JPH02183335A - Quality diagnosis knowledge correcting method - Google Patents

Abstract

PURPOSE:To correct a low knowledge base used for quality diagnosis even by a general technician or an operation technician by correcting a knowledge table, which is composed of specified data, by an editor. CONSTITUTION:The knowledge table is composed of the data, for which data for ranking operation data, evaluation demerit degree for each quality characteristic index and demerit pattern are expressed by the rank block and weight of the respective operation data, and a data file called as the knowledge table is corrected by the editor. The knowledge base to execute the next quality diagnosis processing is automatically generated by the knowledge table. Thus, the knowledge base for one fixed type of the quality diagnosis in a manufacture process can be corrected even by the general technician or operation technician and the fine maintenance of such a quality diagnosis system is easily executed.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a knowledge modification method used for quality diagnosis in manufacturing processes such as steel processes.

Quality diagnosis quickly predicts and evaluates quality characteristics based on current operating conditions, and by taking appropriate action, it contributes to the prevention of quality characteristic defects. It facilitates the maintenance of the knowledge base used for the process and enables permanent use of quality diagnostics.

[Prior Art] Conventionally, quality diagnosis in manufacturing processes such as steel processes has been performed when quality characteristics can be quantitatively predicted using a mathematical model. However, quality characteristics often cannot be predicted quantitatively. At that time, the quality diagnosis system 11 is developed using qualitative or piecemeal empirical rules using a knowledge engineering method such as a rule base.

[Problem to be solved by the invention] However, this new type of quality diagnosis system 11
Currently, there is no established method for quality prediction/diagnosis used for quality prediction and diagnosis, and it differs from system to system. Additionally, the knowledge base used in this new type of quality diagnostic system is typically modified by a KE (Knowledge Engineering Engineer) - often difficult to modify by membrane engineers or operational engineers, and does not allow detailed maintenance. There is this inconvenience.

The present invention provides a knowledge modification method that allows even general engineers and operation engineers to modify the knowledge base used for quality diagnosis based on the empirical rules of manufacturing process experts. The purpose is to provide a method for correcting the original knowledge of diagnostic methods.

[Means for Solving the Problems] The quality diagnosis knowledge adjustment method of the present invention involves ranking a plurality of pieces of operational performance data used for quality prediction in a manufacturing process based on deviations from their target values, and ranking a plurality of pieces of operational performance data used for quality prediction in a manufacturing process based on the deviation from the target value. For each quality characteristic item, the demerit value of the quality characteristic due to the deviation from the target value of the operational data is calculated using the relationship between the rank value of the operational data and the demerit value, and each demerit point value for each quality characteristic item is determined. The evaluation value is obtained by summing up the points and subtracting them from the perfect score, and by calculating the total value of each demerit point for each operational data item and arranging them in descending order, it is possible to display the rating for each quality characteristic item and to calculate the operational data. In quality diagnosis, in which operational instructions are given in order of importance, a knowledge table is created that includes data that ranks operational data, and data that expresses the degree of deterioration in ratings for each quality characteristic item and its deterioration pattern using the rank interval and weight of each operational data. The method consists of modifying a pig file called ``Pig File'' with an editor, and automatically generating a knowledge base from the knowledge table to perform the quality diagnosis process described above.

[Operations] A general engineer or an operation engineer can use an editor to create data that expresses the rank division of operation data in the knowledge table, the degree of deterioration in ratings for each quality characteristic item, and its deterioration pattern using the rank interval and weight of each operation data. Once the correction is made, a knowledge base for the next quality diagnosis process is automatically generated.

Multiple pieces of operational performance data used for quality prediction are ranked according to their deviation from the target value, and the demerit value of each operational data is calculated for each quality characteristic item from the relationship between the rank value and demerit value expressed in the knowledge base. The evaluation value of the quality characteristic is calculated by subtracting the total value from the perfect score value and outputs it. In addition, the sum of each deduction value is calculated for each operation data item, and operation instructions for the operation data are output in descending order of the sum.

[Embodiments] Next, embodiments of the present invention will be described with reference to the drawings.

In this example, knowledge modification for quality diagnosis regarding film characteristics of grain-oriented electrical steel sheets will be described. The quality diagnosis of the coating is carried out by an expert's empirical rule based on the operational performance data of the decarburization line, and the coating performance is obtained after about 10 days. Therefore, the occurrence of coating defects during this period is prevented by this quality diagnosis.

FIG. 1 shows the configuration of a capsular diagnosis system using a knowledge base adjusted by the knowledge modification method of the present invention. Operation performance data is input and quality is evaluated using an inference engine based on this knowledge base. Evaluation values and the like are output in an easy-to-understand manner as a display of quality diagnosis results.

FIG. 2 shows an overall flowchart of this coating quality diagnosis method.

In step 1 of FIG. 2, the operating performance data of the decarburization line is ranked according to its deviation from the target value. Each range of this ranking is given for each steel type standard. Figure 3 shows an example of the range of each rank for ranking knowledge for film prediction. In this example, 10 types of operational performance data are stratified into up to seven ranks depending on the deviation from the target value.

In step 2 of FIG. 2, a deduction value is calculated for each coating quality characteristic item to be diagnosed. For details on this calculation method,
Next, a description will be given with reference to FIG.

Simply put, peroxide nodes are a coating quality characteristic.

Evaluate insufficient oxidation, decarburization abnormalities, etc. FIG. 3 shows the score prediction knowledge for peroxide node A, and the operation pattern in which this quality defect occurs is shown by a thick line. This quality characteristic is evaluated on a scale of 80 points, and when the pattern exhibits the largest deviation, the predicted score is 40 points.

In addition, the defective contribution rate of each operational data is 70%.

They give 5%, 10%, etc.

Using this knowledge, the deviation from the target value of each operational data (
It is possible to find out how much points are subtracted by the rank value) from the quality characteristic score. For example, if the performance of No. 2 in the operation performance data is rank 6, the demerit value of the evaluation value due to the deviation of the target value in this part is 28 points as the calculation result of (80-40)X70・/100 It is calculated as follows. For ranks smaller than 6, the multiplication coefficient from 0 to 1.0 is
Calculated using the specified function, and the demerit value is multiplied by that value. The multiplication coefficient is selected from a plurality of functions shown in FIG. 7, depending on the actual phenomenon.

Therefore, when the rank is 5 in the previous example, if the function No. 3 in FIG. 7 is selected, the deduction value will be 22.2 as the result of dividing 28 points XQog43. In this way, the demerit value resulting from the deviation of the operation performance data from the target value for one film quality characteristic is determined.

Similar predictive knowledge is used for other coating quality characteristics.

Referring again to FIG. 2, in step 3, the sum of the demerit points of the operational performance data is calculated for each film quality characteristic item, and the sum is subtracted from the respective perfect score values. Further, in the next step 4, a deduction value for each coating quality characteristic item is calculated for each operation data item.

An example of the calculations performed in steps 3 and 4 is shown in FIG.

In Figure 4, the vertical direction represents coating quality characteristic items;
The horizontal direction indicates operational data items.

Each numerical value in the center indicates the demerit value for each item, the total demerit value of the operation performance data for each coating film quality characteristic item is shown at the right end, and the total demerit value for each operation data item is shown at the bottom end. has been done. The larger this value is, the more important it is for improving quality characteristics, so when giving operational instructions for operational data, the operational data items with the largest values are listed first.

What is shown in FIG. 5 is what is called a knowledge table in which the coating prediction knowledge shown in FIG. 3 is expressed as a data file. At the beginning of this table, data for ranking the operational data is expressed.

apl indicates the operation data item, 5 indicates the number of rankings, and the following indicates the range of rankings. * represents more than or less than. Other operational data are expressed in the same way.

Next, a list of quality characteristic items is expressed. That is, peroxidation A, peroxidation B, insufficient oxidation A, etc.

Next, the score deterioration and its deterioration pattern are expressed for each quality characteristic item. The example of Peroxide A shows that the score deteriorates from 80 points to 40 points. “1” of dp2
, 3.6.5.6.70.0'', 1 is the number of patterns, 3 is the multiplication coefficient function equation number, 6 is the number of ranks, the next 5 and 6 are the rank intervals of worsening patterns, 70,0 is the weight (defect contribution rate).

This knowledge table is modified using an editor by an operating engineer or a general engineer.

The knowledge base shown in FIG. 6 is generated by the knowledge base generation program using this knowledge table as input data.

Using the IF-THEN method, each condition is judged, the operational performance data is ranked, the deduction points are calculated for each quality characteristic item, and the total deduction points are calculated.

Each rule for calculating the total deduction points for each operation data is generated.

[Effects of the Invention] As explained above, the present invention enables operation engineers and general engineers themselves to modify the knowledge base of a routine quality diagnosis of a manufacturing process, thereby facilitating detailed maintenance of this quality diagnosis system. becomes easier.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of a system implementing the present invention. FIG. 2 is an overall flowchart of the coating quality diagnosis method of the embodiment. FIG. 3 is a map showing the coating prediction knowledge of the embodiment. FIG. 4 is a map showing the predicted demerit point value and the total value in the example. FIGS. 5 and 6 are maps showing the correction knowledge table and the generated knowledge base of the embodiment, respectively. FIG. 7 is a map showing the assignment of calculation formulas for multiplication coefficients in the embodiment. Side bone-

Claims (1)

[Claims] In the manufacturing process, multiple pieces of operational performance data used for quality prediction are ranked based on the deviation from the target value, and the results are calculated based on the deviation of the operational data from the target value for each of the multiple quality characteristic items to be diagnosed. The demerit value of a quality characteristic is determined by using the relationship between the rank value of the operational data and the demerit value, and the evaluation value is determined by summing the demerit value for each quality characteristic item and subtracting it from the full score value. ,
By calculating the total value of each demerit point for each operational data item and arranging them in descending order, it is possible to rank operational data in a quality diagnosis that displays scores for each quality characteristic item and provides operation instructions for operational data in order of importance. Modify with an editor a data file called a knowledge table, which is made up of data expressing the degree of deterioration of ratings for each quality characteristic item and its deterioration pattern using rank intervals and weights for each operation data, and use the knowledge table to diagnose the above quality. A quality diagnosis knowledge correction method characterized by automatically generating a knowledge base for processing.