JPH0573319A - Data collecting method for diagnostic system - Google Patents

Abstract

PURPOSE:To obtain the data collecting method for diagnostic system capable of effective collection of data necessary for the diagnosis in an expert system performing the fault diagnosis of the large-scale institution. CONSTITUTION:Data successively collected from a diagnosis object 1 through a multiplexer 2 is collected to a data collecting device 3. In the case of performing a fault diagnosis in an inference engine 5, the questions for the diagnosis are divided into groups and the data related to all the questions of the group to which the data necessary for the question belongs is called. The diagnostic data suitable for the diagnosis is segmented with the use of the condition logical expression and the condition continuation time by a data processing task 9 to be stored in a diagnostic data memory 10. Thus, the stored diagnostic data becomes the general one with high possibility to be used more than once, and the number of transmissions and the transmission time can be reduced from the data collector.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a diagnostic system data collection method for collecting data necessary for diagnosis from an object to be diagnosed in an expert system for diagnosing equipment and the like.

[0002]

2. Description of the Related Art In a control system for performing sequence control, program control, etc. on equipment such as various plants, electronic devices, etc., and an expert system for performing fault diagnosis, tendency diagnosis, etc. on these control objects as diagnostic targets, Control and diagnosis are performed based on a large number of digital and analog data obtained from various sensors, switches, meters, actuators, etc. provided in the controlled object.

These digital and analog data are once collected by a data collecting device called a data logger.

Conventionally, the following three methods have been considered as methods for the expert system to collect data necessary for diagnosis.

(1) The expert system and the data collecting device share a computer so that the expert system can obtain necessary data at any time.

(2) Call the required data from the data collection device according to the diagnostic procedure of the expert system.

(3) A large amount of data is constantly transmitted and stored in the computer of the expert system so that the necessary data can be obtained when needed, as in the above (1).

[0008]

The method of (1) above is
When the diagnosis target is a large-scale facility or the like, the load on the computer for the diagnosis increases, and it is difficult to collect data while performing the diagnosis.

In the above method (2), since it takes time from when the required data is requested until the data is sent, the time required for the diagnosis becomes long.

In the above method (3), since a very small part of a large number of data that is generally transmitted becomes data required for diagnosis, useless data transmission and storage must be performed. It should be noted that it is not possible to select only the necessary data in advance, because which data will become clear as the diagnosis progresses.

Further, in the expert system, when one or more data are used for diagnosis, which part of the time change of the data is used as the diagnosis data effective for the diagnosis is determined by the diagnostic item (specifically, In the case of failure diagnosis, it depends on the question items etc. for inferring the cause of failure). For example, when analog data indicating the current value of the motor is used, whether the starting current portion of the change in the motor current is used as the diagnostic data or the stable current portion after the starting is used as the diagnostic data depends on the question item.

Conventionally, in order to extract diagnostic data, it has been necessary to individually create a large number of programs for determining which part of the data is to be extracted in response to a question for each of a large number of data. Was needed.

The present invention has been made to solve the above problems, and particularly in an expert system for diagnosing a large-scale equipment failure, a diagnostic system capable of efficiently collecting information necessary for the diagnosis. The purpose is to provide a method of collecting data.

[0014]

According to a first aspect of the present invention, a large number of data obtained from a diagnosis target are accumulated, and a large number of questions are divided into a plurality of groups. When specified, the data about all the questions of the group to which the question belongs is collected.

The invention of claim 2 collects a large number of data, performs arithmetic processing on each of the collected data based on a preset conditional logical expression, and accumulates the processed data. is there.

[0016]

In the invention of claim 1, for example, by grouping questions related to each other,
Collecting all the data for one group increases the probability of using those data, and the number of transmissions can be significantly reduced as compared with calling the required data one by one for each question content.

According to the second aspect of the present invention, it is checked whether or not the condition based on the generalized conditional logical expression is satisfied with respect to the raw data obtained from the object to be diagnosed, and the time when the condition is satisfied continues for a certain time as a reference. Then, the necessary data portion is determined and the portion is used as diagnostic data.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a block diagram showing the mutual relationships among various work parts of a program, a knowledge database, and a hardware part in an expert system.

In FIG. 1, the diagnosis target 1 of the expert system is also a control target of a control system (not shown), and is, for example, rolling equipment. A fixed number of digital data and analog data can be obtained from the diagnosis target 1. Digital data is ON for various sensors, switches, etc.
The analog data is current, voltage, rotation speed, etc. of the motor or the like.

A fixed number of these data obtained from the diagnosis object 1 are sent to the multiplexer 2 for a predetermined time (for example, 1).
0 msec) is sequentially sampled, collected by the data collection device 3, and stored in the memory. This collected data is used for control of the control system and for diagnosis of the expert system.

The diagnostic knowledge 4 is a knowledge database having a plurality of questions for performing a failure diagnosis, failure candidate conclusion data, a relationship between answers to each question and the above conclusion candidates, and the like. The inference engine 5 selects the question of the diagnostic knowledge 4,
It is a knowledge database that infers the cause of failure while sequentially selecting questions according to the answers to the questions.

Answer control knowledge 6 is divided into groups of questions,
The relation between each question and the data, the conditional logical expression for processing (cutting out) data to be described later to create diagnostic data,
Condition duration, conversion method to obtain the answer by further converting the diagnostic data into a form suitable for the answer, conversion method to convert the diagnostic data into display data suitable for the answer, operation of equipment to the operator necessary for data collection It is a knowledge database having knowledge data such as a data processing method for each instruction and question.

FIG. 3 shows an example of grouping of questions described in the answer control knowledge 6 and the relationship between each question and data.

In FIG. 3, questions 1, 2, 3, ...
Are divided into groups 1, 2, 3, ... M each consisting of a plurality of questions that are related to each other. Each question 1-n
The necessary data corresponding to is the data required to answer the question. As the necessary data, one data may be used in a plurality of groups as shown in FIG.

In the example of FIG. 3, the same data 5 is used for question 5 and question 6. It is assumed that this data 5 is, for example, data indicating a change in motor current. Further, the content of question 5 is "whether the peak value of the starting current of the motor current exceeds the reference value?", And the content of question 6 is "whether the stable current of the motor current exceeds the reference value?" ..

As described above, even if the same motor current data is used, the data may be used differently depending on the question. In this case, the necessary data is cut out from the same data depending on the question. Therefore, the answer control knowledge 6 describes the data processing method for each question.
The processing method is given to a data processing task 9 described later.

The answer task 7 performs various processes for obtaining an answer to the question selected by the inference engine 5. The transmission control task 8 performs a process of calling required data from the data collection device 3 in response to a request from the reply task 7 and sending the data to a data processing task 9 described later.

The data processing task 9 processes the data sent from the transmission control task 8 by using the conditional logical expression and the conditional duration given from the answer control knowledge 6 in response to the question, thereby processing the diagnostic data. create. The diagnostic data memory 10 stores the diagnostic data created by the data processing task 9.

The automatic conversion task 11 is the diagnostic data memory 1
The diagnostic data stored in 0 is converted into a form suitable for the answer based on the conversion method given from the answer control knowledge 6 according to the question, and the answer is obtained, and the answer task 7 is passed.

The display conversion task 12 is the diagnostic data memory 1
The diagnostic data accumulated in 0 or the data passed from the answer task 7 is displayed on the screen based on the display method given from the answer control knowledge 6 according to the question, and the answer entered by the operator is displayed according to the display. Pass to answer task 7.

The operation command task 13 displays the operation command on the screen according to the instruction of the answer task 7.

Next, the operation of the above configuration will be described.

When a failure occurs in the diagnosis object 1, the failure diagnosis is executed. The inference engine 5 executes the diagnostic knowledge 4 according to the inference.
The questions are sequentially selected from and the answer task 7 is asked for an answer.
The answer task 7 refers to the answer control knowledge 6 to check whether the diagnostic data necessary for the question is stored in the diagnostic data memory 10. If there is no necessary diagnostic data, the transmission control system 8 is instructed to collect the data. Alternatively, if necessary, the operation command task 13 is caused to display an instruction to operate the diagnosis target 1 to obtain necessary data. In response to this, a part of the diagnosis target 1, for example, the motor is manually rotated in the normal direction, and the data specified during that time is collected.

The transmission control task 8 selects and calls the instructed data from the data collection device 3, and at this time calls the necessary data of all the questions of the group to which the question using the instructed data belongs. The data of the called group is sent to the data processing task 9, where the data necessary for diagnosis is cut out as described later, and the diagnostic data is cut out and stored in the diagnostic data memory 10.

This diagnostic data is sent to the automatic conversion task 11 or the display conversion task 12 according to the question. The automatic conversion task 11 further processes the diagnostic data by the conversion method passed from the answer control knowledge 6 to the transmitted diagnostic data according to the question, and the processing result is, for example, YES.
Alternatively, a NO answer is sought and passed to the answer task 7.

When the operator needs to display the diagnostic data on the screen and judge it, the display conversion task 12 displays the diagnostic data on the screen based on the screen display method passed from the answer control knowledge 6 in response to the question. indicate. For example, when analog data such as voltage and current contains noise, it may be better to display it on a screen and judge it with human eyes. Then, the display conversion task 12 receives the input of the answer of the operator who judges by looking at the display data,
The answer is passed to the answer task 7.

The answer task 7 passes the answer input from the automatic conversion task 11 or the operator to the inference engine 5, and the inference engine 5 makes an inference according to the answer and obtains the cause of failure from the conclusion candidates. If no conclusion is obtained, select the next question from diagnostic knowledge 4.

By repeating the above operation for each question, one conclusion candidate is finally selected and the cause of failure is determined.

As described above, in this embodiment, when a large amount of data is collected and stored in the data collection device 3, when one question requires certain data, all questions in the group to which the question belongs are collected. The data to be used is called, sent to the data processing task 9, processed into diagnostic data, and then stored in the memory 10.

Therefore, the memory 10 stores diagnostic data which are related to each other and therefore have a high probability of being used at least once. That is, since the data with a high utilization rate is collected at one time, the number of data transmissions is significantly reduced as compared with the conventional method in which only required data is called each time. Further, in general, there is no great difference between the time required to send or send one piece of data and the time taken to send a plurality of pieces of data from each other. , It is much more efficient in time to retrieve all the data of a group of related questions.

Next, the data cutting processing in the data processing task 9 will be described.

The data collected by the data collecting device 3 is
It is a fixed large number of raw data obtained at fixed time intervals (10 msec) from sensors, switches, etc. provided in the diagnosis target 1, and it is necessary to convert this raw data into diagnosis data suitable for diagnosis. Here, data indicating the current value of the motor will be described as an example of one piece of analog data.

FIG. 2 shows changes in raw data of the motor current value. FIG. 2B shows a reference value, and FIG. 2A shows an actual current flowing with respect to the reference value.

From these raw data, the data processing task 9 first detects that this condition is satisfied, on condition that the reference value exceeds the 100% level. Next, it is detected that this detection time is set as the reference time t ₀ and the above condition continues for a predetermined time (here, 5 seconds until time t ₃ ).

Then, using the condition continuation portion for 5 seconds as the reference portion, the data portion for 2 seconds from the reference time t ₀ to the time point t ₁ of the reference portion is cut out as the starting current data. Also, cut out the data portion of 2 seconds until t ₁ ~t ₂ point of the reference portion as a stabilizing current data.

These starting current data and stable current data are stored in the memory 10 as diagnostic data, and are used in response to a question when making a fault diagnosis.

In the case of digital data, for example, the combination of ON / OFF states of each of a plurality of switches is used as a condition to detect the satisfaction of the condition, and the necessary diagnostic data is obtained by using the part where the condition continues for a predetermined time as a reference part. Cut out.

The conditional logical expression and the conditional duration required for performing the above-mentioned clipping are described in the answer control knowledge 6 in a generalized form in response to the question. Therefore, for example, it is possible to detect the reference time using the same conditional logical expression and the condition continuation time and obtain the diagnostic data capable of responding to a plurality of questions only by using different cutout times. Can be converted.

[0049]

According to the present invention, since the number of times of data transmission from the data collecting device to the expert system is greatly reduced and the data transmission time is shortened, the data collecting device and the expert system are different. Data can be efficiently collected using a computer, and failure diagnosis can be promptly executed. Also,
Since the process of creating diagnostic data necessary for diagnosis from raw data can be generalized, it is possible to save the time and labor for creating many programs as in the conventional case.

[Brief description of drawings]

FIG. 1 is a configuration diagram of an embodiment of the present invention.

FIG. 2 is a characteristic diagram illustrating a process of cutting out diagnostic data from raw data.

FIG. 3 is a diagram showing an example of grouping of questions described in answer control knowledge and a relationship between each question and data.

[Explanation of symbols]

 1 diagnosis target 3 data collection device 4 diagnostic knowledge 5 inference engine 6 answer control knowledge 9 data processing task 10 diagnostic data memory

Claims (2)

[Claims] 1. When a failure to be diagnosed occurs, one question is selected from a plurality of questions prepared in advance to obtain an answer, inference is performed according to the answer, and the next question is selected according to the inference result. Alternatively, in a diagnostic system in which a failure cause is obtained from a plurality of failure cause candidates prepared in advance, a large number of data indicating the state of the diagnosis target obtained from the diagnosis target is collected and accumulated, and The data required to answer each of the above-mentioned questions are selected from the data of the above, and a knowledge database in which each question is described is provided, and the above-mentioned questions are divided into a plurality of groups. When the required data is specified from the data according to the question, refer to the above knowledge database and collect the data required for all the questions of the group to which the above question belongs. Data collection method of the diagnostic system to be butterflies. 2. Collecting a large number of data indicating the state of this diagnostic object obtained from the diagnostic object, detecting that a preset conditional logical expression is satisfied for each of the collected data, and said conditional logical expression Data collection for a diagnostic system characterized by investigating whether or not it is established for a certain period of time, determining the necessary data part from the established part, and extracting and storing the data of the determined part Method.