JP2719339B2 - Fault diagnosis device - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a failure diagnosis device for concluding the cause of a failure of a numerically controlled machine tool, a sequence controller, or any other device.

2. Description of the Related Art In a failure diagnosis apparatus, a knowledge base in which rules of thumb based on expert knowledge are stored in an external storage device, and the cause of the failure is pursued by a question-and-answer that answers a question item constituting a rule. The rule is composed of a condition part and a conclusion part indicating a conclusion when the condition part is satisfied. For example, as rule No. 1, [if (A and B) and (C or D) and E and
(F or G), then ... "was incorporated into the knowledge base. Here, A to G are answers determined according to the question items. The if ... part is the conditional part, then
... is the conclusion. The question and answer method through the dialogue is performed in the order of registration for the question items constituting the condition part of the rule. Therefore, if the way of organizing the conditional sentence of the rules is bad, a situation arises in which a redundant question is asked and it is difficult to reach a conclusion. At this point, there is a problem that the use of the failure diagnosis apparatus becomes inconvenient because the user cannot modify or add the rule.

"Problem to be Solved by the Invention" The present invention has been made in order to solve the above-mentioned problems, and is a failure which can automatically rearrange the question items of the rule for which an efficient conclusion can be obtained. It is an object to provide a diagnostic device.

"Means for Solving the Problems" A failure diagnosis apparatus according to the present invention, which meets the above objects, includes a condition unit,
A rule consisting of a conclusion part indicating a conclusion when the condition part is satisfied is stored as a knowledge base in an external storage device, and a failure diagnosis device that performs a failure diagnosis based on this knowledge base includes a condition part in a certain rule. A usage count means for counting the number of times the constituent question item is used in how many rules, and the number of true or false whether the answer to the question item is set to true or false in each rule An answer counting means for obtaining the answer, an evaluation means for giving an evaluation value for evaluating the ease of obtaining an answer to each question item, a table for storing the number of times of use, the number of true and false answers and an evaluation value, The number of uses stored in the table,
An importance calculating means for calculating the importance of a question item from the number and evaluation value of each of true and false answers, and a sorting means for rearranging the question items in the rule in descending order of importance. I do.

[Operation] According to the above configuration, when the user modifies or adds a rule, the condition part of the rule is arranged for efficient conclusion. How to organize is as follows: (1) How many rules make up the question items that make up the conditional
(2) Whether each question item is set as “true” or “false” in the rule, the number of each item, and (3) ease of answering the question item are given. Information about three items is stored in the table. Then, the importance of each question item is calculated based on the stored information, and the question items are rearranged in descending order of importance.

"Example" Next, the present invention will be described specifically with reference to an example shown in Figs. 1 to 6. FIG. 1 is a block diagram showing the configuration of the failure diagnosis device of the present embodiment.

The failure diagnosis device 20 is composed of a computer system, and executes a failure diagnosis program.
, A RAM 2 for storing various information, a ROM 7 for storing a processing program, a keyboard 3 and a mouse 4 for inputting data and the like, a CRT 5 for displaying diagnosis results and the like, and an external storage for storing rules as a knowledge base. Device 6;
The keyboard 3 and the mouse 4, the CRT 5, and the external storage device 6 are connected to the CPU 1 via the interface 9.

The RAM 2 has a question item storage area n for storing question items.
And an importance storage area nI for storing the importance of the question item,
A number storage area m for storing the number of times the question item n is used in all rules, a true frequency storage area T for storing the number of rules for which the answer of the question item is set to true, and an answer for the question item A false frequency storage area F for storing the number of rules set to false and an evaluation value storage area E for storing an evaluation value of ease of obtaining an answer to a question item are provided.

The external storage device 6 serving as a knowledge base stores rules in which the experience of the expert is ruled. FIG. 2 and FIG. 3 show the state at the time of rule creation. The rules shown in FIG. 2 are based on the states of the input / output elements of the sequence controller 31 to be diagnosed via the interface 10,
The expert's knowledge of specifying a failure element from information such as an error code and an interrupt cycle output from the numerical controller 30 is composed of various condition parts and a conclusion part specifying the failure element when the condition parts are satisfied. Is given in the form.

For example, as shown in Fig. 2, the rule is that the input / output element status is "contact 4R1P is ON" or "contact 4R2P is O
`` N '', `` contact 4R3P is ON '', the cycle without interruption is `` main arm inversion '', the error code is `` 03 '', and the fault element is identified when the condition is satisfied. Contact 4R3P is abnormal. " The relationship between the condition part and the conclusion part expresses the expert's empirical knowledge in the form of rules. The rules are expressed using a computer language (Prolog language) as shown in FIG. 3, and the rules expressed using the Prolog language are as follows:
It is stored in the external storage device 6.

[Operation] Next, the operation of the embodiment based on the above configuration will be described.

When creating and registering rules, the user uses the keyboard 3 and the mouse 4 shown in FIG.
The rules shown in FIG. 3 are registered and stored in the external storage device 6 via the. The external storage device 6 may be a device in which rules are registered by a device separate from that shown in FIG.

The rules stored in the external storage device 6 are read into the question item area n of the RAM 2, the importance is calculated by executing a program described later, the question items are rearranged in the order of importance, and again the external storage device 6 is stored.

When the user modifies, deletes, or adds a rule, the importance of the question items is calculated in the same manner, and the question items are rearranged in the order of importance. When performing the failure diagnosis, the failure diagnosis is performed according to the rule in which the reordering of the question items is completed, and the user inputs an answer to the question items displayed on the CRT 5 using the keyboard 3 and the mouse 4. A question-and-answer session is conducted in an interactive manner, and the cause of the failure is pursued. Questions and answers are made in the order of registration for the question items constituting the condition part of the rule. However, for a question item that has been answered once, the answer is saved and the next time the answer is referred to, no question is asked.

In the present embodiment, the method of calculating the importance nI of the question item is as follows. That is, importance = (parameter m of the number of appearances in the rule of the question item) + (the number of rules T or F set among the rules set as true t / false f) +
(Evaluation value E of ease of obtaining an answer), which is represented by the following equation (1).

nI = m + min (T, F) + E (1) For example, assuming that the question item "main power supply voltage is 200 V" appears in 20 rules (is used), Fifteen rules set the answer to the above question item as "true", five rules set the answer to the above question item to "false", and ease of obtaining the answer Is easy (= 4), the importance is calculated as follows.

Importance nI = 20 + min (15,5) + 4 = 29 The answer is very easy (= 5), easy (= 4), normal (= 3),
Evaluation is made in five stages: difficult (= 2) and very difficult (= 1).

The question item in which the second term in the above equation (1) is zero (= 0) is automatically excluded from the question item by the failure diagnosis device.

min (T, F) = 0 This is because a question item for which all answers are true or false has no meaning as a question item.

Further, the procedure of the calculation of the above equation (1) will be described with reference to the flowchart shown in FIG. When the program shown in FIG. 4 starts, in step 101, 1 is set to the question item n. In step 102, the question item n is called. In step 103, the number m of times that the question item n appears in the rule is counted.
In step 14, the number (T) in which the answer to the question item n is true (t) and the number (F) in which the answer is false (f) are counted. In step 105, the easiness E of obtaining the answer to the question item n is stored. In step 106, it is determined whether or not the number m is one. If a positive result is obtained, the process branches to step 107 to set the importance nI to one.
If the determination result in step 106 is negative, it is determined in step 108 whether the true number (T) is zero (= 0). If a positive result is obtained in step 108, the question item n is deleted in step 109, and the process proceeds to step 112. If a negative result is obtained in step 108, the process proceeds to step 110, where it is determined whether or not the number F of rules for which the answer to the question item is false is zero (= 0). When the determination result in step 110 is affirmative, the process branches to step 109 to delete the question item n, and when a negative result is obtained, the process proceeds to step 111. Then, in step 111, the importance nI of the question item n is calculated from m, T, F, and E. In step 112, it is determined whether or not the item is the last question item. When a positive result is obtained, the process branches to step 114, and when a negative result is obtained, the process proceeds to step 113. In step 113, m, F, T, and E are cleared, and the process returns to step 102. Steps
At 114, the question items are rearranged in the order of the question items with the highest importance (= nI) for each rule, and the result is END.

After executing the program of this embodiment shown in FIG.
The rule that was in the state shown in FIG.
As shown in the figure, unnecessary question items are deleted, and each question item is rearranged in order of importance. The rule shown in FIG. 3 in the Prolog expression is set to the state shown in FIG. 6 after the execution of the program of this embodiment.

"Effects" As described above, the failure diagnosis device of the present invention uses the number of times the question item is used in how many rules,
The failure diagnosis device automatically determines whether the answer is true or false, based on the number of true and false answers, and the importance calculated from the evaluation value that evaluates the ease of obtaining the answer to the question item. Since the question items in the condition part of the rule are rearranged, the user can easily add, correct, and delete rules without reducing the efficiency of failure diagnosis. Further, since the failure diagnosis apparatus of the present invention has a table for storing information such as the number of times of use, the minimum value, and the evaluation value, each question item can be rearranged by changing the information in the table. In addition, there is an excellent effect that a failure diagnosis device can be provided according to the use condition of the user.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a fault diagnosis device according to an embodiment of the present invention, FIG. 2 is an explanatory diagram showing a state when a rule for concluding a fault element is created, and FIG. FIG. 4 is a flowchart showing a program according to the present embodiment, FIG. 5 is an explanatory diagram showing a rule state after the program is executed, and FIG. 6 is a diagram after the program is executed. FIG. 4 is an explanatory diagram showing a knowledge base in which rules are expressed in a computer language. 1 ... CPU, 2 ... RAM, 3 ... keyboard, 4 ... mouse, 5 ... CRT, 6 ... database.

Claims (1)

(57) [Claims] 1. A failure diagnosis which stores a rule comprising a condition part and a conclusion part indicating a conclusion when the condition part is satisfied in an external storage device as a knowledge base, and performs a failure diagnosis based on the knowledge base. In the device, a usage count means for counting the number of times the question items constituting the condition part are used in how many rules, and whether the answer of the question item is set to true or false in each rule Answer counting means for obtaining true and false numbers, evaluation means for giving an evaluation value for evaluating the ease of obtaining an answer for each question item, and the number of uses and the number of true and false answers and the evaluation value. A table to be stored, importance calculating means for calculating the importance of the question item from the number of times of use, the number of true and false answers, and the evaluation value stored in the table; Fault diagnosis apparatus characterized by comprising a rearranging means for rearranging the question items.