JPH0651334B2 - Injection molding technology support expert system - Google Patents

Description

The present invention relates to an injection molding technology support expert system.

[Conventional technology]

When injection molding is performed, it is difficult to set the optimum molding conditions because the molding conditions greatly differ depending on the size and shape of the molded product in the mold. Generally, the work is performed based on the abundant experience and judgment of a skilled worker, but it takes a considerable amount of time and cost to train an expert of this work.

In recent years, in order to improve the present situation and simplify the molding condition setting work, an attempt to use a computer system equipped with a database and a knowledge base has been proposed, and its realization is expected. Prior arts prior to the filing of the present application include JP-A-63-209917 (hereinafter referred to as Conventional Example 1) and JP-A-63-209918 (hereinafter referred to as Conventional Example 2).

Since these conventional examples are important for understanding the features of the present invention, their outline will be described with reference to FIGS. 8 and 9.

Conventional Example 1 When a defect event and its degree of a molded product are input, defect cause data is inferred according to the diagnosis rule data accumulated in the knowledge base, and the diagnosis result is output. The degree of influence of the diagnosis result is inferred from the degree of the defective event, and the molding condition is corrected (FIG. 8).

Conventional Example 2 When predetermined data is input from a keyboard, a molding condition is inferred using a database / knowledge base, and a difficult-to-guess part is computer-controlled and an injection molding machine in which a sensor is attached to each part is used. The trial shooting is repeated by fully automatic control to acquire data and improve the accuracy of inference.
As a result, initial conditions are set and a molded product is produced. Then, the diagnostic expert system is used to analyze the cause of the defect and make final fine adjustments for the molded product to set the final conditions (FIG. 9).

[Problems to be Solved by the Invention]

According to the study by the inventor of the present application, it was found that the above-described conventional example has some points to be improved as described below.

Regarding Conventional Example 1 (1) Judgment of the degree of a defective event with respect to a defective event is likely to vary from person to person.

(2) It is difficult to properly set the certainty for the diagnosis result.

(3) Although the diagnosis result is displayed with the degree of influence of the diagnosis result, when it is executed by looking at it, variations easily occur among people. In addition, when it is executed, if there is no significant improvement, even if data is input again, only the same diagnostic result as the previous time may be displayed.

(4) Since there are many combinations of defects and defect causes, it is difficult to analyze the cause of defects.

Regarding Conventional Example 2 (1) At the time of setting the initial condition, it is necessary to perform trial shooting to improve the accuracy, which complicates the operation.

(2) The specific contents of the diagnostic expert system are unclear.

[Means for Solving the Problems]

The present invention has been made with the purpose of using an injection molding machine that is normally used to eliminate unclear judgments as much as possible and to give a worker practical instructions for achieving the purpose.

To this end, the present invention analyzes and models the experience and knowledge of experts in injection molding technology, and embodies this model in the most refreshing and simplified form.

That is, by analyzing the work procedure (thinking procedure) of an expert in injection molding technology, the initial conditions are set with reference to past similar molded product data, the actual molded product is observed, and the first The defect that should be most remedied (or the improvement effect can be expected most) is selected, the most effective remedy for that defect is intuitively presented and executed, and the improvement effect is determined. While confirming, it was found that this work was repeated by trial and error. On the contrary, neither the analysis of the cause of the defect, the derivation of the measures based on the detailed analysis, nor the attempt to eliminate a plurality of defects at once are performed.

The present invention which realized this expert model as a system, as shown in FIG. 2, is a part for setting the initial condition by using the data of the similar molded product as it is as much as possible, and an intuitive trial for each defect. It has a part to erroneously guide and present measures.

[Action]

Initial conditions using the data of similar molded products are presented by the functions of the initial molding condition setting means and the molding condition suggesting means. Further, the function of the molding condition inference means always presents a positive measure against the defect of the molded product.

[Embodiment] Next, an embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a diagram showing the configuration of an embodiment of an injection molding technology support expert system of the present invention, FIG. 3 is a diagram for explaining the outline of the present embodiment, and FIG. A flow chart showing the overall work procedure when performing work,
FIG. 5 is a flow chart showing the procedure of countermeasure inference judgment in this embodiment, and FIGS. 6 and 7 are examples of screens on the input means X3 (see FIG. 1) for inputting a molding defect, respectively. It is a figure which shows an example of the countermeasure presentation screen with respect to a defective input.

First, the overall configuration of this embodiment is as shown in FIG.
As shown in the right column of FIG. 3, the countermeasure inference unit E includes a control unit that determines an overall policy such as which defect should be preferentially processed when a plurality of defects occur, and an individual control unit. It has a specialized department for professionally handling defects and a rule base.
Rule-based knowledge is “if premise then action / conclusion”
It is described in the format. The control unit and the specialized unit are constructed by using an object-oriented programming method that refers to the rule base and exchanges messages with each other to achieve the purpose. Although a concrete measure proposal procedure will be described later, conceptually, in the specialized department, measures (priority measures A, B, C) are prioritized for each defect (for example, defect A, defect B, defect C). ..) is used to describe the table. Although not shown, the countermeasure inference unit E further has a blackboard (black board function) capable of writing data such as hypotheses, intermediate results, inference results, and deleting unnecessary data when executing inference. is doing.

Next, the operation will be described.

First, an outline of the work procedure will be described. As shown in the left column of FIG. 3, input is performed from the keyboard (step
400), initial molding conditions are presented (step 500), followed by molding (step 600), input of defective state (step 700), and presentation of countermeasures against this (step 800). After this, steps 600 to 800 (process RI) are repeated by trial and error until the defect is resolved.

Next, a more detailed description will be given with reference to FIGS. 1 and 4. It should be noted that in FIG. 4, the human motion is shown by double lines in order to distinguish the human motion from the system motion for easy understanding.

First, the operator uses the display / input means X1 to display the weight of the molded product,
Data such as projected area, molding material code and machine code are input (step 1).

Then, the initial molding condition setting unit A searches the material master file M1, the machine master file M2, and the similar molding master file M3 to create the current molding condition file D2 (step 2).

Next, the molding condition proposing section B reads the current molding condition file D2 (and a countermeasure file D3 described later), and displays it on the display means X.
2 (step 3).

Next, the operator sets molding conditions with reference to the display screen (step 4), and performs test firing to obtain a molded product S2 (step 5).

Next, the operator sets the defective state of the molded product S2 to the defective input means X.
Input from 3 (step 6). In this case, for example, a screen display as shown in FIG. 6 is made, and the operator inputs a code indicating the corresponding state according to the item.

Next, the molding failure determination section C receives the input failure status data and creates a molding failure status file D4 (step 7). At this time, generational management is performed to record the change in the defective molding state.

Next, the operator inputs the molding conditions (set value and measured value) when the above-mentioned molded article is shot by trial from the display / input means X4 (step 8). At this time, since the molding condition grasping section D displays the current molding condition D2 on the screen, it is convenient for the operator to input only the changed parameters.

Next, the molding condition grasping section D registers the input set values and measured values in the current molding condition file D2 and updates the contents (step 9).

Here, when a non-defective product is obtained, the defective molding determination unit C notifies the success example storage unit F of the fact (step 11), and the success example storage unit F sets the current molding condition D2 to the similar molding master file M.
Register in 3 (step 16).

If there is a defect, the countermeasure inference unit E is activated, the current molding condition D2 and the molding defect status D4 are read, the countermeasure is inferred (step 13), and the countermeasure plan is registered in the countermeasure plan file D3 (step 14).

Next, the molding condition proposing section B reads and displays the current molding condition file D2 and the countermeasure file D3 (step 15),
The process is repeated by returning to step 4. In step 15, the measure is displayed on the display means X2 in the format as shown in FIG. 7, for example.

Next, a procedure for inferring and presenting countermeasures in this embodiment will be described with reference to FIG.

When the contents of the molding defect status file D4 and the current molding condition D2 are read, the countermeasure inference unit E first determines whether a plurality of types of defects have occurred (step 21). If it is a single defect, a high-priority countermeasure for that defect is proposed (step 22), and in the case of multiple defects, the control unit decides which defect should be prioritized for countermeasures (step 23). The part is the identified defect (for example, defect A in FIG. 3).
For example), the measure with the highest priority (measure A in FIG. 3) is output (step 24). After that, the number of times the same countermeasure is output is counted (step 25).

Next, based on the countermeasure, the result of the trial shot being performed and the defective state being input again is determined as follows.

That is, it is determined whether or not the defect state has been resolved for the defects identified in steps 23 and 24 (step
26), if it has been resolved, proceed to step 32. If not solved, it is judged whether or not there is an improvement in the defective state (step 27), it is judged whether the output count is the preset upper limit value, the improvement is recognized, and the countermeasure output frequency is the upper limit value. If not, the same measure as that output in step 24 is output again (step 29), and the process returns to step 25. here,
The upper limit of the number of outputs for the same measure is
It may not be preferable from the viewpoint of processing efficiency to repeatedly output only the same measure many times. Therefore, if the defect is not resolved by repeating the process a predetermined number of times, the process is shifted to the next measure.

If no improvement is found, or if the output count is at the upper limit, determine if there is a next measure (step 3
0), the next measure (measure A in FIG. 3) is output (step 31), the process returns to step 25 and the process is repeated.

If there is no next countermeasure, the countermeasure for the defect is discontinued and the countermeasure for the next defect is entered. That is, it is determined whether all the defects have been resolved, and if they have not been resolved, an instruction to move to the countermeasure for the defect with the next highest priority is given (step
33), and returns to step 24 to repeat the process.

〔The invention's effect〕

As described above, according to the present invention, the following effects can be obtained.

(1) Workers can set molding conditions based on the initial setting conditions and measures proposed by this expert system (adding their own ideas), and even inexperienced people can set appropriate conditions. Also, the burden of setting work is reduced.

(2) Analyzing the cause of defects from the defective state of molded products,
Since the thinking model actually performed on site by the work expert is embodied by directly presenting the most effective measures without performing detailed analysis, a realistic response is possible.

(3) Since unclear concepts such as certainty and influence are not used, it is easy to prevent variations in judgment. In addition, since the user's work can be performed in a form of interacting with the system, his / her work can be easily grasped and organized, and the ability of the worker can be expected to improve.

(4) A normal injection molding machine can be used, introduction to the site is easy, and the cost is reasonable.

[Brief description of drawings]

FIG. 1 is a diagram showing the configuration of an embodiment of an injection molding technology support expert system of the present invention, FIG. 2 is a diagram for explaining the outline of the present invention, and FIG. 3 is an outline of the present embodiment. FIG. 4, FIG. 4 is a flow chart showing the overall work procedure when performing work using this embodiment, FIG. 5 is a flow chart showing the procedure of countermeasure inference judgment in this embodiment, FIGS. 6 and 7. FIG. 8 is a diagram showing an example of a screen on the input means X3 when a molding defect is input, and an example of a countermeasure presentation screen for the defect input. FIGS. 8 and 9 are Conventional Example 1 and Conventional Example 2, respectively. It is a figure for explaining the contents of. A: Initial molding condition setting section, B: Molding condition suggesting section, C: Molding failure judgment section, D: Molding condition grasping section, E: Countermeasure reasoning section, F: Success example accumulating section, X1 ... ... Display / input means, X2 ... display means, X3 ... defective input means, X4 ... display / input means, S1 ... injection molding machine, S2 ... molded product, M1 to M3 ... master file, D1. D4 ... Data file.

Claims (3)

[Claims] 1. An input means for inputting data necessary for setting conditions of injection molding, and using the input data, data for similar molded products stored in a similar molding master file is searched and obtained. The initial molding condition setting means for setting the initial molding condition by using the data about the similar molded product and the current molding condition file, the display means for displaying the set initial molding condition, and the actual molding condition. Defect input means for inputting the defective state of the molded product obtained by the above, molding defective judgment means for creating a defective molding condition file in response to the inputted defective state, and molding conditions for the actual molding Input means for inputting, molding condition grasping means for receiving the molding conditions input from the input means and updating the current molding condition file, and contents of the current molding condition file And the contents of the molding failure status file, and infers the countermeasure against the defect and creates a countermeasure plan file, and the contents of the current molding condition file when a non-defective product is obtained by actual molding. , An injection molding technology support expert system having a success example accumulating means for accumulating in the similar molding master file. 2. The countermeasure inference means, when a plurality of defects occur at the same time, determines a defect to be processed preferentially, for example, determines a general policy for countermeasures against molding defects, and an individual control unit. The "if precondition then" is used to identify the specialized department that deals with defects and creates countermeasures for them, and the knowledge that the control unit and the specialized department use.
2. The injection molding technology according to claim 1, further comprising a rule base described in the "action / conclusion" format, wherein the control unit and the specialized unit exchange messages with each other to execute countermeasure inference. Support expert system. 3. The injection molding technology support expert system according to claim 2, wherein the countermeasure inference means outputs the countermeasure determined to have a high priority by trial and error in consideration of the direction of improvement with respect to the proposed countermeasure. .