JPS63209917A - Injection molding support expert method - Google Patents

Abstract

PURPOSE:To contrive to optimize the injection molding conditions by a method wherein a specified system to infer data on cause of failure from the degree, data and the like of bad phenomena is incorporated in an injection molder. CONSTITUTION:Bad phenomena (short shot, flash and the like) of a molded item are inputted from a key board 20. An inference system consisting of a knowledge base 25 and an inference section 26 infers the degree of influence of the cause of failure to the bad phenomena from the inputted bad phenomena. The degree of influence of the result of diagnosis is shown on a CRT display 28 through an input-output section 21 on the basis of the degree of certainty to the result of diagnosis obtained through diagnosis rules and the degree of bad phenemena. Through the display 28, an operator can surely and easily infer the optimum injection molding conditions on interactive mode.

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention] (Industrial Application Field) The present invention relates to an injection molding support expert system for optimizing molding conditions of an injection molding machine.

(Prior Art) There are various types of injection molding machines, but this injection molding machine has a screw set injection molding machine shown in FIG. 6 as an example. A cylinder 4 is provided with an insertion opening 3 for a screw 2 formed at its end. The screw 2 is rotated by a hydraulic motor 5 and is moved inside the cylinder 4 by an injection cylinder 6 in the forward arrow (a) direction and the backward arrow (b) direction. Further, the cylinder 4 is provided with a hopper 7, from which various plastic materials 8 are charged, and a heater 9 is provided to heat the cylinder 4 to a predetermined high temperature. on the other hand,
A mold □10 is attached to the nozzle 1.

In such an injection molding machine, the mold 1° is first clamped and attached to the nozzle 1. On the other hand, when the plastic material 8 is introduced from the hopper 7 while the screw 2 is rotating and being heated by the heater 9, the plastic material 8 is fluidized 8a. In this state, the plastic material 8a is injected into the mold 10 by advancing the screw 2 with the injection cylinder 6.

After this, pressure is further applied into the mold 1o by the screw 2 so that the plastic 8a is injected. Then, when the molded product 11 is sufficiently solidified in the cooling process, the mold 10 is removed and the next molding is started.

By the way, even if the molded product 11 is molded in the above process, defects will occur, and these defects include insufficient filling, cracking,
There are many things such as sink marks, warpage, bends, twists, breaks, cracks, crazing, silver streaks, etc. The causes of these defects are, for example, insufficient fluidity of the plastic material due to insufficient filling, low injection pressure, etc., or when there is no place for air to escape within the mold 10.

If a defect occurs in this way, the molding conditions of the injection molding machine must be changed or corrected.

Therefore, this setting change/correction is carried out by an expert who is skilled in the art of molding this injection molding machine, by looking at the type and degree of the defect, and determining the cause of the defect through years of empirical trial and error. Therefore, it is difficult to operate an injection molding machine unless you are an expert.

However, in order to actually train experts, it is necessary to accumulate experience over a long period of time and become familiar with injection molding machines and the molding operations of various plastics, making it difficult to obtain experts. Furthermore, when a new injection molding machine is introduced, it is necessary to provide training courses or dispatch personnel to other factories to train experts. In addition, it is difficult to optimize the molding conditions when molding using engineering plastic materials, which are particularly well developed, and it is difficult to train experts.

(Problems to be Solved by the Invention) As described above, in order to optimize the molding conditions of an injection molding machine, an expert who is familiar with the injection molding machine is essential, and it is completely difficult for a beginner to do so.

Therefore, an object of the present invention is to provide an injection molding support expert system that can reliably and easily optimize molding conditions even with little experience.

[Structure of the Invention] (Means for Solving the Problems) The present invention provides an injection molding support expert system that supports an injection molding machine, in which a defective event of a molded product molded by an injection molding machine and the degree of the defective event are detected. At least, defective event data, cause event data, and diagnosis rule data for each combination of defective event data and combinations of the defective event data are accumulated for each type of plastic material for injection molding. Based on the knowledge base in which the degree of certainty for the diagnosis result obtained from the rule data is accumulated, and the failure event and its degree inputted by the input means, the diagnosis result of the cause event for the defective 11J event and this result is prepared according to the diagnosis rule data of the knowledge base. This is an injection molding support expert system that attempts to achieve the above object by including an inference section that infers the degree of influence of the diagnosis result.

(Function) By providing such a means, when a defective event of a molded product and its degree are input from the input means, the inference section calculates the defective event data or defective event according to the diagnostic rule data stored in the knowledge base. A diagnosis result is obtained by inferring failure cause data from the combination of , and the degree of influence of the diagnosis result is inferred from the degree of certainty for the diagnosis result obtained from the diagnosis rule data and the degree of the failure event -0 (Example) Below, An embodiment of the present invention will be described with reference to the drawings. Note that the same parts as in FIG. 5 are given the same reference numerals, and detailed explanation thereof will be omitted.

FIG. 1 is an overall configuration diagram of the injection molding support expert system. Defect events of the molded product 11, such as defective filling, failure, and the degree thereof, are manually entered from the keyboard 20, and these data are stored in a temporary memory 23 through the input/output section 21 according to commands from the main control section 22. . The main control unit 22 instructs the inference system 24 to perform inference, and also controls data exchange with other systems. Now, the inference system 24 has a function of diagnosing the cause event of the defective event based on the inputted defective event and its degree, and also inferring the degree of influence of this diagnosis result, that is, the degree of influence of the defective cause on the defective event. be. Specifically, the knowledge base 25 and the reasoning unit 26
It is composed of. The knowledge base 25 stores data on defective events, causal events, and diagnostic rules for each type of plastic material, as well as reliable data on diagnostic results obtained using these diagnostic rules. To specifically explain the contents of these data, as shown in Figure 2, various plastic materials such as polycarbonate (PC), polyetherimide (PE I'), polyether ether ketone, etc. (PEEK), ABS resin, polyphenylene sulfide (PPS) etc. S2・
...Sa and material data DI, D2...Dq are stored. Diagnostic knowledge 11s1. s2...SQ are the aforementioned defective events.

This includes causal event 1 diagnosis rules and reliable type data, and defective events include filling defects (as shown in Figure 3).
al), Paris (a2), Weld (a3).

Cracks (a4), foreign matter (an-1), 9 threads (an), greasy burns, discoloration, deformation, sink marks, air bubbles, etc. are accumulated for each address 1 to -0. In addition, (al) to (a
n) shows an example of a code when accumulating these defective events, and although these defective events are not shown, they are judged as the presence or absence of a defect, and when it is determined that there is a defect, the code (al ) to (an) are to be read. Next, the cause event is low resin temperature (dl) as shown in FIG.

High resin temperature (d2), low injection pressure (da), high injection pressure (d4), abnormal raw material (do-1)-, high injection speed (da), low mold temperature, long or short injection time, cooling vIIJ The length, length, etc. of 1 to 1 kb are stored in each address. Note that (al) to (an) indicate code examples when accumulating these causal events, and although these causal events are not shown, they are judged as defective 1g (resin temperature is low or high) and normal, respectively. When a defective state is determined, the codes (dl) to (dn) are read out. Next, the diagnostic rules are for inferring the cause event from the input defective event and calculating the degree of influence of the diagnosis result, and diagnostic rules r1 to rp are stored for each address 21 to I2p, respectively. For example, the rule number ``1'' stores the diagnosis result for a weld defect, that is, the resin temperature is low, and the sure type 0.8 for this diagnosis result is stored, and the rule number ``3'' stores the filling defect and weld defect. The diagnosis result that the resin temperature is low when the failure occurs is stored, as well as its certainty type 0.9.In other words, the diagnosis result and its certainty type for each defective event and its combination are stored.By the way,
The knowledge data presented above is acquired based on empirical trial and error by experts while following actual accident cases.

The inference unit 26 infers the causal event from the inputted defective event according to the diagnostic rule and outputs it as a diagnosis result, and also calculates the degree of influence on the diagnosis result from the degree of the inputted bad event and the certainty type of each diagnostic rule. It has a calculation function. Note that the inference unit 26 has a function of comparing the degree of influence on the diagnosis result with an inference threshold value inputted in advance, and omitting diagnosis results below the threshold value. Also, knowledge base 25
A CRT display 27 with a control function for writing and deleting data in the knowledge base 25 is connected to the knowledge base 25, so that each data in the knowledge base 25 can be changed, corrected, and added to. 28 is a CRT display.

Next, the operation of the system configured as described above will be explained. The plastic material 8a fluidized by the action of the injection molding machine is injected into the mold 1o, and after cooling, the mold 10 is removed from the mold 10 as a molded product 11. A judgment is made by the operator. As a result of this determination, for example, if there is a weld defect, the type of plastic material used, for example polyetherimide (PEI), is determined, as well as the weld defect and the degree of this weld defect, for example 0.6, and the inference threshold 0. 3 is pressed manually from the keyboard 20. The main restraining portion 22 is made of these polyetherimides (PE
I) Data on weld defects and their degree are stored in the temporary memory 23 and passed to the inference section 26. Now, this inference section 26 first compares it with a predetermined value of the failure degree in the case where polyetherimide (PEI) is used and a weld failure occurs, and if it is greater than this predetermined value, then polyetherimide (PEI) is used. PEI) to address area B in the knowledge base 25, and this f! 4. Execute inference according to the diagnostic rules stored in area B. Therefore, the inference unit 26 searches for all diagnostic rules including weld defects, and as a result of this search, as shown in FIG. 5, rule number @r1. Find r3. Diagnostic rule "3" then asks the operator in an interactive manner whether there is a filling defect event, and if there is no such event, diagnostic rule r3 is rejected and diagnostic rule r1 is selected. Using r1, a cause event corresponding only to a weld defect, that is, a4 low fat temperature is searched for, and its certainty level of 0.8 is read out.

As a result, the inference unit 26 multiplies the degree of certainty by 0.8 and the degree of weld failure by 0.6 to calculate the influence degree of 0.48, and confirms that the degree of influence is greater than the inference threshold and determines the cause of the low resin temperature. The event and its influence degree of 0.48 are sent to the main control unit 22.

In this way, the cause event of low resin temperature and the influence degree of 0.48 are transmitted to C through the input/output section 21 by the command from the main control section 22.
It is sent to the RT display 28 and displayed as a diagnosis result. The operator judges the change/correction of the molding conditions and the degree of the change/correction based on the displayed low resin temperature and influence degree of 0.48. In this case, since the resin temperature is 8 degrees lower, for example, the heating temperature of each heater 9 is increased,
The degree of increase in this heating temperature is determined based on the degree of influence, and the molding conditions are changed.

In addition, as defective events, insufficient filling and its degree is 0.7, weld defect and its degree is 0.8, and the recommendation threshold is 0.
．． 8 is entered manually, the inference unit 26 uses the diagnostic rules rl, r2.
Search r3 and find insufficient filling and its degree 0 as a defective event
．． 1, weld defect, its degree 0.8, and inference threshold 0.9 are key human inputs.

Then, the inference unit 26 uses diagnostic rule number r1 in the same manner as in the case of only the weld defect. r2. Search r3. Then, due to the combination-OR relationship, the diagnostic rule rl: 0,8 x O,8-0,64r2: 0
．． 7 x O, 8-0, 56r3: 0.7 x O, 8
Using xO,9-0,504, 0.64+ 0.56- (0,84x O,56) error 0.8420.842 +0.504-(0,842x
O,504) -0,922 is obtained. Thus, if the resin temperature is low, the influence degree is 0.922, which is sent to the CRT display 28 through the input/output unit 21 and displayed as a diagnosis result, and the operator can change or modify the molding conditions and the degree of the change or modification from this display. to decide. In this way, diagnostic results are output corresponding to each defective event and its combination.

By the way, when a defective event that is not already stored in the knowledge base 25 appears, the causal event and diagnostic rule for this defective event are acquired from the inference unit 26 by manually inputting keys on the keyboard 20. A message indicating that there is no such information is issued and displayed on the CRT display 28. Therefore, the operator examines this defective event, its cause event, and the diagnostic rule, and adds new knowledge to the knowledge base 25 from the CRT display 27 with a control function, thereby enriching the knowledge.

In this way, in the above-mentioned embodiment, the inference unit 26 infers the cause of the defect from the defective event and the combination of the defective events according to the diagnostic rules stored in the knowledge base 25 based on the defective event of the molded product 11 and its degree, and makes a diagnosis. The configuration is such that the influence of the diagnosis result is inferred in an interactive manner with the operator based on the certainty of the diagnosis result obtained by the diagnosis rule and the degree of defective events, so even beginners who are not experts can understand the type of defective event and the degree of influence of the defective event. The cause of the failure can be determined from the degree of occurrence, and countermeasures can be taken to deal with this cause event. The diagnosis can also be immediately and reliably displayed and output for each type of plastic material. Therefore, it is possible to easily set the optimum molding conditions in the injection molding machine even for engineering plastics, etc., which require poor molding conditions and are not familiar with injection molding machines, and there is no need to take the trouble of training experts.

Note that the present invention is not limited to the above-mentioned embodiment, and may be modified without departing from the spirit thereof. For example, computers and engineering workstations (EWS) of various scales can be used to connect with CACI and CAε systems to accumulate shape data for each molded product in a knowledge base, which can be used to communicate with mold flow analysis systems, etc. The configuration may be such that data is exchanged, or when a defective event is input, an optimal causal event is inferred from shape data and material data for various plastic materials.

[Effects of the Invention] As described in detail above, according to the present invention, it is possible to provide an injection molding support expert system that can reliably and easily optimize molding conditions even with little experience.

[Brief explanation of the drawing]

FIG. 1 is an overall configuration diagram showing one embodiment of an injection molding support expert system according to the present invention, FIGS. 2 to 5 are schematic diagrams showing the contents of each data accumulated in the knowledge base of the system, and FIG. The figure is a configuration diagram of a complete screw injection molding machine. 11... Molded product, 20... Keyboard, 21...
Input/output section, 22... Main tllJI 11 section, 23...
Memory, 24... Reasoning system, 25... Knowledge base, 26... Reasoning unit, 28... CRT display. Applicant's agent Patent attorney Takehiko Suzue Figure 2 Figure 3 Figure 4 Figure 5

Claims (1)

[Claims] In an injection molding support expert system that supports an injection molding machine, an input means for inputting a defective event of a molded product molded by the injection molding machine and the degree of the defective event;
At least, defective event data, cause event data, and diagnostic rule data for each combination of the defective event data and the combinations of the defective event data are accumulated for each type of plastic material for injection molding, and the diagnostic results obtained from these diagnostic rule data are In response to the knowledge base in which the degree of certainty is accumulated, and the defective event and its degree inputted by the input means, the diagnosis result of the cause event for the defective event and the diagnosis result are determined according to the diagnostic rule data of the knowledge base. An injection molding support expert system characterized by comprising an inference section for inferring the degree of influence.