JP2635567B2 - Plant operation guide device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention] (Industrial application field) The present invention relates to a plant operation guide device for instructing an operator to perform an appropriate and flexible operation guide in response to a plant state during plant operation. .

(Prior art) The current plant has a very complicated system due to its increase in size and diversification of functions, and it is not possible for an operator to accurately grasp and operate the entire plant during normal and abnormal situations. Is becoming an ever more burdensome. For the purpose of reducing the burden, automation of plant operation and enhancement of an operation guide presentation function for operators have been attempted in recent years.

(Problems to be Solved by the Invention) However, in constructing these automation and operation guide functions, conventionally, a certain scenario of plant operation or abnormal event occurrence is assumed, and means for dealing with the scenario are loaded into the system. Due to the approach taken, it is necessary to ensure the integrity of the whole automation or guide logic at the time of design when building up the system, and to respond to the occurrence of events that have never been experienced before. Some problems have been pointed out, such as difficulty in flexible response. As a method to overcome these problems,
In recent years, many applications of expert systems using knowledge databases have been tried. The purpose of these expert systems is to flexibly respond to plant operation events by loading the experience and knowledge of operators, knowledge of system configuration, and the like as a database. However, in most cases, the databases that form the basis of these expert systems and the inference engines that process the databases are all diverted from those developed for dialogue-based systems assuming human intervention. The response time of the system to changes in the external world requires several seconds to several tens of seconds, and is required for operation guide systems that require a response within at least one second, and plant control automation systems that require a faster response on the order of milliseconds. Difficult to apply.

FIG. 2 shows an example of a conventional system. In this system, plant information input to a computer from a sensor 2 of a plant 1 via a plant input / output device 3 is stored in a storage device 13 as a process database 5. Man-machine input device is used for knowledge of operators and plant structure
11 and input to the knowledge database 6 by the knowledge database construction device 9 via the man-machine input / output processing device 10.
Is stored in the storage device. In the central processing unit 8, the inference engine 7 estimates and identifies the plant state based on the process database 5 and the knowledge database 6, and selects an optimal operation guide. The operation guide message thus determined is output to the operator by the operation guide output device 12 via the man-machine input / output device 10. In the conventional expert system, real-time processing is practically difficult because a language having a low processing speed such as LISP or PROLOG is used for the description of the knowledge database 6 and the inference engine 7.

[Configuration of the invention]

(Means for Solving the Problems) The configuration of the present invention is roughly divided into two parts. One is an off-line computer system whose main purpose is to build a knowledge database. Here, a knowledge database for the entire plant is built via a man-machine system. For this purpose, the knowledge database and inference engine are, for example, LISP
It is described in a language that focuses on ease of system construction, such as and PROLOG. The other component is an online computer connected to this offline computer system via a transmission line. The main point of this system is to execute an operation guide or control in response to changes in the plant, and the database and the inference engine in this case operate the database and the inference engine of the offline computer system most efficiently in a real-time environment. It is assumed that it is described in a language such as FORTRAN, C, or assembler.

This online computer system has a man-machine interface for presenting an operator guide.

(Operation) An operation guide system construction function centered on a user-friendly man-machine system that can easily build an appropriate plant operation guide system and easily change and add a knowledge database follows changes in the real-time of the plant. A plant operation guide system that presents an operation guide in a timely manner and also has an online operation guide function can be realized.

(Embodiment) FIG. 1 shows the configuration of the embodiment. The embodiment comprises an online computer and an offline computer. The online computer inputs from the information sensor 2 of the plant 1 at every moment through the process input / output device 3, and inputs this information to the online process database in the online computer / storage device 113. 1
Store in 05. The online computer further has an online knowledge database 106, and an online inference engine 107 runs in the online computer central processing unit 108. The online man-machine input device 111 and the online operation guide output device 112 are connected to an online computer central processing unit 108 via an online man-machine input / output processing device 110. This connection is interposed in the online knowledge database construction device 109. The off-line computer system is connected to the on-line computer via a transmission device 15 and a database conversion device 14, and has an off-line process database 5 and an off-line knowledge database 6 in an off-line computer storage device 13; , The offline inference engine 7 runs. The off-line man-machine input device 11 and the off-line operation guide output device 12 are connected to the off-line computer central processing unit 8 via the off-line man-machine input / output processing device 10 and the off-line knowledge database construction device 9.

Next, the operation of the present invention will be described with reference to FIG. First, an offline plant operation guide device is created on an offline computer as follows. First, knowledge of a plant is interactively input by a knowledge engineer, an operator, a designer, or the like using an off-line man-machine input device such as a CRT, a keyboard, and a mouse. This knowledge database construction data is written to the offline knowledge database 6 via the offline man-machine input / output processing device 10 and the offline knowledge database construction device 9 in a predetermined description format. Offline inference engine 7
Uses the latest database data transmitted from the knowledge database 6 and the online computer and written in the offline process database to select an optimal plant operation guide message related to the operation of the plant at present or in the future, and converts this information to the offline. Offline operation guide output device 1 such as CRT via man-machine input / output processing device 10
2. To the operator in the form of a message, graphic display, etc. The operator carries out plant operation based on this message.

Here, the information stored in the off-line knowledge database 6 mainly includes plant operation operations, plant dynamic characteristics, various empirical rules, and calculation knowledge of the structure and operation of the plant and the plant equipment that the operator has. In the work of the database, the database work can be efficiently performed by using the offline knowledge database construction device 9. This database uses linguistic expressions suitable for knowledge description such as LISP and PROLOG. The information in the off-line knowledge database 6 is converted by the data converter 14 into an expression such as FORTRAN or C language suitable for on-line real-time processing, and then down-line loaded to the on-line knowledge database 106 via the transmission device 15. . The online inference engine 107 makes inferences using both this data and the online process database 105 composed of the information of the sensor 2 taken in from the plant 1 via the process input / output device 3 and an operation guide message optimal for the current plant state. And displays it in the form of a message or graphic information on an online operation guide output device 112 such as a CRT via the online man-machine input / output processing device 110 to guide the operator. Further, changes and additions to the online knowledge database 106 can be made by using an online man-machine input device 111 such as a keyboard and an online knowledge database construction device 109. On the other hand, the online knowledge database 106 after the change is added can be transferred to the offline database 6 via the transmission device 15 and the database conversion device.

One embodiment of the data conversion is described below. The knowledge database in the off-line computer system is given by a rule. For example, it is assumed that rule A has the knowledge that "if x = a, y = b", and in the off-line computer, this knowledge is expressed as follows. .

A = (x: a, y: b)-The data conversion unit executes the following conversion F.

The online computer adds this a to the execution source statement.

By connecting the offline computer and the online computer, and sharing the main functions of each in the form of a knowledge database construction and an online operation guide, it has the advantage of flexible and easy knowledge database construction, and is sufficient for plant changes. A system that can present a driving guide to the operator at a speed that can be followed has been realized, thereby reducing the burden on the operator and ensuring safe and flexible driving even if it is not a seasoned operator. Cost reduction is realized by high efficiency operation and labor cost saving.

In the case of a large-scale system, there is a possibility that responsiveness may be delayed due to an increase in processing load when processing the entire system with a set of online computer systems. In order to avoid this, a distributed configuration as shown in FIG. 4 is adopted. Here, the plant 1 is divided into N relatively mutually independent subsystems 1 to N, and an operation guide is executed for each subsystem. 1 for subsystem
N online operation guide computer systems are installed corresponding to .about.N. Each online computer system has the same configuration and the first online computer system
Taking 21 as an example, the configuration is a central processing unit 211,
It comprises an operation guide output device 212, a man-machine input device 213, a process input device 214, a storage device 215, a transmission device 216 to an offline computer, and an online knowledge database 218 relating to the subsystem 1. Offline computer system whose main purpose is to build a knowledge database for all plants
303 has the following configuration. A central processing unit 31, an operation guide output device 32, a man-machine input device 33, a storage device 35, a transmission device 36 with an online computer group, a data conversion device 37, and a knowledge database 38. This knowledge database includes sub-databases 1 to N (381 to 3) corresponding to each subsystem.
8N), and in addition to this, a common knowledge database on operation while maintaining the balance of the entire plant
There are 380. Subdatabases 1 to N (381 to 38N)
Is produced by the offline computer system, and is converted into real-time language data by the data conversion device 37, and is stored in the online computer system group storage device (216-2N6) via the transmission device 36 and the transmission device (216-2N6). 215-2N5)
Downline loading to. Each online computer system uses this knowledge database (218-2N8) and subsystems (11-1N) to process input / output devices (214-2N).
The operation guide is presented to the operator via the operation guide output device (212 to 2N2) based on the state of the subsystem captured through N4).

Knowledge database held by each subsystem (218 to 2
In N8), if it is insufficient for inference, this fact is notified to the offline computer, and the offline computer identifies the operation guide from the viewpoint of harmonization of the whole plant, and the result is used again for the relevant online computer system. And is presented to the operator by the operation guide output device (212 to 2N2).

In addition, due to changes on the plant side and accumulation of operation knowledge,
Changes and additions to the knowledge database are made through the man-machine input device (213-2N3) of each online computer system, and the contents of this updated database (218-2N
8) is periodically transferred to the knowledge database of the offline computer 303, and the equivalence of the knowledge on the offline side and the online side is maintained.

As an embodiment of the present invention, it is possible to perform appropriate control of the plant by outputting a plant actuator activation signal via the process input / output system based on the inference result, not just the operation guide output. .

〔The invention's effect〕

In an operation guide system based on a knowledge database and an inference engine, by dividing into an offline computer suitable for building a system knowledge database and an online computer excellent in follow-up to process changes,
The flexible operation guide expert system can coexist with the appropriate operation guide system according to the change of the plant, and real-time property, which is a problem common to current expert systems, can be obtained. This greatly reduces the load on the operator when performing unusual operations, realizes safe, reliable and efficient operation of the plant, and virtually eliminates the occurrence of plant damage and the like due to erroneous operations. .

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is a block diagram showing a conventional embodiment, FIG. 3 is an explanatory diagram of the operation of the present invention, and FIG. FIG. 2 is a block diagram illustrating a distributed processing system). DESCRIPTION OF SYMBOLS 1 ... Plant 2 ... Sensor 3 ... Process input / output device 5 ... Process database 6 ... Knowledge database 7 ... Inference engine 8 Central processing unit

Claims (2)

(57) [Claims] 1. An off-line computer system having a knowledge database construction means for constructing knowledge about a plant using a language which is easy for an operator to understand, an offline knowledge database for storing knowledge constructed by the knowledge database construction means, A process database that stores process information input by a user, an online knowledge database that stores plant knowledge in a language suitable for real-time processing, and an operation guide that is optimal for a plant state is selected and output based on the process database and the online knowledge database. An on-line computer system having inference engine means for performing transmission, a transmission means for connecting the off-line computer system and the on-line computer system, and an operator of the off-line computer system understanding Plant operation guide apparatus characterized by comprising a database converting means for converting the information of the off-line knowledge database written in pancreatic language to language appropriate for real-time processing stored in the online knowledge database of the online computer system. 2. The plant operation guide device according to claim 1, wherein said on-line computer system has a divided configuration corresponding to a specific subsystem, and selects and outputs an operation guide of each subsystem. If the inference is insufficient in each of the online knowledge databases of the plurality of online computers, the offline computer system comprises a plurality of online computers. A plant operation guide device for transferring a guide to a corresponding online computer.