JPS61245210A - Plant failure cause estimation method - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

Detailed Description of the Invention [Field of Application of the Invention] The present invention provides a method for detecting a failure when a single failure (one cause of failure) occurs in a system consisting of a plurality of devices such as an atomic couplant or a chemical plant. The equipment that causes the
This paper relates to a method for estimating plant failure causes based on information from a limited number of sensors.

[Background of the invention]

The present inventor has already proposed JP-A-58-197513 as a system for estimating the cause of plant failure. One of the features of this method is that it uses the detection time of the abnormal signal, the calculation start time, the failure propagation time, and the normal signal to narrow down the estimated causes. However, this method does not take into account the possibility that the failure propagation time changes depending on the severity of the failure. Therefore, there was a possibility that the cause of the failure could be incorrectly estimated.

[Purpose of the invention]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems of the prior art, it is an object of the present invention to develop a plant that can accurately estimate the true cause of failure even when the failure propagation time between equipment constituting the plant changes depending on the degree of failure. The object of the present invention is to provide a failure cause estimation method.

[Summary of the invention]

In order to achieve the above object, the present invention calculates the failure rate for each device in a system consisting of a plurality of devices, and the direct failure propagation relationship (propagation direction, maximum propagation time, and minimum propagation relationship) between two devices. time and propagation probability), and the cause of the failure is estimated based on the abnormality information obtained from the sensor, its detection time, and normality information. Japanese Patent Publication No. 58-19751
In No. 3, a fixed value was given in advance to the failure propagation time. Therefore, in the present invention, which had the problems described in the prior art section, maximum and minimum values are given to the failure propagation time,
The feature is that the cause of the failure can be correctly estimated if the failure propagation time is within a given range.

[Embodiments of the invention]

Hereinafter, one embodiment of the present invention will be described in detail with reference to FIGS. 1 to 11.

FIG. 1 shows the configuration of an embodiment of a plant system that implements the plant failure cause estimation method according to the present invention.

In FIG. 1, a plant 101 includes a plurality of component devices 102 and sensors 103 for detecting the status of some of the components. These sensors 103 monitor the operating status of each component 102, such as flow rate, temperature,
A signal 104 such as a frequency is detected, and a detection signal 105 is output to an abnormality detection device 107 comprising an abnormality detector 106 corresponding to each sensor 103'. The abnormality detection device 107 stores in advance a reference signal for determining whether each signal is normal or abnormal, and compares the reference signal with the detection signal 105 to determine whether each detection signal 105 is normal or abnormal. It is output to the failure cause estimation device 109. Failure cause estimation device 1
In step 09, the signal 108 is checked at regular intervals, and if the signal changes from a normal signal to an abnormal signal, the time is stored. Operator console 110 calculation instruction signal 1
11 is output, the failure cause estimation device 109

In accordance with the calculation instruction signal 111, the signal 108, the time when the normal signal changed to the abnormal signal, and the initial data input device 112
A signal 113 corresponding to the failure propagation direction, maximum propagation time, minimum propagation time, propagation probability, and failure rate of each device input from
Based on this, the cause of failure is estimated or the estimated causes of failure are prioritized, and the signal 114 is output and displayed on the display device 115.

Note that once the initial data is entered, there is no need to re-enter it unless the data is changed.

FIG. 2 is a flowchart showing an example of the flow of processing in the failure cause estimation device 11109 of FIG.

Hereinafter, an example of application of the present invention to a plant in which the failure propagation relationship of plant component equipment is represented by the network shown in FIG. 3 will be described.

Figure 3 is a network in which the nodes represent plant component equipment, the direction of the arrow represents the direction in which the effects of a failure will directly spread, and the nodes surrounded by squares are devices that can use sensors to determine whether or not the equipment is abnormal. represents.

When the influence of a failure directly spreads between two devices in such a plurality of devices (19 in Figure 3), the direction of the spread, the maximum propagation time, and the minimum propagation time are given, and this relationship is expressed as It is expressed in a matrix as shown in FIG. For example, 100 in the 1st row and 2nd column of matrix A indicates that the influence of the failure will directly spread from device 1 to device 2, and 90 in the 1st row and 2nd column of matrix B indicates that a maximum of 100 seconds will elapse at that time. , indicates that at least 90 seconds have elapsed, which means that the direct propagation time of the failure effect from device 1 to device 2 is between 90 seconds and 100 seconds. A blank space means that there is no direct influence, and is treated as ψ in calculations.

Furthermore, the propagation direction and propagation probability are given when a failure directly propagates between two devices, and this relationship is expressed in a matrix as shown in FIG. For example, 1 in the first row and second column of matrix C indicates that the probability that the influence of a failure will directly spread from device 1 to device 2 is 1. A blank space means that there is no direct influence, and is treated as O in calculations.

Initial data input bag shown in Figure 1! The maximum failure propagation time input from 112 to the failure cause estimation device 109 is matrix A in FIG.
The minimum failure propagation time is input in the form of matrix B in FIG. 5, and the propagation probability is input in the form of matrix C in FIG. 6.

Now, in the plant where the fault propagation relationship is represented by the network in Figure 3, the equipment that is emitting the abnormal signal is identified as 2.
3, 9. 18. 19% equipment outputting a normal signal 7.
10, 11, 15.17. Also, the elapsed time from a certain reference time to the abnormality detection time is t2 = 0 seconds, respectively.
t, =100 seconds, t, =200 secondst tzs=30
0 seconds 1 llm = 300 seconds, and an instruction signal to start calculation from the operator console 110 is sent to the failure cause estimation device 10.
Assume that the elapsed time until input to 9 is t, = 300 seconds.

First, in step 116 of FIG. 2, a failure propagation time matrix for normal information is calculated. Here, in the matrix C, all values of the elements of the value A corresponding to elements having a value less than 1 are replaced with ψ. The results are shown in FIG. This means that there is no ripple effect, and in calculation, ψ
Using this matrix, find the shortest failure propagation time from all devices to the device that is outputting a normal signal. This is a solution method for finding the known shortest distance and shortest path (Masao Iri, Takashi Furubayashi: Network Theory 9th Science and Technology Union Publishing Co., Ltd.)
1976), pp-a 4-37.

(see ). The obtained propagation time is expressed as a matrix as shown in Fig. 8. For example, 410 in the 1st row and 1st column of matrix E is.

The time required for the influence of the failure to spread from device 1 to device 7 is 410 seconds at most, and the blank space means that there is a possibility that the influence will not spread, and is treated as ω in calculations.

Next, in step 117 of FIG. 2, the fault propagation relationship is updated based on the normality information. Here, in matrix B, all rows and columns corresponding to devices that are outputting normal signals are set to gradients. The result is shown in Figure 9. This calculation is
This corresponds to deleting the entry and exit points of devices that are outputting normal signals on the network shown in FIG.

Next, in step 118 of FIG. 2, a fault propagation time matrix and a propagation probability matrix for abnormality information are calculated.

Here, we use matrix F to calculate the shortest failure propagation time from all devices to the device that is emitting the abnormal signal in step 1.
Obtain in the same manner as 16. The obtained propagation time is expressed as a matrix as shown in Fig. 10. For example, 90 in the 1st row and 1st column of the matrix G indicates that it will take at least 90 seconds if the influence of a failure spreads from device 1 to device 2, and the blank part means that there is no spillover. is treated as ψ,
Furthermore, when determining matrix G, the shortest time fault propagation path is simultaneously determined, and matrix C is used to determine the failure propagation probability from all devices to the device issuing the abnormal signal. The obtained spread probability is expressed as a matrix as shown in FIG. 11. For example, 0.5 in the 4th row and 5th column of matrix H is found as follows: 0 When finding matrix G, the shortest time fault propagation path from device 4 to device 19 is 4 → 3 → 8 → 9. As you can see, h4. is = C4,3'X C3,I
X Cs,s = 0, 5 X I X 1 =0.
A blank part of the 0 matrix H, which is 5, means that there is no failure propagation, and is treated as O in calculations.

Next, in step 119 of FIG. 2, the cause of the failure is estimated based on the abnormality information. It is presumed that one of the two is the cause of the failure. On the network shown in Figure 3, this calculation involves tracing the arrow leading to the device where an abnormality has been detected in reverse, without passing through the device that is emitting a normal signal, and selecting a common device on each route. corresponds to

Next, in step 120 of FIG. 2, the estimated cause of failure is checked based on the failure rate. Here, the failure rate of the failure cause estimated in step 119 is checked, and if the failure rate is 0, the equipment is estimated. In one example that is excluded from the cause of failure,
It is assumed that the failure rate of both devices 1 and 2 is not 0.

Next, in step 121 in FIG. 2, the estimated cause of the failure is checked based on the failure propagation time.

Here, the rows in which at least one element has a value other than φ in one matrix E, i.e., 1, 2, 3, 5, 6° 7.10.1
F, 13, 14, 15, 17. Equipment 1 compatible with rows,
2, 3, 5, 6, 7, 10° 11. 13, 14, 15.
Common devices 1 and 2 for estimated failure causes 1 and 2 in step 17 and step 120 are determined. Next, from the matrix G, the shortest failure propagation time gx, x = 90p gz, a = 180* gi, *= from 1 device 1 to abnormal devices 2.3, 9, 18, 19
270egx, □s= 360v gt, zs= 3
Find 60. Furthermore, gl・z+ to -ta
t gl, co+ 1. -12pgz+-+je-ta
w gt, ta+ja-t1m* gt, ts+t,-
t, and find the maximum value t, ,=39Q.

tfl is device 2.tfl when device 1 is assumed to be the cause of the failure.
3, 9, 18, and 19, from a matrix E indicating at least the time that is considered to have elapsed from the occurrence of a failure to the start of calculation, from device 1 to normal device 7,
10, 11, 15. The shortest failure propagation time e, which passes only through the path with failure propagation probability 1 leading to 17, is 4102 'L+
1□=100°e,, 1=200, el, 1. =
410 , el, , find = 3QQ @ 61+? t
sx+tos et+txs eIHL! 1e1. l
, and find the minimum value t, , 1=100.

t. 1 assumes that device 1 is the cause of the failure, at the latest 100 seconds after the occurrence of the failure, device 7゜10.11,
145.17 indicates that an abnormality occurs. 1. engineering and t. , and ta, > j−
If it is 1, an abnormality should occur in either device 7.10, 11°15.17, but it shows a normal value, so device 1 is considered not to be the cause of the failure, and device s1 is assumed to be the cause of the failure. If ttt<t-□, devices 7, 10
, 11°, 15.17, it will take more time to determine whether or not an abnormality occurs in any of them, and it cannot be determined at this point that it is not the cause of the failure, so device 1 is included in the probable cause of failure. In this example, tt,=390)t
, □ = 100, so device 1 is excluded from the probable cause of failure.For device 2, i*= is calculated in the same manner as for device 1.
Since 300<t, ,=310, device 2 is included in the probable cause of failure.

Next, in step 122 of FIG. 2, the estimated causes of failure are prioritized. This step is executed when there are multiple estimated failure causes in step 121. Here, the device k in which the abnormality was detected earliest is selected from among the plurality of abnormality detection devices. Next, from the matrix H, the probability of failure propagation from the estimated failure cause Q to the equipment is calculated. Find the product P1 of the estimated failure cause Q given by the initial data and the failure rate P1.
・Calculate h ah. In this embodiment, the estimated cause of the failure in step 121 is 1 of the device 2.
Therefore, step 122 is not executed.

According to the above-mentioned embodiment, (1) the cause of failure can be estimated even if the number of sensors that can be installed is limited in a plant consisting of multiple pieces of equipment, and (2) there are multiple possible causes of failure. (3) The fault propagation time is calculated by removing the fault propagation relationship of devices that are emitting normal signals, so the accurate It is possible to estimate the cause of failure.

(4) To use the maximum and minimum values of failure propagation time,
This has the effect of allowing accurate estimation of the cause of failure.

Note that the fault propagation time matrix for normal information is calculated in advance in step 116 of FIG. By removing the column corresponding to the anomaly detection device from the input data,
A matrix E can also be created.

〔Effect of the invention〕

As explained above, according to the present invention, (a) it is possible to estimate the cause of failure even in a system that includes equipment in which sensors and abnormality detection devices are not installed, and (b) when there are multiple estimated causes of failure, each failure rate of component equipment,
Using the failure propagation probability, it is possible to prioritize the possibility of being the cause of the failure. It is possible to estimate the cause of the failure. (d) Since the maximum and minimum values of the failure propagation time are used,
It is possible to accurately estimate the cause of a failure.As a result, it becomes possible to estimate the cause of failure, which was not possible with conventional methods, and it has the effect of making it easier to take countermeasures when a failure occurs.

[Brief explanation of drawings]

1st y! I is a configuration diagram of an embodiment of a plant system using the failure cause estimation method of the present invention, FIG. 2 is a diagram showing an example of processing by the failure cause estimation device of FIG. FIG. 4 is a diagram showing the maximum direct fault propagation time matrix, FIG. 5 is a diagram showing the minimum direct fault propagation time matrix, and FIG. 6 is a diagram showing the direct fault propagation probability matrix. FIG. 7 is a diagram showing a direct failure propagation time matrix for normal information, FIG. 8 is a diagram showing a failure propagation time matrix for normal information,
FIG. 9 is a diagram showing the direct failure propagation time matrix for abnormality information, FIG. 10 is a diagram showing the failure propagation time matrix for abnormality information, and FIG.
The figure shows a failure propagation probability matrix. (Continued from page 91) Inventor Shigeru Hashimoto Male Hitachi University Mikacho 5-chome or inside the factory

Claims (1)

[Claims] In a plant failure cause estimation method that estimates equipment that is the cause of a failure based on information from sensors corresponding to limited equipment in a system consisting of multiple equipment, 2 is the initial input.
Failure propagation direction between devices, maximum propagation time, minimum propagation time,
Based on the propagation probability and failure rate of each device, abnormality and normal information input from sensors and their detection times,
means for calculating a fault propagation time matrix for normal information, means for updating a fault propagation relationship based on the normal information, means for calculating a fault propagation time matrix and a propagation probability matrix for the abnormal information, The apparatus is characterized by comprising means for estimating the cause of failure based on the failure rate, means for checking the estimated cause of failure based on the failure rate of each of the devices, and means for prioritizing the estimated cause of failure based on the failure propagation time. A system for estimating the cause of plant failure.