JPH07117429B2 - Plant abnormality diagnosis device - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a plant abnormality diagnosis device that detects an abnormality in a plant such as a power generation plant, estimates the cause of this abnormality, and guides the operator. .

[Conventional technology]

FIG. 2 is a block diagram showing a conventional plant abnormality diagnosing device disclosed in, for example, Japanese Patent Application No. 1-2980008, in which 1 is an electronic computer, which is a process input for reading a process amount from a power plant 2. Unit 3, an abnormality detection processing unit 4 that detects an abnormality from the process amount, an abnormal event detected by the abnormality detection processing unit 4, and a process input unit 3
An abnormal cause inferring unit 5 that infers an abnormal place and an abnormal cause from the data input in the above, and knowledge and a failure cause and an event that qualitatively and quantitatively express the characteristics at a normal time created for each constituent device of the plant The knowledge base 6 is composed of the knowledge of the causal relationship of the above, and the display processing unit 8 for outputting the occurrence event, the failure location / cause, the countermeasure, etc. to the CRT display 7.

Next, the operation will be described. The process input unit 3 inputs data from the power generation plant 2 and converts quantitative process values into qualitative data such as “slow increase” and “constant” used by the abnormality cause inference unit 5. The abnormality detection processing unit 4 monitors the process amount input and processed by the process input unit 3 and detects that an abnormality has occurred when the process amount deviates from the normal range. The anomaly cause inference unit 5 operates when an anomaly is detected by the anomaly detection processing unit 4, infers a failure location / cause based on the detected anomalous event and a qualitative / quantitative process amount, and displays the inference result. Processing unit 8
Is displayed on the CRT display 7. The reason for the abnormality is inferred by the knowledge accumulated in the knowledge base, which is created qualitatively and quantitatively for the characteristics of the plant in each plant, and the knowledge of the cause of failure and the causal relationship between events.

[Problems to be Solved by the Invention]

Since the conventional plant abnormality diagnosing device is configured as described above, when the knowledge on the knowledge base 6 is the knowledge with a low degree of perfection, an incorrect diagnosis is made by the inference by the abnormality cause inference unit 5, There was a problem such as confusing plant operators due to the absence of a guide.

The present invention has been made to solve the above problems, and it is possible to evaluate knowledge with a low degree of perfection by using plant data, save the plant data when the plant is abnormal, and use this for knowledge evaluation. An object of the present invention is to obtain a plant abnormality diagnosing device.

[Means for Solving the Problems]

A plant abnormality diagnosis apparatus according to the present invention comprises an abnormality data saving mechanism for saving abnormal data of a plant at the time of a plant abnormality, and an inference processing unit for evaluation for offline evaluation of knowledge using the saved abnormality data. It was done.

[Action]

The abnormal data table mechanism according to the present invention saves abnormal data when a plant is abnormal in an abnormal file or the like, and uses this abnormal data in parallel during online inference when the plant is abnormal to make an evaluation inference processing unit. In this way, knowledge with a low degree of completeness, such as newly added knowledge, is evaluated, and the degree of completeness of the knowledge is increased before it is used as knowledge for online inference.

Example of Invention

An embodiment of the present invention will be described below with reference to the drawings. First
In the figure, 1 is a computer system, 3 is a power plant 2 as a plant, and a process input section for reading a process amount from a data reproducing mechanism 93. 91 is an abnormal day file 92 in case of an abnormal case, save the abnormal data. Abnormal data save mechanism, 11 is an inference processing unit, 61
Is an online knowledge base, 93 is an evaluation data file 94
The data reproduction mechanism for reproducing data by using, 12 is an inference processing unit for evaluation, 62 is a knowledge base for evaluation, 81 and 82 are abnormal events that have occurred with respect to CRT displays 71 and 72 as display means, Display processing units 81 and 82 for displaying the cause of abnormality and guidance on corrective actions.

Further, both the online inference processing unit 11 and the evaluation inference processing unit 12 detect an abnormality from the abnormality detection processing unit 4, the abnormal event detected by the abnormality detection processing unit 4 and the data input by the process input unit 3. It is composed of a location and an abnormality cause inference unit 5 which infers an abnormality cause.

Next, the operation of the above embodiment will be described. In the process input unit 3, while inputting data from the evaluation data file 94 via the power generation plant 2 and the data reproduction mechanism 93, "slow increase" in which quantitative process values are used in each of the abnormality cause inference units 5, Convert to qualitative data such as "constant". The abnormality cause inference unit 5 in the online inference processing unit 11 operates when an abnormality is detected in the abnormality detection processing unit 4, and detects the abnormal event and the qualitative and quantitative process amount from the power plant 2. The inference result and the cause of the abnormality are inferred, and the inference result is displayed by the display processing unit 81 and displayed on the CRT display 71. That is, inference of the cause of abnormality of the power plant 2 is based on the knowledge that qualitatively and quantitatively expresses the characteristics at the time of normal, which is created for each constituent device of the plant accumulated in the online knowledge base 61, and the cause of failure and The knowledge of the causal relationship of.

On the other hand, in the abnormal data saving mechanism 91, the process input unit 3
When the operator recognizes that the data input from the device is necessary, such as when an abnormality occurs, all the data for a fixed time until the requested time is saved in the abnormal data file 92. The abnormal data file 92 can be used by transferring it to the evaluation data file 94 as one of the plurality of evaluation data files 94. Further, the data reproducing mechanism 93 reproduces the data by using the evaluation data file 94.

Next, the inference processing unit for evaluation 12 can switch and use the data input from the power generation plant 2 through the process input unit 3 and the data reproduced by the data reproduction mechanism 93. Perform the same inference processing. In this case, the reasoning of the anomaly is based on the evaluation knowledge base.
It is based on the knowledge accumulated in 62. The evaluation knowledge base that has been sufficiently evaluated by the evaluation inference processing unit 12 can be used by transferring the file to the online knowledge base 61.

〔The invention's effect〕

As described above, according to the present invention, the abnormal data saving mechanism saves the abnormal data of the plant, the abnormal data is used to evaluate the knowledge offline, and the evaluation result is used as the online inference knowledge. With this configuration, when knowledge is added or changed, the knowledge can be verified and evaluated sufficiently and easily before it can be used online. As a result, the plant abnormality diagnosis by online inference can be performed with high reliability. There is an effect that what can be realized is obtained.

[Brief description of drawings]

FIG. 1 is a block diagram showing a plant abnormality diagnosing apparatus according to an embodiment of the present invention, and FIG. 2 is a block diagram showing a conventional plant abnormality diagnosing apparatus. 2 is a plant (power generation plant), 11 is an online inference processing unit, 12 is an inference processing unit for evaluation, 71 is a display means (CRT display), and 91 is an abnormal data saving mechanism. In the drawings, the same reference numerals indicate the same or corresponding parts.

Claims (1)

[Claims] 1. A plant abnormality including an on-line inference processing unit that infers an abnormal place and an abnormality cause of the plant from a process value input from the plant, and a display unit that displays a processing result and the like by the online inference processing unit. The diagnostic device is provided with an abnormal data saving mechanism for saving the abnormal data of the plant, and an inference processing unit for evaluation that evaluates knowledge off-line by using the abnormal data saved by the abnormal data saving mechanism. Characteristic plant abnormality diagnosis device.