment M. The history data 300 may be generated automatically or manually by the manufacturing equipment M. The generated history data 300 is stored in the manufacturing history DB3.

[0017] The setting unit 12 sets various data. For example, the setting unit 12 sets the operation state of the manufacturing equipment M for the operation data 200 based on the analysis data 100 acquired by the acquisition unit 11. The setting unit 12 transfers the operation data 200 in which the operation state is set to the detection unit 13.

[0018] The detection unit 13 detects various types of data. For example, the detection unit 13 detects an abnormality in the manufacturing equipment M based on the operation data 200 in which the operation state is set by the setting unit 12. The detection unit 13 outputs output data 500 including the detection result of the abnormality to the outside. The output data 500 may be a text file or an image file in a predetermined format (e.g., CSV, JSON, JPEG).

[0019] The manufacturing equipment DB2 is a database that stores operation data 200 of the manufacturing equipment M. The manufacturing history DB3 is a database that stores history data 300 of the manufacturing equipment M. The manufacturing equipment DB2 and the manufacturing history DB3 may be text files in a predetermined format (e.g., CSV, JSON). The manufacturing equipment DB2 and the manufacturing history DB3 may be integrated into one database or text file.

[0020] 2 is a flowchart showing an operation example of the anomaly detection device 1 according to the first embodiment. The anomaly detection device 1 may execute each step of this operation example in response to a start instruction from a user. Alternatively, the anomaly detection device 1 may execute each step of this operation example at any time interval (e.g., one day, one week, one month).

[0021] (Step S1) First, the acquiring unit 11 acquires the analysis data 100. The acquiring unit 11 may acquire the analysis data 100 from only the motion data 200, or from the motion data 200 and the history data 300, in whole or in part. The acquiring unit 11 may acquire a predetermined range of the analysis data 100 according to analysis parameters input by a user or another system. The acquiring unit 11 may determine an analysis target range (data acquisition range) of the motion data 200 by referring to the acquired motion data 200 or history data 300. The analysis target range of the motion data 200 may be included in the analysis parameters (see FIGS. 3 and 4).

[0022] In particular, the acquisition unit 11 may add a range adjustment amount to the analysis parameter. The acquisition unit 11 may acquire analysis data 100 in a range wider or narrower than a predetermined range based on the analysis parameter according to the analysis parameter to which the range adjustment amount has been added, or may acquire analysis data 100 in a range shifted from the predetermined range. For example, the acquisition unit 11 acquires analysis data 100 in a range wider than a predetermined range based on the analysis parameter. This allows the acquisition unit 11 to prevent the analysis data 100 required for analysis from being missed, and to prevent the correspondence between the operation data 200 and the history data 300 from being missed.

[0023] (Step S2) Next, the setting unit 12 sets the operation state of the manufacturing equipment M. Specifically, the setting unit 12 sets the operation state of the manufacturing equipment M for the operation data 200 based on the analysis data 100 acquired in step S1. The setting unit 12 may set the operation state of the manufacturing equipment M for all or a part of the operation data 200 (see FIGS. 5, 6, and 7).

[0024] In particular, the setting unit 12 may set a predetermined operation state for a period from the time when the value of the operation data 200 exceeds a threshold to the time when it falls below the threshold. Alternatively, the setting unit 12 may set a predetermined operation state for a period from the time when the value of the operation data 200 falls below a threshold to the time when it exceeds the threshold. This allows the setting unit 12 to appropriately set an operation state for the operation data 200 in accordance with fluctuations in the value of the operation data 200.

[0025] (Step S3) Finally, the detection unit 13 detects an abnormality in the manufacturing equipment M. Specifically, the detection unit 13 detects an abnormality in the manufacturing equipment M based on the operation data 200 including the operation state set in step S2. The detection unit 13 may detect an abnormality in the manufacturing equipment M using a known statistical method or machine learning method. The detection unit 13 may execute a different abnormality detection method for each operation state of the manufacturing equipment M according to analysis parameters input by a user or another system. The detection unit 13 outputs output data 500 including the detection result of the abnormality in the manufacturing equipment M to the outside (see FIGS. 8, 9, 10, 11, and 12).

[0026] The detection unit 13 may calculate the "degree of abnormality" of the operation data 200 as a detection result of an abnormality in the manufacturing equipment M. The degree of abnormality indicates the magnitude or likelihood of the abnormality. The degree of abnormality may be a continuous value or a discrete value. For example, "degree of abnormality: 0" indicates "no abnormality" and "degree of abnormality: 1" indicates "abnormality". In this case, the closer the degree of abnormality is to "1", the higher the degree of abnormality of the operation data 200. Alternatively, the degree of abnormality may be a categorical value such as "abnormality", "caution", or "warning".

[0027] The detection unit 13 may calculate the degree of abnormality according to the period or frequency during which the value of the operation data 200 exceeds a threshold. The threshold may be a predetermined value, or a value calculated from the average and standard deviation of the values ​​of the operation data 200 within a predetermined range (e.g., three times the standard deviation).

[0028] The detection unit 13 may calculate the degree of abnormality as a probability value in a chi-square distribution from the average and standard deviation of the values ​​of the motion data 200 based on the Hotelling theory. The smaller the probability value, the more abnormal the value of the motion data 200 is (the variation that occurs rarely). Alternatively, the detection unit 13 may calculate a value from 0 to 1 as the probability value in the 1-chi-square distribution. The detection unit 13 may calculate the probability value in the chi-square distribution under the definition that the smaller the value, the higher the degree of abnormality.

[0029] The detection unit 13 may preprocess the operation data 200 and calculate the degree of abnormality based on the preprocessed operation data 200. First, the detection unit 13 may perform noise removal, moving average, or section average as preprocessing. This allows the detection unit 13 to reduce the influence of noise or unintended abrupt changes contained in the operation data 200 on the calculation of the degree of abnormality. Second, the detection unit 13 may perform difference, cumulative sum, differentiation, or integration as preprocessing. This preprocessing is effective when the amount of change in the value of the operation data 200, rather than the absolute value of the value of the operation data 200, is related to the abnormality. Third, the detection unit 13 may perform arithmetic operations on a specific data string of the operation data 200 as preprocessing. For example, an actual measurement value of the pressure when the manufacturing device M processes a product may be recorded as the operation data 200. The detection unit 13 divides the actual measurement value by a set value of the pressure set in the manufacturing device M. As a result, the detection unit 13 converts the actual measurement value into a value indicating what percentage of the set pressure the product was processed at. Therefore, the detection unit 13 can easily detect an abnormality in the manufacturing equipment M by threshold processing.

[0030] The detection unit 13 may use a waveform (normal waveform) of the normal operation data 200 to calculate the degree of abnormality for the waveform (target waveform) of the operation data 200 to be processed. For example, the detection unit 13 calculates the degree of deviation between the normal waveform and the target waveform as the degree of abnormality. The degree of deviation may be a simple difference or RMSE (Root Mean Squared Error). Alternatively, the detection unit 13 may associate two waveforms of different lengths with each other by DTW (Dynamic Time Warping). If the difference between the two waveforms exceeds a reference value, the detection unit 13 may calculate the degree of abnormality to be high.

[0031] The detection unit 13 may add labels to the normal motion data 200 and the abnormal motion data 200. The detection unit 13 may train a machine learning model (e.g., a perceptron, a support vector machine, a decision tree, or a neural network) by supervised learning using the normal motion data 200 and the abnormal motion data 200 to which the labels have been added. The detection unit 13 may classify the motion data 200 to be processed into a "normal" or "abnormal" class (abnormal mode) using the trained machine learning model.

[0032] The detection unit 13 may train a machine learning model to calculate the degree of anomaly of normal operation data 200 as "0" and the degree of anomaly of abnormal operation data 200 as "1". The detection unit 13 may use the trained machine learning model to calculate the degree of anomaly of the operation data 200 to be processed so as to regress to "0" or "1". The detection unit 13 may use various regression methods (e.g., simple regression, multiple regression, kernel regression, support vector regression, neural network regression).

[0033] The detection unit 13 may calculate features of the normal motion data 200 by using a plurality of types of normal motion data 200. The detection unit 13 may train a machine learning model by unsupervised learning using the calculated features. The detection unit 13 may calculate, as an anomaly degree, a deviation of the motion data 200 to be processed from the normal motion data 200 by using the trained machine learning model.

[0034] The multiple types of normal operation data 200 may be data that has been determined to be normal by a human being. Alternatively, the multiple types of normal operation data 200 may be data from a period when the yield of the manufacturing equipment M was relatively good, or data from a period when no trouble occurred with the manufacturing equipment M.

[0035] In supervised learning or unsupervised learning, the detection unit 13 may train the machine learning model using the operation state of the manufacturing equipment M set in the operation data 200, in addition to the operation data 200. This enables the detection unit 13 to acquire appropriate parameters for each operation state of the manufacturing equipment M, and train the machine learning model using the acquired parameters.

[0036] The detection unit 13 may detect an abnormality in the operation data 200 by performing a time series analysis on the operation data 200. The detection unit 13 may use various models (e.g., an autoregressive model, an autoregressive integrated moving average model, a hidden Markov model, a Long Short Term Memory (LSTM), and a Transformer).

[0037] 3 is a diagram showing an example of operation data 200 according to the first embodiment. Table 210 shows operation data 200 corresponding to one manufacturing device M. Table 210 includes items "Time", "Data 1", "Data 2", "Data 3", ... "Data N" in the column direction, and includes records related to each item in the row direction.

[0038] The item "Time" is the time when each record was recorded. According to table 210, each record is recorded at 10 second intervals. The recording interval may be any time interval or may be an irregular time interval. Instead of "Time", a sequential number (e.g. 1, 2, 3...) indicating the order or time interval of each process by manufacturing equipment M may be recorded.

[0039] Items "Data 1" to "Data N" are various types of operation data 200. According to table 210, N types of data are recorded (N is any natural number). Each data may be a measurement value of temperature, pressure, current, or voltage measured by a sensor at a predetermined location of manufacturing equipment M. Each data may be a continuous value or a discrete value.

[0040] When multiple manufacturing equipment M exists, operation data 200 similar to table 210 may be acquired for each of the multiple manufacturing equipment M. The manufacturing equipment DB2 may store table data or file data corresponding to each acquired operation data 200. Alternatively, table 210 may include a data string related to an identifier (e.g., equipment number) that specifies each of the multiple manufacturing equipment M.

[0041] The acquisition unit 11 may determine the analysis target range by referring to the operation data 200 (table 210). As described above, the manufacturing equipment M intermittently processes the product. A certain type of operation data 200 changes in response to the intermittent processing by the manufacturing equipment M. First, when the operation data 200 indicates a measured value of pressure, the measured value rises with the start of the pressurization processing by the manufacturing equipment M and falls with the end of the pressurization processing. Second, when the operation data 200 indicates a measured value of current or voltage, the measured value rises with the start of the processing using electric power by the manufacturing equipment M and falls with the end of the processing. Pressure, electric power, current, voltage, etc. are examples of the operation data 200.

[0042] The acquiring unit 11 may determine a period from the time when the value of the operation data 200 reaches or exceeds a threshold value to the time when the value reaches or falls below the threshold value as the analysis range of the operation data 200. When the value or change of the operation data 200 satisfies a predetermined condition, the acquiring unit 11 may determine a period satisfying the condition as the analysis range of the operation data 200. That is, the acquiring unit 11 may determine the analysis range in response to intermittent processing by the manufacturing equipment M.

[0043] After determining the analysis range of the action data 200 based on the history data 300, the acquiring unit 11 may correct the influence of rounding of the value of the history data 300 by using the change in the value of the action data 200. This method is effective in cases where the start time and end time recorded in the history data 300 are rounded.

[0044] Fig. 4 is a diagram showing an example of history data 300 according to the first embodiment. Table 310 in Fig. 4(A) shows history data 300 according to one example. Table 320 in Fig. 4(B) shows history data 300 according to another example. Tables 310 and 320 are history data 300 corresponding to one manufacturing device M. Each time the manufacturing device M processes a product, a record is recorded in tables 310 and 320.

[0045] Tables 310 and 320 include items "Index," "ID," "Start time," and "End time" in the column direction, and records related to each item in the row direction. In particular, table 320 further includes an item "Work content" in the column direction.

[0046] The item "Index" is a number (e.g., processing sequence) that uniquely specifies each process performed intermittently by manufacturing equipment M. The item "ID" is the number of the product processed by manufacturing equipment M (e.g., component number, lot number, serial number, batch number). The items "Start time" and "End time" are the start time and end time of the process corresponding to the "Index" and "ID". The period between the start time and end time of the process is the "processing time" of the process.

[0047] The "start time" and "end time" may be rounded to any granularity (e.g., one minute, one hour). In particular, if the history data 300 is recorded manually, the acquisition unit 11 may round the history data 300 to any granularity.

[0048] The item "Work content" indicates the specific content of the process performed by the manufacturing equipment M. "Work content" may be "manufacturing", "maintenance", "preparation", or "cleanup". In particular, "maintenance", "preparation", and "cleanup" are processes that do not target products with a specific ID, and are therefore also referred to as "non-processing". The period between the start time and end time of a non-processing is the "non-processing time" of that non-processing.

[0049] When multiple manufacturing equipment M exists, history data 300 similar to table 310 or 320 may be acquired for each of the multiple manufacturing equipment M. The manufacturing history DB 3 may store table data or file data corresponding to each acquired history data 300. Alternatively, table 310 or 320 may include a data string related to an identifier (e.g., equipment number) that specifies each of the multiple manufacturing equipment M.

[0050] When one manufacturing device M manufactures multiple types of products, history data 300 similar to table 310 or 320 may be acquired for each of the multiple types of products. The manufacturing history DB 3 may store table data or file data corresponding to each acquired history data 300. Alternatively, table 310 or 320 may include a data string related to an identifier (e.g., product number) that specifies each of the multiple types of products.

[0051] The acquiring unit 11 may determine the analysis target range of the action data 200 by referring to the history data 300 (table 310 or 320). The acquiring unit 11 may determine the analysis target range based on the "start time" and "end time" corresponding to the "ID" in the table 310 or 320. When multiple IDs correspond to the analysis target range, the acquiring unit 11 may select at least one ID from the multiple IDs. When multiple analysis target ranges exist, the detecting unit 13 may detect an abnormality in the action data 200 corresponding to the multiple analysis target ranges.

[0052] The acquiring unit 11 may specify at least one "ID" in the table 310 or 320 related to the history data 300 as the analysis target range. For example, the acquiring unit 11 acquires the history data 300 corresponding to a predetermined ID in the table 310 or 320. The acquiring unit 11 acquires the operation data 200 corresponding to the start time and end time by using the start time and end time included in the acquired history data 300. This method is effective in cases where the detection unit 13 wants to output an anomaly detection result for the operation data 200 when the manufacturing equipment M processes a specific product.

[0053] Fig. 5 is a diagram showing a first example of an operating state according to the first embodiment. Table 410 in Fig. 5(A) shows a status code set for each point (time) in the operating data 200. Table 420 in Fig. 5(B) shows a correspondence between status codes and operating states according to one example. Table 430 in Fig. 5(C) shows a correspondence between status codes and operating states according to another example.

[0054] Table 410 includes items "time" and "status code" in the column direction, and records related to each item in the row direction. The item "time" in table 410 corresponds to the item "time" in table 210 related to operation data 200 of manufacturing equipment M (see FIG. 3). The item "status code" in table 410 corresponds to the item "status code" in tables 420 and 430.

[0055] According to table 410, manufacturing equipment M alternates between a "non-processing state" and a "processing state" over time. Some records in table 410 record "time" and "state code: 0." That is, at this time, the operating state of manufacturing equipment M is a "non-processing state." Other records in table 410 record "time" and "state code: 1." That is, at this time, the operating state of manufacturing equipment M is a "processing state."

[0056] The "status code" in table 410 may be replaced by the "operation status" in table 420. That is, table 410 may include names (status names) related to the operation status of manufacturing equipment M. The "status code" in table 410 may include more than the two types of operation status (0: non-processing, 1: processing) shown in table 420. Specifically, the "status code" in table 410 may include five types of operation status (0: non-processing, 1: processing, 2: pre-preparation, 3: cleanup, 4: maintenance) shown in table 430.

[0057] Fig. 6 is a diagram showing a second example of an operating state according to the first embodiment. Tables 440 and 450 in Fig. 6(A) and 6(B) include data strings corresponding to two operating states, "STATE 1" and "STATE 2", respectively. Table 460 in Fig. 6(C) includes one data string indicating four operating states.

[0058] The tables 440 and 450 each include items "Time", "Status 1", and "Status 2" in the column direction, and each include records related to each item in the row direction. The item "Time" of the tables 440 and 450 corresponds to the item "Time" of the table 210 related to the operation data 200 of the manufacturing equipment M (see FIG. 3).

[0059] In each record of table 440, a numerical value "0" or "1" is recorded to indicate "Status 1" or "Status 2." A record in which the numerical value "0" is recorded means that the manufacturing equipment M is not in the corresponding operating state at the corresponding time (not applicable). A record in which the numerical value "1" is recorded means that the manufacturing equipment M is in the corresponding operating state at the corresponding time (applicable).

[0060] Each record in table 450 records a value between "0.00" and "1.00" indicating "State 1" or "State 2." A value closer to "0.00" means that there is a higher probability that the manufacturing equipment M is not in that operating state at that time. A value closer to "1.00" means that there is a higher probability that the manufacturing equipment M is in that operating state at that time. In other words, each value between "0.00" and "1.00" indicates that the manufacturing equipment M is in a state transitioning from one state to another (a transition state). Alternatively, each value indicates that the operating state of the manufacturing equipment M is uncertain.

[0061] Table 460 includes items "time" and "status code" in the column direction, and records related to each item in the row direction. The item "time" of table 460 corresponds to the item "time" of table 210 related to operation data 200 of manufacturing equipment M (see FIG. 3).

[0062] Each record in table 460 records a numerical value indicating a "status code." In one example, each numerical value indicates four types of operation states (0: non-processing, 1: pre-preparation, 2: processing, 3: cleanup). In another example, each numerical value indicates four types of operation states (0: non-processing, 1: start-up, 2: processing, 3: fall-down). According to table 460, manufacturing equipment M switches between the four types of operation states in succession.

[0063] 7A and 7B are diagrams showing examples of setting operation states according to the first embodiment. Graph 220 in FIG. 7A and FIG. 7B shows changes over time in time-series data 220A related to operation data 200. In graph 220, the horizontal axis shows "time" and the vertical axis shows "operation values" of manufacturing equipment M. Time-series data 220A may be any data among "Data 1" to "Data N" in table 210 (see FIG. 3).

[0064] 7(A), the setting unit 12 may set a threshold value TH for the time series data 220A. For example, the setting unit 12 sets a "processing state" as the operation state of the manufacturing equipment M for each time point when the time series data 220A is equal to or greater than the threshold value TH. The setting unit 12 sets a "non-processing state" as the operation state of the manufacturing equipment M for each time point when the time series data 220A is less than the threshold value TH.

[0065] Specifically, the setting unit 12 identifies intersections P1, P2, P3, P4, P5, and P6 where the value of the time series data 220A reaches the threshold value TH. The setting unit 12 identifies that the value of the time series data 220A is equal to or greater than the threshold value TH over each of the closed intervals [P1,P2], [P3,P4], and [P5,P6]. The setting unit 12 sets a "processing state" as the operating state of the manufacturing equipment M for each of the closed intervals [P1,P2], [P3,P4], and [P5,P6].

[0066] On the other hand, the setting unit 12 determines that the value of the time-series data 220A is less than the threshold value TH over each of the open intervals (P2, P3) and (P4, P5). The setting unit 12 sets the “non-processing state” as the operating state of the manufacturing equipment M for each of the open intervals (P2, P3) and (P4, P5).

[0067] The setting unit 12 may set a plurality of thresholds for the time series data 220A. For example, the setting unit 12 may set a first threshold and a second threshold greater than the first threshold for the time series data 220A. First, the setting unit 12 may set "rising" as the operating state of the manufacturing equipment M for a section in which the value of the time series data 220A exceeds the first threshold until it exceeds the second threshold. Second, the setting unit 12 may set "intermediate" as the operating state of the manufacturing equipment M for a section in which the value of the time series data 220A exceeds the second threshold. Third, the setting unit 12 may set "falling" as the operating state of the manufacturing equipment M for a section in which the value of the time series data 220A falls below the second threshold until it falls below the first threshold. That is, the setting unit 12 may set an operating state corresponding to a predetermined condition for a section in which the time series data 220A satisfies the predetermined condition.

[0068] As shown in FIG. 7(B), the setting unit 12 may set the operation state of the manufacturing equipment M for each time point of the time series data 220A by associating the time series data 220A with the history data 300. For example, the setting unit 12 refers to the "start time" and "end time" of each process for the table 310 or 320 related to the history data 300 (see FIG. 4). The setting unit 12 sets the "processing state" as the operation state of the manufacturing equipment M for each time point of the time series data 220A that is between the "start time" and the "end time" of each process. The setting unit 12 sets the "non-processing state" as the operation state of the manufacturing equipment M for each time point of the time series data 220A that is not between the "start time" and the "end time" of each process.

[0069] Specifically, the setting unit 12 identifies the start time S1 and end time E1 of the first process, the start time S2 and end time E2 of the second process, and the start time S3 and end time E3 of the third process from the table 310 or 320. The setting unit 12 sets a "processing state" as the operating state of the manufacturing equipment M for each closed interval [S1, E1], [S2, E2], and [S3, E3] between the start time and end time of each process. On the other hand, the setting unit 12 sets a "non-processing state" as the operating state of the manufacturing equipment M for each open interval (E1, S2) and (E2, S3).

[0070] The setting unit 12 may set a predetermined state according to the "operation content" in the table 320 for each time point of the time series data 220A between the start time and end time of each process (see FIG. 4). For example, the setting unit 12 divides the closed interval [S1, E1] into a plurality of small intervals, and sets an operation state (e.g., rising, middle, falling) of the manufacturing equipment M for each small interval. The setting unit 12 may set each operation state by setting a plurality of thresholds for the time series data 220A.

[0071] When the time series data 220A satisfies a predetermined condition, the setting unit 12 may set an operation state corresponding to the predetermined condition. First, the setting unit 12 may set "pre-preparation" for the five minutes immediately preceding the start time of each process. Second, the setting unit 12 may set "cleanup" for the ten minutes immediately preceding the end time of each process. Third, the setting unit 12 may set "maintenance" for a time that is not a processing time of one hour or more (non-processing time).

[0072] FIG. 8 is a diagram showing an example of anomaly detection for each operating state according to the first embodiment. Graph 220 in FIG. 8(A) shows the operating states (processing state, non-processing state) of manufacturing equipment M set in time series data 220A. Graph 220 in FIG. 8(A) is the same as graph 220 in FIG. 7(B). Graph 230 in FIG. 8(B) shows each of time series data 221, 222, and 223 in the "processing state" cut out from time series data 220A. Graph 240 in FIG. 8(C) shows each of time series data 224 and 225 in the "non-processing state" cut out from time series data 220A.

[0073] The detection unit 13 identifies each piece of time series data in the “processing state” of the manufacturing equipment M in the time series data 220A. Specifically, the detection unit 13 identifies time series data 221 in the closed interval [S1, E1], time series data 222 in the closed interval [S2, E2], and time series data 223 in the closed interval [S3, E3]. The detection unit 13 extracts each of the identified time series data 221, 222, and 223 from the time series data 220A and superimposes them on the graph 230.

[0074] In the graph 230, the start points (start times) of the time series data 221, 222, and 223 are aligned to the same time. This shows the differences in the end points (end times) of the time series data 221, 222, and 223. At the same time, the differences in the changes over time of the "motion numerical values" of the time series data 221, 222, and 223 are shown.

[0075] The detection unit 13 can selectively detect an abnormality in the manufacturing equipment M in the "processing state" based on the graph 230. For example, the detection unit 13 can detect in which time series data an abnormality has occurred by comparing the time series data 221, 222, and 223 with each other. Specifically, the detection unit 13 can detect that an abnormality has occurred in the time series data 223 having a peak value. That is, the detection unit 13 selectively detects an abnormality in the manufacturing equipment M during product processing, which can contribute to product quality control and reduce the impact on subsequent processes after the product processing.

[0076] Furthermore, the detection unit 13 can selectively detect an abnormality in the manufacturing equipment M in each operation state (e.g., rising, middle, falling) obtained by subdividing the "processing state." This allows the detection unit 13 to improve the manufacturing process of the product by the manufacturing equipment M.

[0077] Meanwhile, the detection unit 13 identifies each time series data in the “non-processing state” of the manufacturing equipment M in the time series data 220A. Specifically, the detection unit 13 identifies time series data 224 in the open interval (E1, S2) and time series data 225 in the open interval (E2, S3). The detection unit 13 extracts the identified time series data 224 and 225 from the time series data 220A and superimposes them on the graph 240.

[0078] In the graph 240, the start points (start times) of the time series data 224 and 225 are aligned to the same time. This shows the difference between the end points (end times) of the time series data 224 and 225. At the same time, the difference in the change over time of the "operation numerical values" of the time series data 224 and 225 is shown.

[0079] The detection unit 13 can selectively detect an abnormality in the manufacturing equipment M in a "non-processing state" based on the graph 240. For example, the detection unit 13 can detect in which time series data an abnormality has occurred by comparing the time series data 224 and 225 with each other. Specifically, the detection unit 13 can detect that an abnormality has occurred in the time series data 225 having a peak value. An abnormality in the manufacturing equipment M in a non-processing state may be related to, for example, deterioration of the manufacturing equipment M. Since the detection unit 13 selectively detects an abnormality in the manufacturing equipment M while the product is not being processed, it can detect deterioration of the manufacturing equipment M and contribute to condition-based maintenance (CBM) of the manufacturing equipment M.

[0080] Furthermore, the detection unit 13 can selectively detect an abnormality in the manufacturing equipment M in each operation state (e.g., maintenance, cleanup) obtained by subdividing the “non-processing state.” This allows the detection unit 13 to contribute to the appropriate management of the manufacturing equipment M.

[0081] Of course, the detection unit 13 may change the detection method for an abnormality of the manufacturing equipment M according to the operating state of the manufacturing equipment M. For example, the detection unit 13 sets a threshold value for each of the time series data 221, 222, and 223 in the "processing state" of the manufacturing equipment M so as not to allow even a relatively small fluctuation in value. On the other hand, the detection unit 13 sets a threshold value for each of the time series data 224 and 225 in the "non-processing state" of the manufacturing equipment M so as to allow a relatively small fluctuation in value. In this way, the detection unit 13 may change the setting criteria for the threshold value for abnormality detection according to the operating state of the manufacturing equipment M. Therefore, the detection unit 13 can flexibly detect abnormalities in the manufacturing equipment M.

[0082] 9 is a diagram showing a first example of anomaly detection according to the first embodiment. Graphs 250 in FIG. 9(A) and FIG. 9(B) show changes over time in time-series data 250A related to operation data 200. In this example, it is assumed that manufacturing equipment M temporarily suspends an ongoing process due to an emergency stop, and then resumes the process. In this case, operation data 200 of manufacturing equipment M changes in a sawtooth pattern in the order of peak, valley, and peak, as shown in time-series data 250A.

[0083] 9(A), the setting unit 12 sets a "processing state" for each closed interval [S1, E1] and [S2, E2] of the time series data 250A. On the other hand, the setting unit 12 sets a "non-processing state" for the open interval (E1, S2) of the time series data 250A. That is, the setting unit 12 sets each operation state for the time series data 250A in the order of "processing state, non-processing state, processing state."

[0084] At this time, during the period (E1, S2) of the "non-processing state", a fluctuation such as a temporary drop in the value of the time-series data 250A occurs. Since the fluctuation is caused by an emergency stop of the manufacturing equipment M, there is a demand to distinguish the fluctuation from fluctuations in the "processing state". Alternatively, there is a demand to detect the fluctuation as an abnormality and notify a supervisor that it is a fluctuation in the "non-processing state".

[0085] Therefore, when the operating state of the manufacturing equipment M changes in the order of "processing state, non-processing state, processing state" and anomaly detection is performed during the "non-processing state", the detection unit 13 adjusts the anomaly degree for that period. For example, since the fluctuation due to an emergency stop is under management, the detection unit 13 adjusts the anomaly degree to be low when the priority of monitoring the fluctuation is to be lowered. Alternatively, the detection unit 13 adjusts the anomaly degree to be high when the priority of monitoring the fluctuation due to an emergency stop is to be increased. Of course, the detection unit 13 may output information indicating that an anomaly occurred in the non-processing state together with the normal, unadjusted anomaly degree.

[0086] The detection unit 13 may adjust the degree of abnormality by adding or multiplying a predetermined value to the normal degree of abnormality. Alternatively, the detection unit 13 may adjust the degree of abnormality by changing parameters such as the abnormality detection method and thresholds.

[0087] 9(B), it is assumed that the setting unit 12 sets a "processing state" for the closed interval [S1, E2] of the time-series data 250A. In this case, a "processing state" is also set for the open interval (E1, S2) included in the closed interval [S1, E2]. The detection unit 13 detects an abnormality in the open interval (E1, S2) by applying the anomaly detection method in the "processing state" to the open interval (E1, S2) instead of the anomaly detection method in the "non-processing state".

[0088] Fig. 10 is a diagram showing a second example of anomaly detection according to the first embodiment. A graph 260 in Fig. 10(A) shows a change over time in time series data 260A related to the motion data 200. A graph 270 in Fig. 10(B) shows a change over time in time series data 270A different from the time series data 260A.

[0089] The setting unit 12 sets a "non-processing state" for the period up to the start time S1 of the time series data 260A and 270A, and sets a "processing state" for another period [S1, E1] following the period. That is, the setting unit 12 sets each operation state for the time series data 260A and 270A in the order of "non-processing state, processing state."

[0090] In this case, if a fluctuation occurs in the time series data 260A or 270A during a period of the "non-processing state", the fluctuation may affect the fluctuation of the time series data 260A or 270A during the subsequent period of the "processing state".

[0091] Therefore, the detection unit 13 may change the method or criteria for detecting an abnormality from the magnitude (amount of variation) of the variation in the "processing state" depending on the magnitude (amount of variation) of the variation in the "non-processing state". For example, the detection unit 13 sets a stricter threshold value for the amount of variation in the "processing state" as the amount of variation in the "non-processing state" increases. In other words, the detection unit 13 increases the detection sensitivity for abnormalities in the manufacturing equipment M in the "processing state" compared to the detection sensitivity for abnormalities in the manufacturing equipment M in the "non-processing state".

[0092] 10(A), the detection unit 13 sets a first threshold value TH1 on the graph 260 and detects an abnormality in the time series data 260A in the “processing state.” Since the first threshold value TH1 is set so as to exceed the peak value of the time series data 260A, the detection unit 13 does not detect an abnormality in the time series data 260A.

[0093] As shown in FIG. 10(B), the detection unit 13 sets a second threshold value TH2 on the graph 270 and detects an abnormality in the time series data 270A in the "processing state". Here, an abnormality is detected in the "non-processing state" before the "processing state". Therefore, the detection unit 13 sets a second threshold value TH2 smaller than the first threshold value TH1 so as to strictly detect an abnormality in the "processing state". Since the second threshold value TH2 is set to be lower than the peak value of the time series data 270A, the detection unit 13 detects an abnormality in the time series data 270A. In this way, the detection unit 13 can strictly detect an abnormality in the manufacturing equipment M in the "processing state".

[0094] Conversely, if an abnormality occurs in the "non-processing state" following the "processing state", unlike a normal non-processing state (maintenance, etc.), it is assumed that the manufacturing equipment M has deteriorated due to the processing of the manufacturing equipment M in the "processing state". Therefore, the detection unit 13 may change the method of detecting an abnormality in the "non-processing state" when the operating state of the manufacturing equipment M changes in the order of "processing state, non-processing state".

[0095] For example, as a method of detecting an abnormality in the "non-processing state", the detection unit 13 uses a method of detecting a trend (e.g., monotonically increasing or decreasing) in the operation data 200 for a longer period than usual. Alternatively, if the tendency that deterioration of the manufacturing equipment M has on the operation data 200 is known, the detection unit 13 may use a method of detecting the tendency. In this way, the detection unit 13 can detect deterioration of the manufacturing equipment M in the "non-processing state".

[0096] Fig. 11 is a diagram showing a third example of anomaly detection according to the first embodiment. A graph 280 in Fig. 11 shows a change over time in time-series data 280A related to the action data 200.

[0097] The setting unit 12 sets a "processing state" for the closed interval [S1, E1] of the time series data 280A. The setting unit 12 divides the closed interval [S1, E1] into three small intervals and sets each operation state for each divided small interval. Specifically, the setting unit 12 sets a "rising state" for the closed interval [S1, C1], a "middle state" for the closed interval [C1, C2], and a "falling state" for the closed interval [C2, E1].

[0098] The detection unit 13 may change the anomaly detection method according to each operation state. For example, the detection unit 13 detects anomalies more strictly for the time series data 280A in the "intermediate state" than for the time series data 280A in the "rising state" or "falling state". Alternatively, the detection unit 13 may weight the degree of anomaly calculated from the "intermediate state" more heavily and weight the degree of anomaly calculated from the "rising state" or "falling state" less heavily. The detection unit 13 may calculate the sum of the degrees of anomaly weighted for each operation state as the overall degree of anomaly of the time series data 280A.

[0099] Fig. 12 is a diagram showing a fourth example of anomaly detection according to the first embodiment. Graph 290A in Fig. 12(A) shows changes over time in time series data 291A and 292 related to the operation data 200. Graph 290B in Fig. 12(B) shows changes over time in time series data 291B and 292 related to the operation data 200. Each of the time series data 291A, 291B, and 292 is time series data in a "processing state".

[0100] In general, the processing time required for a manufacturing device M varies depending on the processing volume required for processing the product. The processing volume corresponds to the physical amount (e.g., weight, length, area, volume, concentration) of the product being processed. The processing volume may be included in the history data 300 (see FIG. 4).

[0101] Hereinafter, an example will be described in which the processing amount is length. Simply put, when the manufacturing equipment M processes a product that is twice as long as a normal product, the manufacturing equipment M requires twice as much processing time as a normal product.

[0102] In graph 290A, time series data 291A is waveform data when manufacturing equipment M processes a product with length L [m]. Time series data 292 is waveform data when manufacturing equipment M processes another product with length 2L [m].

[0103] The setting unit 12 may convert the time series data 291A and 292 based on the product throughput, and the detection unit 13 may detect an abnormality in the manufacturing equipment M based on the converted time series data 291A and 292. Specifically, the setting unit 12 may normalize the waveform length of the time series data 291A and the waveform length of the time series data 292 using the product throughput (L [m]) related to the time series data 291A and the product throughput (2L [m]) related to the time series data 292. The detection unit 13 may detect an abnormality in the time series data 291A and 292 by comparing the two normalized waveform lengths with each other.

[0104] That is, the setting unit 12 may use the processing amount of the product to normalize the operation data 200. The detection unit 13 may detect an abnormality in the two types of operation data 200 by comparing the two types of operation data 200 with each other.

[0105] For example, when the time series data 292 is the reference, the setting unit 12 extends the length of the waveform of the time series data 291A in the time direction (horizontal axis direction) by two times. The extended time series data 291A corresponds to the time series data 291B of the graph 290B.

[0106] In graph 290B, the detection unit 13 compares the time series data 291B and 292 with each other to detect an abnormality in the time series data 291B and 292. The time series data 291B is considered to be waveform data when the manufacturing equipment M processes a product having a length of 2L [m]. In other words, the detection unit 13 assumes that the manufacturing equipment M processes two products having the same length (2L [m]), and compares the time series data 291B and 292 relating to the two products with each other.

[0107] As a result of the comparison, the detection unit 13 does not detect any abnormality in the time series data 291B and 292. That is, the detection unit 13 determines that the difference in the waveform length between the time series data 291A and 292 is caused by the difference in the product throughput. Therefore, the detection unit 13 does not detect any abnormality in the time series data 291A and 292.

[0108] In this way, the detection unit 13 normalizes the length of the waveform of the time-series data according to the processing volume of the product, thereby enabling highly accurate detection of anomalies in the operation data 200 when the processing time varies according to the processing volume. This allows the detection unit 13 to detect as an anomaly a case where the processing time exceeds or is shorter than the standard due to a work delay, a work omission, or the like.

[0109] The above process can be restated as follows. The history data 300 includes the processing amount of the process on the product. The setting unit 12 sets the processing state as the operation state of the manufacturing equipment M for a predetermined period of the operation data 200. The detection unit 13 detects an abnormality in the manufacturing equipment M based on a comparison between the processing amount in the processing state and the length of the predetermined period. In particular, the detection unit 13 normalizes the length of the predetermined period by the processing amount, and detects an abnormality in the manufacturing equipment M based on a comparison between the processing amount and the normalized length of the predetermined period.

[0110] According to the first embodiment described above, the anomaly detection device 1 can flexibly detect anomalies in the manufacturing equipment M depending on the operation state of the manufacturing equipment M. For example, the anomaly detection device 1 can detect anomalies more strictly in the operation data 200 in a state (processing state) in which the manufacturing equipment M is processing a product than in the operation data 200 in a state (non-processing state) in which the manufacturing equipment M is not processing a product.

[0111] In particular, the anomaly detection device 1 can selectively detect anomalies in the operation data 200 of the manufacturing equipment M in a desired operating state. The anomaly detection device 1 can improve the detection performance for anomalies in the manufacturing equipment M by controlling the anomaly detection method for each operating state of the manufacturing equipment M.

[0112] Second embodiment 13 is a block diagram showing an example of a functional configuration of the anomaly detection device 1 according to the second embodiment. The anomaly detection device 1 according to the second embodiment includes an estimation unit 14A in addition to the respective configurations according to the first embodiment.

[0113] The estimation unit 14A estimates various data. For example, the estimation unit 14A estimates the cause of an abnormality based on a detection result of the abnormality of the manufacturing equipment M output from the detection unit 13. The estimation unit 14A outputs output data 500 including the estimation result of the abnormality cause to the outside. That is, the output data 500 includes the detection result of the abnormality from the detection unit 13 and the estimation result of the abnormality cause from the estimation unit 14A.

[0114] Fig. 14 is a flowchart showing an example of the operation of the anomaly detection device 1 according to the second embodiment. The example of the operation in Fig. 14 includes step S4A in addition to the steps (S1, S2, S3) in the example of the operation in Fig. 2.

[0115] (Step S4A) Finally, the anomaly detection device 1 uses the estimation unit 14A to estimate the cause of the anomaly in the manufacturing equipment M. Specifically, the estimation unit 14A estimates the cause of the anomaly in the manufacturing equipment M based on the detection result of the anomaly in the manufacturing equipment M detected in step S3. The estimation unit 14A outputs output data 500 including the estimation result of the cause of the anomaly to the outside.

[0116] First, the estimation unit 14A may prepare a list of multiple candidate abnormality causes for the operation data 200 of the manufacturing equipment M. The list may be prepared for each manufacturing equipment M or each product, or may be prepared for each state of the abnormality detection result (magnitude of abnormality, abnormality mode). The estimation unit 14A may output at least one abnormality cause that corresponds to the operation data 200 to be processed, among the multiple abnormality causes included in the list.

[0117] Secondly, the estimation unit 14A may estimate the cause of the abnormality of the manufacturing equipment M by associating the operation data 200 of the manufacturing equipment M with the history data 300. The history data 300 includes the model number, processing conditions, etc. of the product processed by the manufacturing equipment M. The estimation unit 14A may estimate the model number, processing conditions, etc. of the product corresponding to the operation data 200 to be processed as the cause of the abnormality.

[0118] Thirdly, the estimation unit 14A may estimate another operation data 200 correlated with the operation data 200 to be processed as the cause of the abnormality. For example, it is assumed that the operation data 200 to be processed indicates the "pressure" of the manufacturing equipment M, and the other operation data 200 indicates the "temperature" of the manufacturing equipment M. When the pressure fluctuation correlates with the temperature fluctuation, the estimation unit 14A estimates that the cause of the abnormality in the pressure of the manufacturing equipment M is the temperature of the manufacturing equipment M.

[0119] Fourth, the estimation unit 14A may estimate the cause of the abnormality in the action data 200 according to the order or combination of the action states set in the action data 200 to be processed.

[0120] According to the second embodiment described above, the anomaly detection device 1 has the same effects as those of the first embodiment. Furthermore, the anomaly detection device 1 can contribute to appropriate management of the manufacturing equipment M by estimating the cause of an anomaly in the manufacturing equipment M.

[0121] Third embodiment 15 is a block diagram showing an example of a functional configuration of an anomaly detection device 1 according to the third embodiment. The anomaly detection device 1 according to the third embodiment includes a generation unit 14B in addition to the respective configurations according to the first embodiment.

[0122] The generating unit 14B generates various types of data. For example, based on the detection result of an abnormality in the manufacturing equipment M output from the detecting unit 13, the generating unit 14B generates display data 600 for displaying the detection result on a display device. The generating unit 14B outputs the generated display data 600 to the outside. The display data 600 may be a text file or an image file in a predetermined format (e.g., CSV, JSON, JPEG).

[0123] Fig. 16 is a flowchart showing an example of the operation of the anomaly detection device 1 according to the third embodiment. The example of the operation in Fig. 16 includes step S4B in addition to the steps (S1, S2, S3) in the example of the operation in Fig. 2.

[0124] (Step S4B) Finally, the anomaly detection device 1 generates display data 600 by the generation unit 14B. Specifically, the generation unit 14B generates the display data 600 based on the detection result of the anomaly of the manufacturing equipment M detected in step S3. The generation unit 14B outputs the display data 600 including the detection result of the anomaly to the outside.

[0125] First, the generating unit 14B may represent the motion data 200 to be processed by a line graph.

[0126] Secondly, the generating unit 14B may express, by characters, the action state set in the action data 200 to be processed. The generating unit 14B may superimpose the action state on the action data 200 to be processed.

[0127] Thirdly, the generation unit 14B may express, by characters, the detection result (degree of abnormality) of the abnormality of the manufacturing equipment M. The generation unit 14B may express, by an image showing the appearance of the manufacturing equipment M, the location where the abnormality has occurred in the manufacturing equipment M.

[0128] Of course, the display data 600 may include the various data described above. Specifically, the display data 600 may include the tables 210, 310, 320, 410, 420, 430, 440, 450, or 460. The display data 600 may include the graphs 220, 230, 240, 250, 260, 270, 280, 290A, or 290B. The display data 600 may include the time series data 220A, 221, 222, 223, 224, 225, 250A, 260A, 270A, 280A, 291A, 291B, or 292.

[0129] According to the third embodiment described above, the anomaly detection device 1 exhibits the same effects as those of the first embodiment. Furthermore, the anomaly detection device 1 generates display data 600 including the detection result of an anomaly in the manufacturing equipment M, thereby contributing to appropriate management of the manufacturing equipment M.

[0130] The anomaly detection device 1 according to the other embodiment may include both the estimation unit 14A and the generation unit 14B. In this case, the anomaly detection device 1 may generate display data 600 including an estimation result of the cause of the anomaly in the manufacturing equipment M. According to the other embodiment, the anomaly detection device 1 achieves the same effects as the second and third embodiments.

[0131] 17 is a block diagram showing an example of a hardware configuration of the anomaly detection device 1 according to each embodiment. The anomaly detection device 1 includes, as its components, a CPU 111, a RAM 112, a ROM 113, a storage 114, a display device 115, an input device 116, and a communication device 117. Each component is communicatively connected to one another via a bus (BUS) which is a common signal communication path.

[0132] The CPU 111 is a processor that executes various processes according to a program. The CPU 111 uses a predetermined area of ​​the RAM 112 as a working area. The CPU 111 realizes each part (the acquisition part 11, the setting part 12, the detection part 13, the estimation part 14A, and the generation part 14B) of the anomaly detection device 1 by reading and executing a program stored in the ROM 113 or the storage 114. The CPU 111 is an example of a processing part.

[0133] The RAM 112 is a memory that stores various types of data in a rewritable manner. For example, the RAM 112 is a Synchronous Dynamic Random Access Memory (SDRAM). The RAM 112 is an example of a storage unit.

[0134] The ROM 113 is a memory that stores various types of data in a non-rewritable manner and is an example of a storage unit.

[0135] The storage 114 is various types of storage media (e.g., magnetic storage media, semiconductor storage media, optical storage media), or may be a drive device that writes or reads various types of data to the storage media under the control of the CPU 111. The storage 114 is an example of a storage unit.

[0136] Display device 115 is a device that displays various data. For example, display device 115 is an LCD (Liquid Crystal Display). Display device 115 displays various data according to control by CPU 111. In particular, display device 115 may display a display image based on display data 600. Display device 115 is an example of a display unit.

[0137] The input device 116 is a device that accepts input of various types of data. For example, the input device 116 is a mouse or a keyboard. The input device 116 accepts an operation input entered by a user as an electrical signal, and transfers this electrical signal to the CPU 111. In particular, the input device 116 may accept input of analysis parameters from the user. The input device 116 is an example of an input unit.

[0138] The communication device 117 is a device that communicates various data with an external device. The communication device 117 communicates with the external device via a network in accordance with the control of the CPU 111. In particular, the communication device 117 may receive analysis parameters from another system. The communication device 117 is an example of a communication unit.

[0139] The various processes performed by the anomaly detection device 1 may be executed by a general-purpose computer (e.g., a personal computer, a microcomputer, an arithmetic device). For example, the general-purpose computer stores programs corresponding to the various processes in a storage medium, and reads and executes the stored programs. Alternatively, the general-purpose computer reads and executes programs from an external storage medium connected via a network (e.g., a LAN, the Internet). This allows the general-purpose computer to execute processes similar to those of the anomaly detection device 1.

[0140] The storage medium may be a magnetic disk (e.g., a flexible disk, a hard disk), an optical disk (e.g., a CD-ROM, a CD-R, a CD-RW, a DVD-ROM, a DVD±R, a DVD±RW, a Blu-ray (registered trademark) Disc), a semiconductor memory, or a storage medium similar thereto. The storage medium may store a program downloaded from a network. Of course, multiple programs may be stored in multiple storage media.

[0141] Instead of a single computer, a system consisting of multiple computers, an OS (operating system), database management software, MW (middleware), or other entity may execute the same processing as the anomaly detection device 1.

[0142] Each unit of the anomaly detection device 1 (the acquisition unit 11, the setting unit 12, the detection unit 13, the estimation unit 14A, and the generation unit 14B) may be present in the cloud or on-premise.

[0143] Although some embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their modifications are included in the scope and spirit of the invention, and are included in the scope of the invention and its equivalents described in the claims. [Explanation of symbols]

[0144] 1... anomaly detection device, 2... manufacturing device DB, 3... manufacturing history DB, 11... acquisition unit, 12... setting unit, 13... detection unit, 14A... estimation unit, 14B... generation unit, 100... analysis data, 111... CPU, 112... RAM, 113... ROM, 114... storage, 115... display device, 116... input device, 117... communication device, 200... operation data, 210, 310, 320, 410, 420, 430, 440, 450, 460...table, 220, 230, 240, 250, 260, 270, 280, 290A, 290B...graph, 220A, 221, 222, 223, 224, 225, 250A, 260A, 270A, 280A, 291A, 291B, 292...time series data, 300...history data, 500...output data, 600...display data

Claims

1. an acquisition unit that acquires analysis data including at least operation data of a manufacturing device that intermittently processes at least one product and history data related to a manufacturing history of the product processed by the manufacturing device; a setting unit that sets an operating state of the manufacturing apparatus for the operation data based on the acquired analysis data; A detection unit that detects an abnormality in the manufacturing apparatus based on the operation data and the set operation state; An abnormality detection device comprising:

2. The detection unit changes a detection method for detecting an abnormality in the manufacturing apparatus according to the operating state. The anomaly detection device according to claim 1 .

3. The setting unit sets a first state as the operation state for each time point when the operation data is equal to or greater than a threshold value, and sets a second state different from the first state as the operation state for each time point when the operation data is less than the threshold value. The anomaly detection device according to claim 1 .

4. the historical data includes start and end times of processing the product; the setting unit sets a processing state as the operation state for each time point of the operation data between the start time and the end time, and sets a non-processing state as the operation state for each time point of the operation data that is not between the start time and the end time. The anomaly detection device according to claim 1 .

5. The history data includes details of processing operations performed on the product, the setting unit sets a predetermined state corresponding to the work content as the operation state for each time point of the operation data between the start time and the end time. The anomaly detection device according to claim 4.

6. The detection unit detects an abnormality in the manufacturing apparatus based on a sequence or combination of the operation states set in the operation data. The anomaly detection device according to claim 1 .

7. the setting unit sets a non-processing state for a first period of the motion data, and sets a processing state for a second period following the first period; The detection unit increases a detection sensitivity for an abnormality of the manufacturing apparatus in the second time period to be higher than a detection sensitivity for an abnormality of the manufacturing apparatus in the first time period. The anomaly detection device according to claim 6.

8. The setting unit converts the operation data and the set operation state based on a processing amount of the process for the product, The detection unit detects an abnormality in the manufacturing apparatus based on the converted operation data and the converted operation state. The anomaly detection device according to claim 1 .

9. the setting unit sets a processing state as the operation state for a predetermined period of the operation data; The detection unit detects an abnormality in the manufacturing apparatus based on a comparison between the processing amount in the processing state and the length of the predetermined period. The anomaly detection device according to claim 8.

10. The setting unit normalizes the length of the predetermined period according to the amount of processing; The detection unit detects an abnormality in the manufacturing equipment based on a comparison between the processing amount and the normalized length of the predetermined period. The anomaly detection device according to claim 9.

11. acquiring analytical data including at least one of operational data of a manufacturing device that intermittently processes at least one product and history data relating to a manufacturing history of the product processed by the manufacturing device; setting an operating state of the manufacturing apparatus for the operation data based on the acquired analysis data; Detecting an abnormality in the manufacturing apparatus based on the operation data and the set operation state. Anomaly detection methods.

12. On the computer, An acquisition function for acquiring analysis data including at least operation data of a manufacturing device that intermittently processes at least one product and history data related to a manufacturing history of the product processed by the manufacturing device; A setting function for setting an operating state of the manufacturing apparatus for the operation data based on the acquired analysis data; A detection function that detects an abnormality in the manufacturing apparatus based on the operation data and the set operation state; An anomaly detection program that achieves this.

Images