JPH01177199A - Development supporting device for abnormality diagnostic system - Google Patents

Abstract

PURPOSE:To develop a high-speed diagnosing system having high diagnostic accuracy by executing an automating processing by a computer using the applicability of respective diagnostic element techniques and knowledge concerning the combination and integration of the respective techniques. CONSTITUTION:An interface mechanism is composed of an inference integrating part 101 and five applicability deciding parts 110, 120, 130, 140, and 150 to decide the applicability of an upper and lower limit checking technique, a balance checking technique, a redundant signal comparing technique, a model comparing technique, and a knowledge engineering technique respectively, and the respective parts of the inference mechanism are integrated through a working memory area 103 to store the processed results of the respective parts. Consequently, the applicability of the respective techniques can be judged precisely, and the optimal combination and integration of the respective techniques can be attained by the knowledge of a knowledge-base operation. Thus, the support concerning the various techniques and their integration can be attained, the level of which is almost the same as that by a skilled technician.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a support device applied to the development of an abnormality diagnosis system for plants and machinery.

[Prior art] Various elemental methods have been developed and applied in abnormality diagnosis of plants and machinery, such as upper and lower limit checking methods, balance check methods, redundant signal comparison methods, model comparison methods, and knowledge engineering methods. There is. However, there is no objective standard as to which elemental method should be used to make the required diagnosis with the required speed and certainty, and the selection is left entirely to human judgment.

Next, we will provide an overview of conventional diagnostic elemental technologies.

(1) Upper and lower limit checking method is a method of detecting abnormalities by monitoring process quantities such as pressure and temperature measured in plants and machinery to see if they deviate from predetermined normal ranges. be. Rather than the amount of process itself,
There are also improvements, such as monitoring the rate of change.

(2) Balance check method is to monitor whether multiple related l[M are balanced, such as mass balance and energy balance, such as inflow flow rate and outflow flow rate must be balanced. This is a method for detecting anomalies and identifying locations where anomalies occur. A static balance check is usually performed.

(3) Redundant signal comparison method involves measuring the same physical quantity with multiple measuring instruments, and by performing the same calculation using multiple different algorithms, multiple redundant signals and signals can be obtained for one quantity. , is a method of diagnosing abnormalities in measuring instruments, computers, and arithmetic methods by comparing these.

(4) The model comparison method uses a model that shows the behavior of plants and machinery under normal conditions, and diagnoses abnormalities by monitoring the difference between the calculated value of the physical quantity calculated using this model and the calculated value of that physical quantity. This is a method to do this.

A dynamic characteristic model is usually used as the model.

(5) Knowledge engineering method is the computerization of the knowledge and thinking methods using knowledge used by anomaly diagnosis experts in detecting anomalies, identifying causes, and determining countermeasures in the form of knowledge bases and inference programs, respectively. This is a method that uses computers to perform the diagnostic process performed by experts.

[Problem that the invention attempts to solve] The drawbacks of the conventional method are listed below.

(1) The required speed cannot be achieved because an inappropriate method has been selected.

(2) An incorrect diagnosis is made because an inappropriate method is selected.

(3) Because it was not possible to select an appropriate method, it was not possible to detect an abnormality, identify the cause, or decide on countermeasures, which could lead to serious damage or accidents.

[Means for solving the problem] The present invention uses an upper and lower limit checking method as a diagnostic element method.
We will focus on the balance check method, redundant signal comparison method, model comparison method, and knowledge engineering method. We support the development of an optimal abnormality diagnosis system by converting knowledge about combinations and integration into a knowledge-paced system and using inference techniques of knowledge engineering while interacting with users.

[Effect] With the above knowledge, it is possible to accurately judge the applicability of each method,
The knowledge of the above-mentioned knowledge pacing enables optimal combination and integration, and has the effect of providing support similar to that of a skilled professional engineer regarding various methods and their integration.

[Example] 1) Equipment configuration An embodiment of the invention is shown in FIG.

A central processing unit (hereinafter abbreviated as CPU) 1 is connected to a shared memory 3, an input device 4, and an output device 5 via a data bus 2. Furthermore, distributed processing units (hereinafter abbreviated as DPUs) 10, 20, 30, 40 and 50 are connected to the data bus 2, and each DPU has a memory (MA) that stores knowledge regarding the characteristics of diagnostic methods. ) 11
, 21 , 31 , 41 and 5-1, and memory (MB) 12, 22.32 for storing past usage experience.
42 and 52 are in succession. CPU 1 has a memo IJ (MC) that stores knowledge regarding method integration.
e is connected.

The CPU 1 inputs data from the input device 4 via the data bus 2, and each DPU 10, 20, jO, 4
0 and 50 via the data bus 2. Each DPU J O, 20, J O, 40 and 50
In response to this command, memory 11, 21, and 31 respectively
．． Inference processing is performed using the knowledge bases stored in 41 and 51 and the empirical data stored in memories 12, 22, 32. 42 and 52, and the respective processing results are shared via the data bus 2. Output to memory 3. The CPU 1 integrates the contents of the shared memory 3 by referring to the knowledge base stored in the memory 6, and outputs the integrated result to the output device 5 via the data bus 2, and also writes it to the shared memory 3 at the same time. Each DPU10.20
, 30, 40 and 50 refer to this and update the contents of memories 12, 22, 32, 42 and 52, respectively, which store usage experience.

2) Processing details of the inference mechanism As shown in FIG. 2, the inference mechanism includes an inference control unit 101 and the application of the upper and lower limit check method, the balance check method, the redundant signal comparison method, the model comparison method, and the knowledge engineering method. Five applicability determining units 110, 120, 1
30, 140, and 150 are integrated via a working memory area 103 that stores the processing results of each section.

The processing of the inference control unit is executed on the CPU I using the knowledge in the knowledge base memory 6. Five applicability determination parts 1
10, 120, 130, 140 and 150 process knowledge in knowledge base memories 11, 21, 31, 41 and 51, and experience data memories 12, 22, 32, .
Using the data in 42 and 52, DPU 10,
Runs on 20, 30, 40 and 50.

” The working memory area 103 is provided on the shared memory 3.

(1) Processing of the inference control unit The inference control unit 101 executes the following processing upon startup of this device.

(1.1) Each applicability determination unit 110 , 120 , 130
At 140 and 150, an execution command for the applicability determination process described in (2) is issued.

(1.2) Check whether the applicability determination process for all elemental methods has been completed, and if it has been completed, proceed to steps (1, 6) described below.
fly to

(1, 3) If the above steps (1, 6) have not been completed, check whether there is a question request from each applicability determination section (steps (2, 4) described below), If there is no question request, the process returns to Stella 7' (1, 2).

(1, 4) If there is a question request here, the question is issued via the output device 5 and the answer is received via the input device 4.

(1, 5) Write the answer to the working memory area 103.

(1, 6) According to the method integration knowledge base, a draft of the integrated abnormality diagnosis system is created in the working memory area 103 by referring to the degree of applicability of each element method written in the working memory area 103.

(1, 7) At the same time as outputting the draft created in steps (1, 6) via the output device 5, the user is asked whether or not the draft needs to be modified.

(1, 8) Through the input device 4, input whether or not correction is necessary and correction items if correction is necessary.

(1, 9) Record the final draft with necessary modifications in the working memory area 103.

(1, 10) The final draft created in Stella f (1, 9) above is output via the output device 5.

' (1, 11) Finally, each applicability determination unit 110, 1
20° 130. Issue an execution command for the experience data update process to 140 and 150.

(2) Processing of applicability determination unit Applicability determination unit 110, 120° 130 for each diagnostic element method
．． 140 and 150 have the same processing content, except that the knowledge base and experience data used are different.

(2, 1) When the inference control unit 101 receives an execution command, the following steps (2, 2) or (2) are performed depending on whether it is an applicability determination process or an experience data update process.
, 8).

(2, 2) Applicability determination is performed using the applicability determination knowledge base for three purposes: abnormality detection, cause identification, and countermeasure determination.

(2, 3) First, select condition items for determining applicability.

(2, 4) At this time, if there is a condition for which no value is given, the inference control unit 1
Send a question request to 01 and get the value.

(2, 5) Once values are given to all conditions, the applicability of each elemental method to abnormality detection, cause identification, and countermeasure determination is determined using knowledge regarding the applicability determination method. Here, the applicability means +1.0 is the most appropriate, -
1, 0 is the most inappropriate, 0 is undeterminable, +1.0 to -1.
Values between 0 indicate the degree of appropriateness/inappropriateness depending on the size.

(2, 6) Extract past experience data from the experience data memory.

(2,7) The degree of applicability determined in step (2,5) above,
The experience data obtained in steps (2, 6) above is written to the working memory area 103, and the applicability determination process is completed.

(2, 8) Update the number of times each element method is used and the number of times it is combined with the detailed method with reference to the final plan in the working memory.

3) Knowledge base configuration example The knowledge base is a knowledge base that determines the applicability of the upper and lower limit check method, balance check method, redundant signal comparison method, model comparison method, and knowledge engineering method, and combinations among these methods. It consists of a knowledge base related to the integration of methods and techniques.

(1) Applicability judgment knowledge base Examples of the applicability judgment knowledge base for each diagnostic element method are shown in Table 1.
It is shown in Table 5. This knowledge base consists of knowledge about:

a) Condition items for determining the applicability of elemental methods This includes the prerequisites for applying the method, the usage environment such as a computer,
It is classified into the measurability of necessary data and the definition of normal values that are the basis for abnormal diagnosis.

b) A method of determining the degree of applicability of the method by combining the condition items mentioned in the previous section. This method uses a pair of determination conditions and degree of applicability when the conditions are met for each purpose of anomaly detection, cause identification, and countermeasure determination. Multiple items are given. However, the determination conditions are determined in order from the top, and if one is found that holds true, the degree of applicability of that pair is determined as the determination result, and the process ends. In other words, any decision conditions below this are ignored.

(2) Method integrated knowledge pace An example of method integrated knowledge pace is shown in Table 6. This knowledge base is knowledge for integrating the results of determining the applicability of each element method using the knowledge base described in item (1) above.

4) Contents of Working Memory Area The working memory area 103 consists of a question and answer data section 201 and an integrated system data section 202, as shown in FIG.

(1) As shown in FIG. 3, the question and answer data section 201 can store multiple pairs of question data and response data. The ones that are not shown are those that have a question request.
The response data also contains a value when a response to the question is given.

(2) The integrated system data section 202 further includes a draft section 203.
and final draft section 204, each having the same data structure.

The data structure is as shown in Figure 3. a) Name of method used for anomaly detection, degree of application, and number of times used b) Name of method used for cause identification, degree of applicability, and number of times used C) Name of method used to determine countermeasures, degree of applicability, and number of times used can be stored.

Table 1 Example of knowledge base for determining applicability of upper and lower limit checking method 2nd
Example of knowledge pace for determining applicability of table balance check method
3 Table Example of knowledge pace for determining applicability of redundant signal comparison method No. 4
Table 5 Table 6 Example of Element Method Integrated Knowledge Pace [Effects of the Invention] By using the development support device for an abnormality diagnosis system according to the present invention described above, the following effects are achieved.

(Li) Applicability of each diagnostic element method and their combination/
By using knowledge about integration and performing automated processing using a computer, it is possible to develop a diagnostic system with high diagnostic accuracy and high speed.

(2) When a new method is devised, device functions can be easily expanded by simply creating a knowledge base for applicability determination regarding the method and adding a portion of the integrated knowledge base.

(3) Even when the applicability judgment result of the one-child method or the integrated method result is unfavorable, related knowledge is localized to the applicability judgment knowledge pace or integrated knowledge pace of each method. Maintenance and improvement become easier.

Data for this maintenance and improvement can be retrieved from the experience data memory.

[Brief explanation of the drawing]

Figure M1 is a block diagram showing the system configuration in one embodiment of the present invention, Figure 2 is a block diagram showing the configuration of the inference mechanism in the above embodiment, and Figure 3 is a working memory in the above embodiment. FIG. 3 is a diagram showing the configuration of an area. 1...Central processing unit (CPU), 2...Data bus (DATA-BUS), 3...Shared memory, 4...
- Input device, 5... Output device, 6... Knowledge pace memory (MC), 10°20.30, 40.50... Distributed processing device, 11°21.31, 41.51... Knowledge Pace Memory (MA), 12.22, 32e42.52
...Experiential data memory (MB), 701...Inference control department, 103...Working memo +), 110...
Upper and lower limit check, method applicability determining unit, 120... Balance check method applicability determining unit, 130... Redundant signal comparison method applicability determining unit, 140... Model comparison method applicability determining unit, 150.・Knowledge Engineering Method Applicability Judgment Department, 201・
...Question response 1f, 7'' (-evening section, 202...Integrated system data section, 2Q3...Draft section, 204...Final draft section. Figure 1

Claims (1)

[Claims] First storage means of the same number as the number of the above-mentioned elemental methods storing knowledge regarding the applicability of the abnormality diagnosis elemental methods such as the upper and lower limit check method, the balance check method, the redundant signal comparison method, the model comparison method, and the knowledge engineering method. and a number of distributed processing devices equal to the number of the first storage means that calculate the application effectiveness of each elemental method using the knowledge stored in these storage means, and store the experience of using each of the above methods. 1st above for
a central processing unit that integrates the calculation results of the distributed processing device, an input device that inputs information to the central processing unit, and the central processing unit; 1. A development support device for an abnormality diagnosis system, comprising: an output device for outputting information on processing results; and a bus for transmitting and receiving information between each of the devices.