JPH01121944A - Distributed trouble detector - Google Patents

Abstract

PURPOSE:To enable each management device to independently estimate trouble of the other subsystems by providing each of plural subsystems with the management device and estimating the state of the subsystem corresponding to each management device in accordance with information from this subsystem and communicating information. CONSTITUTION:A management device 3a of a subsystem Si2a detects abnormality of an observation quantity Qi6 of its own subsystem. As the result, the state of this subsystem is estimated in accordance with a manipulated variable Ii5 and the observation quantity Qi6. The management device 3a of the subsystem Si2a receives an estimated input Uj9 from a management device 3b of a subsystem Sj2b through a communication line 4 and compares it with an output Yi8 to discriminate trouble/normalcy of the adjacent subsystem Sj2b in accordance with disaccord/coincidence resulting from this comparison. When the management device 3a of the subsystem Si2a judges that the subsystem Sj2b is in trouble, its message is transmitted to a management device 3c of another subsystem 2c.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a system failure detection device.

[Conventional technology]

System management is 1. For example, as shown in the Proceedings of the 1981 Electrical Engineers of Japan Federation Conference (+), pages t-49 to 1-52, "Recent Trends in Automatic Main Engine Control of Thermal Power Plants," etc. There is a management device that serves as a host, and the host monitors the entire system either directly or at the top of a hierarchical structure, using a centralized system management device.

Under such a management device, it is natural to have a centralized fault detection device in which the host management device makes the final conclusion, and one that accumulates and organizes causal data and applies multivariate analysis or pattern recognition. There are many methods for this purpose, including those that model the internal structure of the system and apply control theory.

[Problem that the invention seeks to solve]

Conventional centralized failure detection devices are configured as described above, so when it becomes necessary to degrade or expand the system, the management system must also be changed, and the system must be stopped each time. There were problems such as the need for

This invention was made to solve the above problems, even if the structure of the system changes.

The object of the present invention is to obtain a highly expandable failure detection device that requires only some additions or deletions to the existing portions of the management system, and does not require any fundamental changes.

[Means for solving problems]

The distributed fault detection device according to the present invention divides a system into a plurality of subsystems, places a management device in each subsystem, and enables information exchange in the same form as the combination of subsystems. By configuring each of the management devices by connecting them with a communication line, estimating the status of each subsystem from the information obtained from the subsystem in each management device, and communicating the estimation results with each other. , each management device can independently estimate failures in other subsystems.

[For production]

The failure detection device according to the present invention divides and models a target system into several subsystems according to function or structure, and arranges a management device in each of the subsystems. Each management device estimates the state of its subsystem only from information from that subsystem, exchanges the results with each other,
Each management device independently correctly estimates the failure/normality of other subsystems. With distributed management in this way, changes in system structure can be made relatively easily by simply adding subsystems and their corresponding management devices to or deleting them from the existing system. .

〔Example〕

DESCRIPTION OF THE PREFERRED EMBODIMENTS A distributed fault detection device according to an embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing an example of this embodiment, and in FIG. 2, 11) is a target system;
(2a), (2b) and (2C) are subsystems divided into several parts according to their functions or areas;
(sa), (3b), and (5C) are management devices corresponding to the respective subsystems (2a), (2b), and (2G), and (4) is their management device (3a).
, (3b).

This is a communication line between (3C). This communication line (4) connects each management device (3a), (3b), (5C) so that information can be exchanged in the same manner as the combination of subsystems (2a), (2b), (2C). It is meant to connect.

In this embodiment, they are connected in a ring shape. FIG. 2 is a model of one subsystem 5i (2a), for example,
In the figure, (51 is the manipulated variable ii * +61 of the subsystem Si (2a) is its observed quantity 'i, (71 is the input ui r from another subsystem +81 is the output yi to the other subsystem). FIG. 3 is a model of coupling between subsystems. In FIG.
b) input from (output yit from subsystem Si
81).

Next, the operation will be explained. 4 and 5 both show the distributed fault detection device shown in FIG. 1.

This is a flowchart showing an example of the work procedure performed by the management device of each subsystem. Based on this flowchart,
The operation of the failure detection device will be explained.

The management device (3a) of the subsystem Si (2a) is.

When an abnormality of observation fl Oi +61 in each subsystem is detected (block αυ), the operation amount 1it51
The state of each subsystem is estimated from the observed amount oi161 (block α2). From this manipulated variable i, 151 and observed quantity 0ii61, state quantity xi, input ui from other adjacent subsystems (for example, subsystem (2C))
+71° The output yi181 to another adjacent subsystem (for example, subsystem (2b)) is the following equation +11.
+21. Estimated by +31.

rx = fi(Xi, it, ui, i)
""" ulO1=gi(xi, ii, uilt)
・・・・・・ (21yi=hi(xi, ii, u
iIt) ...... (Similarly to 31, the management device (3b) of the subsystem Sj (2b) estimates the state of each subsystem Sj (2b), and based on the input from the subsystem 5i (2a) ujf91 is estimated.

Next, the management device (3a) of the subsystem Si (2a) receives the estimated input ujt91 from the management device (3b) of the subsystem Sj (2b) through the communication line (4), and outputs the output yi (by comparison with 81). Based on the mismatch/coincidence, the failure/normality of the adjacent subsystem 5j (2b) is determined (block σ3).This result.

When the management device (3a) of subsystem Si (2a) determines that subsystem Sj (2b) is in failure,
The management device (3a) is a subsystem Si (2a
) of the other subsystem (2C) adjacent to the management device (3
C) sends a message to C) that subsystem Sj (zb) is out of order (block α).

The above task involves sending a message to the entire system about a fault detected in its own subsystem.On the other hand, an example of the task when a message is received from another subsystem indicating that a subsystem is faulty is as follows. It is shown in FIG.

The management device (3a) of the subsystem Si (2a) receives a message from the management device (3b) of the subsystem 5j (3b) through the communication line (4) that the subsystem Sk is out of order (block αcj ). At this time, the same determination as in block α3 above is made by subsystem Sj
(2b) That is, this is performed for the subsystem that sent the failure message (block Q[9). This judgment (
When subsystem 5j (2b) is determined to be faulty as a result of block αη), subsystem Sj (
2b) sends a message to the other subsystem (2C) that it is abnormal. When the state of the subsystem (2b) transmitted in block aη is normal, the message of subsystem Sj (2b) is transmitted to other subsystems. In this way.

Each management device (sa), (sb) uses only the result received from the management device of the subsystem for which one determination result was determined to be normal in the determination of block αη.

. . . are each independently connected to other subsystems (2
a).

(2b) It is possible to estimate the failure/normality of...

When expanding the system in such equipment, that part is modeled as a new subsystem, and communication lines are installed in the existing parts of the equipment to enable information exchange in the same form as the combination of subsystems. It can be easily expanded by simply connecting.

This expanded portion is indicated by the dotted line in FIG. Also.

When degenerating the entire system, it is only necessary to separate the subsystems that correspond to the degenerated part, and there is almost no need to change the existing parts.

Specifically, for example, the present invention can be applied to a device for detecting failures of valves, pumps, etc. in a water supply system as an entire system. In this case, the state quantity xi of the subsystem reservoir and subsystem may be water volume, and the manipulated variable ii may be valve operation, but the present invention is not limited to this, and can also be applied to systems such as electric power and energy.

Further, the nine work steps are not limited to the one embodiment described above, and can be realized by other steps.

〔Effect of the invention〕

As described above, according to the present invention, a system is divided into a plurality of subsystems, a management device is placed in each subsystem, and information can be exchanged in the same manner as the combination of subsystems. Connect and configure each management device with a communication line.

Each management device estimates the state of the subsystem from the information obtained from that subsystem, and the management devices communicate the estimation results with each other, so that each management device can independently control the status of other subsystems. By making it possible to estimate failures, even if the structure of the system changes.

This has the effect of providing a distributed fault detection device that can respond to changes relatively easily by making some changes to existing parts.

[Brief explanation of the drawing]

FIG. 1 is a conceptual diagram showing a distributed fault detection device according to an embodiment of the present invention, and FIG. 2 is a block diagram showing the characteristics of one subsystem according to this embodiment. Figure 3 is a block diagram showing the connections between each subsystem, Figure 4 is a flowchart showing the work of estimating the state of each subsystem, and Figure 5 shows how each subsystem communicates estimation results. It is a flowchart showing the work of the latter part. 11)...System, (2a)-(2C)...
sub-system. (3a) to (3C)...management device, (4)...communication line. In addition, in FIG. 1, the same reference numerals indicate the same or equivalent parts. Daisototo Oiwa Masuo Figure 1 1 System 2a-, 2c: Subsystem 3aJc: Management Narrowness 4: Tsui ZV Huge] Bare Figure 2 I I 3rd Stage Figure 4 Figure 5

Claims (1)

[Claims] The system is divided into multiple subsystems, a management device is placed in each of the above subsystems, and each of the above management devices is connected via a communication line so that information can be exchanged in the same manner as the combination of the above subsystems. connect and configure,
Each of the above-mentioned management devices estimates the status of the subsystem from the information obtained from that subsystem, and the above-mentioned management devices mutually communicate the estimation results, so that each management device can independently communicate with other subsystems. A distributed fault detection device characterized in that it is capable of estimating system faults.