TWI824444B - Status determination device and status determination method - Google Patents

Abstract

The present disclosure proposes a status determination device and a status determination method that can improve the accuracy of determining the status of a processing device. The status determination device 50 has a control portion 51 for determining the status of the processing device 10. The control portion 51 obtains the feature amounts of plural batches normally processed by the processing device 10 as reference data; generates a unit space based on the reference data; calculates the Mahalanobis Distance of the feature amounts of each batch included in the reference data in the unit space and deletes the feature amounts of batches whose Mahalanobis Distances is above a deletion threshold value from the reference data; generates an update unit space based on the reference data which is updated; obtains the feature amounts of the batch of the object that determines the status of the processing device as determination object data; calculating the Mahalanobis Distance of the determination object data in the updated unit space; outputs a result that the state of the processing device 10 is abnormal if the Mahalanobis Distance is above a determination threshold value.

Description

Status determination device and status determination method

The present disclosure relates to a status determination device and a status determination method.

In the past, there was a state determination device that provided useful information to the operator in a work machine to identify a fault location (for example, see Patent Document 1). [Prior patent documents] [Patent Document]

Patent Document 1: Japanese Patent Application Publication No. 2013-199804

[Problem to be solved by the invention]

When a malfunction is predicted to occur in the processing device, it is determined that the status of the processing device is abnormal. When it is determined that the status of the processing device is abnormal, the processing device is stopped and inspected. If the judgment is wrong, opportunities will be lost due to meaningless stops of the processing equipment. Therefore, in order to reduce the frequency of unnecessary stops of the processing device, it is necessary to improve the accuracy of determining the status of the processing device.

Therefore, an object of the present invention is to provide a state determination device and a state determination method that can improve the accuracy of determining the state of a processing device. [Means used to solve problems]

An implementation form of the present disclosure that solves the above problems is as follows. [1] A state determination device, including: a control unit that determines the state of a processing device, wherein the control unit: obtains feature quantities of a plurality of batches normally processed by the processing device as reference data; based on the reference data, generates unit space , which represents the characteristics of the batch group normally processed by the processing device; calculate the Mahalanobis Distance (Mahalanobis Distance) of the characteristic quantities of each batch contained in the reference data in the unit space, and delete it from the reference data The characteristic quantity of the batch whose Mahalanobis distance is above the deletion threshold; based on the updated reference data, generate this unit space as the updated unit space; obtain the characteristic quantity of the batch that is the object of determining the status of the processing device as Determine the target data; calculate the Mahalanobis distance of the judgment target data in the update unit space; when the Mahalanobis distance is greater than the judgment threshold set based on the update unit space, output the decision The state of this processing device is the result of an abnormal condition. [2] A state determination device, including: a control unit that determines the state of a processing device, wherein the control unit: obtains feature quantities of a plurality of batches normally processed by the processing device as reference data; based on the reference data, generates unit space , which represents the characteristics of the batch group normally processed by the processing device; calculate the Mahalanobis distance of the characteristic amount of the batch newly obtained in the unit space to add to the reference data, and divide the Mahalanobis distance The feature values of the batches whose distance is less than the additional threshold value are added to the reference data; based on the updated reference data, the unit space is generated as the update unit space; the feature values of the batch targeted for determining the status of the processing device are obtained as Determine the target data; calculate the Mahalanobis distance of the judgment target data in the update unit space; when the Mahalanobis distance is greater than the judgment threshold set based on the update unit space, output the decision The state of this processing device is the result of an abnormal condition. [3] A state determination device, including: a control unit that determines the state of a processing device, wherein the control unit: obtains feature quantities of a plurality of batches normally processed by the processing device as reference data; based on the reference data, generates unit space , which represents the characteristics of the batch group normally processed by the processing device; calculate the Mahalanobis distance of the characteristic quantities of each batch contained in the reference data in the unit space, and delete the Mahalanobis distance from the reference data The feature quantities of the batch whose Bis distance is greater than the deletion threshold; calculate the Mahalanobis distance of the feature quantities of the batch newly obtained to add to the reference data in the unit space, and divide the Mahalanobis distance into The feature quantity of the batch that does not meet the additional threshold value is added to the reference data; based on the updated reference data, the unit space is generated as the update unit space; the feature quantity of the batch targeted for determining the status of the processing device is obtained as the determination Object data; calculate the Mahalanobis distance of the judgment object data in the update unit space; when the Mahalanobis distance is greater than the judgment threshold set based on the update unit space, output the judgment The state of the processing equipment is the result of abnormal conditions. [4] The state determination device of [3] above, wherein the control unit sets the deletion threshold and the addition threshold so that the number of batches removed from the reference data is equal to the number of batches added to the reference data. . [5] The state determination device according to any one of [2] to [4] above, wherein: the Mahalanobis distance of feature quantities of batches of more than a predetermined number newly acquired for adding to the reference data If the value is continuously equal to or higher than the addition threshold, the control unit adds the feature values of batches exceeding the predetermined number to the reference data. [6] A state determination method, executed by a state determination device to determine the state of a processing device, includes: obtaining the characteristic quantities of a plurality of batches normally processed by the processing device as reference data; based on the reference data, generating a unit space, in which Represents the characteristics of the batch group normally processed by the processing device; calculates the Mahalanobis distance of the characteristic quantities of each batch contained in the reference data in the unit space, and deletes the Mahalanobis distance from the reference data The characteristic amount of the batch whose distance is greater than the deletion threshold; based on the updated reference data, generate the unit space as the updated unit space; obtain the characteristic amount of the batch that is the object of determining the status of the processing device as the determination target data; calculate the Update the Mahalanobis distance of the judgment target data in the unit space; when the Mahalanobis distance is greater than the judgment threshold set based on the updated unit space, output a judgment that the status of the processing device is The result of an abnormal state. [7] A state determination method, executed by a state determination device to determine the state of a processing device, includes: obtaining the characteristic quantities of a plurality of batches normally processed by the processing device as reference data; based on the reference data, generating a unit space, in which Indicates the characteristics of the batch group normally processed by the processing device; calculates the Mahalanobis distance of the characteristic amount of the batch newly obtained to add to the reference data in the unit space, and divides the Mahalanobis distance into The feature quantity of the batch that reaches the additional threshold value is added to the reference data; based on the updated reference data, the unit space is generated as the update unit space; the feature quantity of the batch that is the object of determining the status of the processing device is obtained as the determination object data; calculate the Mahalanobis distance of the judgment target data in the update unit space; when the Mahalanobis distance is greater than the judgment threshold set based on the update unit space, output the judgment of the processing The status of the device is the result of an abnormal condition. [8] A state determination method, executed by a state determination device to determine the state of a processing device, includes: obtaining the characteristic quantities of a plurality of batches normally processed by the processing device as reference data; based on the reference data, generating a unit space, in which Represents the characteristics of the batch group normally processed by the processing device; calculates the Mahalanobis distance of the characteristic quantities of each batch contained in the reference data in the unit space, and deletes the Mahalanobis distance from the reference data Feature quantities of batches whose distance is greater than the deletion threshold; calculate the Mahalanobis distance of feature quantities of newly acquired batches in the unit space to be added to the reference data, and add the Mahalanobis distance if the Mahalanobis distance is less than the deletion threshold. The feature quantity of the batch with the added threshold value is added to the reference data; based on the updated reference data, the unit space is generated as the update unit space; the feature quantity of the batch targeted for determining the status of the processing device is obtained as the determination target data ; Calculate the Mahalanobis distance of the determination target data in the update unit space; when the Mahalanobis distance is greater than or equal to the determination threshold set based on the update unit space, output a determination of the processing device The state is the result of an abnormal state. [Invention effect]

According to an embodiment of the present disclosure, the accuracy of determining the status of the processing device can be improved.

[Outline of state determination system 1 according to an embodiment] The state determination system 1 (refer to FIG. 1 ) or the state determination device 50 (refer to FIG. 1 ) according to an embodiment of the present disclosure determines the state of the processing device 10 (refer to FIG. 1 or 2 ) installed in a factory that manufactures products. The processing device 10 processes the received member in at least part of the steps of manufacturing the product, and sends it out as a processed product.

The processing device 10 may stop due to a malfunction during processing. In this case, the processed product that was previously in processing becomes a defective product. The production of defective products causes losses to the factory. Furthermore, more work procedures are required for the recovery work of the processing device 10 than for the maintenance work of the processing device 10 . Therefore, failure of the processing device 10 during processing will increase the factory's losses.

The status determination system 1 or the status determination device 50 of an embodiment of the present disclosure determines the possibility of a failure before the processing device 10 fails during processing and prompts the processing device 10 to stop or perform maintenance operations. Thereby, the loss caused by the processing device 10 can be reduced.

Hereinafter, a structural example of the state determination system 1 and the state determination device 50 according to an embodiment will be described.

[System architecture example] As shown in FIG. 1 , a state determination system 1 according to an embodiment includes a state determination device 50 and a sensor 60 . The sensor 60 measures various data about the status of the processing device 10 and outputs it.

The state determination device 50 includes a control unit 51 , an output unit 53 , and an input unit 54 . The control unit 51 provides control and processing capabilities for executing various functions of the status determination device 50 . As will be described later, the control unit 51 determines whether the state of the processing device 10 is an abnormal state.

The control unit 51 may include at least one processor. The processor executes programs that implement various functions of the control unit 51 . The processor can also be implemented as a single integrated circuit. Integrated circuit is also called IC (Integrated Circuit). The processor can also be implemented by a plurality of communicatively connected integrated circuits and discrete circuits. The processor may also be implemented according to various other known technologies.

The control unit 51 may include a storage unit. The storage unit may include an electromagnetic storage medium such as a magnetic disk, or may include a memory such as a semiconductor memory or a magnetic memory. The storage may also include non-transitory computer-readable media. The storage unit stores various information, programs executed by the control unit 51, and the like. The storage unit may also serve as a working memory of the control unit 51 . At least part of the storage unit may be configured independently from the control unit 51 .

The control unit 51 obtains measurement data from the sensor 60 . The control unit 51 may obtain measurement data from the sensor 60 based on a standard such as RS485. The control unit 51 may also be connected to the sensor 60 through a communication interface such as LAN (Local Area Network). The control unit 51 is not limited to the above example, and may be connected to the sensor 60 in various ways.

The output unit 53 outputs the information acquired from the control unit 51 . The output unit 53 may output information to notify the operator or maintenance person of the processing device 10 . The output unit 53 may include a display device. The display device may include, for example, a liquid crystal display, an organic EL (Electroluminescence) display, or an inorganic EL display, but is not limited thereto and may also include other devices. The output unit 53 may display the information acquired from the control unit 51 on the display device as text, images, etc., and notify the surroundings of the information.

The output unit 53 may be provided with a light source such as an LED (Light Emission Diode) or a halogen lamp. The output unit 53 may turn on or off the light source based on the information acquired from the control unit 51 to notify surrounding operators or maintenance personnel. The output unit 53 may include a buzzer such as a piezoelectric buzzer or an electromagnetic buzzer, or a speaker that emits a predetermined sound. The output unit 53 may cause a buzzer to sound or a speaker to sound based on the information obtained from the control unit 51 to notify surrounding operators or maintenance personnel.

The output unit 53 may output information to the processing device 10 . The output unit 53 may also be equipped with a communication module to be communicably connected to the processing device 10 .

The input unit 54 includes an input device for accepting operations or inputs from an operator or a maintenance person responsible for the management status determination device 50 . The input device may include, for example, a keyboard or physical keys, or may include a touch panel, a touch sensor, or a directional device such as a mouse. When the input device is a touch panel or a touch sensor, it may be configured integrally with the display of the output unit 53 . The input device may be, for example, a microphone that accepts audio input. The input unit 54 is not limited to the above example as an input device, and may include other types of devices.

[Specific example of processing device 10: wire saw device] In this embodiment, the processing device 10 is assumed to be a wire saw device. As shown in FIG. 2 , the processing device 10 as a wire saw device is provided with a wire group 16 that stretches the wire 12 between a plurality of rollers 14 in a parallel and reciprocating manner. The processing device 10 includes a workpiece holding mechanism 18 that holds the workpiece W and moves the workpiece W in a direction that pushes the workpiece W toward the wire group 16 . The processing device 10 is provided with a pair of nozzles 20 that supply the abrasive to the area pressed by the workpiece W on the wire group 16 . The processing device 10 cuts the workpiece W using the wire group 16 . The workpiece W is assumed to be a block of silicon or the like (a single crystal silicon rod cut into blocks). The processing device 10 delivers silicon or other sliced wafers obtained by cutting the workpiece W as a processed product.

The wire 12 is wound around a set of wire spools 38A and 38B. The thread 12 is drawn from one thread reel 38A to the other thread reel 38B through the guide roller 32 and the roller 14 and the like.

The line reels 38A and 38B are each rotated by the drive motor 36 . The drive motor 36 drives and rotates the line spools 38A and 38B, whereby the line 12 can be extended from the line spool 38A on one side and moved to the line spool 38B on the other side through the guide roller 32 and the roller 14 and the like. The wire 12 moves via a tension applying device including a swing arm 33, a swing roller 34, and the like. As the wire 12 moves through the tension applying device, tension is applied to the wire 12 . Line 12 moves via touch roller 35 . When the touch roller 35 is extended from the line reels 38A and 38B or wound around the line reels 38A and 38B, it moves following the position of the moving line 12 .

The wire 12 is spirally wound around a plurality of rollers 14 multiple times. The spirally wound wires 12 form a wire group 16 arranged in a direction perpendicular to the roller axis direction X between the rollers 14 . The roller 14 has a structure in which polyurethane is pressed into the periphery of a steel cylinder and grooves are cut at certain intervals on the surface. The wires 12 are embedded in grooves cut on the surface of the roller 14, whereby the wire group 16 can move stably.

The direction of movement of the wire 12 is controlled by the direction of rotation of the drive motor 36 . The wire 12 can also be controlled to move in one direction, or back and forth as needed. The amount of tension applied to the wire 12 can be set appropriately. The moving speed of the wire 12 can be set appropriately.

The abrasive supplied to the wire group 16 from the nozzle 20 is stored in the abrasive storage tank 40, and is sent from the abrasive storage tank 40 to the nozzle 20 through the abrasive cooler 42 that regulates the temperature of the abrasive.

The sliced wafers sent out as processed products from the wire saw device are further processed in steps such as polishing, and the final wafers are shipped.

[Construction example of state determination device 50] Hereinafter, an example of a configuration in which the state determination device 50 determines whether the state of the processing device 10 is an abnormal state will be described. In the following description, the processing device 10 is a wire saw device.

〈Abnormal status of the wire saw device〉 The abnormal state of the wire saw device corresponds to a state in which the probability of the wire saw device stopping during cutting is increased. The wire saw device sometimes stops due to the wire 12 being broken. In the wire saw device, there are also cases where the wire 12 is constantly stopped offline.

Several factors are considered as to why a scroll saw unit stops. Known factors include the following. For example, if an abnormality occurs in the supply of abrasive, the wire 12 may be easily disconnected. For example, abrasive adheres to the sensor that detects the breakage of the wire 12, causing the wire saw device to incorrectly detect the breakage of the wire 12 and stop. For example, an abnormality in the amount or temperature of the abrasive in the abrasive storage tank 40 may cause the wire saw device to stop. On the other hand, the wire saw device stops due to unknown factors that do not conform to the known factors.

The control unit 51 of the state determination device 50 acquires measurement values of various data items obtained from the wire saw device. Data items obtained from the wire saw device include, for example, abrasive flow rate, wire tension, or cooling water flow rate. The data items obtained from the wire saw device are not limited to these examples and may include other various items.

The wire saw device cuts one block to produce sliced wafers as one batch of processing. The control unit 51 can acquire a plurality of measurement values during the processing of one batch based on each data item. The wire saw device takes a predetermined time to cut one block into a plurality of sliced wafers. In other words, the processing of one batch takes a predetermined time. The control unit 51 may acquire the measurement value of each data item every 15 seconds, for example, during the processing of one batch. In this way, the control unit 51 can obtain multiple measurement values of each data item during the processing of one batch. The control unit 51 calculates the average value and the standard deviation of the plurality of measured values obtained in the processing of each batch. The control unit 51 may divide the period from the start to the end of the processing of each batch into the first processing period to the n-th processing period, and then calculate the average and standard deviation of the plurality of measurement values obtained in each processing period. For example, the control unit 51 may regard the period from 0 mm to 10 mm as the cutting distance of the block in the processing of each batch as the first processing period, and calculate the average value of the plurality of measurement values obtained during the first processing period. standard deviation. The average value and the standard deviation of the plurality of measured values calculated in the first to nth processing periods of each batch are respectively called the first to nth characteristic quantities of each batch. The first to n-th feature quantities of each batch are also collectively referred to as feature quantities of each batch. The characteristic quantity of the batch calculated based on the measurement value of each data item is simply called the characteristic quantity of the batch of each data item. The control unit 51 may calculate a plurality of feature quantities based on each data item of each batch as described above, or may calculate a single feature quantity.

The control unit 51 determines whether the wire saw device is in a state with a low probability of stopping or a state in which the wire saw device is highly likely to stop based on the feature value of the batch of each data item of the wire saw device. A state in which the probability of the wire saw device stopping is low is also called a normal state. A state in which the wire saw device is likely to stop is also called an abnormal state.

〈About the MT method (Maharanobis Taguchi method)〉 The status of the wire saw device may be determined every time a batch of processing is executed. The status of the wire saw device may be determined focusing on only one data item of the wire saw device. In this case, under the assumption that the distribution of the batch feature quantity of one data item follows a normal distribution, it can be determined whether the batch feature quantity obtained when the wire saw device processes the batch is an outlier. When it is determined that the feature value of a certain batch is an abnormal value, it is highly likely that the wire saw apparatus can be determined to be stopped during processing of the next batch. That is, the state of the wire saw device can be determined to be an abnormal state.

When the number of data items obtained from the wire saw device is large, the status of the wire saw device can be determined based on the plurality of data items. In this case, the distribution of feature quantities of each batch of plural data items can also be regarded as following a multivariate normal distribution. The control unit 51 determines whether the feature quantity of the batch of each data item is an abnormal value based on the MT method (Maharanobis Taguchi method), which is one of the multivariate analysis techniques. When the feature quantity of the batch is an abnormal value, , it can be determined that the state of the wire saw device when processing this batch is an abnormal state. Specifically, when a plurality of feature quantities are calculated in one batch, the control unit 51 may determine that the state of the wire saw device when processing the batch is an abnormal state when at least one feature quantity is an abnormal value.

The control unit 51 generates a unit space based on the feature quantities of each batch of plural data items obtained when the state of the wire saw device is a normal state. The unit space is represented by a formula that specifies the distribution of characteristic quantities of batches of each data item. Specifically, the control unit 51 normalizes the feature quantities of the batches of each data item, generates a correlation coefficient matrix using the correlation coefficient between the data items as an element, and calculates the inverse matrix of the correlation coefficient matrix, thereby generating a representation of the unit space. formula. That is, the unit space is specified by the inverse matrix of the correlation coefficient matrix.

The unit space represents the characteristics of the feature quantity distribution of the batch processed when the state of the wire saw device is a normal state. The unit space can also be said to represent the characteristics of the batch group that is normally cut by the wire saw device. The unit space can also be said to represent the characteristics of the batch group normally processed by the processing device 10 . The control unit 51 can calculate the Mahalanobis Distance (MD) as a set of the feature values of the plurality of data items obtained from the wire saw device based on the formula representing the unit space. MD is a concept commonly used in the field of quality engineering, for example, it is explained in the following documents. Literature: Jie Renzhecheng, "Introduction to Mahalanobis Distance", Journal of Quality Engineering Society "Quality Engineering", Vol. 9, No. 1, p. 13-19

MD is an index indicating how close the distribution of feature values of a batch processed when the status of the wire saw device is unknown is close to the distribution of feature values of a batch processed when the status of the wire saw device is normal. It can also be said that the smaller the MD calculated from the feature quantity of a batch, the higher the possibility that the feature quantity of the batch is the feature quantity of the batch processed when the state of the wire saw apparatus is a normal state. On the contrary, it can also be said that the larger the MD calculated from the feature quantity of a batch, the higher the possibility that the feature quantity of the batch is the feature quantity of the batch processed when the state of the wire saw device is an abnormal state. .

Therefore, the control unit 51 calculates the MD of the feature value of the newly processed batch based on the unit space, thereby making it possible to determine whether the state of the wire saw device when the feature value of the batch is obtained is normal or abnormal. Hereinafter, the data of the wire saw apparatus used to generate unit space, including the feature quantities of the batches processed in the past predetermined period, are also referred to as reference data. In addition, the data obtained when the wire saw device processes a new batch is used to determine the status of the wire saw device when processing the batch, so it is also called determination target data. The control unit 51 calculates the MD of the determination target data in the unit space generated based on the reference data, and determines the state of the wire saw device. The control unit 51 may generate the unit space based on the reference data, or may obtain information about the unit space from an external device. When the control unit 51 calculates a plurality of feature quantities for one data item based on the data obtained from the wire saw device in one batch of processing, the control unit 51 calculates a set of the plurality of feature quantities calculated for each of the plurality of data items. MD of the data. That is, the control unit 51 regards a set of data having a plurality of feature quantities calculated for each of a plurality of data items as the reference data and the determination target data. For example, when 30 feature quantities are calculated for each of 40 data items in one batch of processing, one set of data includes 1,200 feature quantities.

〈Generation of unit space based on cause data items〉 As mentioned above, the scroll saw unit may stop due to various factors. The factors causing the wire saw device to stop may include factors known to the user of the wire saw device and unknown factors. Known factors are classified into known failure modes. Known failure modes are also called known modes.

When the probability of the wire saw apparatus stopping becomes high, the characteristic quantity of the batch of the data item corresponding to the factor causing the stop may be an abnormal value. The control unit 51 determines whether the feature quantity of the batch of the data item corresponding to the factor causing the wire saw device to stop is an abnormal value, thereby determining that the possibility of the wire saw device stopping is high. The factors that cause the wire saw device to stop are also called stopping factors. The data items corresponding to the stopping factors are also called cause data items. There can be multiple factors that cause the wire saw device to stop. Therefore, there are cause data items corresponding to plural stop factors respectively.

In order to determine whether the feature quantity of the batch of cause data items is an abnormal value, the control unit 51 may generate a unit space based on the feature quantity of the batch of cause data items. The characteristic quantities of the batch of cause data items include plural feature quantities of the cause data items acquired from the wire saw device during the batch processing. On the other hand, the feature quantity of the batch of the cause data items does not include the plural feature quantities of items other than the cause data items acquired from the wire saw device in the batch processing. Unit spaces can be created in external installations. Specifically, the control unit 51 acquires the measurement values of a plurality of data items for each batch processed from the start of operation to the stop of the wire saw device, and calculates the characteristic amount of the batch based on the measurement values as reference data. The control unit 51 generates a unit space based on the characteristic amount of the batch processed when the state of the wire saw device is determined to be normal among the acquired reference data.

In order to determine the state of the wire saw device, the control unit 51 acquires the measurement values of each of the plurality of data items of the batch newly processed by the wire saw device, and calculates the characteristic amount of the batch based on the measurement values as the determination target data. The control unit 51 calculates the MD of the feature amounts of the batch of the cause data items among the data acquired as the determination target data for each stop factor in the unit space generated in advance based on the feature amounts of the batch of the cause data items. . The control unit 51 can determine whether the feature amount of the batch of cause data items is an abnormal value based on the calculated MD. When the feature quantity of the batch of the cause data item is an abnormal value, it is highly likely that the wire saw apparatus can be determined to be stopped during the processing of the next batch. That is, when processing the target batch for which the determination target data is acquired, the state of the wire saw device can be determined to be an abnormal state.

Here, the unit space hypothesis is also generated based on the feature quantity of the batch of data items other than the cause data items and not corresponding to the stop factor. Data items that do not correspond to a stopping factor are also called non-cause data items. When the unit space is generated based not only on the characteristic quantity of the batch of causal data items but also on the characteristic quantities of the batch of non-causal data items, the unit space also represents a tendency that is not related to the stop of the wire saw device. As described above, the unit space is generated based on the correlation between the feature quantities of the batches of each data item. In this way, the unit space generated based on the feature quantities of batches of non-causal data items contains less relevant relationships and is more lengthy than the unit space generated based only on the feature quantities of batches of causal data items. . The MD calculated in a lengthy unit space becomes very insensitive to changes in the feature quantities of the batch of causal data items due to the influence of the feature quantities of the batch of non-causal data items. On the other hand, if a unit space is generated based on only the causal data items, the control unit 51 calculates the MD in the unit space, thereby detecting changes in the feature quantities of the batches of the causal data items with high sensitivity.

When the wire saw device processes one batch, the control unit 51 acquires the measurement values of each data item of the batch, and calculates the characteristic amount of the batch based on the measurement values as the determination target data. The control unit 51 calculates the MD of the feature amounts of the batch of cause data items included in the judgment target data in the generated unit space. When the calculated MD is less than the threshold value, the control unit 51 determines that the state when the wire saw device processes one batch is a normal state. When the calculated MD is equal to or greater than the threshold value, the control unit 51 determines that the state when processing one batch is an abnormal state.

〈Generation of unit space corresponding to known pattern〉 The failure modes that cause the wire saw device to stop may include known failure modes or unknown failure modes. In the following, the failure mode causing the wire saw device to stop is assumed to include at least two known modes. The control unit 51 generates the unit space based on the cause data items corresponding to each known pattern. A unit space can be generated for each known pattern. When the unit space is generated for each known pattern, it is linked to each known pattern.

In the unit space corresponding to each known pattern, the control unit 51 calculates the MD of the feature amount of the batch corresponding to the cause data item of each known pattern included in the judgment target data obtained when the wire saw device processes one batch. When the MD calculated for a certain known pattern is equal to or greater than a predetermined threshold, the control unit 51 determines that the wire saw apparatus is highly likely to stop in the known pattern when processing the next batch. The predetermined threshold can be set to the same value in each known mode, or can be set to different values.

At least two known modes that cause the wire saw device to stop include a first known mode and a second known mode. The known pattern is not limited to the first known pattern and the second known pattern, but may also include a third known pattern, for example. Also, please note that identifiers such as "1st" and "2nd" do not indicate the quality of a known pattern. The data item corresponding to the factor causing the wire saw device to stop in the first known mode is also referred to as the first cause data item. The control unit 51 generates a unit space based on the feature amount of the batch of the first cause data item. The unit space generated based on the characteristic quantity of the batch of the first cause data items is also called the first unit space. The data item corresponding to the factor causing the wire saw device to stop in the second known mode is also called the second reason data item. The control unit 51 generates a unit space based on the feature amount of the batch of the second cause data item. The unit space generated based on the feature quantity of the batch of the second source data items is also called the second unit space. The cause data items are not limited to the first cause data item and the second cause data item, but also include, for example, a third cause data item corresponding to the factor that stopped in the third known mode. The unit space is not limited to the first unit space and the second unit space, but also includes, for example, a third unit space generated based on the characteristic amount of the batch of the third cause data item. In addition, identifiers such as "1st" and "2nd" do not indicate the quality of the cause data item or the unit space.

The control unit 51 calculates the MD of the feature amount of the batch of the first cause data item included in the judgment target data obtained when the wire saw device processes one batch in the first unit space. The control unit 51 calculates the MD of the feature amount of the batch of the second cause data item included in the judgment target data in the second unit space. When the MD calculated in the first unit space is equal to or greater than a predetermined threshold, the control unit 51 can determine that the possibility of stopping the wire saw device in the first known mode becomes high. When the MD calculated in the second unit space is equal to or greater than a predetermined threshold, the control unit 51 can determine that the possibility of stopping the wire saw device in the second known mode becomes high. When the control unit 51 determines that one of the first known mode and the second known mode is more likely to stop, it may not determine the other mode. Even if the control unit 51 determines that the possibility of stopping in at least one mode is high, it may determine that the possibility of stopping in another mode becomes high. It is assumed that when it is determined that the possibility of stopping in a plurality of failure modes increases, the user can perform maintenance work uniformly for the target failure modes.

Each of the first cause data item and the second cause data item includes plural data items. Both the first cause data item and the second cause data item include data items not included in the other cause data item. In other words, the first cause data item and the second cause data item do not have a mutually inclusive relationship. In addition, the data items included in the first cause data item and the data items included in the second cause data item are not exactly the same. Assuming that the first cause data item and the second cause data item are inclusive or identical, the first known mode and the second known mode can also be said to be failure modes that are not independent of each other, but mutually exclusive. associated failure modes of the same kind. Both the first cause data item and the second cause data item include data items not included in the cause data item of the other, whereby the control unit 51 can detect at least two known events that occur independently of each other with high accuracy. model.

〈Data used for unit space generation〉 As described above, the control unit 51 generates the unit space based on the batch feature amount of the cause data item of the batch processed during the period when the status of the wire saw device is normal among the reference data. An example of the data included in the reference materials will be described below.

The state of the wire saw device. After the parts replacement work, disassembly and assembly work, or cleaning work of the wire saw device is performed, and during the processing of several batches to dozens of batches after the operation is started, the parts may be replaced or assembled. An unstable state in which parts do not operate smoothly. After that, the state of the wire saw unit may become a stable state during the processing of dozens to hundreds of batches after the parts begin to operate smoothly. In addition, the cleaning state of the wire saw device deteriorates due to aging of parts or accumulation of abrasives, which increases the failure rate and may lead to an unstable state in a few batches to dozens of batches before failure occurs.

The control unit 51 acquires the measurement value of each data item in each batch processed from the start of operation to the stop of the wire saw device, and calculates the characteristic amount of each batch as reference data based on the measurement value. The feature amount of each batch is calculated based on the measured values of N data items from D1 to DN. In this embodiment, the control unit 51 calculates a plurality of feature quantities for each data item. The control unit 51 may calculate one feature value for each data item.

As illustrated in the table in Fig. 3, the characteristic amount of the batch calculated as the reference data is linked to the batch and data items. The rows of the table in FIG. 3 represent N data items from D1 to DN. The columns of the table of Figure 3 represent batches. The batch is further divided into three periods: P1, P2, and P3 according to the processing time. P1 corresponds to the M1 batches processed after the wire saw device is restarted. P3 corresponds to the M3 batches processed before the batch when the wire saw device is stopped. P2 corresponds to M2 batches not included in P1 and P3. P2 can also be said to correspond to the batch processed while the state of the wire saw device is in a stable state.

In the period P1, the control unit 51 calculates the plurality of feature quantities of each of the M1 batches from D1P1_1 to D1P1_M1 based on the measurement values of the data items D1 of each of the M1 batches. In the period P1, the control unit 51 calculates the plurality of feature quantities of each of the M1 batches from DNP1_1 to DNP1_M1 based on the measurement values of the data items DN of each of the M1 batches. In the period P2, the control unit 51 calculates the plurality of feature quantities of each of the M2 batches from D1P2_1 to D1P2_M2 based on the measurement values of the data items D1 of each of the M2 batches. In the period P2, the control unit 51 calculates the plurality of feature quantities of each of the M2 batches from DNP2_1 to DNP2_M2 based on the measurement values of the data items DN of each of the M2 batches. In the period P3, the control unit 51 calculates the plurality of feature quantities of each of the M3 batches from D1P3_1 to D1P3_M3 based on the measurement values of the data items D1 of each of the M3 batches. In the period P3, the control unit 51 calculates the plurality of feature quantities for each of the M3 batches from DNP3_1 to DNP3_M3 based on the measurement values of the data items DN of each of the M3 batches.

It is assumed that the wire saw device is a target for acquiring the reference data illustrated in the table of FIG. 3 and is stopped in the known mode during the batch processing after period P3. In other words, the batches in period P3 correspond to the last M3 batches processed by the wire saw device before the wire saw device stopped in the known mode.

The control unit 51 generates the unit space based on the feature values of the batch in the period P2 when the wire saw device state becomes a stable state among the feature values of the batches included in the acquired reference data. Here, it is assumed that some of the feature values of the batch are excluded from the feature values of each batch in the period P2 used to generate the unit space. When a plurality of feature quantities are calculated for a batch to be excluded, it is assumed that all feature quantities of the batch are excluded. The reason why the feature quantities of some batches are excluded will be described later. The resulting unit spaces are linked to known patterns.

〈Status determination of wire saw device based on MD〉 The control unit 51 acquires the batch feature quantity of each data item during the batch processing of the wire saw device as the determination target data. In the unit space generated in advance based on the reference data, the MD of the feature amount of the batch acquired as the judgment target data is calculated.

〈〈Verification of suitability of unit space〉〉 Here, it is necessary to create a unit space so that the state determination device 50 can determine whether the wire saw device is in a stable state or a state in which the possibility of failure increases based on the MD of the feature amount of the batch. That is, it is necessary to create a unit space so that the MD of the feature quantity of the batch in a state where the wire saw apparatus is likely to stop becomes higher than a predetermined threshold value, and that the MD of the batch feature quantity of the wire saw apparatus in a stable state is The MD does not meet the established threshold.

In order to confirm that the generated unit space satisfies the above conditions, MD is calculated based on the feature quantities of each batch included in the reference data illustrated in FIG. 3 . When a complex feature quantity is calculated for each batch, MD is calculated based on the data in which the complex feature quantity of each batch is one set. FIG. 4 shows a bar graph in which the MD values calculated based on the feature quantities of each batch included in the reference data are arranged in the order of processing by the wire saw device. The MD calculated from the feature quantities of each batch is also called the MD of each batch. The horizontal axis represents batches. The vertical axis represents the MD value of each batch. The range of batches included in each of the periods P1, P2, and P3 divided based on the state of the wire saw device is displayed along the horizontal axis. The number of batches included in period P1 (M1), the number of batches included in period P2 (M2), and the number of batches included in period P3 (M3) are 10 batches, 143 batches and 40 batches respectively. The ratios of the batches included in periods P1, P2, and P3 to the total number of batches are approximately 5%, approximately 75%, and approximately 20%. The ratio of batches included in each period is not limited to this ratio and can be changed appropriately.

The bar graph shown on the far right side of FIG. 4 represents the MD of the stopped batch 71 processed when the wire saw apparatus is stopped. In Figure 4, the predetermined threshold is represented by MD_T. If the vertical bar indicating the MD value of each batch extends above the dotted line indicating the value of MD_T, the MD is larger than the predetermined threshold. The control unit 51 sets the MD_T value so that the MD of each batch included in period P2 is less than MD_T, and the MD of at least one batch included in period P3 is above MD_T. The batch with MD above MD_T represents anomaly detection batch 72. By setting MD_T in this manner by the control unit 51, it can be determined that the state of the wire saw device is an abnormal state before the wire saw device stops during batch processing and a loss occurs.

The MD of the lots included in period P2 tends to be smaller than the MD of the lots included in periods P1 and P3. The reason for this is that the unit period is generated based on the feature amount of the batch included in the period P2. The feature quantity of each batch used to generate the unit space will expand the unit space in accordance with the feature quantity of each batch itself. Therefore, in a certain unit space, when the MD of the feature quantity used to generate the batch of the unit space is calculated, the calculated MD naturally becomes a very small value. On the contrary, when the MD of a feature quantity that is not used to generate a batch in a certain unit space is calculated, the possibility that the calculated MD becomes a very large value becomes high. In other words, the wider the distribution of the feature quantities of the batch used to generate the unit space, the smaller the MD calculated in the generated unit space becomes.

As described above, some of the feature quantities of the batch are excluded from the feature quantities of the batch in the period P2 used to generate the unit space. A batch that obtains measurement values in order to calculate a feature quantity in a unit space is assumed to be called an intra-unit space batch 73 . In order to calculate a batch that obtains measurement values without generating feature quantities in the unit space, it is assumed to be called a unit space out-of-unit batch 74 . In other words, the unit space is based on the feature quantity of the batch 73 within the unit space, and is not generated based on the batch 74 outside the unit space. The MD for generating a batch per unit space as described above is naturally calculated with a small value. Therefore, the MD of the batch 74 outside the unit space tends to have a larger value than the MD of the batch 73 within the unit space.

Here, the batch outside the unit space 74 is a batch processed during the period P2 when the status of the wire saw device is normal, although the characteristic quantity calculated based on the measured value is not used to generate the unit space. Therefore, a predetermined threshold is set so that the MD of the batch 74 outside the unit space does not exceed the predetermined threshold. If the MD of the lot 74 outside the unit space is set to be equal to or higher than a predetermined threshold, even if the state of the wire saw device is a normal state, it will easily be misjudged as an abnormal state. Therefore, by setting the predetermined threshold value to a value larger than the MD of the lot 74 outside the unit space, the control unit 51 can easily avoid determining that the status of the wire saw device is normal even though the status of the wire saw device is normal. unusual circumstances. In addition, when the threshold value has been set, it can be verified whether the unit space is appropriately set based on the MD value of the batch that is outside the unit space (outside unit space batch 74 ).

As described above, the control unit 51 generates the unit space based on the characteristic amount of each batch processed from the start of the operation of the wire saw device to the stop in the past predetermined period. That is, the control unit 51 acquires the feature quantity of the past processing batch at the time when the wire saw device is stopped as reference data, and generates the unit space based on the reference data. In the unit space generated in this way, the control unit 51 can obtain the feature value of the batch newly processed by the wire saw device as the determination target data, calculate the MD of the batch, and determine the state of the wire saw device.

〈〈Comparative example of data item selection〉〉 The unit space that is subject to the above-mentioned suitability verification is generated based on knowing the characteristic amount of the batch of data items that have a large contribution to the known pattern. Here, as a comparative example, the calculation in the unit space based on non-causal data items that have almost no contribution to the known pattern or have a small contribution is explained for each data item of the batch processed until the known pattern is stopped. Examples of MD values. Figure 5 shows a bar chart in which the MD values of each batch are arranged in the order of processing by the wire saw device in the unit space generated based on non-causal data items. Explanations on the meanings of the horizontal and vertical axes, and the meanings of periods P1, P2, and P3 are the same as those in Fig. 4 and will be omitted.

When comparing the MD values of each batch illustrated in FIG. 5 , the difference between the MD of the batch 74 outside the unit space in the period P2 and the MD of each batch in the period P3 is small. When MD_T indicating the predetermined threshold value is set low, the possibility that the wire saw device is misjudged as an abnormal state although it is a normal state increases. When MD_T is set to a high value, even though it is an abnormal state with a high possibility of stopping the wire saw device, it is not likely to be determined to be an abnormal state. If the state of the wire saw device is an abnormal state but is not determined to be an abnormal state, the opportunity to stop the wire saw device preventively is lost, and the possibility of stopping due to malfunction during batch processing increases. In the example of FIG. 5 , no matter how MD_T is set, the possibility of misjudgment of the status of the wire saw device increases.

Looking at the examples of FIGS. 4 and 5 , the value of the MD of the batch outside the unit space 74 in the period P2 subtracted from the MD of at least part of the batches in the period P3 is equal to or greater than the predetermined value, thereby improving the performance of the wire saw. The accuracy of determining the status of the device. The predetermined value can be set appropriately. The control unit 51 can appropriately select a cause data item from a plurality of data items so that the MD of the batches included in periods P2 and P3 satisfies the above conditions, and generates a unit space based on the cause data item. The control unit 51 sets the predetermined threshold value to be smaller than the MD of at least a part of the batches in the period P3 and larger than the maximum value of the MD of the batches 74 outside the unit space in the period P2. This makes it easier for the control unit 51 to avoid a situation in which the wire saw device is determined to be abnormal even though the state of the wire saw device is normal.

If the state determination device 50 generates the unit space based on all data items, the unit space may also be generated based on non-cause data items that have little or low contribution to factors that cause the wire saw device to stop. In this way, it becomes difficult for the MD value of the batch calculated in the unit space generated based on all the data items to reflect the change in the state of the wire saw device. Therefore, if the state of the wire saw device is determined based on the MD value of the batch calculated in the unit space generated by all data items, the determination accuracy can become as low as the example of FIG. 5 .

〈〈Update of unit space〉〉 As described above, the control unit 51 of the state determination device 50 generates the unit space based on the reference data acquired from the wire saw device. However, the characteristic amount of each batch of the data items of the processing device 10 such as the wire saw device changes as the operation time increases or as time passes. For example, when cooling water is used in the processing device 10, the temperature of the cooling water may vary depending on the season. Furthermore, for example, the feature quantity of each batch may change due to changes in the parts of the processing device 10 over time.

The criterion for determining the state of the wire saw device may change due to changes in the characteristic quantity of each batch of data items of the wire saw device. The status determination device 50 of this embodiment can update the unit space by updating the reference data used to generate the unit space, so that the determination standard matches the change in the characteristic amount of each batch of data items of the wire saw device. The following is an example of a structure for updating reference data.

〈〈Deletion of reference data〉〉 The control unit 51 calculates the MD value of the feature quantity of each batch included in the reference data in the unit space generated based on the reference data. The control unit 51 regards a plurality of feature values calculated for each of a plurality of data items in a batch as one set of feature values, and calculates one MD value for one batch based on this set of feature values. When the value of MD is large, the feature amount of the batch becomes a factor that expands the unit space. If the unit space is too large, it will become difficult to detect that the status of the wire saw device is abnormal.

Therefore, when the MD value of the feature quantity of each batch is equal to or greater than a predetermined threshold, the control unit 51 deletes the feature quantity of the batch from the reference data. If the feature values of the batch include feature values of a plurality of data items, the control unit 51 deletes the feature values of all data items in the batch from the reference data. If the feature values of the batch include a plurality of feature values for each data item, the control unit 51 deletes all the feature values of each data item from the reference data. When the MD value of the feature quantity of each batch is less than the predetermined threshold, the control unit 51 leaves the feature quantity of the batch in the reference data. The control unit 51 determines whether to delete all the plurality of feature values of the plurality of data items to be batched from the reference data or to leave them in the reference data. The established threshold is also called the deletion threshold. The control unit 51 may determine whether to delete from the reference data based on the feature values of all batches included in the reference data, or may determine whether to delete from the reference data based on the feature values of a part of the batches.

FIG. 6 illustrates MD values of feature quantities of 20 batches acquired from the wire saw device in chronological order. The horizontal axis of FIG. 6 represents the symbols assigned to each batch in chronological order. The vertical axis represents the MD value of the feature quantity of each batch. In the vertical axis, the deletion threshold is represented by MD_D. The control unit 51 determines that the MD values of the feature quantities of batches No. 1 to No. 10 and No. 12 in order of time are less than the deletion threshold and remain in the reference data. On the contrary, the control unit 51 determines that the MD values of the feature quantities of batches No. 11, No. 13 to No. 20 in chronological order are equal to or higher than the deletion threshold, and deletes them from the reference data.

The control unit 51 generates a unit space based on the reference data from which a part of the feature quantities of the batch has been deleted, thereby updating the existing unit space. The unit space being updated is also called the update unit space. The update unit space generated by deleting the feature values of batches with large MD values can be narrower than the unit space generated by simply deleting the feature values of 20 batches from the reference data in chronological order. In other words, the control unit 51 can control not to expand the unit space excessively by appropriately deleting the feature values of a part of the batch from the reference data, and therefore can suppress a decrease in sensitivity due to seasonal changes or time changes. As a result, the abnormality detection performance of the state determination device 50 is improved or is less likely to decrease.

When the control unit 51 generates the update unit space based on the reference data, the control unit 51 may use the various methods described above as methods of generating the unit space. For example, the control unit 51 may generate a unit space by dividing the reference data for which feature values of the batch have been deleted or added into data within the unit space and data outside the unit space.

〈〈〈Addition of reference materials〉〉〉 The control unit 51 operates the wire saw device to acquire the feature value of the newly processed batch. The control unit 51 may add the feature value of the newly acquired batch to the reference data. Thereby, the characteristic amount of the batch corresponding to the state change of the wire saw device can be reflected in the unit space. As a result, the criterion for determining the state of the wire saw device can be changed in response to changes in the state of the wire saw device.

The control unit 51 calculates the MD value of the newly acquired feature quantity in the existing unit space. The control unit 51 regards a plurality of feature values calculated for each of a plurality of data items in a batch as one set of feature values, and calculates one MD value for one batch based on this set of feature values. When the value of MD is large, the feature quantity of the batch becomes a factor that expands the unit space. If the unit space is too large, it will be difficult to detect abnormal status of the wire saw device. Therefore, the control unit 51 can add only the feature values of the batch with a small MD value to the reference data.

Therefore, when the MD value of the newly acquired feature quantity of a batch is less than the predetermined threshold, the control unit 51 adds the feature quantity of the batch to the reference data. When the feature amount of the batch includes feature amounts of a plurality of data items, the control unit 51 adds all the feature amounts of all the data items of the batch to the reference data. When the feature amount of the batch includes a plurality of feature amounts for each data item, the control unit 51 adds all the feature amounts of each data item to the reference data. On the other hand, when the MD value of the newly acquired feature quantity of a batch is equal to or greater than a predetermined threshold, the control unit 51 does not add the feature quantity of the batch to the reference data. The control unit 51 determines whether or not to add all the plurality of feature quantities of the plurality of data items in the batch to the reference data. The predetermined threshold used to determine whether to newly add feature quantities to the reference data is also called the additional threshold. The additional threshold may be set to the same value as the update threshold, or may be set to a different value.

The control unit 51 may determine whether to acquire the feature quantities of a plurality of batches at once, calculate the MD value of the feature quantities of each batch, and add it to the reference data.

FIG. 7 illustrates MD values of feature quantities of 20 batches newly acquired from the wire saw device. The horizontal axis of FIG. 7 represents the number marked on each batch according to the order of processing by the wire saw device. The vertical axis represents the MD value of the feature quantity of each batch. In the vertical axis, the update threshold is represented by MD_A. The control unit 51 determines that the MD value of the feature quantity of batches No. 11 to No. 19 according to the processing order is less than the addition threshold, and adds it to the reference data. On the contrary, the control unit 51 determines that the MD value of the feature value of batches No. 1 to No. 10 and No. 20 according to the processing order is equal to or higher than the addition threshold value, and does not add it to the reference data.

The control unit 51 generates a unit space based on the reference data of the added feature value of the new batch, thereby updating the existing unit space. It means that the update unit space generated by adding the feature quantities of batches with small MD values can be less likely to expand than the unit space generated by simply adding 20 batches of feature quantities to the reference data. In other words, the control unit 51 can control not to expand the unit space excessively by appropriately adding the feature values of the new batch to the reference data, and therefore can suppress the decrease in sensitivity due to seasonal changes or time changes. As a result, the abnormality detection performance of the state determination device 50 is improved or is less likely to decrease.

〈〈〈Exception processing for additional reference data〉〉〉 Due to changes in the status of the wire saw device, the characteristic values of the batch may move. When only the feature quantities of batches whose MD values are less than the addition threshold are added to the reference data, there may be cases where the update cannot be performed so that the unit space moves according to the movement of the feature quantities of the batch. Therefore, as an exception to the addition of reference data, the control unit 51 may add the feature values of a new batch to the reference data based on the following conditions.

The control unit 51 calculates the MD value of the feature quantity of each batch newly processed by the wire saw device. The control unit 51 determines whether the MD value of the feature amount of each batch is less than the additional threshold value. In principle, the control unit 51 adds only the feature values of batches whose MD values are less than the addition threshold to the reference data.

Here, as an exception process, the control unit 51 may add the feature values of these batches to the reference data when a predetermined number or more of batches whose MD values are equal to or higher than the addition threshold occur continuously. This is because it is statistically unimaginable that the data will continuously deviate from the predetermined range (in this embodiment, the range is determined based on the unit space) for more than a predetermined number of batches. For example, it is statistically unimaginable that the engineering capability index (Cp or CpK, etc.) used in quality management would continuously deviate beyond a predetermined number of batches. Here, the predetermined number may be set to a value of 2 or more. The predetermined number can be set, for example, so that the probability that the MD value exceeds the predetermined number and continues to be above the additional threshold is less than 0.3% (that is, the probability of 3σ).

As can be seen from the above, the phenomenon that the MD value continues to be above the additional threshold value beyond a predetermined number is considered to reflect the characteristics of each batch rather than a sudden change in the characteristic amount of the batch processed by the wire saw device. resulting from changes in quantitative tendencies. Therefore, when the control unit 51 executes the above-mentioned exception processing, the change in the tendency of the feature quantity of the batch processed by the wire saw device can be reflected in the unit space. As a result, the accuracy of abnormality detection by the state determination device 50 can be improved.

〈〈〈Deletion and addition of reference data〉〉〉 The control unit 51 may perform an operation of deleting a part of the feature quantity of a batch from the reference data and an operation of adding a feature quantity of a new batch to the reference data in a coordinated manner, and then generate the unit based on the reference data newly obtained by performing these operations. space.

In the example of Figure 6, the feature quantities of 11 batches are deleted from the reference data. In the example of Fig. 7, the feature values of 9 batches are added to the reference data. As a result, the number of batches constituting the reference data is reduced by 2. In order to maintain a constant number of batches constituting the reference data, the control unit 51 may set a deletion threshold and an addition threshold respectively so that the number of batches to be deleted matches the number of batches to be added.

For example, in the example of FIG. 6 , the control unit 51 may reduce the deletion threshold so that the MD value of the feature amount of batch No. 5 does not satisfy the deletion threshold. With this, the number of batches deleted from the reference data becomes 10. On the other hand, in the example of FIG. 7 , the additional threshold can be increased so that the MD value of the feature quantity of batch No. 1 does not satisfy the additional threshold. As a result, the number of batches added to the reference data becomes 10. As a result, the control unit 51 can make the number of deleted batches and the number of added batches equal.

The control unit 51 may increase both the deletion threshold and the addition threshold, thereby making the number of deleted batches and the number of added batches equal. The control unit 51 may reduce both the deletion threshold and the addition threshold, thereby making the number of deleted batches equal to the number of added batches. By increasing or decreasing both the deletion threshold and the addition threshold, the state of the wire saw device becomes easier to be reflected in the unit space. As a result, the abnormality detection performance of the state determination device 50 is improved or is less likely to deteriorate.

〈〈Status determination of wire saw device based on generated or updated unit space and determination threshold〉〉 The control unit 51 calculates the characteristic amount of the batch of the wire saw device whose stop time point is unknown based on the measurement value obtained when the device processes the batch, acquires it as the determination target data, and determines that the wire saw device is stopped based on the determination target data. possibility becomes higher.

Specifically, the control unit 51 calculates the MD of the newly processed batch in the unit space generated or updated based on the feature values of the past processed batches. The control unit 51 determines, based on the calculated value of MD, whether or not the possibility that the wire saw device will stop during the processing of the next batch or later of the batch increases.

When the MD is less than the predetermined threshold, the control unit 51 determines that the possibility of the wire saw device stopping is low, that is, the state of the wire saw device is a normal state during processing of the batch. When MD is equal to or greater than a predetermined threshold, the control unit 51 determines that the wire saw device is likely to stop, that is, the state of the wire saw device is an abnormal state during the processing of this batch. A predetermined threshold used to determine the state of the wire saw device is also called a determination threshold.

The control unit 51 may display the message that the wire saw device is in an abnormal state on the output unit 53 or notify the user through sound output or the like. Based on this notification, the user can prevent the wire saw device from processing the next batch, stop it preventively, and perform maintenance operations. The control unit 51 may output a message to the wire saw device that the state of the wire saw device is an abnormal state, and preventively stop the wire saw device.

〈〈Determination of abnormal conditions based on known patterns〉〉 When the control unit 51 confirms that the wire saw device is stopped in a known mode, it may obtain the feature values of the batches processed from the start of the operation to the stop as reference data. Data containing characteristic quantities of batches processed before the known mode was stopped is also known as known mode data. The control unit 51 may also generate or update the unit space based on known pattern data. The unit space based on known model data is also called known model unit space.

The control unit 51 calculates the MD of the batch processed by the wire saw device whose status is unknown in the known pattern unit space. When the MD of the batch calculated in the known mode unit space is equal to or higher than a predetermined threshold, the control unit 51 can determine that the wire saw device is in an abnormal state in which the possibility of stopping the wire saw device in the known mode increases. An abnormal state in which the possibility of the wire saw device stopping in a known pattern becomes high is also called an abnormal state based on a known pattern.

When the known pattern includes a plurality of failure modes such as a first known mode or a second known mode, the control unit 51 may, based on the reference data including the data of the batch processed before stopping each failure mode, Generate or update the unit space corresponding to each failure mode. For example, the control unit 51 generates or updates a first unit space corresponding to a first known pattern, or a second unit space corresponding to a second known pattern. The control unit 51 can calculate the MD of the batch processed by the wire saw processing device in the unit space corresponding to each failure mode. When the MD of the batch calculated in the unit space corresponding to each failure mode is equal to or greater than a predetermined threshold, the control unit 51 may determine that the wire saw device is more likely to stop in the failure mode in which the MD is equal to or greater than the predetermined threshold. abnormal status. The thresholds used to determine the MD values of the first unit space and the second unit space are also called a first threshold and a second threshold, respectively. The first threshold and the second threshold may be different values or may be the same value. The states in which the wire saw device is likely to stop in the first known mode and the second known mode are also called a first abnormal state and a second abnormal state respectively. The threshold value is not limited to the first threshold value and the second threshold value, and may include, for example, a third threshold value set to determine the MD of the third unit space. The abnormal state is not limited to the first abnormal state and the second abnormal state, and may include, for example, a third abnormal state corresponding to a state with a high possibility of stopping in the third known mode. Also, please note that identifiers such as "1st" and "2nd" do not indicate the threshold or the quality of the abnormal state.

The control unit 51 may determine whether the state of the wire saw device is an abnormal state based on a known pattern, a first abnormal state, or a second abnormal state, and output the result of the determination to the output unit 53 . The control unit 51 can notify users such as an operator of the wire saw device or a maintenance person of the determination result. Based on the notified determination result, the user performs various operations such as inspection operations, parts replacement operations, or repair operations of the wire saw device. Thereby, the user can properly maintain the wire saw device so that the wire saw device does not malfunction during batch processing.

〈〈Determination of the possibility of unexplained stoppage〉〉 When the wire saw device stops due to an unknown reason, the control unit 51 may obtain the feature values of the batches processed from the start of the operation to the stop as reference data. Data including the characteristic quantities of batches whose pre-processing was stopped due to unknown reasons is also called unexplained data. The control unit 51 may also generate or update the unit space based on unexplained data. Unit spaces based on unexplained data are also called unexplained unit spaces. The control unit 51 may select all data items obtained from the wire saw device to generate or update the unexplained unit space, or may select part of the data items to generate or update the unexplained unit space.

The control unit 51 calculates the MD of the batch processed by the wire saw device whose status is unknown in the unknown unit space. The control unit 51 may determine that the wire saw device is in an abnormal state in which the possibility of stopping the wire saw device due to an unknown cause increases when the MD of the batch calculated in the unexplained unit space is equal to or higher than a predetermined threshold. The threshold used to determine the MD value of the batch calculated in the unexplained unit space is also called the unexplained threshold. The unexplained threshold is assumed to be a predetermined threshold used to determine the MD value of the batch calculated in the known pattern unit space, or the same value as the first threshold or the second threshold, but may be set to a different value.

When the control unit 51 determines that there is a high possibility that the wire saw device will stop due to an unknown reason, the control unit 51 may further perform effectiveness analysis to narrow the range of factors causing the wire saw device to stop. Validity analysis corresponds to analyzing which data item among the data items used to generate or update the cause unit space has a large impact on the value of MD. The control unit 51 can perform validity analysis using various techniques. The control unit 51 can determine the data item that increases the MD value of the batch calculated in the unexplained unit space through validity analysis. A data item that increases the MD value of the batch calculated in the unexplained unit space is also called a candidate cause item. A specific example of the method of determining candidate cause items through effectiveness analysis will be described later.

The control unit 51 may determine the cause candidate items and output them to the output unit 53, thereby notifying users such as the operator of the wire saw device or the maintenance person in charge of the cause candidate items. The user performs various operations such as inspection operations, parts replacement operations, or repair operations of the wire saw device based on the data items determined as cause candidate items. Thereby, the wire saw device can be properly maintained before failure during batch processing.

〈Change in MD value of judgment target data due to update of unit space〉 As described above, the control unit 51 updates the unit space by adding the feature values of the batch newly processed by the wire saw device to the reference data, or deleting the feature values of the earlier batch from the reference data. In the following, in order to explain the effect of updating the unit space by the method of this embodiment, the case where the control unit 51 does not update the unit space, the case where the unit space is updated by the method of the comparative example, and the case where the unit space is updated by the method of this embodiment are described below. The MD value of the situation comparison judgment target data.

In the following description, it is assumed that the unit space is generated or updated based on the reference data of the first wire saw device. It is assumed that the judgment target data is the feature quantity of the batch processed by the second wire saw device. That is, for convenience, the device for acquiring the reference data and the device for acquiring the determination target data are different. By making the device that acquires the reference data and the device that acquires the determination target data different, the MD value is calculated with a larger value. By calculating the value of MD with a large value, it is easier to compare the value of MD in each case.

First, when the unit space is not updated, the MD value of the determination target data calculated in the unit space is shown in FIG. 8 . The horizontal axis of Fig. 8 represents the batch. The vertical axis represents the MD value of each batch.

Next, as a method of the comparative example, the control unit 51 deletes the feature values of 20 batches from the reference data in sequence, regardless of the MD value of the feature values of each batch, and adds the feature values of the newly processed 20 batches. . An example of the MD value of the determination target data calculated in the unit space updated by the method of the comparative example is shown in FIG. 9 . The horizontal axis of Fig. 9 represents the batch. The vertical axis represents the MD value of each batch.

Next, as a method of this embodiment, the control unit 51 deletes 11 batches of feature values whose MD value is above the deletion threshold from the reference data in sequential order, and adds 20 newly processed batches. Among the feature values of 9 times, the MD value is less than the additional threshold value. An example of the MD value of the determination target data calculated in the unit space updated by the method of this embodiment is shown in FIG. 10 . The horizontal axis of Fig. 10 represents the batch. The vertical axis represents the MD value of each batch.

Comparing Figures 8 and 9, the MD value of each batch in Figure 9 becomes smaller. The smaller MD value of each batch means that when the unit space updated by the method of the comparative example is used for judgment, the detected status of the wire saw device is smaller than when the unit space is not updated. Abnormal sensitivity decrease.

On the other hand, comparing FIG. 8 and FIG. 10 , the MD value of each batch in FIG. 10 is larger. What means that the MD value of each batch becomes larger is that when the unit space updated by the method of this embodiment is used for determination, the wire saw device is detected more quickly than when the unit space is not updated. The status is an abnormal sensitivity increase.

As described above, the control unit 51 of this embodiment deletes old reference data or adds new reference data by considering the MD value of the feature quantity of the batch, thereby being able to cope with data that changes due to seasonal changes or aging. Therefore, It is possible to improve the sensitivity of abnormality detection of wire saw equipment using unit space.

Furthermore, as an exception process in the method of this embodiment, when the MD value exceeds the addition threshold and a predetermined number or more consecutive batches occur, the control unit 51 adds the batches that meet the condition to the reference data. An example of the MD value of the determination target data calculated in the unit space and updated based on the reference data using exception processing in the method of this embodiment is shown in FIG. 11 . The horizontal axis of Fig. 11 represents the batch. The vertical axis represents the MD value of each batch.

Comparing Fig. 8 and Fig. 11, the MD value of each batch in Fig. 11 becomes larger. That is, the MD value of each batch does not become smaller. Moreover, comparing FIG. 10 and FIG. 11 , the MD value of each batch in FIG. 11 becomes smaller. Based on these comparison results, it can be seen that by using exception processing to add the feature values of a new batch to the reference data, the unit space will not become too narrow, and the sensitivity for detecting whether the status of the wire saw device is abnormal will not decrease. . That is, the unit space is updated to move in accordance with the tendency of the feature quantity of the batch newly processed by the wire saw device. As a result, it is possible to improve the accuracy of detecting whether the state of the wire saw device is abnormal.

<Summary> As described above, according to the state determination system 1 and the state determination device 50 of this embodiment, before the processing device 10 such as a wire saw device fails, it can be determined whether or not there is an abnormal state in which the possibility of the processing device 10 failure increases. Thereby, the processing device 10 becomes less likely to malfunction during the processing. If the processing device 10 fails during the processing, great losses will occur. Because the processing device 10 becomes less likely to malfunction during the processing, losses can be reduced.

[Step example of status determination method] The control unit 51 of the state determination device 50 can execute, for example, a state determination method including the steps of the flowcharts illustrated in FIGS. 12 , 13 , and 14 . The control unit 51 executes the illustrated state determination method and can determine whether the state of the wire saw device is an abnormal state. The state determination method may be implemented as a state determination program executed by the control unit 51 . The steps shown in Figures 12 and 13 are examples. It can also be changed appropriately.

The control unit 51 generates or updates the unit space (step S1). The control unit 51 generates or updates the unit space based on the reference data of the wire saw device. When the wire saw device is stopped in a known mode, the control unit 51 can generate or update the unit space corresponding to the known mode based on reference data obtained as data of the batch processed before stopping. When the wire saw device stops due to an unknown reason, the control unit 51 can generate or update a unit space corresponding to the unknown reason based on the reference data obtained as the data of the batch processed before the stop.

FIG. 13 illustrates an example of steps for generating a unit space as a subroutine. The control unit 51 acquires the reference data (step S11 in FIG. 13 ). The reference data includes the characteristic amount of the batch calculated based on the measured values of each batch processed from the past operation of the wire saw device until it was stopped. The control unit 51 may obtain the feature value of the batch processed by the wire saw device before stopping in a known mode such as the first known mode or the second known mode as reference data. The control unit 51 may also obtain the feature value of the batch processed before the wire saw device stopped due to unknown reasons as reference data.

The control unit 51 selects a cause data item from the reference data (step S12). The reference data includes a plurality of data items that can be obtained from the wire saw device. The control unit 51 may select a data item corresponding to a predetermined failure mode (a factor that causes the wire saw device to stop) as the cause data item. For example, the control unit 51 may select data items corresponding to the first known pattern and the second known pattern as the first cause data item and the second cause data item, respectively. When the control unit 51 obtains as reference data the characteristic values of the batch processed before the wire saw device stopped due to unknown reasons, it may select all the cause data items.

The control unit 51 classifies the reference data into data within the unit space (data of the batch 73 within the unit space) and data outside the unit space (data of the batch 74 outside the unit space) (step S13).

The control unit 51 generates a unit space based on the data in the unit space (step S14). The control unit 51 may generate a first unit space, a second unit space or an unexplained unit space as the unit space.

The control unit 51 determines whether the generated unit space is appropriate (step S15). Specifically, the control unit 51 calculates the MD of the batch 73 within the unit space and the MD of the batch 74 outside the unit space, respectively, in the generated unit space. For example, the control unit 51 may determine that the generated unit space is appropriate when the difference between the MD of the lot 73 within the unit space and the MD of the lot 74 outside the unit space is less than a predetermined value. When the control unit 51 determines that the generated unit space is inappropriate (step S15: NO), the control unit 51 returns to step S13. The control unit 51 may return to step S12. When the control unit 51 determines that the generated unit space is appropriate (step S15: YES), the subroutine process illustrated in FIG. 13 ends and returns to step S2 in FIG. 12 .

The control unit 51 can generate a unit space corresponding to a plurality of known patterns in the subroutine process of the unit space generation step illustrated in FIG. 13 . The control unit 51 may generate an unexplained unit space corresponding to an unexplained stop.

Moreover, as a subroutine, FIG. 14 shows an example of the procedure of updating the unit space.

The control unit 51 calculates the MD value of the feature quantity of each batch included in the reference data for generating the unit space (step S21 ). The control unit 51 determines whether the MD value of the feature amount of each batch is equal to or higher than the deletion threshold (step S22 ). When the MD value is equal to or greater than the deletion threshold (step S22: YES), the control unit 51 deletes the feature amount of the batch from the reference material (step S23). When the value of MD is not greater than the deletion threshold (step S22: NO), that is, when the value of MD is less than the deletion threshold, the control unit 51 skips step S23 so that the feature values of the batch are not removed from the reference data. Delete and proceed to step S24. The control unit 51 may execute steps S22 and S23 on the feature values of a part of the batches included in the reference data, or may execute the steps S22 and S23 on the feature values of all the batches. The control unit 51 can appropriately delete the feature amount of the batch from the reference data by executing steps S21 to S23.

The control unit 51 acquires the feature quantity of the batch newly processed by the wire saw device, and calculates the MD value of the acquired feature quantity of the batch (step S24 ). The control unit 51 determines whether the MD value of the feature amount of the batch is less than the additional threshold (step S25 ). When the MD value is less than the addition threshold (step S25: YES), the control unit 51 adds the feature value of the batch to the reference data (step S26).

When the MD value is not less than the additional threshold value (step S25: NO), that is, when the MD value is equal to or higher than the additional threshold value, the control unit 51 determines whether there are consecutive MDs of feature amounts of more than a predetermined number of batches. The value is above the additional threshold (step S27). When the MD values of the feature quantities of more than the predetermined number of consecutive batches are not equal to or greater than the additional threshold (step S27: NO), the control unit 51 skips step S26 and proceeds to step S28 so that the feature quantities of the new batches are It is not added to the reference data. When there are more than a predetermined number of consecutive batches whose feature values have MD values that are equal to or greater than the additional threshold (step S27: YES), the control unit 51 proceeds to step S26 and adds a predetermined number of consecutive batches whose MD values are equal to or greater than the additional threshold. The feature quantities of the batch are appended to the reference data. The control unit 51 executes steps S24 to S27, thereby appropriately adding the feature amount of the batch to the reference data.

The control unit 51 generates an update unit space based on the reference data (step S28 ). After the control unit 51 executes step S28, it ends the subroutine processing illustrated in FIG. 14 and returns to step S2 in FIG. 12.

Returning from the subroutine processing of FIG. 13 or FIG. 14 executed as step S1 of FIG. 12 , the control unit 51 obtains the determination target data (step S2 of FIG. 12 ).

The control unit 51 calculates the MD of the lot included in the determination target data (step S3). The control unit 51 may calculate the MD of the batch included in the determination target data for each of the plurality of unit spaces generated or updated in step S1. The control unit 51 can calculate the MD of the batch included in the determination target data in the first unit space. The MD of the batch included in the judgment target data is calculated in the first unit space, which is also called the MD of the first unit space. The control unit 51 may calculate the MD of the batch included in the determination target data in the second unit space. The MD of the batch included in the judgment target data is calculated in the second unit space, which is also called the MD of the second unit space. The control unit 51 may calculate the MD of the batch included in the determination target data in the unexplained unit space. The MD of the batch included in the judgment target data is calculated in the unexplained unit space, which is also called the MD of the unexplained unit space.

The control unit 51 determines whether the MD of the first unit space is equal to or greater than the first threshold (step S4). For example, the control unit 51 may set the first threshold when the first unit space is generated in step S1. The first threshold value can be set so as to detect that the state of the wire saw device becomes an abnormal state based on the first known pattern, and to prevent erroneous detection of an abnormality when the state of the wire saw device is normal.

When the MD of the first unit space is equal to or greater than the first threshold (step S4: YES), the control unit 51 proceeds to step S8. When the MD of the first unit space is not greater than the first threshold (step S4: NO), that is, when the MD of the first unit space is less than the first threshold, the control unit 51 determines whether the MD of the second unit space is is above the second threshold (step S5). For example, the control unit 51 may set the second threshold when the second unit space is generated in step S1. The second threshold value can be set so as to detect that the state of the wire saw device becomes an abnormal state based on the second known pattern, and to prevent erroneous detection of an abnormality when the state of the wire saw device is normal.

When the MD of the second unit space is equal to or greater than the second threshold (step S5: YES), the control unit 51 proceeds to step S8. When the MD of the second unit space is not greater than the second threshold (step S5: NO), that is, when the MD of the second unit space is less than the second threshold, the control unit 51 determines whether the MD of the unexplained unit space is is above the unknown cause threshold (step S6). For example, the control unit 51 may set the unexplained threshold value when an unexplained unit space occurs in step S1. The unexplained threshold value can be set so as to detect that the state of the wire saw device becomes an abnormal state of unknown cause, and at the same time, when the state of the wire saw device is normal, an abnormality will not be erroneously detected.

When the MD of the unexplained unit space is not equal to or higher than the unexplained threshold (step S6: NO), that is, when the MD of the unexplained unit space is less than the unexplained threshold, the control unit 51 ends the steps of the flowchart in FIG. 12 .

When the MD of the unexplained unit space is equal to or higher than the unexplained threshold (step S6: YES), the control unit 51 determines a cause candidate item (step S7), which is data corresponding to factors that increase the possibility of the wire saw device stopping. project. The control unit 51 may determine candidate cause items through validity analysis. Specific examples of validity analysis will be described later.

After step S7 , or when it is determined that MD is equal to or higher than the threshold in step S4 or S5 , the control unit 51 preventively stops the wire saw device (step S8 ). Thereby, the wire saw device as the processing device 10 can be maintained through inspection work, parts replacement work, repair work, and the like before a failure occurs.

According to the state determination method described above, it can be determined whether the state of the wire saw device is an abnormal state before the wire saw device as the processing device 10 breaks down. Thereby, the wire saw device as the processing device 10 becomes less likely to malfunction and stop during batch processing. If the processing device 10 fails during batch processing, a large loss will occur such that the batch being processed needs to be discarded. Since the processing device 10 is less likely to malfunction during batch processing, losses can be reduced.

[Other implementation types] In the above-mentioned embodiment, the structure in which the state determination device 50 determines the state of the wire saw device as the processing device 10 has been explained. The processing device 10 whose status the status determination device 50 determines is not limited to the wire saw device, and may determine the status of other devices such as a polishing device.

Although the implementation forms of the present disclosure have been described with reference to the drawings and examples, it should be noted that those with ordinary skill in the technical field to which the present invention belongs can make various modifications or changes with reference to the present disclosure. Therefore, these modifications or changes are also included in the scope of the present disclosure. For example, the functions included in each component or step can be rearranged without logical contradiction, and multiple components or steps can be combined into one or divided. The implementation form of the present disclosure is mainly explained based on the device, but the implementation form of the present disclosure can also be implemented by a method including steps executed by each component of the device. The implementation form of the present disclosure can also be implemented by a method, a program executed by a processor of the device, or a storage medium that stores the program. The scope of this disclosure should be understood to include these contents.

The drawings contained in this disclosure are schematic diagrams. The proportions will not necessarily match reality. [Industrial Applicability]

According to the embodiments of the present disclosure, the accuracy of determining the status of the processing device 10 can be improved.

1: Status determination system 10: Processing device 12: line 14:Roller 16: Line group 18: Workpiece holding mechanism 20:Nozzle 32: Guide roller 33: Swing arm 34: Swing roller 35:Touch wheel 36: Drive motor 38A: Line reel 38B: Line reel 40:Abrasive storage tank 42:Abrasive cooler 50: Status determination device 51:Control Department 53:Output Department 54:Input part 60: Sensor 71:Stop batch 72: Anomaly detection batch 73: Batch within unit space 74: Batch outside unit space W: workpiece (block) X:Roller axis direction

FIG. 1 is a block diagram showing an architectural example of a state determination system according to an embodiment. FIG. 2 is a schematic diagram showing a frame example of a wire saw device as a processing device. FIG. 3 is a table showing an example of data items obtained from the wire saw device. FIG. 4 is a graph showing an example of the MD of each batch calculated based on the unit space of the cause data item. FIG. 5 is a graph showing an example of the MD of each batch calculated in the unit space of the comparative example. FIG. 6 is a graph showing an example of MD of feature quantities of candidate batches excluded from the data in the unit space. FIG. 7 is a graph showing an example of MD of feature amounts of candidate batches added to the unit space data. FIG. 8 is a graph showing an example of the MD value of the determination target data calculated based on the unit space before update. FIG. 9 is a graph showing an example of the MD value of the determination target data calculated based on the unit space updated by the method of the comparative example. FIG. 10 is a graph showing an example of the MD value of the determination target data calculated based on the unit space updated by the method of this embodiment. FIG. 11 is a graph showing an example of the MD value of the determination target data calculated based on the unit space updated by adding exception processing according to the method of this embodiment. FIG. 12 is a flowchart showing an example of steps of a state determination method according to an embodiment. FIG. 13 is a flowchart showing an example of subroutine processing for generating a unit space. FIG. 14 is a flowchart showing an example of subroutine processing for updating the unit space.

1: Status determination system

10: Processing device

50: Status determination device

51:Control Department

53:Output Department

54:Input part

60: Sensor

Claims (8)

A status determination device, including: The control unit determines the status of the processing device, Among them the control department: Obtain the characteristic quantities of multiple batches normally processed by the processing device as reference data; Based on the reference data, a unit space is generated that represents the characteristics of the batch group normally processed by the processing device; Calculate the Mahalanobis Distance (Mahalanobis Distance) of the feature quantities of each batch contained in the reference data in the unit space, and delete the batches whose Mahalanobis distance is above the deletion threshold from the reference data. characteristic quantity; Based on the updated reference data, generate the unit space as the updated unit space; Obtain the feature quantity of the batch targeted for determining the status of the processing device as the determination target data; Calculate the Mahalanobis distance of the determination object data in the updated unit space; When the Mahalanobis distance is equal to or greater than the determination threshold set based on the updated unit space, a result is output that determines that the state of the processing device is an abnormal state. A status determination device, including: The control unit determines the status of the processing device, Among them the control department: Obtain the characteristic quantities of multiple batches normally processed by the processing device as reference data; Based on the reference data, a unit space is generated that represents the characteristics of the batch group normally processed by the processing device; Calculate the Mahalanobis distance of the feature quantities of the batch newly acquired in order to add to the reference data in the unit space, and add the feature quantities of the batch whose Mahalanobis distance does not meet the addition threshold to the reference in the data; Based on the updated reference data, generate the unit space as the updated unit space; Obtain the feature quantity of the batch targeted for determining the status of the processing device as the determination target data; Calculate the Mahalanobis distance of the determination object data in the updated unit space; When the Mahalanobis distance is equal to or greater than the determination threshold set based on the updated unit space, a result is output that determines that the state of the processing device is an abnormal state. A status determination device, including: The control unit determines the status of the processing device, Among them the control department: Obtain the characteristic quantities of multiple batches normally processed by the processing device as reference data; Based on the reference data, a unit space is generated that represents the characteristics of the batch group normally processed by the processing device; Calculate the Mahalanobis distance of the feature quantities of each batch contained in the reference data in the unit space, and delete the feature quantities of the batches whose Mahalanobis distance is above the deletion threshold from the reference data; Calculate the Mahalanobis distance of the feature quantities of the batch newly acquired in order to add to the reference data in the unit space, and add the feature quantities of the batch whose Mahalanobis distance does not meet the addition threshold to the reference in the data; Based on the updated reference data, generate the unit space as the updated unit space; Obtain the feature quantity of the batch targeted for determining the status of the processing device as the determination target data; Calculate the Mahalanobis distance of the determination object data in the updated unit space; When the Mahalanobis distance is equal to or greater than the determination threshold set based on the updated unit space, a result is output that determines that the state of the processing device is an abnormal state. Such as the status determination device of claim 3, wherein: The control unit sets the deletion threshold and the addition threshold so that the number of batches removed from the reference data is equal to the number of batches added to the reference data. Such as requesting the status determination device of any one of items 2 to 4, wherein: When the Mahalanobis distance of the feature values of the batches of more than a predetermined number newly acquired for adding to the reference data is continuously at or above the addition threshold, the control unit removes the batches of more than the predetermined number from the The feature quantity of is appended to the reference data. A status determination method, executed by a status determination device to determine the status of a processing device, includes the following steps: Obtain the characteristic quantities of multiple batches normally processed by the processing device as reference data; Based on the reference data, a unit space is generated that represents the characteristics of the batch group normally processed by the processing device; Calculate the Mahalanobis distance of the feature quantities of each batch contained in the reference data in the unit space, and delete the feature quantities of the batches whose Mahalanobis distance is above the deletion threshold from the reference data; Based on the updated reference data, generate the unit space as the updated unit space; Obtain the feature quantity of the batch targeted for determining the status of the processing device as the determination target data; Calculate the Mahalanobis distance of the determination object data in the updated unit space; When the Mahalanobis distance is equal to or greater than the determination threshold set based on the updated unit space, a result is output that determines that the state of the processing device is an abnormal state. A status determination method, executed by a status determination device to determine the status of a processing device, includes the following steps: Obtain the characteristic quantities of multiple batches normally processed by the processing device as reference data; Based on the reference data, a unit space is generated that represents the characteristics of the batch group normally processed by the processing device; Calculate the Mahalanobis distance of the feature quantities of the batch newly acquired in order to add to the reference data in the unit space, and add the feature quantities of the batch whose Mahalanobis distance does not meet the addition threshold to the reference in the data; Based on the updated reference data, generate the unit space as the updated unit space; Obtain the feature quantity of the batch targeted for determining the status of the processing device as the determination target data; Calculate the Mahalanobis distance of the determination object data in the updated unit space; When the Mahalanobis distance is equal to or greater than the determination threshold set based on the updated unit space, a result is output that determines that the state of the processing device is an abnormal state. A status determination method, executed by a status determination device to determine the status of a processing device, includes the following steps: Obtain the characteristic quantities of multiple batches normally processed by the processing device as reference data; Based on the reference data, a unit space is generated that represents the characteristics of the batch group normally processed by the processing device; Calculate the Mahalanobis distance of the feature quantities of each batch contained in the reference data in the unit space, and delete the feature quantities of the batches whose Mahalanobis distance is above the deletion threshold from the reference data; Calculate the Mahalanobis distance of the feature quantities of the batch newly acquired in order to add to the reference data in the unit space, and add the feature quantities of the batch whose Mahalanobis distance does not meet the addition threshold to the reference in the data; Based on the updated reference data, generate the unit space as the updated unit space; Obtain the feature quantity of the batch targeted for determining the status of the processing device as the determination target data; Calculate the Mahalanobis distance of the determination object data in the updated unit space; When the Mahalanobis distance is equal to or greater than the determination threshold set based on the updated unit space, a result is output that determines that the state of the processing device is an abnormal state.

Images