JPH04143859A - Casting schedule planning system - Google Patents

Abstract

PURPOSE:To enable a person in charge to shorten the time needed for production of a casting schedule plan table and also to decrease the production mistakes by choosing the proper one of several patterns abased on the order pattern knowledge of the melting material and deciding the order of the melting materials and the melting amount of each material. CONSTITUTION:The requested frame number of a casting date is calculated for each parts name and the load time of each core type is obtained for each parts name to change the requested frame number. Then a combination of division patterns is decided based on the division pattern combination knowledge 4b and the frame number is obtained. The proper one of several patterns is chosen based on the order patter knowledge 4c of the melting material, and the order of the melting materials is decided together with the melting amount of each material. Then the parts name order of materials and its requested number are changed based on the material order pattern knowledge 4d. As a result, a casting schedule plan table is obtained in a short time together with reduction of production mistakes.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a casting schedule planning system using an expert system.

[Conventional technology]

In recent years, production planning using expert systems has been actively researched and developed, and its practical application is being promoted.

The creation of a casting schedule at a foundry, that is, the determination of the casting product name for each casting line, the number of casting frames, and the casting order, is partially done by computer, but it still relies on the hands of experienced production staff. I owe a lot to my work.

To date, the following two methods have been proposed for automating production planning, that is, as automatic scheduling methods.

1] Operations research methods such as linear programming or a method of performing simulation using a mathematical model applying queuing theory are available. This method is effective when the number of products to be handled and the number of constraints are small and there are few changes to them.

2] Scheduling method using an expert system This method has been actively researched and developed in recent years as a planning type expert system, and some of it has been put into practical use. Most of the scheduling by this expert system can be classified into the following two types.

■ If a wide variety of products are flowing through many processes, for example, in what order should the products be passed through each process in order to meet the customer's delivery deadline?

■ If there are multiple processing facilities in the same process, how should we allocate the various products to which processing facilities and in what order to improve the utilization rate of the facilities? (JP-A-Sho) However, when creating a casting schedule plan in the casting process, it is necessary to use production technology know-how, such as the melting order of each material in consideration of energy conservation,
When using split models (hereinafter referred to as "split patterns"), it is necessary to consider the compatibility of their combinations, changing the number of frames depending on the load time for each core model, determining the casting order for each product name within the same material, etc. This is different from the above classification. Since these production technology know-how change with changes in the surrounding environment, it has been difficult to computerize them in conventional systems. Said Japanese Unexamined Patent Publication No. 64-716
Publication No. 52 (hereinafter referred to as a comparison system) discloses scheduling using an expert system, and when the current casting schedule planning system is incorporated into this comparison system, the result will be as follows. First is the combination compatibility that uses split patterns, but in the comparison system, the combination partner is included in the master. However, in this system, only what can be called the compatibility of the opponent is registered, so the number of combinations can be much larger, and this plan will not work unless you can create many combinations. Next, in the comparison system, the plan is compared with a pre-registered pattern and a schedule is created, whereas in this system, although there is a basic pattern, the pattern is generally close to fixed. Since it is a product, it can be registered, but there are an infinite number of order patterns of product names under that pattern, so it is almost impossible to register. For this reason, the pattern itself is not registered in the product name order of this system. By registering what can be called pattern trends in the form of rules,
We are ready to handle whatever comes our way.

In this way, the current system cannot be used as a conventional system.

[Problem that the invention seeks to solve]

Currently, plans are created manually, which causes the following problems.

1) The amount of information handled is increasing due to the increasing number of products being produced in small quantities. As a result, it takes a long time for the person in charge to create the casting schedule, and the possibility of making mistakes is also increasing.

2) If the current expert system performs the combinations of division patterns required when creating a casting schedule,
The maintenance of master data becomes complicated and the system may not be practical.

This invention was made to solve the above-mentioned problems. It is possible to create a casting schedule table in a short time, reduces the possibility of making mistakes, and makes it possible to create a casting schedule table that can withstand practical systems. The purpose is to obtain a planning system.

Means for Solving Problems c] The present invention is a casting schedule planning system for a casting line that adopts a multi-material and disassembly pattern method, and includes a system for calculating the required number of casting dies for each product name. The first method, the second method of determining the load time for each product name and core machine type and changing the required number of slots, and determining the combination of division patterns based on the knowledge of division pattern combinations, The third method is to calculate the number of frames, and the fourth method is to select an appropriate one from among several patterns based on the knowledge of the order pattern of melted materials and determine the order of melted materials and the amount of melted material for each material. and a fifth means for changing the order of product names within the same material and the requested number thereof based on knowledge of the order pattern within the same material.

〔Example〕

Embodiments of the present invention will be described in detail below with reference to the drawings.

Fig. 1 is a system configuration diagram of the present invention, Fig. 3 is an example of a product master, Fig. 3 is a core master, Fig. 4 is an example of division pattern combination knowledge, Fig. 5 is an example of dissolution order knowledge, FIG. 6 shows all processing means in the workstation, FIG. 7 shows a division pattern combination determination procedure, and FIG. 8 shows an example of a casting schedule table.

1 of the system configuration diagram of the present invention shown in FIG. 1, a host computer receives a request from a customer and determines the number of casting requests for each product name on a daily basis. 2 is a workstation;
A casting schedule is created based on the data input from the host.

In the workstation 2, 3 is a database section, 4 is a knowledge data section, and 5 is an inference engine. The database unit 3 stores a product master 3a, a core master 3b, casting request data 3C1 and casting schedule planning data 3d. As shown in Figure 2, the product master 3a contains product information such as product injection stacking and number of core inserts. 3b
The core master is as shown in Figure 3, including the number of points used, ST, etc.
Contains information about the core used when casting a product. 3
The casting request data of C stores the number of daily casting requests from the host and the number of casting frames of the casting mill. The casting schedule data 3d is where the data of the casting schedule table output from the workstation is temporarily stored. The knowledge data section 4 includes knowledge 4a about changing the number of frames due to overloading of cores, knowledge 4b about combinations of division patterns, melting order knowledge 4C1, and casting order knowledge 4d.

Knowledge 4a of changing the number of slots due to core load time exceeding If it is over, change the number of slots according to your knowledge.

The knowledge 4b of the division pattern combination constitutes the third means of the present invention, and is obtained by using a model of the division pattern method.
When casting, it indicates the compatibility of two product names when combined into a single frame. What is the compatibility here?
Conditions for combining division patterns, such as whether the number of frames is the same or the material is the same. For example, A in Figure 4.
OI and Ac1 have reason code l, so they are not compatible.
etc.

Melting order knowledge 4C constitutes the fourth means of the present invention. Since products of different materials are cast in one day, it is necessary to consider energy saving of the melting furnace, ease of component adjustment, etc. Several types of effect casting patterns have been developed that take these into account.

Currently, the pattern is represented by four patterns as shown in Figure 5, and after that, by inputting the information of the day, we select which of the patterns is the best.At this time, only one pattern can be used. However, as the number of required frames changes, the amount of dissolved material decreases, and F of pattern 1 in Fig. 5 is determined.
, may be reduced from three times to two times.

The casting order knowledge 4d constitutes the fifth means of the present invention, and when there are multiple product names to be cast in the same material, the knowledge for determining the order and the separation for each material are provided for each melting furnace. This is the knowledge needed to adjust the number of slots in order to match the number of slots.

The inference engine performs inference using the knowledge unit 4 based on the casting request data 3C.

Next, the overall processing procedure of the present invention will be explained using FIG. 6.

1) The number of casting requests by day and product name is obtained from the host computer and stored in the casting request data 3C in accordance with the customer's requirements of the basic data of the casting schedule plan.

2) Find the number of casting frames for each product name based on the host data, and find the melting amount for each material based on that number and the unit weight of the frame of the product master 3a.

3) Find the core load man-hours for each core molding machine based on the above data and the core master 3b, and if it is larger than the predetermined value, use the keyboard and CRT that serve as the man-machine interface 6. The number of casting frames is corrected manually or by using the knowledge 4a for changing the number of casting frames due to overloading of the core.

The number of casting frames for each product name is determined by the above operations, and based on this, division patterns are combined using the division pattern combination knowledge 4b, and the number of frames is also determined. This method is illustrated in the flowchart of FIG.

Based on the above data and daily information, one pattern is selected from among the patterns shown in FIG. 5 using the dissolution order knowledge m4c, and the dissolution amount for each material is determined.

Once the pattern and melting amount for each material have been determined, use the casting order knowledge 4d to adjust the order of product names and the number of frames.
A casting schedule table is created and stored in the casting schedule data 3d.

7) Output the final stored casting schedule data 3d to the printer and create a table as shown in FIG.

In addition, at various locations, the person in charge can change or add data in an interactive manner from the man-machine interface 6 in the system configuration diagram of the present invention shown in FIG. Furthermore, the pattern shown in FIG. 5 can also be changed using the keyboard of the man-machine interface 6, depending on the material.

By using the above system, the working time was reduced to 173 to 115 hours compared to conventional manual work.

〔Effect of the invention〕 Since the present invention has the above configuration, it has the following effects.

1) By automating the creation of casting schedule plans, the work time for the person in charge was reduced to 173 to 115 hours compared to the previous manual work. 2) Because an expert system was used, the work was previously a specialized job. Anyone can now do things.

3) Because the expert system is used, knowledge can be modified more easily than with conventional systems, and production technology know-how can be preserved in an easy-to-see format.

[Brief explanation of the drawing]

Figure 1 is a system configuration diagram of the present invention, Figure 2 is an example of a product master, Figure 3 is a core master, Figure 4 is an example of division pattern combination knowledge, Figure 5 is an example of melting order knowledge, 6
The figure shows the entire processing procedure in the workstation, FIG. 7 shows the division pattern combination determination procedure, and FIG. 8 shows an example of the casting schedule table. 1: host computer, 2: workstation, 3
: Database part, 4: Knowledge data part, 5: Inference engine, 6: Man-machine interface, 7: Printer,
3a: Product master, 3b: Core master, 3C: Casting request data, 3d: Casting schedule planning data, 4a: Knowledge of changing the number of frames due to core loading time exceeding, 4b Knowledge of two-part pattern combination, 4C: Melting order Knowledge, 4d: Casting order knowledge. Figure Figure Figure Figure 8

Claims (1)

[Claims] A casting schedule planning system for a casting line that employs multiple materials and a split pattern method, comprising: a first means for calculating a required number of slots on a casting date for each product name; , a second means for determining the load time for each core machine type and changing the required number of slots, and a third means for determining the combination of division patterns and determining the number of slots based on division pattern combination knowledge. A fourth method is to select an appropriate one from among several patterns based on recognition of the order pattern of dissolved materials and solve the order of dissolved materials and the amount of dissolution for each material, and the order within the same material. A casting schedule planning system comprising: a fifth means for changing the order of product names within the same material and the requested number thereof based on pattern recognition.