JP2004355095A - Production management method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a production management method for accurately and quantitatively analyzing a bottle neck which is not affected by the forgotten input etc. <P>SOLUTION: This production management method for executing the unitary management of production lines configured of a plurality of processors, and for completing a product is provided to calculate a difference between a time when the processing of a certain product is ended by a first processor and a time when the processing of the product is successively started by a second processor, that is, the standby time of each device of a certain product in order to find out a device being bottle neck. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a production management method capable of finding a bottleneck device (process) in a production line represented by a semiconductor production line.
[0002]
[Prior art]
Generally, in a semiconductor manufacturing line, a product is sequentially processed in a plurality of processes, and after all processes are shipped as a finished product. The work-in-progress status of each product, the process flow of each product, the operation status of each device, and the like are centrally managed by a production management system.
In such a production line, shortening the production time is an extremely effective method that leads to a reduction in in-process inventory, an increase in the amount of shipment, and the like, leading to an improvement in productivity.
Therefore, it has been attempted to find a place that hinders the progress of manufacturing (hereinafter referred to as a bottleneck) and improve the processing capacity of that part.
[0003]
A conventional method for searching for such a bottleneck will be described with reference to FIGS. FIG. 12 shows a state definition of, for example, a CVD apparatus by a conventional bottleneck analysis method.
By "in the product process" in the figure shows a state of depositing a product (wafer) on the insulating film (S i _O2), or fix the, for example fault "under maintenance" operations such as to ensure the gas flow rate Therefore, the product cannot be processed. Further, "waiting" is a state other than the above, that is, a state in which nothing is performed. In particular, an apparatus that has a shorter “standby” time has less room and is determined to be a bottleneck, so attention is paid to this standby time. Note that such a state definition is performed for each device.
[0004]
FIG. 13 shows an example of an apparatus state transition when the state definition as shown in FIG. 12 is performed for a CMP apparatus. The time advances from the left to the right of the line, and the leftmost of the line is 0:00, the rightmost is 24:00, and one line indicates the state of the apparatus for one day.
To create an apparatus state transition diagram as shown in FIG. 13, an apparatus state management table as shown in FIG. 14 is required on the production management system. In FIG. 14, the start date and time and the end date and time indicate the date and time when the processing of the product starts and ends. The state time is a value obtained by subtracting the start date and time from the end date and time.
[0005]
Since the production management system has a start / end event when processing a product, the event is used to create data of “product in process”.
The same applies to “during maintenance”. Further, “standby” refers to a time in a state that is neither “under product processing” nor “under maintenance”, and is found and created from the apparatus state management table. FIG. 14 shows data of a CMP apparatus, and data of other apparatuses are created in the same manner.
FIG. 15 is a graph in which “waiting” time is accumulated for one week for each device from the data of each device as shown in FIG. 14 and arranged in ascending order from the left. From this graph, the largest bottleneck is a stepper. I conclude that there is.
[0006]
[Problems to be solved by the invention]
In the conventional bottleneck analysis method described above, the bottleneck is searched for by focusing on the standby time of the manufacturing / inspection apparatus for processing the product. Could not be found as a neck.
In addition, although maintenance-related occurrence events are collected by the production management system, input to the production management system is manually performed by the line operator, so that input is forgotten or omitted (especially short-time maintenance). Even if a bottleneck is analyzed in such a celebration, data collection will be inaccurate, resulting in erroneous improvement activities for the wrong equipment, and accurate reduction of manufacturing time. There was a problem that it was not possible.
[0007]
An object of the present invention is to provide an accurate and quantitative bottleneck analysis method that is not affected by forgetting to input or the like by adopting a new bottleneck analysis method that focuses on product transport time.
[0008]
[Means for Solving the Problems]
The present invention relates to a production management method for centrally managing a production line composed of a plurality of processing devices to complete a product. By calculating the difference from the time at which processing is started by the apparatus, that is, the waiting time of each product for each apparatus, the apparatus that is the bottleneck is found.
[0009]
Further, the present invention provides a production management method for centrally managing a production line composed of a plurality of processing devices to complete a product, wherein a time at which the processing of a certain product is completed by the first processing device, By calculating the difference from the time when the processing is completed by the processing device, that is, the transit time of each product for each device, the device that is the bottleneck is found.
[0010]
BEST MODE FOR CARRYING OUT THE INVENTION
Embodiment 1 FIG.
FIG. 1 is a schematic diagram showing an example of a semiconductor manufacturing line to which the first embodiment of the present invention is applied. In the figure, a server 1 of a production management system is usually installed outside a line, and a line is connected via a communication cable 2. It is connected to a plurality of client PCs 3 installed inside. The client PC 3 is a PC for providing management information to line operators and engineers, and a part of the PC 3 may be installed in an office or the like other than the line. On the other hand, in the line, there are a plurality of manufacturing apparatuses and inspection apparatuses 4, products A, B,..., An inter-bay transport system 5, an in-bay transport vehicle 6, and the like, and a line operator 7.
[0011]
The client PC 3 and the manufacturing apparatus and the inspection apparatus 4 are also connected by the communication cable 2 and exchange various kinds of information. As a typical apparatus for manufacturing the above semiconductor, a stepper 10, a CMP apparatus 11, a CVD apparatus 12, and an etcher 13 are exemplified here.
The product is transported between these devices based on the instruction information, and is completed through a plurality of processing steps. For example, the product A is processed in the order of the stepper 10 → the etcher 13 → the CVD apparatus 12, and the product B is processed in the order of the CVD apparatus 12 → the CMP apparatus 11. Needless to say, these orders only indicate a part of the entire manufacturing process.
[0012]
The method of transporting products is roughly classified into automatic and manual. A portion in which the devices are grouped together is referred to as a bay, and a bay-to-bay transport system 5 that transports between the bays is automated by mechanization. The transport vehicle 6 transports products to individual devices in the bay, and is wirelessly controlled by the interbay transport system 5. The parts that are not automated are manually conveyed by the line operator 7.
In this way, the production management system manages the status of the work in progress of each product, the process flow of each product, the operation status of each device, and the like.
[0013]
Next, FIG. 2 shows a definition for time analysis by the bottleneck analysis method of the present invention. The present invention is characterized in that the time analysis of the product itself is performed instead of the conventional time analysis of the device itself. The horizontal axis in FIG. 2 shows the flow of time, and shows the flow of the product from the time when the processing of the stepper 10 is completed (10:00 on February 1) with the product A as an example.
[0014]
After the processing by the stepper 10 is completed, the product A is transported to the etcher 13, and after arriving at the etcher 13, prepares and stands by until processing by the etcher 13 starts. The processing in the etcher 13 lasts for one hour from the processing start (11:00) to the processing end (12:00). Further, the product A is transported to the CVD device 12, and after arriving at the CVD device 12, prepares or waits until the processing in the CVD device 12 starts. In the above process, the waiting time in the etcher 13 is defined as a time (1 h in this case) from the time when the processing of the stepper 10 ends (10:00) to the start of the processing in the etcher 13 (11:00), and similarly. The waiting time in the CVD apparatus 12 is defined as the time (3h in this case) from the time when the processing in the etcher 13 ends (12:00) to the start of the processing in the CVD apparatus 12 (15:00).
[0015]
FIG. 3 is a data management table for managing the waiting time. In the figure, the process number defines the order in which the products are processed, and starts from 1 and has a width of about 500 to 1000 depending on the type. . The processing start and processing end relate to the processing of the product, and the waiting time is automatically calculated by subtracting the processing end time of the previous process from the processing start time.
According to a concrete calculation using FIG. 3, the 120th waiting time 1h for the product A is a value obtained by subtracting 2/1 10:00 from 2/1 11:00.
[0016]
FIG. 4 is a table summarizing the waiting time for each device. FIG. 5 is a graph showing the totaling result of the bottleneck analysis method according to the first embodiment. The data of FIG. 4 is obtained by accumulating the above waiting time for each device for one week, arranged in descending order from the left. is there. From this graph, it can be concluded that the largest bottleneck is etcher P.
[0017]
Next, FIG. 6 is a flowchart illustrating a method of storing data in the data management table according to the first embodiment, and describes how the production management system operates from FIG. 3 to FIG.
First, it is assumed that a certain product starts processing in a certain device (step 100). At that time, the processing start date and time, the product number, and the like are also recorded in the data management table (step 101). When the processing of the product is completed (step 102), data such as the processing end date and time is recorded (step 103). If the product has the next process, it is transported there (step 104), otherwise the product is completed (step 105). This is repeated for all the manufacturing apparatuses and inspection apparatuses that process the product.
[0018]
FIG. 7 is a flowchart showing a method for obtaining the cumulative value of the waiting time. First, a period to be accumulated is specified (step 200). For example, one week from February 1, 2003.
Next, by designating a certain device, matching data is extracted from the data management table of FIG. 3 (step 201) and recorded in the data totaling table of FIG. 4 (step 202). Then, an accumulated value is obtained (step 203) and recorded in a totaling result as shown in the graph of FIG. 5 (step 204). This is repeated for the next device, and is executed until there is no more device to be targeted.
[0019]
In this way, it is possible to omit the product transfer time and the inaccuracy of the maintenance time (the present invention is included in the waiting time, so that there is no influence even if there is an input error, etc.), which has occurred during the conventional bottleneck analysis. It will be possible, and you will be able to find the bottleneck accurately.
Incidentally, as a method of remedying the bottleneck found as described above, a review of the product transport route (for example, searching for a shorter route and switching to it), increasing the number of the devices, and maintenance of the devices It is conceivable to reduce the time, shorten the product processing time of the device, or the like.
[0020]
Embodiment 2 FIG.
Embodiment 2 is an improvement of Embodiment 1.
In the first embodiment, the method of finding a bottleneck using the accumulated value of the waiting time has been described. In this case, if the waiting time of a specific product is extremely long, the accumulated value also becomes large, and several unique data items If so, it will determine the bottleneck. In this embodiment, an average value is used together to prevent such a situation. Therefore, by using the average value of the waiting time in place of the accumulated value of the waiting time, it is possible to more accurately grasp the bottleneck.
[0021]
Embodiment 3 FIG.
The third embodiment is a further improvement of the first embodiment. FIG. 8 shows a definition for time analysis in the bottleneck analysis method according to the third embodiment. In the third embodiment, the time obtained by adding the processing time of the product to the waiting time in the first embodiment, that is, the accumulated value of the passing time of the product to the processing device is obtained.
FIG. 9 is a graph showing a total result of the bottleneck analysis method according to the third embodiment, and shows an analysis result obtained by performing the same calculation as in FIG. In this example, the largest bottleneck in the graph can be determined to be the etcher.
[0022]
This method enables more accurate analysis than that of the first embodiment, but shortening the processing time of the product may impair the quality of the product and involves technically difficult problems. This analysis is desirably performed after the bottleneck improvement countermeasures according to the first embodiment are completed.
[0023]
Embodiment 4 FIG.
Embodiment 4 is an improvement of Embodiment 3.
In the third embodiment, the method of finding a bottleneck using the cumulative value of the transit time has been described. In this case, however, as in the second embodiment, when the transit time of a specific product is extremely long, the cumulative value increases. If there are several pieces of unique data, a bottleneck will be determined accordingly. In this embodiment, an average value is used together to prevent such a situation. Therefore, by using the average value of the transit time in place of the accumulated value of the transit time, it is possible to more accurately grasp the bottleneck.
[0024]
Embodiment 5 FIG.
An embodiment for performing storage monitoring will be described.
FIG. 10 shows a data management table for managing a waiting time when performing storage monitoring. In the manufacturing process, there is a storage monitoring part as shown in the rightmost column in the figure. This storage monitoring part indicates that there is a limit on the waiting time from the end of the process with the start mark to the start of the process with the end mark. For example, the storage time limit from the pre-diffusion processing to the diffusion processing is generally 3 hours. This is because if left unprocessed for 3 hours or more, the product will deteriorate.
[0025]
In a process in which such storage monitoring is performed, the waiting time of a product may be longer than usual depending on the state of a device in a subsequent process. FIG. 11 shows a data totaling table in consideration of storage monitoring, in which a waiting time is corrected. For example, the waiting time in the row of the processing start date and time of 2/1 11:00 in the figure should be 1h (see FIG. 4). However, since storage monitoring is performed, 1 h is added to the waiting time when the product A is processed by the CVD apparatus in the subsequent process, and is set to 4 h. By changing the waiting time in this way, the waiting time of the CVD apparatus is increased, and the accuracy of the bottleneck is further improved.
[0026]
【The invention's effect】
As described above, according to the production management method of the present invention, by adopting a new analysis method using the waiting time of a product for each device, an accurate and quantitative analysis of a bottleneck which is not affected by forgetting to input etc. It has the effect that can be done.
[Brief description of the drawings]
FIG. 1 is a schematic diagram showing an example of a semiconductor manufacturing line to which the present invention is applied.
FIG. 2 is a definition diagram for time analysis by a bottleneck analysis method according to Embodiment 1 of the present invention.
FIG. 3 is a data management table for managing a waiting time.
FIG. 4 is a table summarizing waiting time for each device.
FIG. 5 is a graph showing a total result of the bottleneck analysis method in the first embodiment.
FIG. 6 is a flowchart showing a method of storing data in a data management table according to the first embodiment.
FIG. 7 is a flowchart illustrating a method of calculating a cumulative value of a waiting time.
FIG. 8 is a definition diagram for time analysis by a bottleneck analysis method according to the third embodiment.
FIG. 9 is a graph showing a total result of the bottleneck analysis method according to the third embodiment.
FIG. 10 is a data management table for managing a waiting time.
FIG. 11 is a data aggregation table in consideration of storage monitoring.
FIG. 12 is a diagram in which a state of, for example, a CVD apparatus is defined by a conventional bottleneck analysis method.
FIG. 13 is a diagram showing an example of a device state transition when the definition in FIG. 12 is used.
FIG. 14 is a diagram showing an apparatus state management table in a conventional bottleneck analysis method.
FIG. 15 is a graph showing a result of aggregation by a conventional bottleneck analysis method.
[Explanation of symbols]
1 Production management system, 2 communication cable, 3 client PC, 4 manufacturing equipment or inspection equipment, 5 bay transport system, 6 transport vehicle, 7 line operator.

Claims (7)

In a production management method for centrally managing a production line composed of a plurality of processing devices to complete a product, a time when a certain product finishes processing by a first processing device, and a process in which the product is continuously processed by a second processing device. A production management method characterized by finding a bottleneck device by calculating a difference from a start time, that is, a waiting time of each product for each device. 2. The production management method according to claim 1, wherein a waiting time for each product is calculated for each device, and the waiting time is calculated for each device to find a bottleneck device. 2. The production management method according to claim 1, wherein a waiting time for each product is calculated for each device, and the waiting time is calculated for each device to find a bottleneck device. In a production management method for centrally managing a production line composed of a plurality of processing devices to complete a product, a time when a certain product finishes processing by a first processing device, and a process in which the product is continuously processed by a second processing device. A production management method characterized by finding a bottleneck device by calculating a difference from an end time, that is, a transit time of a certain product for each device. 5. The production management method according to claim 4, wherein a waiting time of each product is calculated for each device, and the waiting time is calculated for each device to find a bottleneck device. 5. The production management method according to claim 4, wherein a waiting time for each product is calculated for each device, and the waiting time is calculated for each device to find a bottleneck device. 5. The production management method according to claim 1, wherein a waiting time is transferred when storage monitoring is performed, and a waiting time of each device of a certain product is calculated.

Images