JP7223947B2 - Manufacturing condition calculation device, manufacturing condition calculation method, and manufacturing condition calculation program - Google Patents

Description

The present invention relates to anomaly detection technology for chemical plants, steel plants, semiconductor manufacturing facilities, general industrial product manufacturing facilities, etc. In particular, the anomaly detection accuracy of plants and manufacturing equipment is improved based on the operation results of a plurality of past operations. The present invention relates to a production condition calculation device, a production condition calculation method, and a production condition calculation program for improving production conditions.

Chemical plants often use multiple devices (such as reactors) to produce the same product. In this case, it is necessary to appropriately manage the operating state of each device so that the products produced are uniform.
However, when multiple devices are operated in parallel under the same conditions, even if the catalog specifications of each device are the same, there are individual differences in the actual operation results, and the operation results differ for each processing head in the device. In addition, even if production is performed under the same conditions, the results of operation may differ due to the difference in aging of equipment, and the results of operation may vary greatly due to the influence of previous processes. It is difficult to accurately manage the state in which the manufactured products are uniform for each apparatus due to the effects of such apparatus differences, processing head differences, temporal change differences, and previous processes.

Similar problems occur not only in chemical plants, but also in semiconductor manufacturing facilities, metal refining plants, water treatment facilities, gas plants, and the like. Moreover, such problems can arise not only in facilities that manufacture something, but also in facilities that change the characteristics of intangible objects, such as power generation and substation facilities. Hereinafter, the term "plant" includes not only a facility that manufactures something, but also a facility that changes the properties of intangible objects.
In addition to complex facilities, the term "plant" also includes single facilities that behave differently depending on operating conditions.

In a conventional plant control information generation device, the relationship between past samples is used to estimate the cause of an abnormality in a plant.
More specifically, in Patent Document 1, as one of the conventional methods for estimating the cause of an abnormality in a production process such as a device manufacturing process or a material process, for example, pattern recognition using the Mahalanobis distance is used to detect an abnormality. A method for diagnosing the cause is disclosed.
FIG. 10 shows a flow diagram of the method for diagnosing the cause of abnormality disclosed in Patent Document 1. As shown in FIG.

As shown in FIG. 10, first, an appropriate amount of predetermined data (for example, temperature, pressure, etc.) in the steady state of the machine to be diagnosed is measured and used as reference data, and the feature amount of these reference data is demand. Then, the feature values are normalized, a correlation matrix is obtained based on the correlations of all combinations between the normalized feature values, and then an inverse matrix (that is, Mahalanobis space) of the correlation matrix is obtained (step 1). Next, the Mahalanobis distance (MD) of the machine to be diagnosed is obtained (step 2). Next, MD is compared with a preset threshold (generally 0 to 3) (step 3), and if MD is greater than the threshold, the maximum distance element value is identified (step 4). Then, the feature amount corresponding to the maximum value of the distance element values is replaced with the average value of the feature amounts of the reference data (step 5). Next, MD is obtained again using the replaced feature amount (step 2). Next, the newly determined MD is compared with a threshold (step 3), and if the MD is greater than the threshold, the maximum distance element value is identified (step 4). Then, in the same manner as described above, replacement processing is performed (step 5) and repeated until MD becomes smaller than the threshold. Then, when the MD becomes smaller than the threshold, the distance element values are totaled, and the value when the distance element value reaches the maximum is set as the value "final effect" indicating the cause of abnormality of the feature amount.

JP 2004-227279 A

However, in the configuration of the conventional patent document 1, since it is necessary to process the data of individual samples as a processing flow, there is a problem that the calculation takes a long time when the number of samples becomes enormous. In addition, the size relationship between the process control data estimated as the cause of the abnormality and the process control data in the normal state (direction of abnormality) and the degree of deviation (degree of abnormality) are unknown, so information on the direction and degree of abnormality There is a problem that it is difficult to formulate an appropriate countermeasure against abnormality based on the diagnosis.
In particular, in general industrial product manufacturing equipment where the parameters that determine the manufacturing process are not comprehensively measured, there are cases where non-measurement items in the previous process have a dominant effect on the process. However, it may detect that the operating state of the apparatus has changed in a better direction than before, and there is a problem that the use of abnormality detection is insufficient.

An object of the present invention is to solve the above-mentioned problems, and to quantitatively determine how to deal with equipment failures, even in general processing equipment where the parameters that determine the manufacturing process are not comprehensively measured. It is an object of the present invention to provide a production condition calculation device, a production condition calculation method, and an anomaly detection program.

In order to achieve the above object, the manufacturing condition calculation device of the present invention divides actual measurement data related to product manufacturing into unit analysis periods, and for each combination of measurement items and devices included in the measurement data and the unit analysis periods a histogram feature quantity calculation unit that calculates a histogram feature quantity of the measurement result included in the actual measurement data; and a Mahalanobis distance for each combination of the measurement item, the device, and the unit analysis period based on the histogram feature quantity a Mahalanobis distance calculation unit that calculates a total Mahalanobis distance calculation unit that calculates the total Mahalanobis distance based on the total value of the Mahalanobis distance for each combination of the measurement item and the device in the most recent analysis target period, the measurement item, a process capability index calculation unit that calculates a process capability index for each combination of the device and the unit analysis period; An aggregated process capability index calculator for calculating a process capability index, and an anomaly detection display unit for displaying the aggregated Mahalanobis distance and the aggregated process capability index by dividing them.

The manufacturing condition calculation method of the present invention divides the actual measurement data related to the manufacturing of the product into each unit analysis period, and includes in the actual measurement data for each combination of measurement items and devices included in the actual measurement data and the unit analysis period. A histogram feature amount calculation step of calculating a histogram feature amount of a measurement result; and a Mahalanobis distance calculation step of calculating a Mahalanobis distance for each combination of the measurement item, the device, and the unit analysis period based on the histogram feature amount. , a total Mahalanobis distance calculation step of calculating the total Mahalanobis distance based on the total value of the Mahalanobis distance for each combination of the measurement item and the device in the most recent analysis target period; A process capability index calculation step of calculating a process capability index for each combination, and an aggregation step of calculating an aggregated process capability index based on the aggregate value of the process capability index for each combination of the measurement items and the devices in the most recent analysis target period. A capability index calculation step and an anomaly detection display step of dividing and displaying the aggregated Mahalanobis distance and the aggregated process capability index.

The production condition calculation program of the present invention causes a computer to divide actual measurement data related to product manufacturing into unit analysis periods, and for each combination of measurement items and devices included in the measurement data and the unit analysis period, the measurement data and a Mahalanobis distance for calculating the Mahalanobis distance for each combination of the measurement item, the device, and the unit analysis period based on the histogram feature amount. a calculating step, an aggregated Mahalanobis distance calculating step of calculating an aggregated Mahalanobis distance based on an aggregated value of the Mahalanobis distance for each combination of the measurement item and the device in the most recent analysis target period, the measurement item, the device, and the unit A process capability index calculation step of calculating a process capability index for each combination of analysis periods, and calculating an aggregated process capability index based on the aggregate value of the process capability index for each combination of the measurement items and the devices in the most recent analysis target period. and an abnormality detection display step of dividing and displaying the aggregated Mahalanobis distance and the aggregated process capability index.

As described above, according to the manufacturing condition calculation device, the manufacturing condition calculation method, and the manufacturing condition calculation program of the present invention, it is possible to quantitatively determine how to deal with equipment failures.

1 is a block diagram of a manufacturing condition calculation device according to Embodiments 1 and 2 of the present invention; Block diagram of the processing unit of the manufacturing condition calculation device according to the first and second embodiments 4 is a flow chart showing processing of the manufacturing condition calculation device according to Embodiment 1. FIG. 4 shows an example of data registered in the database according to Embodiment 1; A diagram showing an example of simultaneously representing the degree of change and the degree of soundness according to the first embodiment Flowchart showing processing of the manufacturing condition calculation device according to the second embodiment Flowchart showing details of part of the processing of the manufacturing condition calculation device according to the second embodiment FIG. 11 shows an example of data registered in a database according to Embodiment 2; A diagram showing an example of simultaneously expressing the degree of change and the degree of soundness according to the second embodiment Flowchart showing a conventional abnormality cause diagnosis method

(Embodiment 1)
FIG. 1 is a block diagram of a manufacturing condition calculation device 1 according to Embodiment 1 of the present invention. Figure 2
2 is a block diagram of a processing unit 166 of the manufacturing condition calculation device 1. FIG. The manufacturing condition calculation apparatus 1 according to Embodiment 1 can be realized by computer system hardware and a program executed on the computer system. It should be noted that the manufacturing condition calculation device 1 shown here is merely an example, and other configurations are also available.

<Equipment>
As shown in FIG. 1, the manufacturing condition calculation apparatus 1 includes a computer 120 , a monitor 122 , a printer 124 , a keyboard 126 and a mouse 128 all connected to the computer 120 .

Computer 120 includes a DVD-ROM (Digital Versatile Disk Read-Only-Memory) drive 150 and a semiconductor memory port 152 . Computer 120 further includes a bus 142 connected to DVD-ROM drive 150 and semiconductor memory port 152 , a CPU 140 and a ROM 144 that store a boot-up program for computer 120 , all connected to bus 142 . Computer 120 also includes RAM 146 , hard disk drive 148 , and network interface 154 . RAM 146 provides a work area used by CPU 140 and is a storage area for programs executed by CPU 140 . Hard disk drive 148 stores audio data, acoustic models, language models, lexicons, and mapping tables. Network interface 154 provides connection to network 164 .

Referring to FIG. 2, manufacturing condition calculation apparatus 1 includes processing unit 166 configured by CPU 140 executing a program. The processing unit 166 includes an actual measurement data recording unit 168, an analysis period setting unit 170, a histogram feature amount calculation unit 172, a Mahalanobis distance calculation unit 174, an aggregated Mahalanobis distance calculation unit 176, a process capability index calculation unit 178, It includes an aggregate process capability index calculator 180 and an anomaly detection display 182 .

Software for realizing the processing unit 166 of the manufacturing condition calculation apparatus 1 according to the first embodiment is distributed in the form of object code or script recorded on a medium such as the DVD-ROM 162 or the semiconductor memory 160, and is distributed in the form of a DVD-ROM drive. 150 or via a readout device such as semiconductor memory port 152 to computer 120 and stored on hard disk drive 148 . When CPU 140 executes a program, the program is read from hard disk drive 148 and loaded into RAM 146 . An instruction is fetched from an address specified by a program counter (not shown) and executed. CPU 140 also reads data to be processed from hard disk drive 148 and stores the results of the processing also on hard disk drive 148 . A hard copy of the estimated product molecular weight is output by printer 124 .
Since the general operation of computer 120 is well known, detailed description is omitted.

Regarding the software distribution method, the software does not necessarily have to be recorded on a storage medium. For example, software may be distributed from another computer connected to a network. Alternatively, a portion of the software may be stored on hard disk drive 148 and the remaining portion of the software may be retrieved to hard disk drive 148 via a network and integrated during execution.
Also, the distribution form of software is not limited to object code. It may be in the form of a script as described above, or it may be distributed in the form of a source program that is converted into object code by an appropriate compiler installed in the computer 120 .

Typically, modern computers utilize common functions provided by an operating system (OS) to accomplish functions in a controlled manner according to desired objectives. Therefore, even if the program does not include general functions that can be provided by the OS or a third party and only specifies the combination of the execution order of general functions, the program as a whole can achieve the desired purpose. It is obvious that the program is included in the scope of the present invention as long as it has

<Flow>
Next, a flowchart of data analysis processing in the manufacturing condition calculation device 1 will be described with reference to FIG.

<S1>
First, in step S<b>1 , the measured data recording unit 168 of the processing unit 166 adds measured data to a database (hereinafter simply referred to as “DB”) built in the computer 120 . The actual measurement data when registered for each product to be manufactured includes the product type and serial number of the product to be manufactured, and the manufacturing time, manufacturing equipment, manufacturing head, and manufacturing results in each manufacturing process.
The manufacturing result is the manufacturing result in each manufacturing process, such as the width of the product to be manufactured and the length of the product to be manufactured.

<S2>
Next, in step S2, the analysis period setting unit 170 sets the unit analysis period to, for example, 6 hours based on the analysis target/data/evaluation candidate. Here, there is a production lot as one that conforms to the unit analysis period. This lot is a certain quantity, for example, 1,000 pieces or 10,000 pieces. used to Since the quantity of this production lot is set so as to easily reflect the actual situation of each manufacturing site, the lot number that manages the production lot will be explained as an index of the unit analysis period in the drawings in the future explanation.

<S3>
Next, in step S3, the histogram feature amount calculation unit 172 divides the actual measurement data for each unit analysis period, calculates the histogram feature amount of the measurement result for each combination of the device/processing head/lot number/measurement item, and stores the data in the DB. to register.
Specifically, the histogram feature amount is the mean value, median value, standard deviation, mode value, kurtosis, skewness, etc. of the measurement results. Generally good results are often obtained by calculating the feature amount by combining the median, standard deviation, kurtosis, and skewness.

FIG. 4 shows an example of the result of calculating the histogram feature amount for each unit analysis period from the measured data from the process to the average.
In Embodiment 1, the histogram feature amount of the measurement result is calculated for each combination of target process, apparatus, processing head, lot number, and measurement item.

First, a process is a process in which an individual to be manufactured is processed from materials to a finished form. For example, in the manufacturing process of a battery, it includes processes such as kneading, coating, winding, can insertion, welding, liquid injection, sealing, initial inspection, aging, and shipping inspection. In Embodiment 1, only the process 1, for example, only the winding process is partially shown while the manufacturing process including the processes 1 to 10 is assumed.

Next, the equipment shown in FIG. 4 is the name for the machine that processes the product to be manufactured. Depending on the process, some processes have only one device, while others have multiple devices. In the first embodiment, it is assumed that there are four apparatuses in process 1, and partial results of apparatus 1 to apparatus 4 are displayed.

The processing head shown in FIG. 4 is the name of the processing head. In order to increase productivity in an apparatus, a plurality of parts that perform the same processing may be provided. In Embodiment 1, it is assumed that each apparatus in process 1 has two processing heads, and partial results of processing heads 1 to 2 are displayed.

Lot No. shown in FIG. 4 is a name that uniquely identifies a lot. This lot is a certain quantity, for example, 1,000 pieces or 10,000 pieces. A unit used to measure The number of production lots is set so as to easily reflect the actual situation of each manufacturing site. In Embodiment 1, the lot number that manages the manufacturing lot is used in the description as an index according to the unit analysis period.

The measurement items shown in FIG. 4 are items measured for the purpose of guaranteeing the results of processing in the apparatus and items measured for the purpose of recording variations in processing conditions. be. For example, assuming a winding process, measurement items include the height of the wound body, the diameter of the wound body, the amount of deviation in the winding, the winding tension, and the winding correction amount. In Embodiment 1, it is assumed that there are measurement items from A to G, and partial results of A and D are displayed.

Then, after stratifying the actual measurement data for each combination of the process, equipment, processing head, lot number according to the unit analysis period, and measurement items, the histogram feature values are the number of constituents N, the median value of the measurement results, the standard deviation, Kurtosis, skewness, and average are calculated. At least one of the median, standard deviation, kurtosis, skewness, and average may not be included in the histogram feature amount, or calculated values other than these may be calculated as the histogram feature amount.

Here, the skewness is calculated using the following formula (1).

Moreover, the kurtosis is calculated using the following formula (2).

<S4>
Next, in step S4, the Mahalanobis distance calculation unit 174 calculates the Mahalanobis distance MD for each combination of apparatus, processing head, lot number, and measurement item, including the histogram feature values registered in the past by the histogram feature value calculation unit 172. Calculate and register in DB.
Specifically, among the histogram feature quantities, a combination of median, standard deviation, kurtosis, and skewness can be used to calculate the Mahalanobis distance based on the data for the past month, and generally good results can be obtained. There are many things.

FIG. 4 shows an example of the result of calculating the Mahalanobis distance MD.
Here, the Mahalanobis distance MD is calculated using the following equation (3), assuming p dimensions in which there is no correlation between different variables. In Embodiment 1, the Mahalanobis distance MD is calculated with i=1 as the median, i=2 as the standard deviation, i=3 as the kurtosis, and i=4 as the skewness.

<S5>
Next, in step S5, the total Mahalanobis distance calculation unit 176 calculates the Mahalanobis distance MD calculated by the Mahalanobis distance calculation unit 174 by calculating A_MD is calculated and registered in the DB.
Specifically, the most recent analysis target period is set to 3 days, and the average value of the Mahalanobis distance MD for each combination of the device, processing head, and measurement item in this analysis target period is calculated as the total Mahalanobis distance A_MD. results are often obtained.
FIG. 4 shows an example of a result of calculating aggregated Mahalanobis distance A_MD using data of Mahalanobis distance MD. For example, the Mahalanobis distances MD of combinations of the device 1, the machining head 1, and the measurement item A in the lot numbers AAAA001, AAAA004 to AAAA008, etc. during the analysis target period are 0.91, 2.43, 1.71, and 3.72, respectively. , 2.32, 2.14, . The total Mahalanobis distance calculation unit 176 calculates 2.21 of these average values as the total Mahalanobis distance A_MD.

<S6>
Next, in step S6, the process capability index calculation unit 178 divides the measured data for each unit analysis period, calculates the process capability index for each combination of the device/processing head/lot number/measurement item, and registers it in the DB. .
Specifically, generally good results are often obtained by using the process capability index Cpk that takes into account the bias in the average.
FIG. 4 shows an example of the result of calculating the process capability index Cpk using the upper standard value UCL and the lower standard value LCL of the measurement items.
Here, the process capability index Cpk is calculated using the following equation (4), with UCL as the upper standard value and LCL as the lower standard value for each measurement item.

<S7>
Next, in step S7, the aggregated process capability index calculation unit 180 calculates the process capability index Cpk calculated by the process capability index calculation unit 178 for each combination of the device/processing head/measurement item in the latest analysis target period. The total process capability index A_Cpk is totaled and registered in the DB.
Specifically, the most recent analysis target period is set to 3 days, and the average value of the process capability index Cpk for each combination of the device/processing head/measurement item in this analysis target period is calculated as the total process capability index A_Cpk. Good results are generally obtained.
FIG. 4 shows an example of the result of calculating the total process capability index A_Cpk using the data of the process capability index Cpk. For example, the process capability index Cpk of the combination of the device 1, the machining head 1, and the measurement item A in the lot numbers AAAA001, AAAA004 to AAAA008, etc. during the analysis target period is 3.41, 3.72, 3.57, 2.41, 3.72, 3.57, 2.57, and 2.57, respectively. 9, 3, 3.42, . . . The aggregated process capability index calculator 180 calculates 3.34 of these average values as the aggregated process capability index A_Cpk.

<S8>
Next, in step S8, the abnormality detection display unit 182 combines the summed Mahalanobis distance A_MD and the summed process capability index A_Cpk and outputs them to the monitor 122 or the like for confirmation by the operator at the manufacturing site.
FIG. 5 shows an example of the result of graph display with the aggregated Mahalanobis distance A_MD representing the degree of change on the horizontal axis and the aggregated process capability index A_Cpk representing the soundness on the vertical axis.
For example, the lower right area of FIG. 5 indicates that there is a possibility that a sudden abnormality is indicated because the soundness level is low and the recent degree of change is large. Therefore, the operator immediately investigates the current situation and considers whether to implement countermeasures.
On the other hand, since the lower left area has a low degree of health and a small degree of recent change, it indicates that there is a possibility of indicating a chronic abnormality. For this reason, although the operator has a low priority for efforts, it is necessary to continuously consider improvement plans.
Furthermore, the upper right area indicates that the degree of soundness is high, but the degree of recent change is large. For this reason, the operator needs to continuously monitor for sudden abnormalities.
The upper left region has a high degree of soundness and a small degree of recent change, so the priority of efforts is extremely low.

In the process of going through this flow, even in general processing equipment where the parameters that determine the manufacturing process are not comprehensively measured, while evaluating changes in the measurement items as a statistical model, Relevant data directly linked to deterioration can be displayed. By checking the display results from the above perspective, the operator will be able to detect changes in material properties that are difficult to measure and record, as well as abnormalities that are unlikely to be affected by previous processes. Decisions can be made quantitatively.
It should be noted that the present invention is not limited to the above-described embodiment, and can be implemented in various other modes.

A second embodiment of the present invention will be described. The differences between the first embodiment and the second embodiment are the processing contents of the total Mahalanobis distance calculation unit 176 and the abnormality detection display unit 182 .

<Flow>
A flowchart of data analysis processing in the manufacturing condition calculation device 1 will be described with reference to FIGS. 6 and 7. FIG. Note that the flow of the second embodiment is the same as that of the first embodiment up to step S7. Here, in the second embodiment, the process of step S5 may not be performed.

<S11>
After the process of step S7 is performed, as shown in FIG. 6, in step S11, the aggregated Mahalanobis distance calculator 176 calculates the device/processing head/lot based on the process capability index Cpk and the aggregated process capability index A_Cpk. No. The corrected Mahalanobis distance MD' is calculated for each combination of measurement items. The process of step S11 includes the processes of steps S21 to S23, as shown in FIG.
<S21>
In step S21, the aggregated Mahalanobis distance calculator 176 determines whether the process capability index Cpk exceeds the aggregated process capability index A_Cpk.
<S22>
When the aggregated Mahalanobis distance calculation unit 176 determines that the process capability index Cpk exceeds the aggregated process capability index A_Cpk, in step S22, the value obtained by multiplying the Mahalanobis distance MD by "-1" is calculated as the corrected Mahalanobis distance MD'. .
<S23>
When the aggregated Mahalanobis distance calculation unit 176 determines that the process capability index Cpk does not exceed the aggregated process capability index A_Cpk, in step S23, the value obtained by multiplying the Mahalanobis distance MD by "1" is calculated as the corrected Mahalanobis distance MD'. .

<S12>
Next, after the process of step S11 is performed, in step S12, the total Mahalanobis distance calculation unit 176 registers the corrected Mahalanobis distance MD' in the DB.
FIG. 8 shows an example of the result of calculating the corrected Mahalanobis distance MD' using the data of the Mahalanobis distance MD.

<S13>
Next, in step S13, the aggregated Mahalanobis distance calculation unit 176 calculates the aggregated corrected Mahalanobis distance A_MD' for each combination of the device/processing head/measurement item in the latest analysis target period with respect to the corrected Mahalanobis distance MD', Register in DB.
Specifically, the most recent analysis target period is set to 3 days, and the average value of the corrected Mahalanobis distance MD' for each combination of the device/processing head/measurement item in this analysis target period is calculated as the total corrected Mahalanobis distance A_MD'. and generally good results are obtained.

FIG. 8 shows an example of the result of calculating the aggregate modified Mahalanobis distance A_MD' using the data of the modified Mahalanobis distance MD'. For example, the corrected Mahalanobis distances MD′ of combinations of the device 1, the machining head 1, and the measurement item A in the lot numbers AAAA001, AAAA004 to AAAA008, etc. during the analysis target period are -0.9, -2.4, -1. 7, 3.72, 2.32. −2.1, . . . The total Mahalanobis distance calculation unit 176 calculates −0.2 of these average values as the total modified Mahalanobis distance A_MD′.

<S14>
Next, in step S14, the abnormality detection display unit 182 combines the total modified Mahalanobis distance A_MD' and the total process capability index A_Cpk and outputs them to the monitor 122 or the like for confirmation by the operator at the manufacturing site.
FIG. 9 shows an example of the results of graph display with the horizontal axis representing the aggregate modified Mahalanobis distance A_MD′ representing the degree of change and the vertical axis representing the aggregate process capability index A_Cpk representing the degree of soundness.
For example, the lower right area of FIG. 9 indicates that the degree of soundness is low and the latest degree of change is large in the positive direction, indicating that there is a possibility of indicating a sudden worsening abnormality. Therefore, the operator needs to investigate the current situation as soon as possible and take countermeasures.
On the other hand, since the lower center area has a low degree of health and a small degree of recent change, it indicates that there is a possibility of indicating a chronic abnormality. For this reason, although the operator has a low priority for efforts, it is necessary to continuously consider improvement plans.
Furthermore, since the lower left area has a low degree of soundness and the degree of recent change is large in the negative direction, there is a possibility that it indicates a sudden improvement abnormality. For this reason, although the operator has a low priority to work on, it is necessary to confirm what kind of changes have been made and stock them as future improvement plans.
Moreover, since the upper region generally has a high degree of soundness, the operator continuously monitors whether the degree of soundness deteriorates.

By confirming the display result from such a viewpoint, the operator can quantitatively determine how to deal with the apparatus abnormality.
It should be noted that the present invention is not limited to the above-described embodiment, and can be implemented in various other modes.

The manufacturing condition calculation apparatus, manufacturing condition calculation method, and manufacturing condition calculation program according to the present invention are difficult to measure and record even in general processing equipment and lines where parameters that determine the manufacturing process are not exhaustively measured. It will be possible to quantitatively evaluate the degree of change due to changes in material properties and equipment fluctuations, making it possible to quantitatively determine changes in manufacturing conditions, and to accelerate the use of conventionally acquired parameters in general factories. It is useful for things like

1 manufacturing condition calculation device 120 computer 122 monitor 124 printer 126 keyboard 128 mouse 140 CPU
142 bus 144 ROM
146 RAMs
148 hard disk drive 150 DVD-ROM drive 152 semiconductor memory port 154 network interface 160 semiconductor memory 162 DVD-ROM
164 network 166 processing unit 168 measured data recording unit 170 analysis period setting unit 172 histogram feature amount calculation unit 174 Mahalanobis distance calculation unit 176 total Mahalanobis distance calculation unit 178 process capability index calculation unit 180 aggregate process capability index calculation unit 182 abnormality detection display unit

Claims (7)

Actual measurement data related to product manufacturing is divided into unit analysis periods, and a histogram feature amount of measurement results included in the actual measurement data is calculated for each combination of measurement items and devices included in the actual measurement data and the unit analysis period. a histogram feature quantity calculator;
a Mahalanobis distance calculation unit that calculates a Mahalanobis distance for each combination of the measurement item, the device, and the unit analysis period based on the histogram feature amount;
an aggregated Mahalanobis distance calculation unit that calculates an aggregated Mahalanobis distance based on an aggregated value of the Mahalanobis distances for each combination of the measurement items and the devices in the most recent analysis target period;
a process capability index calculator that calculates a process capability index for each combination of the measurement item, the device, and the unit analysis period;
an aggregated process capability index calculation unit that calculates an aggregated process capability index based on aggregated values of the process capability index for each combination of the measurement items and the devices in the most recent analysis target period;
A manufacturing condition calculation device, comprising: an anomaly detection display section for displaying the summed Mahalanobis distance and the summed process capability index in a divided manner. 2. The manufacturing condition calculation device according to claim 1, wherein said histogram feature amount calculator calculates at least one of median, standard deviation, kurtosis and skewness of said measurement result as said histogram feature amount. . 3. The manufacturing condition calculation device according to claim 1, wherein said aggregated Mahalanobis distance calculation unit calculates an average value of said Mahalanobis distances for each of said combinations as said aggregated Mahalanobis distance. 4. The manufacturing condition calculation apparatus according to claim 1, wherein said aggregated process capability index calculator calculates an average value of said process capability indexes for each said combination as said aggregated process capability index. The aggregated Mahalanobis distance calculation unit compares the process capability index and the aggregated process capability index for each combination of the measurement item, the device, and the unit analysis period, and if the process capability index exceeds the aggregated process capability index , assigning a negative sign to the Mahalanobis distance corresponding to the process capability index, and assigning a positive sign to the Mahalanobis distance corresponding to the process capability index when the process capability index does not exceed the aggregated process capability index 5. The manufacturing condition calculation device according to claim 1, wherein the total Mahalanobis distance is calculated based on the Mahalanobis distances to which positive and negative signs are assigned. Actual measurement data related to product manufacturing is divided into unit analysis periods, and a histogram feature amount of measurement results included in the actual measurement data is calculated for each combination of measurement items and devices included in the actual measurement data and the unit analysis period. A histogram feature quantity calculation step;
A Mahalanobis distance calculation step of calculating a Mahalanobis distance for each combination of the measurement item, the device, and the unit analysis period based on the histogram feature amount;
an aggregated Mahalanobis distance calculation step of calculating an aggregated Mahalanobis distance based on an aggregated value of the Mahalanobis distances for each combination of the measurement items and the devices in the most recent analysis target period;
a process capability index calculation step of calculating a process capability index for each combination of the measurement item, the device, and the unit analysis period;
an aggregated process capability index calculation step of calculating an aggregated process capability index based on aggregated values of the process capability index for each combination of the measurement item and the apparatus in the most recent analysis target period;
and an abnormality detection display step of dividing and displaying the aggregated Mahalanobis distance and the aggregated process capability index. to the computer,
Actual measurement data related to product manufacturing is divided into unit analysis periods, and a histogram feature amount of measurement results included in the actual measurement data is calculated for each combination of measurement items and devices included in the actual measurement data and the unit analysis period. A histogram feature quantity calculation step;
A Mahalanobis distance calculation step of calculating a Mahalanobis distance for each combination of the measurement item, the device, and the unit analysis period based on the histogram feature amount;
an aggregated Mahalanobis distance calculation step of calculating an aggregated Mahalanobis distance based on an aggregated value of the Mahalanobis distances for each combination of the measurement items and the devices in the most recent analysis target period;
a process capability index calculation step of calculating a process capability index for each combination of the measurement item, the device, and the unit analysis period;
an aggregated process capability index calculation step of calculating an aggregated process capability index based on aggregated values of the process capability index for each combination of the measurement item and the apparatus in the most recent analysis target period;
and an abnormality detection display step of dividing and displaying the aggregated Mahalanobis distance and the aggregated process capability index.

Images