JPH02155006A - Knowledge base type process operating/supporting device - Google Patents

Abstract

PURPOSE:To display not only an instruction in a sentence but also an inference ground to have outputted the instruction by applying inference ground image data to be the set of data, which are obtained with dividing the inference ground into plural hierarchies and made into images, to a computer. CONSTITUTION:The knowledge of specialists concerning the operation and supervision of a process is accumulated in a computer 3, and the knowledge of the specialists are inputted to the computer 3 in a method to input the knowledge as image data 8 arranged in a tree-form as well as in a method to input the knowledge in a sentence format from a console, etc. Further, the inference ground image data are obtained by hierarchizing the inference ground image, and the respective image data are inputted from an image data input device 4. Thus, by using process data inputted from a process data input device 5, the computer 3 operates and supports the process through a CRT 2 according to the accumulated knowledge, and the suitably detailed inference ground can be displayed.

Description

[Detailed description of the invention] [Industrial application field]

The present invention relates to a knowledge-based process operation support device that uses an expert system to provide an operator with appropriate process operation instructions.

[Conventional technology]

FIG. 4 is a block diagram showing the configuration of a conventional knowledge-based process operation support device. In the figure, 2 is a CRT (display device) that gives operation instructions to the operator, 3 is a computer having reasoning means, and 5 is a block diagram showing the configuration of a conventional knowledge-based process operation support device. A process data input device 6 provides process data to the computer 3, and a sensor 6 measures the process data. Further, 1 is an operator who operates and manages a process (not shown), 7 is an expert on process operation and management, and 13 is a knowledge engineer who is familiar with inference means. Next, the operation will be explained. The computer 3 stores the knowledge of the expert 7 regarding the operational management of this process, classified into rules and facts. The knowledge engineer extracts the knowledge of the expert 7 through interviews and the like, organizes it into a format that can be input into the computer 3, and performs input work. When process data is measured by the sensor 6 installed in the process, this process data is converted by the process data input device 5 into data that can be handled by the computer 3 and is input to the computer 3. At the request of the operator 1, the computer 3 uses the input process data and the facts in the accumulated knowledge, according to the rules in the accumulated knowledge, and as appropriate through the screen of the CRT 2. 1 in a conversational format, selects appropriate instructions for operation management, and outputs them to the CRT 2. In this way, support for operational management is provided.

[Problem to be solved by the invention]

Since the conventional knowledge-based process operation support device is configured as described above, only a simple written instruction can be obtained in response to a conversational request from the operator I.
The operator 1 is not notified of the inference basis in the inference process that outputs the instruction, and furthermore, the input of the expert's knowledge into the computer 3 is performed by a knowledgeable person who has specialized knowledge and is familiar with the inference means.・There was an issue that 13 engineers were required. This invention was made to solve the above-mentioned problems, and it is possible to display not only written instructions but also the inference basis for outputting the instructions according to the level of the operator. Changes or additions to the basis can be made using Knowledge.
The purpose is to obtain a knowledge-based process operation support device that can be easily operated without an engineer. Note that a similar prior art is disclosed in Japanese Patent Laid-Open No. 60-61840.

[Means to solve the problem]

The knowledge-based process operation support device according to this invention includes:
Inference means using a knowledge structure tree based on image data on a computer and other inference means, and inference basis image data, which is a collection of data that is an image of the inference grounds in each inference process of these inference means divided into multiple layers. The computer is equipped with an image input device for inputting the knowledge structure tree and the imaged inference grounds into the computer.

[Effect]

The computer according to the present invention inputs a knowledge structure tree, which is the main part of an inference means using a knowledge structure tree, and an imaged inference basis from an image input device, and proceeds with inference using the input knowledge structure tree. , the inference basis visualized as an image is displayed on the display device according to the operator's request.

【Example】

An embodiment of the present invention will be described below with reference to the drawings. 1st
In the figure, 4 is an image data input device that inputs image data into the computer 3, 8 is image data that is an image of the knowledge of an expert regarding process operation management, and the other items are given the same reference numerals as in Figure 4. This is the same as shown in the figure below. FIG. 2 is a block diagram showing the software configuration of the computer 3, where 9 is a man-machine interface section that handles input/output information for the operator, 10 is an inference section that makes inferences, and 11 is a man-machine interface section that handles inference basis image data 12. This is a knowledge image data management unit that hands over the information to the intertooth unit 9. Next, the operation will be explained. The computer 3 stores expert knowledge regarding the operational management of this process. The expert's knowledge is input into the computer 3 by inputting the knowledge in the form of text from a console or the like as in the past, as well as by inputting the knowledge as image data 8 arranged in a tree shape. The nodes and techniques of the tree on this image data correspond to separately stored facts and rules that indicate the relationships between these facts. This correspondence is established by specifying the coordinates of the nodes and techniques on the image. The inference basis image data 12 is obtained by layering the inference grounds as shown in FIG. 3, and inputting each layered layer as image data from the image data input device 4. Here, the hierarchical reasoning basis is one that explains the rules in an easy-to-understand manner even for non-experts, for example, by using illustrations, etc.
The lower the level, the more detailed it becomes. Regarding the correspondence with the rules and the hierarchical structure, a table showing each is prepared in the memory (not shown) of the computer 3. Next, during process operation, process data is measured by the sensor 6, appropriately converted by the process data input device 5, and input to the computer 3. At the request of the operator 1, the computer 3 operates the CRT 2 according to the accumulated knowledge while using the input process data.
Support operations by providing appropriate instructions regarding process operation management through Inference regarding operation management is performed by an inference unit 10a based on knowledge rules, an inference unit 10b using a knowledge structure tree, or a tabular inference unit 10c shown in FIG. Operation R1 activates the inference unit having the currently optimal method among these. Here, the inference unit 10b using the knowledge structure tree has the above tree-shaped image data 8. The inference unit 10b using this knowledge structure tree follows the tree and determines the optimal instructions based on the input plant data and a table specifying the correspondence between rules and coordinates, and then sends the instructions via the man-machine interface unit 9. Display on CRT2. Further, the tabular reasoning unit 10c has a table in which input states and instructions are arranged in a matrix, and selects the most suitable instruction based on the input state at that time. When outputting the basis of inference as requested by operation R1,
A request detected by the man-machine interface section 9 is transmitted to an inference section 10a, 10b, or 10c that performs an inference operation. The inference unit 10a, 10b, or 10c instructs the knowledge image data management unit 11 to provide the image data 12a of the inference basis of the first layer corresponding to the rule, and the knowledge image data management unit 11 instructs the man-machine interface unit 9. This is displayed on the CRT2 via. When the operator 1 requests the output of the inference grounds again, the next layer of image data 12b, 1 illustrating more detailed grounds is displayed.
2c or 12d is displayed on the CRT2. Further, the tree-shaped image data 8 and the inference basis image data 12 can be easily changed or added through the image data input device 4. Note that although the above embodiment describes process operation support, it may also be used for process operation training education, and the same effects as in the above embodiment can be achieved.

【Effect of the invention】

As described above, according to the present invention, a knowledge-based process operation support device is implemented by a computer having a plurality of inference means and inference basis image data which is an image of the inference grounds in the inference process of the inference means divided into a plurality of layers. , and an image data input device for inputting the image data, it is possible to provide support with optimal instructions depending on the situation, and to display reasoning grounds in appropriate detail on the display device according to the operator's request. Moreover, it is easy to change or add inference grounds, and there is an effect that it is easy for the operator to handle.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of a knowledge-based process operation device according to an embodiment of the present invention, FIG. 2 is a block diagram showing the software configuration of the computer shown in FIG. 1, and FIG. 3 is an image of the basis for inference. FIG. 4, which is a conceptual diagram for explaining data hierarchy, is a block diagram showing the configuration of a conventional knowledge-based process operation device. 2 is a CRT (display device), 3 is a computer, 4 is an image data input device, 5 is a process data input device,
6 is a sensor, 10 is an inference unit, 10a is an inference unit based on knowledge rules, 10b is an inference unit using a knowledge structure tree, 10c is a tabular inference unit, and 12 is inference basis image data. In addition, in the figures, the same reference numerals indicate the same or equivalent parts.

Claims (1)

[Claims] A plurality of inference means including an inference means using a knowledge structure tree based on image data, and inference basis image data that is a collection of data obtained by dividing the inference grounds in each inference process of the plurality of inference means into a plurality of hierarchies and converting them into images. an image data input device for inputting the knowledge structure tree and the imaged inference basis into the computer; a display device for the computer to provide instructions for appropriate process operation management to an operator; A knowledge-based process operation support device comprising a sensor and a process data input device for inputting to the computer.