JPH10124134A - Diagnostic system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To more excellently, correctly diagnose the state of a process correctly representing it, not only at the time of fault diagnosis but also at the time of maintenance and management diagnosises by aking a threshold value corresponding to a prescribed process variable as the function of other process variable. SOLUTION: A recognition setting means 105 sets a corresponding node N in an abnormal or normal correspondence state in accordance with the value of the process variable which is inputted to an input part 106 with a database 7. A cause node specifying means 107 specifies an abnormality cause corresponding to the main abnormality cause in accordance with the mutual abnormal/ normal correspondence setting state of the nodes. In this diagnostic system, the threshold value set in accordance with the respective nodes N as against the prescribed process variable is subordinate to another process variable. Besides, the system is constructed so as to easily execute the external input setting of the threshold value based on information which is stored in the, what is called, database 7.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to operation management of plant equipment, preventive maintenance, and identification of the cause of a failure.
The present invention relates to a diagnosis system for so-called process diagnosis. One example of such a diagnostic system utilizes a directed graph model, which uses a directed graph model configured by connecting a plurality of nodes provided corresponding to each process variable by a directed segment. And make the diagnosis. Here, the node types include a node called a failure mode, which is an abnormal state of a process variable to be subjected to threshold processing, and a node called a failure state representing a state (cause) of a failure occurring in a process. These are connected by a directed segment to form a model. The present application relates to a technique relating to threshold setting of a node (mainly a failure mode) to be subjected to threshold processing among such nodes (an example of a model part described later). Now, each node is a process variable (temperature, process, temperature,
The diagnostic system is provided in accordance with a failure propagation graph in which these nodes are connected by a directed line segment.
The cause of the failure is automatically specified.

[0002]

2. Description of the Related Art First, a conventional fault diagnosis method will be described. Conventionally, when an abnormality occurs in equipment and the equipment stops, a failure is detected and the cause of the failure is specified. Further, it is preferable to maintain the process in good condition by performing appropriate maintenance and inspection on the process in advance without waiting for the occurrence of an extreme abnormality (failure) such as the above-mentioned plant stoppage. The only way to monitor process data online is for a specialist at that facility to look at the data and determine any signs of anomaly.

On the other hand, as a method of automatically performing a so-called fault diagnosis, a method of performing a diagnosis using a model in which the form of a plant is described in accordance with a fault propagation system has been proposed. As such a technique, the inventor has proposed a technique of specifying a cause of failure using a so-called directed graph model. A diagnostic system using this method is configured as follows, as shown in FIG. That is, such a diagnostic system includes a directed graph model for the process to be diagnosed in the model storage unit. Here, this directed graph model is formed by connecting a plurality of nodes provided corresponding to each process variable of the process to be diagnosed by a directed segment, and the value of the process variable corresponding to these nodes Can set a node to an abnormal response state if it is in an abnormal state, or a normal response state if it is not in an abnormal state, depending on whether or not it is in an abnormal state with respect to a preset threshold. It is configured. That is,
Each node is turned on when the value of the process variable indicates an abnormal value. For this purpose, corresponding to each node of the directed graph model, a recognition setting for recognizing the value of the process variable in the process to be diagnosed and setting the corresponding failure mode to the abnormal response state or the normal response state. Means are provided. This recognition setting means,
The failure mode ON process is substantially performed. Further, in the above-described directed graph model, an abnormality /
A cause node specifying means for specifying an abnormal cause node corresponding to a main cause of abnormality according to the normal response setting state is provided. The cause node identification means is a so-called failure diagnosis execution means, which is processed by the recognition setting means and is a directed graph model in a state capable of representing the state of the plant to be diagnosed (only the nodes in the abnormality handling state are turned on). (Processed), only the abnormal node is traced to its root side in accordance with the connection state of the directed line segment, and the node closest to the root side is identified as the abnormal cause node. By such a method, an abnormal cause node is specified, and a failure cause registered for the abnormal cause node is diagnosed as a primitive failure cause of the process.

[0004] In the above-described diagnostic system, conventionally, a threshold value that needs to be set for each node has been fixed by an input operation performed in advance. In addition, in the conventional fault diagnosis system, the threshold value is a value related only to the process variable corresponding to the node.

[0005]

The diagnostic system to which the present invention is directed is capable of coping with extreme abnormal conditions such as process stoppages and includes a possibility of being utilized for daily process maintenance management. However, in order to achieve such an object, it is necessary to appropriately set the threshold value of each node in the above-described directed graph model.
However, in the conventional method, since this is set in a fixed state in relation to a directly related process variable, there is room for improvement. Further, as will be described in the following embodiments, for example, when the process to be diagnosed is a fuel cell process, the battery current value corresponding to the load greatly affects each process variable value. However, conventionally, since the threshold value is not set in relation to the values of other process variables, the threshold value needs to be set relatively loosely, and a suitable diagnosis cannot be made.
In setting such a threshold, a so-called “
Many systems rely on the knowledge of experts called experts, and a system that can set and change the threshold value relatively easily and rationally has not yet been established. Accordingly, an object of the present invention is to provide a diagnostic system that can solve the above-described problems.

[0006]

According to a first aspect of the present invention, there is provided a diagnostic system according to the present invention for achieving the above object, using the above-described directed graph model. Threshold value changing means for dynamically changing a threshold value corresponding to a predetermined process variable according to the value of another process variable, wherein a threshold value corresponding to the predetermined process variable is provided. Is a function of other process variables. As described above, a process variable is inherently related to another process variable. Therefore, the judgment of abnormal or normal should also be made dependent on other process variables. Therefore, in the present application, a threshold value changing means is provided, which is dynamically changed in relation to other process variables, and the threshold value at each node is made dependent on the other process variables. .
As a result, the setting of the threshold value can be made more accurate and strict, and not only the failure diagnosis, but also the diagnosis when maintenance and management are required, better and more accurately represent the process situation. , Can now be diagnosed. Here, as an example, taking an example of a predetermined process variable and other process variables, in a fuel cell process, the reformer burner exhaust gas temperature as a predetermined process variable, and the battery current as another process variable Values can be given.

Now, as described in claim 2,
In the above diagnostic system, as the threshold value for the process variable, an upper threshold value and a lower threshold value can be set, and as the node, the upper threshold value and the lower threshold value are separately set. Preferably, a corresponding upper limit node and a lower limit node are provided, respectively, and the directed graph model is configured. By configuring the directed graph model in this way, process variables can be grasped from both the upper limit and the lower limit, failure and maintenance management limits can be appropriately set, and a very useful diagnostic system can be provided. Further, as set forth in claim 3, the threshold value for the predetermined process variable is configured to be set to an external input in association with the value of the other process variable, and the threshold value is set according to the external input setting. Preferably, the threshold value changing means dynamically changes the threshold value based on a relationship with the value of another process variable. The dynamic threshold value setting by the threshold value changing means is performed in a normal state in a preset state (a predetermined process variable associated with the value of another process variable stored in the storage means). The threshold value is automatically set in the diagnostic system and the diagnosis is performed. The operation status of the process and the purpose of using the diagnostic system (the purpose of performing a fault diagnosis in an extremely abnormal situation) Depending on the case (in this case, the threshold is set relatively loosely), or for the purpose of properly maintaining the normal process (in this case, the threshold is set relatively hard), etc. You may want to change the threshold settings on purpose. In such a case, the diagnosis for the purpose of the present application can be performed by setting the external input of the threshold value in relation to other process variables.

In the diagnostic system described above, when a node is regarded as a model portion provided for each process variable, a process having a plurality of process variables is set as a diagnosis target, and the value of each process variable is set in advance. A plurality of model parts that can be set to an abnormal response state when in the abnormal state, and to a normal response state when not in the abnormal state, according to whether or not an abnormal state is compared with a threshold value, A recognition setting means for recognizing a value of the process variable in the target process and setting the corresponding model part to the abnormal response state or the normal response state; The diagnostic system derives the cause of the abnormality according to the corresponding setting state. Accordingly, in a diagnostic system having such a basic configuration, a threshold value changing means for dynamically changing a threshold value corresponding to a predetermined process variable in accordance with the value of another process variable is provided. If the threshold corresponding to is a function of other process variables,
Appropriate diagnosis can be made. This configuration is described in claim 10. The above model portion includes the nodes described above. As a diagnostic system, a diagnostic system having a so-called IF-THEN structure rule and an inference engine is known. In this case,
This model part is used as a condition part or a conclusion part of the IF-THEN rule. That is, the above configuration is
A similar effect can be obtained with a diagnostic system having a THEN structure.

The above description relates to the setting of a threshold in relation to other process variables. However, the setting of a threshold in a diagnostic system does not necessarily require expert knowledge. The present invention relates to a configuration that can be performed without performing. According to a second aspect of the present invention, there is provided a diagnostic system using a directed graph model as described above. Display means for displaying a history of the value of the predetermined process variable based on information from a database that stores the past state of the process variable of the process to be diagnosed, and externally setting the threshold value for the process variable. It is provided with a threshold value input means enabling input setting.
In this diagnostic system, a display means and a threshold value input means are provided in parallel, and an operator looks at least a history of process variables of a process to be diagnosed stored in a database displayed on the display means on a display screen, A threshold can be set and entered. That is, in general, when a process variable is in an extremely abnormal state, it is normal for the worker to notice that the state is different from a normal state. This allows the operator to make settings while looking at the display screen. This makes it possible to set the threshold value relatively easily. FIG. 5 shows an example of such a display screen.
In this example, the history of a process variable to be diagnosed and a predetermined process variable from another process of the same type are displayed as discrete points. Many discrete points are included in a certain group, and two points are included. An abnormal degree of discrete points can be recognized.
In such a case, a threshold value can be set at the boundary between the normal group and the abnormal discrete point.

Further, in the diagnostic system having the above configuration, the display means is configured to be capable of displaying a value of a predetermined process variable in relation to a value of another process variable. It is preferable that the threshold value input means is configured to be able to externally set the threshold value for the predetermined process variable in association with the value of the other process variable. in this case,
The predetermined process variable is displayed in relation to other process variables, and the operator can set a threshold value in the predetermined process variable while determining the relationship between these two variables. In the example shown in FIG. 5 corresponding to the lower part of FIG. 6 described above, the two solid lines and the one-dot chain line in FIG. 5 indicate this threshold. This threshold value is a function of the battery current value on the horizontal axis in FIG. Further, in this case, as described in claim 6, as the threshold value for the process variable, an upper threshold value and a lower threshold value are provided, and as the node,
An upper limit node and a lower limit node respectively corresponding to the upper threshold value and the lower threshold value are provided to constitute the directed graph model, and the upper threshold value and the lower threshold value are set in the threshold value input means. It is preferable that a value and an external input can be individually set. Also in this case, the process variables can be grasped from both the upper limit and the lower limit, and the limit of failure and maintenance can be appropriately set, so that a very useful diagnostic system can be provided.

As described above, when the node is regarded as a model portion provided for each of the process variables, the diagnostic system of the present invention treats a process having a plurality of process variables as a diagnosis target and the value of each process variable is A plurality of models that can be set to an abnormal response state when in the abnormal state and to a normal response state when not in the abnormal state, depending on whether or not there is an abnormal state as compared with a preset threshold value. A recognition setting means for recognizing a value of the process variable in the process to be diagnosed and setting the corresponding model part to the abnormal response state or the normal response state. Mutual abnormalities /
The diagnostic system derives the cause of the abnormality according to the normal response setting state. Therefore, even in the diagnostic system having the basic configuration as described above, the history of the value of the predetermined process variable is displayed based on at least the information from the database storing the past state of the process variable of the process to be diagnosed. It is preferable that the display device further include a threshold value input device that enables external input of the threshold value for the predetermined process variable. Also in this case, for example, in a diagnostic system having an IF-THEN structure, a similar effect can be obtained by appropriately setting the threshold value. This configuration is described in claim 11.

In the examples described above, the origin of the process variable displayed by the display means is not specifically described, but is basically used for the purpose of setting the threshold value. The larger the amount of information to be performed, the better. Therefore, it is preferable to adopt the following configuration to meet this purpose. That is, an input unit is provided for inputting process variable information from other similar processes other than the process to be diagnosed, and the display means displays the value of the predetermined process variable in each process variable information from another similar process. It is made possible. In the case of adopting such a configuration, the operator makes an abnormality / normal judgment while viewing the information from other processes input to the diagnostic system as a display image, and determines the status of the other processes. Can be set based on a sufficient amount of knowledge taking into account Also in this case, more useful diagnosis can be performed if the information from another similar process is history information from another similar process stored in the database.

In general, a diagnostic system is built with a process control system. Therefore, a so-called control target value is set as a process variable from the control system side.
The following configuration attempts to diagnose the state of the process based on such information from the control side. That is,
A third diagnostic system according to the present invention is characterized in that, in the diagnostic system using the above-described directed graph model, when a predetermined process variable has a control target value, as described in claim 8, And a control target value dependent threshold value setting means for determining the threshold value in relation to the control target value. In this diagnostic system, a threshold value for a predetermined process variable is set in relation to a control target value of the process variable. Therefore, the diagnostic system as a whole becomes an obstacle in order to determine whether an abnormality is likely to occur when the control target of the process changes, and to achieve the control target when an abnormality occurs. It can work to verify which process variables are present and what causes are inherent. As a result, the process can be diagnosed from the control side.

Further, in this case as well, an upper threshold value and a lower threshold value are provided as the threshold value for the process variable, and the upper limit value is set as the node. An upper limit node and a lower limit node respectively corresponding to a threshold value and a lower limit threshold value are provided, and the directed graph model is constructed. It is preferable that the threshold value and the threshold value can be individually set. Also in this case, the process variables can be grasped from both the upper limit and the lower limit, and the limit of failure and maintenance can be appropriately set, so that a very useful diagnostic system can be provided.

As described above, when the diagnostic system of the present invention regards a node as a model part provided for each process variable, a process having a plurality of process variables is regarded as a diagnosis target, and the value of each process variable is determined. A plurality of models that can be set to an abnormal response state when in the abnormal state and to a normal response state when not in the abnormal state, depending on whether or not there is an abnormal state as compared with a preset threshold value. A recognition setting means for recognizing a value of the process variable in the process to be diagnosed and setting the corresponding model part to the abnormal response state or the normal response state. Mutual abnormalities /
The diagnostic system derives the cause of the abnormality according to the normal response setting state. Therefore, even in a diagnostic system having such a basic configuration, when a predetermined process variable has a control target value, a control target value-dependent threshold value that determines the threshold value in relation to the control target value. It is preferable to provide setting means. Also in this case, for example, IF-
In a diagnostic system having a THEN structure, a similar effect can be obtained by appropriately setting a threshold value. This configuration is described in claim 12.

As described above, the diagnostic system of the present application is useful, but one suitable application example of such a diagnostic process is a fuel cell process. In such a fuel cell process, the battery current value corresponding to the load is a factor that greatly influences the process diagnosis. Therefore, the other process variables described above include the battery current value in the fuel cell. Selecting a value and setting a threshold for the process variable of the fuel cell process in a manner dependent on this value,
A good diagnostic system for the fuel cell process will be provided.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a diagnostic system 1 according to the present invention will be described below with reference to the drawings. In this example, the diagnostic system 1 is provided for maintenance / management of the fuel cell process 2 and failure diagnosis. Further, in the case of this example, the diagnostic system 1 is configured to accumulate operation information on the fuel cell processes 2 in the plurality of fuel cell plants 3 and perform diagnosis based on the information of the plurality of processes 2. The feature is that is adopted.

First, using the diagnostic system 1 of the present application,
An overview of a system called a process control support system by the inventors will be described with reference to FIG. FIG. 1 shows the overall configuration of the process control support system, and the controlled support side is a fuel cell process 2 in each plant 3. Here, each fuel cell process 2 is first connected to a data collection system 6 by a telephone line 5 via a modem 4. This data collection system 6
Comprises a data collection system computer 6a and data collection system software 6b stored therein. Therefore, the operation status information of the fuel cell process 2 in each plant 3 is collected, accumulated, and stored in the database 7 provided in the data collection system 6.
On the other hand, the real-time information and the accumulated information collected in the data collection system 6 in this way can be sent to the diagnostic system 1 by information transfer by ftp (file transfer protocol). The diagnostic system 1 includes a diagnostic system computer 8a.
And the diagnostic system software 8b stored therein. Therefore, the diagnostic system 1 is capable of grasping the operation information of the fuel cell process 2 in each plant 3 in real time, and at the same time, stores the accumulated information from the data collection system 6 and the desired information including the operation history. It is possible to take out in the state of.

In the diagnostic system 1, the driving information is
It is recognized as a so-called process variable. Further, on the fuel cell side, it is necessary to operate the fuel cell process 2 in accordance with a predetermined purpose. Therefore, for each so-called process variable, the control target value is set on the plant side or the control support value. The control target value is generated and set by the system, but such a control target value is also input to the diagnostic system.

Next, the diagnostic system 1 will be described in more detail. FIG. 2 is a block diagram functionally expressing the diagnostic system 1 of the present application. The diagnostic system 1 basically has the structure of the conventional diagnostic system described above. That is, the diagnostic system 1
Is provided with a model storage unit 101 that stores a directed graph model 100 for at least a process to be diagnosed shown in FIG. 2 (specifically, a specific process of the illustrated fuel cell process 2). Here, the directed graph model 100 is configured as an abnormal propagation graph in which a plurality of nodes N provided corresponding to each process variable of a process are connected by a directed line segment 102. As described above, the node N has a failure mode and a failure state. However, in the example described below (from this part to before another embodiment), the node N is subject to threshold processing. Since only the failure mode is set, the node N in the example of the following embodiment uniquely means the failure mode.

Further, each node N (each failure mode)
According to whether the value of the process variable is in an abnormal state as compared with a preset threshold value, an abnormal response state is set in an abnormal state, and a normal response state is set in a non-abnormal state. It is configured to be possible. The threshold value set corresponding to each node N in this manner is stored in the threshold value storage means 103. On the other hand, in each node N, when this is abnormal, it is necessary to specify the cause. These causes are also registered in advance, and the cause to be specified is stored in the fault cause storage unit 104. Have been.

More specifically, the configuration of the directed graph model 100 will be described with reference to FIGS. 3 and 4. As shown in FIG. 3, a plurality of nodes N fail based on the structure, function, and experience of the fuel cell process 2. They are connected by a directed segment 102 so that propagation can be represented. The figure shows a part of the model 100, but it can be seen that a pair of nodes N is provided corresponding to each process variable. That is,
These pairs of nodes are an upper node Na and a lower node Nb corresponding to the upper threshold and the lower threshold of each process variable. In the figure, the symbols described corresponding to the node N are as follows. N1 (reformer burner exhaust gas temperature), N2 (raw material gas flow rate), N3 (ejector opening), N4 (raw material pressure), N5
(Raw material valve opening) Further, for each node N (each failure mode), FIG.
As shown in (1), a failure state (cause) and an alarm are registered separately, and the cause of the failure can be traced back to the abnormality handling state of the node N. Furthermore, it is configured to generate an alarm corresponding to each node N (resultingly corresponding to a process variable) by an alarm. The occurrence of the alarm at each node N is passed to the display input device (display and keyboard) 111 by the alarm output means 101a, and the image can be displayed. Here, as an example, as a cause of a high-side abnormality of the reformer burner exhaust gas temperature, a thermometer thermocouple failure or the like is registered.

Further, the diagnosis system 1 includes a recognition setting means 105 for recognizing a value of a process variable in a process to be diagnosed and setting a corresponding node N to the abnormal response state or the normal response state. Have been. That is, the recognition setting unit 105 sets the corresponding node N to an abnormal or normal corresponding state according to the value of the process variable input to the input unit 106 of the diagnostic system 1 via the database 7. That is, the process 2 to be diagnosed is specified. For example, when performing a real-time diagnosis, the information input to the input unit 106 is arranged for each process variable, and the corresponding node N is set first. Set to one of the two states described. As a result, when the directed graph model 100 is viewed as a whole, the node N set in one of the abnormal response state and the normal response state causes the directed graph model 100 to
This is established according to the state of the fuel cell process 2.

Next, the fuel cell process 2
Graph model 100 in a state that can represent the state of
Therefore, a configuration for performing diagnosis and failure diagnosis will be described below. That is, the diagnostic system 1 includes the directed graph model 1
At 00, a cause node specifying means 107 for specifying an abnormal cause node corresponding to a main cause of abnormality according to the setting state of the abnormality / normal correspondence between the nodes. By the cause node specifying means 107, for example, while tracing the arrow of the directed graph in reverse only for the node in the abnormal response state, the terminal node N in the abnormal response state is set as a so-called root node (cause node) in the directed graph. Specified. Then, the failure cause stored in the failure cause storage means 104 is transferred to the output side for the cause node N by the diagnosis result output means 107a. The diagnostic system 1 of the present application operates by following the above sequence.

The basic operation has been described above. The diagnostic system 1 of the present application is characterized by the threshold value set corresponding to each node N. The first feature is that the threshold set for a given process variable is dependent on other process variables, and the second feature is that the threshold The system is configured so that external input setting can be easily performed based on so-called information stored in the database 7.

First, a specific description will be given with reference to FIG. 5 for easy understanding. The figure shows the display screen when the operator sets the upper and lower threshold values of the reformer burner exhaust gas temperature (T1) as an example by external input based on the battery current value (IDC). The image to see is shown.
In the figure, the horizontal axis is the battery current value, and the vertical axis is the reformer burner exhaust gas temperature. Then, a plurality of types of discrete points P are shown on the screen. These discrete points are information stored in the database 7 that can represent the operating states of the fuel cell processes 2 in the plurality of plants 3 described above. That is, information on the reformer burner exhaust gas temperature and the battery current value is extracted from the information from each plant 3, and is arranged and displayed as an image. As shown in FIG. 5, the discrete points P shown in FIG. 5 are within a certain tendency range, for example, when the battery current value is 480 A, the reformer burner exhaust gas temperature is about 650 ° F. Burner exhaust gas temperature deviates from the normal range of about 650 ° F. and can be divided into Px. Further, when the battery current value is in a range exceeding 300 A, the operation is generally performed in such a manner that the reformer burner exhaust gas temperature tends to increase. Therefore, it is understood that there is a certain relationship between these two in a normal operation state. From the above situation, in the diagnostic system 1 of the present application, the operator can externally set the threshold value while viewing the image based on the database as shown in FIG. In FIG. 5, a solid line indicates an upper limit threshold value externally set with respect to the reformer burner exhaust gas temperature in relation to the battery current. On the other hand, what is indicated by a chain line is the lower threshold value. In this way, the threshold for a given process variable is set in relation to other process variables.
Naturally, these pieces of information are stored and stored in the threshold value storage unit 103 described above, and are appropriately used in the system.

The situation of such use will be described with reference to the movement in the system. This mainly relates to the above-described first feature of the diagnostic system 1 of the present application. As described above, in the diagnostic system 1 of the present application, the threshold value of the predetermined process variable is determined in advance in relation to other process variables, and this relation is The threshold value is stored in the threshold value storage unit 103 by the input setting operation. Therefore, in a state after this relationship has been set in advance, the diagnosis system 1 performs a diagnosis while automatically dynamically changing the threshold value. To explain this function, as shown in FIG.
Includes other process variable information from the input unit 106,
Automatically referring to predetermined process variable information (for example, upper and lower threshold values shown in FIG. 5) stored in the threshold value storage unit 103 and defined in relation to other process variables. Change means 1 for changing the threshold value
08 is provided. By the function of this means 108,
In each diagnostic stage, the diagnostic system 1 can dynamically change the threshold value of a predetermined process variable based on other process information, and can perform an accurate diagnosis suitable for an operating condition.

On the other hand, with respect to the second feature of the external input setting of the threshold value, the diagnostic system 1 of the present application can employ various input methods. Basically, the system 1 of the present application
Is provided with the database 7 for storing and accumulating the history of each process variable as described above, and is provided with the display means 109 for displaying a change state of the value of the predetermined process variable from the database 7. Threshold value input means 1 capable of externally inputting a threshold value for a process variable
10 is provided. That is, the display interface 109a and the threshold input interface 1 shown in FIG.
10a and display input device 11 provided corresponding thereto
1, the display means 109 and the threshold value input means 110 described above are configured. In the diagnostic system 1 of the present application, the threshold value can be input in the following three types. B. First method of threshold external input setting FIG. 6 shows an image displayed on the display screen of the display means 109 when this method is adopted. In this method, upper and lower threshold values of a predetermined process variable are set by a polygonal line specifying a coordinate point. That is, a relationship graph between a predetermined process variable and another process variable corresponding to FIG. 5 is displayed in a lower portion of the screen. On the other hand, a specified process variable (type specification) can be specified by specifying a target TAG 60 displayed at the top of the screen (type specification), and other process variables can be similarly specified (type specification) at the top of the screen. Specification of related TAG61 to be displayed (type specification)
Is possible. The selection of each TAG 60, 61 is T
This can be performed by specifying a TAG corresponding to each process variable displayed on the AG selection unit 62. Further, the upper and lower thresholds can be set by setting the upper and lower coordinates displayed in the middle of the screen. In this method, the upper threshold value and the lower threshold value of a predetermined process variable can be externally set in a polygonal line format. Also in this case, the operating state of each fuel cell process 2 stored in the database 7 so far can be fully utilized.

FIG. 7 shows an image displayed on the display screen of the display means 109 when this method is adopted. In this method, upper and lower threshold values of a predetermined process variable are set based on a reference line L7. Also in this example, a relationship graph between a predetermined process variable and another process variable corresponding to FIG. 5 is displayed in a lower portion of the screen. On the other hand, the specification of the predetermined process variable is performed by using the target TAG displayed at the top
70 can be specified, and other process variables can be similarly specified by specifying the related TAG 71 displayed at the top of the screen. Each TAG can be selected by specifying a TAG corresponding to each process variable displayed in the TAG selection section 72. Further, the setting of the upper and lower threshold values is based on the reference line L displayed in the middle of the screen.
7 and the setting of the upper limit width and the lower limit width. Therefore, in this method, the upper and lower limits of the predetermined process variable can be set by external input in the form of a reference line. Also in this case, the operation history of each fuel cell process 2 stored in the database 7 can be fully utilized.

C. Third Method for Setting Threshold Value External Input FIG. 9 shows an image displayed on the display screen of the display means 109 when this method is used. In this method, upper and lower threshold values of a predetermined process variable are set by simply setting values, and this is a method which has been conventionally adopted. Also in this case, the specification of the predetermined process variable is performed by specifying the target TAG 90 displayed at the top of the screen.

The threshold value thus set externally is stored in the threshold value storage means 103 as described above, and is read into the threshold value change means 108 as necessary. Then, it is used for setting an appropriate threshold value and further for diagnosis.

Now, in the diagnostic system 1 of the present application,
In addition to the setting of the threshold value based on the above-described external input setting, the threshold value can be set also for a control target value set for a predetermined process variable. What performs this function is the function of the control target value dependent threshold value setting means 112 shown in FIG. 2, and the screen shown in FIG. It is a screen for defining and setting a relation to a value. That is, the operator specifies a predetermined process variable for which the threshold is to be set by specifying the target TAG 80, and sets the control target value to the threshold value by setting the set value TAG 83 to the control target value. It can be set to be a reference value for determination. Furthermore, by inputting the upper limit width and the lower limit width, the upper limit and the lower limit of the threshold can be specified. Setting a threshold value based on the control target value,
The setting of the threshold value of the external input setting reference shown above
They may work alternatively or both may work in parallel. In the case of working in parallel, it is preferable that a strict threshold is given priority. Furthermore, it may be configured so that it can be displayed separately for each purpose.

[Another Embodiment] Another embodiment of the present invention will be described below. (B) In the above embodiment, the case where the fuel cell process is targeted has been described, but the diagnostic system of the present invention is applicable to any process such as a cogeneration process and an absorption refrigeration process. It is. (B) In the above embodiment, the example of the reformer burner exhaust gas temperature is shown as the predetermined process variable, and the example of the battery current is shown as another example of the process variable. Any desired combination such as desulfurizer temperature and battery current, CO transformer temperature and battery current can be selected. Furthermore, as described above, the independent variable to which a predetermined process variable is dependent is described as an independent variable. For example, the battery current, the reformer burner, It is also possible to take the exhaust gas temperature control target value or the like as an independent value. That is,
As other process variables, there may be a plurality of them. In short, it is imperative that the setting of the threshold value be configured so that the operating state of the process is well represented in accordance with the purpose of improving the accuracy of the diagnostic system. (C) In the above embodiment, an example in which the threshold value is set based on the external input setting has been described. However, a configuration in which the system can be automatically set while referring to information from a database. It is preferable that (D) In addition to the purpose of using the above diagnostic system, in addition to the failure diagnosis in an extreme failure state in which a major failure actually occurs and the process stops,
According to this purpose, maintenance of a problematic part of the process (a failure cause part in a state close to a failure) may be performed in advance by separately setting an appropriate threshold value for the maintenance and management of the process. In such a case, while providing a threshold group for investigating the cause of the process stop,
It is preferable to operate the diagnostic system with a threshold group for maintenance. (E) In the above embodiment, an example of a model in which a failure mode is subjected to threshold processing and nodes classified as the failure mode are connected by a directed segment has been described. Nodes classified as faulty states may be connected by directed segments to form a model.

As described above, an accurate diagnosis result can be obtained at the time of the process diagnosis, and it is possible to quickly cope with the abnormality. Furthermore, by dynamically changing the threshold value according to the status of the equipment, it is possible to prevent erroneous diagnosis such as abnormal status even though the status is normal, or normal status even if the status is abnormal. Was more efficient. Furthermore, by setting a threshold value using existing data such as data from another device, it is possible to make an accurate determination and prevent erroneous diagnosis. (F) In the above embodiment, an example in which the threshold value is set in relation to other process variables while looking at the display screen displaying the history information of the process to be diagnosed and other similar processes However, in setting such a threshold, it is not always necessary to set it in relation to other process variables. That is, since the data appearing on the display screen can satisfactorily represent the state of the process, setting a threshold value while referring to the display screen makes it possible to easily set an accurate threshold value.

[Brief description of the drawings]

FIG. 1 is a schematic diagram of an overall configuration of a fuel cell failure diagnosis system employing a diagnosis system of the present application.

FIG. 2 is an explanatory diagram of a functional configuration of the diagnostic system of the present application.

FIG. 3 is a partial detailed view of a directed graph model of a fuel cell process.

FIG. 4 is an explanatory diagram showing a relationship among nodes, fault states, and alarms.

FIG. 5 is a diagram showing a display state of a battery current and a reformer burner exhaust gas temperature.

FIG. 6 is a diagram showing a display screen for setting a threshold by designating coordinate points;

FIG. 7 is a diagram showing a display screen for setting a threshold based on a reference line.

FIG. 8 is a diagram showing a display screen for threshold value setting based on a set value TAG.

FIG. 9 is a diagram showing a display screen for setting a threshold value by a conventional value designation method.

[Explanation of symbols]

 2 Fuel cell process 7 Database 100 Directed graph model 101 Model storage means 102 Directed line segment 105 Recognition setting means 107 Cause node specifying means 108 Threshold changing means 109 Display means 110 Threshold input means 112 Control target value dependent threshold Setting means N node Na upper limit node Nb lower limit node

Continued on the front page (72) Tetsuo Hirono, 1-2-1, Hirano-cho, Chuo-ku, Osaka-shi, Osaka Inside Osaka Gas Co., Ltd. (72) Shomei Yamazaki 4-1-2, Hirano-cho, Chuo-ku, Osaka, Osaka No. Osaka Gas Co., Ltd.

Claims (14)

[Claims] 1. A model storage means for storing a directed graph model for a process to be diagnosed, wherein the directed graph model connects a plurality of nodes provided corresponding to each process variable of the process by a directed segment. Along with being configured as a graph, the value of the process variable is compared to a preset threshold, depending on whether or not the node is in an abnormal state. When the node is not in the abnormal state, the node can be set to a normal corresponding state. The value of the process variable in the diagnosis target process is recognized, and the corresponding node is set to the abnormal corresponding state or the normal state. A recognition setting unit for setting a corresponding state, and in the directed graph model, according to the abnormal / normal corresponding setting state of the nodes. A diagnostic system comprising a cause node identifying means for identifying an abnormal cause node corresponding to a main abnormal cause, wherein the threshold value corresponding to a predetermined process variable is dynamically set according to the value of another process variable. And a threshold changing means for changing the threshold value, wherein the threshold value corresponding to the predetermined process variable is a function of the other process variable. 2. An upper limit threshold and a lower limit threshold can be set as the threshold for the process variable,
2. The directed graph model according to claim 1, wherein an upper node and a lower node respectively corresponding to the upper threshold and the lower threshold are provided as the nodes.
The described diagnostic system. 3. An external input setting of the threshold value for the predetermined process variable in association with a value of the other process variable, wherein the threshold value changing unit is configured to:
3. The diagnostic system according to claim 1, wherein the threshold value is dynamically changed based on a relationship with a value of another process variable. 4. A model storage means for storing a directed graph model for a process to be diagnosed, wherein the directed graph model is connected to a plurality of nodes provided corresponding to respective process variables of the process by a directed segment. Along with being configured as a graph, the value of the process variable is compared to a preset threshold, depending on whether or not the node is in an abnormal state. When the node is not in the abnormal state, the node can be set to a normal corresponding state. The value of the process variable in the diagnosis target process is recognized, and the corresponding node is set to the abnormal corresponding state or the normal state. A recognition setting unit for setting a corresponding state, and in the directed graph model, according to the abnormal / normal corresponding setting state of the nodes. A diagnostic system comprising a cause node identifying means for identifying an abnormal cause node corresponding to a main abnormal cause, based on at least information from a database storing past states of the process variables of the process to be diagnosed. And a threshold value input means for externally setting the threshold value for the predetermined process variable, the display means being configured to display a history of values of the predetermined process variable. 5. The display means is configured to be able to display the value of the predetermined process variable in relation to the value of another process variable, and the threshold input means is configured to display the value of the other process variable The diagnostic system according to claim 4, wherein the threshold value for the predetermined process variable is configured to be set by an external input. 6. An upper limit node corresponding to the upper limit threshold and a lower limit threshold respectively corresponding to the upper limit threshold and the lower limit threshold as the node for the process variable. And the lower limit node is provided to constitute the directed graph model, and the threshold input means is configured such that the upper limit threshold and the lower limit threshold can be externally input separately. 5. The diagnostic system according to 5. 7. An input unit for receiving each process variable information from a similar process other than the diagnosis target process,
The diagnostic system according to any one of claims 4 to 6, wherein the display means is configured to be able to display a value of the predetermined process variable in the process variable information from the other similar process. 8. An abnormal propagation comprising a model storage means for storing a directed graph model for a process to be diagnosed, wherein the directed graph model connects a plurality of nodes provided corresponding to each process variable of the process by a directed segment. Along with being configured as a graph, the value of the process variable is compared to a preset threshold, depending on whether or not the node is in an abnormal state. When the node is not in the abnormal state, the node can be set to a normal corresponding state. The value of the process variable in the diagnosis target process is recognized, and the corresponding node is set to the abnormal corresponding state or the normal state. A recognition setting unit for setting a corresponding state, and in the directed graph model, according to the abnormal / normal corresponding setting state of the nodes. A diagnostic system comprising a cause node specifying means for specifying an abnormal cause node corresponding to a main cause of abnormality, wherein when a predetermined process variable has a control target value, the diagnostic system has a relationship with the control target value. A diagnostic system comprising control target value dependent threshold setting means for determining a threshold. 9. An upper limit node corresponding to the upper limit threshold and a lower limit threshold respectively corresponding to the upper limit threshold and the lower limit threshold as the node. And a lower limit node, respectively, to configure the directed graph model, wherein the control target value dependent threshold input means is configured to be able to input and set the upper threshold and the lower threshold separately. Item 10. The diagnostic system according to Item 8. 10. A process having a plurality of process variables is diagnosed, and a value of each of the process variables is in an abnormal state according to whether or not the value is in an abnormal state as compared with a preset threshold value. A plurality of model parts that can be set to a normal response state when the abnormality is not in the abnormal response state, and recognizes the value of the process variable in the process to be diagnosed, and A diagnostic system comprising: recognition setting means for setting a portion to the abnormal response state or the normal response status; and a diagnostic system for deriving an abnormality cause according to the abnormal / normal response setting status between the model portions. Threshold value changing means for dynamically changing the threshold value corresponding to the variable according to the value of another process variable is provided, and the threshold value changing means corresponding to the predetermined process variable is provided. A diagnostic system wherein the threshold is a function of said other process variable. 11. A process having a plurality of process variables is set as a diagnosis target, and the value of each process variable is in an abnormal state according to whether or not the value is in an abnormal state as compared with a preset threshold value. A plurality of model parts that can be set to a normal response state when the abnormality is not in the abnormal response state, and recognizes the value of the process variable in the process to be diagnosed, and A diagnostic system comprising: recognition setting means for setting a part to the abnormal response state or the normal response state; and a diagnostic system for deriving an abnormality cause in accordance with the abnormality / normal response setting state of the model parts, at least A display means for displaying a history of values of predetermined process variables based on information from a database in which past states of the process variables of the process are accumulated. A diagnostic system comprising: a threshold; and a threshold value input unit that enables the threshold value for the predetermined process variable to be externally set. 12. A process having a plurality of process variables is diagnosed, and the value of each process variable is in an abnormal state depending on whether the value is in an abnormal state as compared with a preset threshold value. A plurality of model parts that can be set to a normal response state when the abnormality is not in the abnormal response state, and recognizes the value of the process variable in the process to be diagnosed, and A diagnostic system comprising: recognition setting means for setting a portion to the abnormal response state or the normal response status; and a diagnostic system for deriving an abnormality cause according to the abnormal / normal response setting status between the model portions. A diagnosis system comprising a control target value dependent threshold value setting means for determining the threshold value in relation to the control target value when a variable has a control target value. 13. The diagnostic system according to claim 1, wherein the process to be diagnosed is a fuel cell facility process. 14. The diagnostic system according to claim 1, wherein the process to be diagnosed is a fuel cell facility process, and the other process variable is a battery current value in a fuel cell.