JPH03138528A - Sensor abnormality detecting method - Google Patents

Abstract

PURPOSE:To detect sensor abnormality at low cost by gathering measured values of various sensors which are arranged in a plant, finding predicted values for the respective sensors by a previously set model by using the measured values, and comparing them with the measured values and detecting the sensor abnormality according to a previously set decision point. CONSTITUTION:The measured values obtained by pressure sensors 1 - 3 and a flow rate sensor 4 are gathered and inputted to a computer. Then the computer uses the model which is set previously and inputted to find the predicted values of the sensors 2 and 3. The computer compares the measured values of the sensors 2 and 3 with their predicted values. The computer decides whether or not the sensors are abnormal on the basis of the decision point which is set previously and inputted according to whether or not the measured values of the sensors 2 and 3 are different from the predicted values. Thus, the sensor abnormality is detected at low cost and the reliability of the plant is improved.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a sensor abnormality detection method applied to a plant.

rPrior art] Conventional methods for detecting sensor abnormalities include a method of duplicating sensors and evaluating based on differences in the values detected by the sensors, or a method of correlating the values detected by each sensor placed in a series of processes. There were methods to evaluate based on relationships.

[Problem to be solved by the invention]

In conventional sensor abnormality detection methods, when duplicating sensors, the cost is high, and when evaluating based on the correlation between the values detected by each sensor, it is difficult to determine whether the abnormality is in the sensor or in the process.

The present invention aims to solve the above problems.

[Failure to solve the problem]

The sensor abnormality detection method of the present invention collects the actual values measured by each sensor from various sensors installed in the plant, and uses the actual values to set the values that each sensor measures in a normal state. The predicted value that each sensor will show is determined using a model formula that shows the relationship between It is characterized by detecting the presence or absence of an abnormality in each sensor based on this.

[Effect]

In the above, a model formula that shows the relationship between the values measured by various sensors installed in the plant under normal conditions is set in advance, and the above model formula is used to determine the relationship between the values measured by each sensor. A predicted value is calculated, and the predicted value and the actual value are compared for each sensor.

To determine the presence or absence of a difference between the predicted value and the actual measured value of each of the above sensors, a judgment procedure is determined in advance, and after comparing the above predicted value and the actual measurement value, the above judgment procedure is determined based on the presence or absence of each difference. Based on this, a determination is made to detect whether or not there is an abnormality in the sensor.

Based on the above, the predicted value is determined using the model formula, and the presence or absence of an abnormality is determined based on the determination procedure.
Sensor abnormalities can be detected at low cost, and plant reliability can be improved.

〔Example〕

An embodiment of the present invention will be explained with reference to FIG.

The present embodiment in which the sensor abnormality detection method of the present invention is applied is an example of application to a pipe 5 having pressure sensors 1, 2° 3 and a flow rate sensor 4 shown in FIG. It is something that is done.

In Fig. 1, first, pressure sensor 1.2 is set at 5TEP1.
, 3 and the flow rate sensor 4 are collected and input into the computer. Next, 5TE
At P2, the computer calculates predicted values for the pressure sensors 2 and 3 using model equations (1) and (2) that have been set and input in advance. Next, at 5TEP3, the computer calculates the actual measured values for pressure sensors 2 and 3 by 5T.
Compare with the predicted value obtained in BP2. The last 5TEP4
Then, the computer determines whether there is a difference between the actual measured value and the predicted value of the pressure sensors 2 and 3, and whether the sensor is abnormal or the process is abnormal, based on a preset and input determination procedure.

In the above, it is assumed that the pressure sensor 1 and the flow rate sensor 4 indicate accurate values, and the predicted value of the pressure sensor 2 at 5TEP2 is the pressure sensor 1 and the flow rate sensor 4.
It is calculated using the following model formula (1) using the actual measured value.

Predicted value of pressure sensor 2 = (1) Further, the predicted value of pressure sensor 3 is determined using the actual measured values of pressure sensor 2 and flow rate sensor 4 using the following model formula (2).

Predicted value of pressure sensor 3 = (2) The predicted values of pressure sensors 2 and 3 are compared with their actual measured values at 8TEP3, and based on whether there is a difference between each predicted value and the actual measured value at 5TEP4, Pressure sensor abnormality or process abnormality is determined based on the preset determination procedure shown in FIG.

Based on the above, the predicted value is determined using the model formula, and the presence or absence of an abnormality is determined based on the determination procedure.
Detection of sensor abnormality becomes possible at low cost, and the reliability of the plant can be improved.0 [Effects of the Invention] The sensor abnormality detection method of the present invention collects actual measured values of various sensors installed in the plant. , calculate the predicted value for each sensor using the same actual measured value using a preset model formula, compare it with the actual measured value, and detect the presence or absence of sensor abnormality based on the preset judgment procedure. Detection becomes possible at low cost, and plant reliability can be improved.

[Brief explanation of the drawing]

FIG. 1 is a flowchart of one embodiment of the present invention, FIG. 2 is a piping diagram to which the above embodiment is applied, and FIG. 3 is an explanatory diagram of a procedure for determining the presence or absence of an abnormality in the above embodiment. 1.2.3...Pressure sensor, 4...Flow rate sensor,
5... Piping. Mature 1m 5TEP T rEP2 2 tails, 3 flashes

Claims (1)

[Claims] The actual values measured by each sensor are collected from various sensors installed in the plant, and a model formula is used that shows the relationship between the values measured by each sensor under normal conditions, which are set in advance using the actual values. Find the predicted value that each sensor will show, compare the predicted value and actual measured value for each sensor, and determine whether there is any difference based on a preset judgment procedure and whether or not there is an abnormality in each sensor. A sensor abnormality detection method characterized by detecting.