JP7428598B2 - Fault diagnosis system, fault diagnosis method, and program - Google Patents

Description

The present invention relates to a failure diagnosis system and the like.

As a technique for estimating the cause of equipment failure, for example, the technique described in Patent Document 1 is known. In other words, Patent Document 1 describes a ``first causal network probabilistically representing the relationship between a plurality of failure types and observed information having a causal relationship with each failure type,'' and ``a first causal network that probabilistically represents the relationship between a plurality of failure types and observation information that has a causal relationship with each failure type.'' A failure diagnosis system is described that includes a plurality of second causal networks that probabilistically represent the relationship between information and failure causes.

Japanese Patent Application Publication No. 2007-328645

In the technique described in Patent Document 1, the probability of occurrence of each failure type is calculated based on the detection values (observation information) of a plurality of sensors installed in a device, and further, failure cause candidates are extracted. However, the technique described in Patent Document 1 does not take into account, for example, the causal relationship between failure types when a plurality of failure types occur simultaneously.

In actual equipment, when a given failure type occurs, that failure type often causes another failure type to occur. In equipment maintenance work, one failure type is repaired, but a higher-order failure type that is the cause of that failure type may remain unrepaired. In such cases, it is difficult for maintenance personnel to notice the existence of higher-level failure types, and the root cause of the failure is not removed, resulting in lower-level failure types that have been repaired occurring again, resulting in time and costs being spent on equipment maintenance. There are circumstances that require it.

Therefore, an object of the present invention is to provide a failure diagnosis system and the like that facilitate maintenance work for equipment.

In order to solve the above-mentioned problems, the present invention provides a communication unit that performs predetermined communication with a terminal operated by a maintenance person who performs maintenance work on equipment, and an event that can cause a failure of the equipment. a failure cause-and-effect model including causal relationships between a plurality of failure modes and a causal relationship between a plurality of inspection items regarding the device and a plurality of the failure modes; a failure mode occurrence index value estimation unit that calculates a predetermined index value indicating the likelihood of occurrence of each of the plurality of failure modes based on the results of the plurality of inspection items; and the failure mode occurrence index value estimation unit. a terminal display information generation processing section that generates information to be displayed on the terminal by associating the estimation results of the test item with the failure cause and effect model; After the index values of each of the plurality of failure modes based on the above are displayed on the terminal in association with the failure causal model, a repaired failure mode that is a predetermined failure mode that has been repaired by the maintenance person is displayed. When information is received from the terminal via the communication unit, the index value for each of the plurality of failure modes is calculated again based on the failure causal model and the information on the repaired failure mode, and the information is sent to the terminal. The display information generation processing unit is characterized in that it associates the index values of each of the plurality of failure modes with the failure causal model and generates again information to be displayed on the terminal.

According to the present invention, it is possible to provide a failure diagnosis system and the like that facilitate maintenance work for equipment.

FIG. 1 is a functional block diagram showing the configuration of a failure diagnosis system according to a first embodiment. FIG. 2 is an explanatory diagram of information included in a failure causal model of the failure diagnosis system according to the first embodiment. FIG. 2 is an explanatory diagram showing a data structure of inspection item data of the failure diagnosis system according to the first embodiment. FIG. 3 is an explanatory diagram showing a data structure of lower failure mode data of the failure diagnosis system according to the first embodiment. FIG. 2 is an explanatory diagram showing a data structure of top-level failure mode data of the failure diagnosis system according to the first embodiment. FIG. 2 is an explanatory diagram showing a data structure of failure mode probability data of the failure diagnosis system according to the first embodiment. FIG. 2 is an explanatory diagram showing a data structure of inspection item-failure mode probability data of the failure diagnosis system according to the first embodiment. FIG. 2 is an explanatory diagram showing a data structure of lower failure mode-higher failure mode probability data of the failure diagnosis system according to the first embodiment. FIG. 2 is an explanatory diagram showing a data structure of initial information stored in a temporary storage unit of the failure diagnosis system according to the first embodiment. FIG. 2 is an explanatory diagram showing a data structure of test item results and failure mode results stored in a temporary storage unit of the failure diagnosis system according to the first embodiment. FIG. 2 is a functional block diagram including a failure mode probability estimator of the failure diagnosis system according to the first embodiment. It is a flowchart showing the flow of processing in the failure diagnosis system according to the first embodiment. FIG. 2 is an explanatory diagram showing a display example of an initial information input screen on a terminal in the failure diagnosis system according to the first embodiment. FIG. 2 is an explanatory diagram showing causal relationships among failure modes and the like displayed on a terminal in the failure diagnosis system according to the first embodiment. FIG. 2 is an explanatory diagram showing a state of a predetermined test item of causal relationship such as a failure mode displayed on a terminal in the failure diagnosis system according to the first embodiment. FIG. 7 is an explanatory diagram showing an example of a screen displayed on a terminal when a lower failure mode is determined and a repair result is transmitted in the failure diagnosis system according to the first embodiment. FIG. 7 is a functional block diagram showing a configuration for extracting a causal relationship between a higher-order failure mode and a lower-order failure mode in a server of a failure diagnosis system according to a second embodiment.

≪First embodiment≫
<Fault diagnosis system configuration>
FIG. 1 is a functional block diagram showing the configuration of a failure diagnosis system 100 according to the first embodiment.
A failure diagnosis system 100 shown in FIG. 1 is a system for diagnosing the cause of a failure of a device M. In addition, examples of equipment M to be maintained include, in addition to air conditioning equipment and refrigeration equipment, refrigerators, railway vehicles, aircraft, power generation equipment, power distribution equipment, medical equipment, inspection equipment, communication equipment, and ATMs (Automatic Teller Machines). However, it is not limited to this. Further, calculating the probability of occurrence of a failure mode of the device M (a predetermined index value indicating the ease with which a failure mode occurs) is also included in the matter of failure diagnosis of the device M. Furthermore, the term "failure mode" refers to an event that can cause a failure of the device M.

As shown in FIG. 1, the failure diagnosis system 100 includes a terminal 10 and a server 20. The terminal 10 is a device that performs predetermined processing based on operations by the maintenance person R, and is carried by the maintenance person R. As such a terminal 10, a tablet, a smartphone, a notebook computer, etc. are used. Note that the maintenance person R is a person who performs maintenance work such as inspection and repair of the equipment M and inputs the results into the terminal 10.

As shown in FIG. 1, the terminal 10 includes an input section 11, a communication section 12, and an output section 13. The input unit 11 is, for example, a touch panel or a keyboard, and has a function of inputting data by the maintenance person R's operation. The communication unit 12 has a function of performing predetermined communication with the server 20 via a network (not shown). The output unit 13 is, for example, a liquid crystal display, and displays predetermined information. In addition, although not shown, the terminal 10 also includes a storage section and a control section.

The data input via the input unit 11 of the terminal 10 includes initial information, the results (states) of inspection items, and the results (states) of failure modes. The initial information is information that is grasped by the maintenance person R at the start of maintenance work on the device M. Such initial information includes the model number and unique ID of the device M. The inspection items are items for which the maintenance person R inspects the equipment M, and are set in advance. Examples of such inspection items include the detection values of each sensor (not shown) installed in the equipment M, but there are also inspection items such as segment display of the control board (not shown). Sometimes it is included.

The failure mode result is the presence or absence of a predetermined failure mode confirmed by the maintenance person R. To explain the case where device M is a refrigerator equipped with a vapor compression type refrigerator, examples of failure modes include frost forming on the evaporator (not shown) and failure of the compressor electric motor (not shown). ) is damaged.

The server 20 has a function of calculating the probability of occurrence of a failure mode of the device M and extracting causal relationships regarding the failure mode based on information received from the terminal 10. Although not shown in the drawings, the hardware configuration of the server 20 includes electronic circuits such as a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), and various interfaces. Then, the program stored in the ROM is read out and expanded to the RAM, and the CPU executes various processes.

As shown in FIG. 1, the server 20 has a functional configuration including a communication section 21, an input reading section 22, a temporary storage section 23, and a failure mode probability estimation section 24 (failure mode occurrence index value estimation section). , a failure causal model 25, an estimation result output unit 26, a causal relationship extraction unit 27, a failure history database 28, and a terminal display information generation processing unit 29.

The communication unit 21 of the server 20 has a function of performing predetermined communication (transmission or reception of information) with the communication unit 12 of the terminal 10. For example, the communication unit 21 receives, as information input into the terminal 10 by the maintenance person R who performs maintenance work on the device M, the results of inspection items and the results of failure modes in addition to the above-mentioned initial information. Each piece of information received by the communication section 21 is output to the input reading section 22 .

The input reading unit 22 reads information input from the communication unit 21 (initial information, results of inspection items, results of failure modes, etc.). The information read by the input reading unit 22 is stored in the temporary storage unit 23.
The failure mode probability estimation unit 24 calculates the probability of occurrence of each failure mode based on the information stored in the temporary storage unit 23 and the information of the failure causal model 25. Although details will be described later, the failure mode probability estimation unit 24 calculates the probability of occurrence of each failure mode using, for example, a probability statistical method based on a Bayesian network. The calculation result of the failure mode probability estimation section 24 is output to the estimation result output section 26.

The failure cause-and-effect model 25 includes information that specifies the causal relationships between multiple failure modes and the causal relationships between multiple inspection items regarding the device M and (some of) the multiple failure modes. Specifically, the failure causal model 25 is a database group that stores predetermined parameters used by the failure mode probability estimating unit 24 to calculate the probability of occurrence of each failure mode, and is set in advance. For example, the failure cause-and-effect model 25 has multiple inspection items and multiple failure modes as nodes, as well as causal relationships between the multiple failure modes and (parts of) the multiple failure modes and multiple inspection items. It is configured as a Bayesian network with causal relationships between them as links.

The calculation results of the failure mode probability estimation section 24 are input to the estimation result output section 26 . Then, the estimation result output unit 26 outputs to the communication unit 21 a predetermined group of pairs of a predetermined failure mode whose occurrence probability is to be estimated and the probability of occurrence of this failure mode.

The causal relationship extraction unit 27 extracts a causal relationship between an inspection item of the device M and a predetermined failure mode, and a plurality of failures based on the information stored in the temporary storage unit 23 and the failure causal model 25. Extract causal relationships between modes. For example, when a Bayesian network is used as the failure causal model 25, the causal relationship extraction unit 27 extracts those that have a causal relationship with each other in the Bayesian network from among predetermined inspection items and a plurality of failure modes.

In addition, the causal relationship extraction unit 27 also has a function of reading out the status of inspection items (inspection results) and failure mode status (failure mode confirmation result by maintenance person R) from the temporary storage unit 23. The causal relationship extraction unit 27 also has a function of referring to the failure history database 28, reading the past maintenance history of the device M, and adding the maintenance history to the extracted causal relationship information. The details of extracting causal relationships and adding maintenance history will be described later.

The failure history database 28 is a database that includes information on the history of failure repairs of the equipment M. Note that the results of inspection items and failure modes regarding the device M are transmitted from the terminal 10 to the server 20 by the maintenance personnel R's operations. In this case, the results of the inspection items and the failure mode of the device M are additionally stored in the failure history database 28 in association with the initial information stored in the temporary storage unit 23.

The terminal display information generation processing section 29 associates the estimation result of the failure mode probability estimation section 24 with the failure causal model 25 and generates predetermined information for displaying on the terminal 10. Information generated by the terminal display information generation processing section 29 is transmitted from the communication section 21 to the terminal 10 via a network (not shown). Note that the terminal display information generation processing section 29 may be included in the terminal 10 instead of the server 20.

FIG. 2 is an explanatory diagram of information included in the failure cause-and-effect model 25.
As shown in FIG. 2, the failure cause and effect model 25 includes inspection item data 25a, lower failure mode data 25b, top failure mode data 25c, failure mode probability data 25d, and inspection item-failure mode probability data 25e. and lower failure mode-higher failure mode probability data 25f. Each of these data is stored in a predetermined storage means (not shown) included in the server 20 (see FIG. 1). Each of these pieces of data will be explained using FIGS. 3 to 8 using an example in which the device M is a refrigerator equipped with a vapor compression refrigerator.

FIG. 3 is an explanatory diagram showing the data structure of the inspection item data 25a.
The inspection item data 25a shown in FIG. 3 is data showing a list of inspection items for the device M (see FIG. 1), and is included in the failure causal model 25 (see FIG. 2). As shown in FIG. 3, the test item data 25a includes test item IDs, test item names, and search item status candidates, which are set in advance. The inspection item ID is a unique ID assigned in advance to the inspection item of the device M.

The inspection item name is associated with the inspection item ID as a predetermined name so that the maintenance person R can understand the contents of the inspection item. For example, the test item ID: I_001 is given the test item name "Is the power supply voltage normal?" Furthermore, the inspection item ID: I_002 is given the inspection item name "Is there enough heat exhaust space?"

The inspection item status candidate is a response candidate when the maintenance person R (see FIG. 1) inputs the inspection result of a predetermined inspection item by operating the terminal 10, and is associated with the inspection item ID. In the example of FIG. 3, two answer candidates (state candidates), "Yes" and "No", are provided for inspection item IDs: I_001 and I_002. Note that the options for the test item state candidates are not limited to two, "Yes" and "No," and three or more predetermined options may be provided. Further, the test item state candidates are not limited to options, but may be a predetermined numerical range. For example, in association with the inspection item name (not shown in Fig. 3), "What is the detected value of the temperature sensor inside the refrigerator?", "-30.0℃ to -20 A predetermined numerical range such as ".0°C" may be set.

FIG. 4 is an explanatory diagram showing the data structure of the lower failure mode data 25b.
Here, before explaining each column in FIG. 4, lower failure modes and upper failure modes will be explained. When there is an abnormality in the device M, a plurality of failure modes (events that cause a failure) that have a causal relationship with each other often occur in a chain. To give a specific example, if the status of the inspection item that the refrigerator "does not cool down" (see Figure 15) is "Yes", the direct cause is "There is frost on the evaporator" (see Figure 15). 15) failure mode may occur. Further, as a cause of the failure mode "There is frost on the evaporator", another failure mode called "drain pan clogging" (see FIG. 15) may occur.

In other words, the root cause is a "drain pan blockage" (see Figure 15), which causes the phenomenon of "frost on the evaporator" and, as a result, the phenomenon that the refrigerator "does not cool" (inspection Item results) may occur. When a plurality of failure modes are considered to occur in a chain manner due to a predetermined causal relationship, a failure mode whose causal relationship is relatively higher (closer to the root cause) is referred to as a higher failure mode. Furthermore, a failure mode whose causal relationship is relatively low-level (far from the root cause) is called a low-level failure mode. Here, the words "lower" and "higher" indicate a relative relationship with other failure modes.

Further, in a predetermined causal relationship regarding a failure of the device M, the highest failure mode is referred to as the highest failure mode. The number of top-level failure mode data is not limited to one, but often exists in plural pieces.
Note that the lower failure mode data 25b in FIG. 4 shows data of a failure mode that is relatively lower than the highest failure mode (see FIG. 2) that is the highest. Therefore, among all the failure modes set in advance, some of them are included in the highest failure mode data 25c (see FIG. 2), and the rest are included in the lower failure mode data 25b (see FIGS. 2 and 4). It is.

The lower failure mode data 25b shown in FIG. 4 is data indicating the content of the lower failure mode, and is included in the failure causal model 25 (see FIG. 2) as described above. As shown in FIG. 4, the lower failure mode data 25b includes a failure mode ID, a failure mode name, and a failure mode state candidate, which are set in advance. The failure mode ID is a unique ID assigned in advance to each failure mode.

The failure mode name is associated with the failure mode ID as a predetermined name so that the maintenance person R can understand the contents of the failure mode. For example, the failure mode ID: FM_001 is associated with the failure mode name "refrigerant shortage." Furthermore, the failure mode name “low pressure pipe clogging” is associated with the failure mode ID: FM_002.

The failure mode state candidate is an answer candidate when the maintenance person R confirms the state of a predetermined failure mode and then inputs it by operating the terminal 10, and is associated with the failure mode ID. In the example of FIG. 4, two answer candidates (state candidates), "Yes" and "No", are provided for the failure mode IDs: FM_001 and FM_002. Note that "Yes" in the failure mode state candidate means that a predetermined failure mode has occurred, while "No" means that the failure mode has not occurred.

Further, the failure mode state candidates are not limited to two, "Yes" and "No", and three or more predetermined options may be provided. Further, the failure mode state candidates are not limited to options, but may be a predetermined numerical range.

FIG. 5 is an explanatory diagram showing the data structure of the highest level failure mode data 25c.
The highest level failure mode data 25c shown in FIG. 5 is data indicating the contents of the highest level failure mode, and is included in the failure causal model 25 (see FIG. 2) as described above. The top failure mode is a predetermined failure mode that is not caused by other failure modes. As shown in FIG. 5, the highest failure mode data 25c includes a failure mode ID, a failure mode name, and a failure mode state candidate, which are set in advance. Each of these items is the same as the lower failure mode data 25b in FIG. 4, so a description thereof will be omitted.

FIG. 6 is an explanatory diagram showing the data structure of the failure mode probability data 25d.
The failure mode probability data 25d shown in FIG. 6 is data indicating the probability that each failure mode will occur, and is included in the failure causal model 25 (see FIG. 2) as described above. As shown in FIG. 6, the failure mode probability data 25d includes a failure mode ID, a failure mode name, a failure mode state, and an occurrence probability.
Note that the failure mode probability data 25d includes both data on the probability of occurrence of the highest failure mode and data on the probability of occurrence of the lower failure modes. In other words, data regarding the probability of occurrence of each of all failure modes is included in the failure mode probability data 25d.

The failure mode ID and failure mode name shown in FIG. 6 are the same as those shown in FIG. 4 and FIG. 5, so the explanation will be omitted. Note that the failure mode IDs: FM_001 and FM_002 shown in FIG. 6 are the same as those shown in FIG. 4 (lower failure mode data 25b). Although not shown in FIG. 6, data on the probability of occurrence of failure mode IDs: FM_003, FM_009, etc. (the highest failure mode data 25c in FIG. 5) is also included in the failure mode probability data 25d.

The failure mode state is a predetermined state (for example, "Yes") in the failure mode, and is a state whose occurrence probability is to be calculated. To give a specific example, regarding failure mode ID: FM_001, the probability that the failure mode state of "refrigerant shortage" (that is, the failure mode state) becomes "Yes" is 0.2 in the example of FIG. There is. Such a failure mode state corresponds to one of the above-described failure mode state candidates (“Yes” or “No” in FIGS. 4 and 5).

The occurrence probability shown in FIG. 6 is the probability that a predetermined failure mode state will occur. A predetermined initial value may be set in advance as this probability. Then, based on the processing of the failure mode probability estimating unit 24, the probability of occurrence of the failure mode may be updated sequentially.

FIG. 7 is an explanatory diagram showing the data structure of the inspection item-failure mode probability data 25e.
The inspection item-failure mode probability data 25e is data indicating the probability that a predetermined state of the inspection item occurs when each failure mode occurs, and as described above, the failure causal model 25 (see FIG. 2) ) included. As shown in FIG. 7, the test item-failure mode probability data 25e includes a failure mode ID, a failure mode name, a failure mode state, a test item ID, a test item name, a test item state, and an occurrence These are set in advance.

The failure mode ID is set to match the ID of the predetermined lower failure mode data 25b (see FIG. 4) or the highest failure mode data 25c (see FIG. 5). The failure mode names and failure mode states are the same as those explained using FIG. 6 (failure mode probability data 25d). Furthermore, the test item IDs and test item names are the same as those explained using FIG. 3 (test item data 25a), so their explanation will be omitted.

The test item state is data indicating a state (for example, "Yes") for which the occurrence probability of a predetermined test item is calculated. The occurrence probability is the probability that an inspection item is in a predetermined test item state when the failure mode is in a predetermined failure mode state. In the example of FIG. 7, when the failure mode state of "refrigerant shortage" is "Yes", the probability that the state of the inspection item "Is the refrigerant overheating" is "Yes" is 0.8. It has become. Note that it is assumed that the causal relationship between inspection items and failure modes is set in advance. Furthermore, in the inspection item-failure mode probability data 25e, it is assumed that failure mode IDs and inspection item IDs are associated in advance based on the above-mentioned causal relationship.

FIG. 8 is an explanatory diagram showing the data structure of the lower failure mode-higher failure mode probability data 25f.
The lower failure mode-higher failure mode probability data 25f is data indicating the probability that each lower failure mode will occur when a predetermined upper failure mode occurs, and as described above, the failure causal model 25 ( (see Figure 2).

Note that the higher-ranking failure modes shown in Figure 8 are not limited to the highest-ranking failure modes described above, but are higher-ranking failure modes relative to other failure modes (the root cause in the causal relationship of failure modes). (close) failure mode. For example, if one of the causes of "refrigerant shortage" (see Figure 8) is "low-pressure pipe cracking" (see Figure 8), the failure mode of "low-pressure pipe cracking" is different from the failure mode of "refrigerant shortage". It is relatively higher than that.

As shown in FIG. 8, the lower failure mode-higher failure mode probability data 25f includes an upper failure mode ID, a higher failure mode name, and a higher failure mode state. Further, the lower failure mode-higher failure mode probability data 25f includes a lower failure mode ID, a lower failure mode name, a lower failure mode state, and an occurrence probability.

The upper failure mode ID is a unique ID assigned in advance to a predetermined upper failure mode. The higher-order failure mode name is associated with the failure mode ID as the name of a predetermined higher-order failure mode. The upper failure mode state is a predetermined state (for example, "Yes") in the upper failure mode.

The lower failure mode ID is the lower failure mode that may be affected by the failure when the upper failure mode becomes a predetermined upper failure mode state (for example, the state of "low pressure pipe cracking" is "Yes"). This is the mode ID. The lower failure mode name is associated with the lower failure mode ID as the name of a predetermined lower failure mode. The lower failure mode state is a predetermined state (eg, "Yes") in the lower failure mode.

The occurrence probability is the probability that a lower-order failure mode becomes a predetermined lower-order failure mode when a higher-order failure mode becomes a predetermined upper-order failure mode. In the example of Fig. 8, when the state of the higher-order failure mode of "low-pressure piping crack" is "Yes", the probability that the lower-order failure mode of "insufficiency of refrigerant" is "Yes" is 0.9. ing. Note that the causal relationship between the upper failure mode and the lower failure mode is set in advance as the lower failure mode-higher failure mode probability data 25f.

Note that the information regarding the ID, name, and state candidate of the lower failure mode shown in FIG. 8 is associated with the lower failure mode data 25b of FIG. 4. On the other hand, the information on the ID/name/state candidate of the higher-order failure mode shown in FIG. 8 is included in either the highest-order failure mode data 25c (see FIG. 5) or the lower-order failure mode data 25b (see FIG. 4). . This is because, as described above, the highest-level failure mode is the highest-level failure mode that is considered to be the root cause of the failure, and the high-level failure mode in FIG. 8 does not necessarily correspond to the highest-level failure mode.

For example, if there is a causal relationship from cause to effect such as "failure mode A → failure mode B → failure mode C → inspection item S", "failure mode A" is included in the highest failure mode data 25c, “Failure mode B” and “failure mode C” are included in the lower failure mode data 25b. In addition, in the lower failure mode-higher failure mode probability data 25f, among the causal relationships described above, when the causal relationship of "failure mode A → failure mode B" is specified, the higher failure mode is "failure mode A ”, and the lower failure mode is “failure mode B”. Furthermore, when the causal relationship of "failure mode B→failure mode C" is specified, the higher-order failure mode is "failure mode B" and the lower-order failure mode is "failure mode C."

FIG. 9A is an explanatory diagram showing the data structure of initial information stored in the temporary storage unit.
As described above, the temporary storage unit 23 (see FIG. 1) includes the initial information of the device M (see FIG. 1) as well as the results of predetermined inspection items and failure modes. Among the various data stored in the temporary storage unit 23 in this way, the data structure of the initial information of the device M is shown in FIG. 9A. As shown in FIG. 9A, the initial information includes an initial information name and initial information content. The initial information name is the name of data such as the model number and unique ID of the device M that is grasped by the maintenance person R side (administrator side) at the time of starting the maintenance work. The initial information content is predetermined data corresponding to the initial information name.

FIG. 9B is an explanatory diagram showing the data structure of test item results and failure mode results stored in the temporary storage unit.
As shown in FIG. 9B, the inspection item and failure mode results include "type", "ID", "name", and "state". “Type” is a symbol used to determine whether the item corresponds to an inspection item or a failure mode. "ID" is identification information for each inspection item or failure mode, and a predetermined ID included in the failure cause-and-effect model 25 (see FIG. 1) is used. "Name" is the name of the inspection item or failure mode, and is associated with the above-mentioned ID. "Status" is the state of the inspection item (for example, "Yes" or "No"), as well as the state of the failure mode (for example, "Yes" or "No"), and is the state of the inspection item (for example, "Yes" or "No"). After being input by operating the terminal 10 (see FIG. 1), the information is received from the terminal 10 via the communication unit 21 (see FIG. 1).

FIG. 10 is a functional block diagram including the failure mode probability estimation section 24 (see also FIG. 1 as appropriate).
In addition, in FIG. 10, arrows indicate the exchange of data between each unit when the failure mode probability estimating unit 24 uses a Bayesian network to calculate the probability of occurrence of a failure mode. As shown in FIG. 10, the failure mode probability estimation unit 24 includes a failure mode probability extraction unit 24a, an inspection item conditional probability extraction unit 24b, a lower failure mode conditional probability extraction unit 24c, and a failure mode conditional probability calculation unit 24a. A portion 24d is provided.

The failure mode probability extraction unit 24a extracts the probability P(FM) of occurrence of a predetermined failure mode, which is a target of probability estimation, from the failure mode probability data 25d (see also FIG. 6). This probability P(FM) is the probability that a predetermined failure mode FM will occur when no conditions are given in advance (that is, it is not a conditional probability), and is set in advance.

The test item conditional probability extraction unit 24b determines whether each test item I is in a predetermined state stored in the temporary storage unit 23 (for example, “ The probability P(I|FM) of the state "Yes" is extracted from the inspection item-failure mode probability data 25e (see FIG. 7). This probability P(I|FM) is a so-called conditional probability and is set in advance. Further, the predetermined state stored in the temporary storage unit 23 refers to the result (state) of the inspection item transmitted from the terminal 10 to the server 20 (see FIG. 1) based on the operation by the maintenance person R (see FIG. 1). ).

The lower failure mode conditional probability extraction unit 24c extracts a lower failure mode MFM that has a causal relationship such that the failure mode FM is a cause when a predetermined failure mode FM that is a target of probability estimation occurs in a temporary storage unit. The probability P (MFM|FM) of a predetermined state (for example, "Yes" state) stored in 23 is extracted from the lower failure mode-upper failure mode probability data 25f (see FIG. 8). Here, the predetermined state stored in the temporary storage unit 23 refers to the failure mode result (refer to FIG. 1) sent from the terminal 10 to the server 20 (see FIG. condition).
Note that if there is no lower-order failure mode MFM that has a causal relationship with the failure mode FM, extraction by the lower-order failure mode conditional probability extraction unit 24c is not particularly performed.

The failure mode conditional probability calculation unit 24d calculates the probability P of failure mode FM occurring when the inspection item and the lower failure mode are in a predetermined state (for example, “Yes” state) stored in the temporary storage unit 23. Calculate (FM|I, MFM). Such processing by the failure mode conditional probability calculation unit 24d is based on each probability extracted by the failure mode probability extraction unit 24a, the inspection item conditional probability extraction unit 24b, and the lower failure mode conditional probability extraction unit 24c. It will be done.

The probability P(I|FM) and the probability P(MFM|FM) described above are conditional probabilities when a predetermined failure mode FM is the cause. The failure mode conditional probability calculation unit 24d performs predetermined extraction with respect to all test items and lower failure modes whose states (for example, “Yes” state) are stored in the temporary storage unit 23. Then, the failure mode conditional probability calculation unit 24d calculates the probability that failure mode FM will occur when the inspection item or the lower failure mode reaches a predetermined state stored in the temporary storage unit 23. For example, the well-known Bayes formula is used for such calculations. The probability P (FM|I, MFM) calculated by the failure mode conditional probability calculation unit 24d is output to the estimation result output unit 26.

Then, as described above, the estimation result output unit 26 outputs a group of pairs of failure modes and probabilities of occurrence of the failure modes to the communication unit 21. This pair group is transmitted to the terminal 10 (see FIG. 1) via the communication section 21 (see FIG. 1) together with the information extracted by the causal relationship extraction section 27 (see FIG. 1), and is sent to the output section 13 of the terminal 10. (See Figure 1). As a result, the maintenance person R who owns the terminal 10 can understand the causal relationship between each inspection item and each failure mode, and can also understand the probability that each failure mode will occur.

Note that the causal relationship extraction unit 27 may refer to the failure history database 28, read the past maintenance history of the device M to be maintained, and add maintenance history information to the extracted causal relationship. For example, the causal relationship extraction unit 27 refers to the past repair history of the device M, and adds the failure mode to the causal relationship information based on the number of results of inspection items that have been performed in the repair history (for example, in a predetermined period in the past). do. In the failure history database 28, if the number of times repairs have been performed in the past is greater than or equal to a predetermined value, and the repair is still being performed for another higher-order failure mode linked in the failure causal model 25 as the cause of the failure. If there is a predetermined failure mode that does not exist, the terminal display information generation processing unit 29 generates information for causing the terminal 10 to display a warning to the effect that repair of another higher-order failure mode should be performed. This information is transmitted to the terminal 10 via the communication unit 21 of the server 20, and displayed in a predetermined manner on the terminal 10. As a result, the maintenance person R can understand that the root cause of the failure of the equipment M has not been removed, so that the failure of the equipment M can be prevented from occurring again.

FIG. 11 is a flowchart showing the flow of processing in the failure diagnosis system (see FIG. 1 as appropriate).
In step S101 in FIG. 11, the server 20 uses the input reading unit 22 to read initial information. As described above, the initial information is input by the maintenance person R's operation via the input unit 11 of the terminal 10. The information read by the input reading section 22 is stored in the temporary storage section 23.

FIG. 12 is an explanatory diagram showing a display example of the initial information input screen on the terminal (see also FIG. 1 and FIG. 9A as appropriate).
In the example of FIG. 12, an input screen for inputting the model number and ID (device ID) of the device M, as well as the maintenance person ID, etc., is displayed on the output unit 13 of the terminal 10. After inputting these initial information, by pressing the send button B1 displayed on the output unit 13, the initial information is sent from the terminal 10 to the server 20. Note that instead of inputting information into the terminal 10 by the maintenance person R, the information entered by the receptionist may be used as appropriate when accepting a failure repair request.

Although not shown in FIG. 11, the server 20 (see FIG. 1) selects the failure causal model 25 corresponding to the device ID based on the initial information stored in the temporary storage unit 23, and extracts the causal relationship. Information on a predetermined causal relationship corresponding to the initial information is transmitted to the terminal 10 via the unit 27 . Information on such causal relationships is received by the communication unit 12 of the terminal 10 and further displayed on the output unit 13.

FIG. 13 is an explanatory diagram showing the causal relationship of failure modes, etc. displayed on the terminal (see FIG. 1 as appropriate).
In the example shown in Figure 13, there is a causal relationship between inspection items such as "Is the power supply voltage normal?", lower-level failure modes such as "Insufficient air volume in the evaporator," and higher-level failure modes such as "Damage in the evaporator blower motor." is displayed in a tree format. In other words, the terminal display information generation processing unit 29 determines the causal relationships between multiple failure modes and the causal relationships between (some of) the multiple failure modes and the results of multiple inspection items on the terminal 10. Information for displaying in a tree shape is generated and this information is transmitted to the terminal 10, and as a result, a predetermined image is displayed on the terminal 10.

Specifically, as shown in FIG. 13, a test item presentation field C1 (node), a causal relationship presentation line L1 (link), a failure mode presentation field C2 (node), and a send button B2 are provided on the terminal 10. (see FIG. 1) is displayed on the output section 13 (see FIG. 1). The test item presentation field C1 includes a test item name field Ca1 and a test item state candidate field Cb1. In the test item name column Ca1, predetermined test item names (see FIG. 3) in the test item data 25a are displayed. In the test item state candidate column Cb1, predetermined test item state candidates (see FIG. 3) in the test item data 25a are displayed. In the example of FIG. 13, candidates "Yes" and "No" are displayed as character strings in the test item state candidate column Cb1, and one of the states is selected by the maintenance person R's operation.

The causal relationship presentation line L1 shown in FIG. 13 is a line that indicates the causal relationship between failure modes and inspection items with arrows, and also indicates the causal relationship between upper failure modes and lower failure modes with arrows. The failure mode presentation field C2 includes a failure mode name field Ca2, a failure mode state candidate field Cb2, and a failure mode state probability display field Cc2. Then, the predetermined failure mode name included in the highest failure mode data 25c (see FIG. 5) or the lower failure mode data 25b (see FIG. 4) is displayed in the failure mode name field Ca2, and the failure mode state candidate is It is displayed in the mode state candidate column Cb2.

In the terminal 10, when inspection items and failure mode states have not yet been input, a predetermined probability set in advance is displayed in the failure mode state probability display column Cc2. Then, after the maintenance person R starts a predetermined maintenance work and inputs inspection items as appropriate, when the send button B2 is pressed, information on the inspection items is transmitted from the terminal 10 to the server 20. At the time of this transmission, it is not necessary for the maintenance person R to input all the statuses of the inspection items of the device M into the terminal 10, and the status may be a part of the inspection items.

In step S102 of FIG. 11, the server 20 uses the input reading section 22 to read the test item results and stores them in the temporary storage section 23. In other words, the server 20 performs predetermined communication with the terminal 10 and receives information including the results of predetermined test items regarding the device M (first process).
Next, in step S103, the server 20 calculates the probability of occurrence of the failure mode, and further extracts the causal relationship between the failure modes. Specifically, the server 20 uses the failure mode probability estimation unit 24 to determine each of the plurality of failure modes based on the failure causal model 25 and the results of the plurality of inspection items (data stored in the temporary storage unit 23). Calculate the probability of occurrence (index value) of (second process).

Further, in step S103, the server 20 uses the causal relationship extraction unit 27 to extract causal relationships such as failure modes based on the information in the temporary storage unit 23, failure causal model 25, and failure history database 28. Note that if the temporary storage unit 23 also stores the results of lower failure modes, that information is also used appropriately in the process of step S103.

In step S104, the server 20 presents the probability of failure mode occurrence and causal relationship to the user. That is, the server 20 causes the terminal display information generation processing section 29 to generate predetermined information for displaying the estimation result of the failure mode probability estimation section 24 on the terminal 10 in association with the failure causal model 25 (third processing). ), and transmits this information to the terminal 10 via the communication unit 21. As a result, the probability of occurrence of each failure mode is displayed on the terminal 10 in association with the failure causal model. After the probability of occurrence of each failure mode is displayed on the terminal 10 in this way, the server 20 sends information on a repaired failure mode, which is a predetermined failure mode that has been repaired by the maintenance person R, via the communication unit 21. Receive from terminal 10. Note that when a predetermined failure mode is repaired by the maintenance person R, that failure mode is referred to as a "repaired failure mode."

FIG. 14 is an explanatory diagram showing how the state of a predetermined inspection item is selected in the causal relationships such as failure modes displayed on the terminal.
In addition, in the example of FIG. 14, unlike in FIG. 13, in addition to inspection items such as "Is the power supply voltage normal?" and "The refrigerator (not shown) does not cool down," there are also inspection items such as "There is frost on the evaporator." Lower-level failure modes and higher-level failure modes such as "drain pan clogging" are displayed in a tree format. Further, FIG. 14 shows a case where "Yes" is selected on the terminal 10 side for the inspection item "Is the power supply voltage normal?" and the send button B2 is further pressed. As a result, information regarding the status of the inspection item is transmitted from the terminal 10 to the server 20.

Then, the server 20 calculates the probability of occurrence of the failure mode, thereby updating the value of the failure mode state probability display column Cc2. Note that the arrangement of the display of inspection items and failure modes on the terminal 10 may be changed as appropriate based on the magnitude of the probability of occurrence of the failure mode.

In step S105 in FIG. 11, the maintenance person R determines whether additional inspection is necessary. For example, if the failure mode can be sufficiently diagnosed from the results of the inspection items that have already been carried out, the maintenance person R determines in step S105 that no additional inspection is necessary. In this way, if there is no need for additional testing (S105: No), the process of the server 20 proceeds to step S106. On the other hand, regarding diagnosis of the failure mode, if the inspection items already performed are insufficient, the maintenance person R determines that additional inspection is required. If an additional inspection is required in this way, the process of the server 20 returns to step S102 based on the operation of the terminal 10 by the maintenance person R.

In step S106, the server 20 uses the input reading unit 22 to read the lower failure mode. Note that the state of the lower failure mode is determined by the maintenance person R by visually observing a predetermined location of the device M, and is inputted through an operation via the input unit 11 of the terminal 10. Note that when the maintenance person R decides which failure mode to check for occurrence or non-occurrence, the high probability of occurrence of each failure mode displayed on the output unit 13 is taken into consideration.

Then, if a failure mode has occurred, the maintenance person R performs a predetermined repair, and inputs into the terminal 10 information that the failure mode has occurred and information regarding the repair. Furthermore, even when a failure mode has not occurred, the maintenance person R inputs that fact into the terminal 10 and presses the send button B2. As a result, information regarding the state of the lower failure mode is transmitted from the terminal 10 to the server 20.

In step S107 of FIG. 11, the server 20 calculates the probability of occurrence of the failure mode and extracts the causal relationship of the failure modes. That is, when the information on the repaired failure mode is received from the terminal 10 via the communication unit 21, the failure mode probability estimation unit 24 of the server 20 calculates a plurality of repaired failure modes based on the failure causal model 25 and the information on the repaired failure mode. The probability of occurrence (index value) of each failure mode is calculated again (fourth process). Further, the server 20 uses the causal relationship extraction unit 27 to extract the causal relationship of the failure mode based on information stored in the temporary storage unit 23, the failure causal model 25, and the failure history database 28.

In step S108, the server 20 presents the probability of failure mode occurrence and causal relationship to the user. That is, the server 20 causes the terminal display information generation processing unit 29 to again generate information for displaying the probability of occurrence (index value) of each of the plurality of failure modes on the terminal 10 in association with the failure causal model 25. (Fifth process). The information generated by the terminal display information generation processing unit 29 in this way is transmitted to the terminal 10 via the communication unit of the server 20 (sixth process). The information generated by the terminal display information generation processing section 29 is then displayed on the terminal 10.

FIG. 15 is an explanatory diagram showing a display example of the terminal screen when the lower failure mode is determined and the repair result is transmitted.
As shown in FIG. 15, the status of the inspection item confirmed by the maintenance person R is input into the inspection item status candidate column Cb1, and the status of the failure mode confirmed by the maintenance person R is input into the failure mode status candidate column Cb2. is input. Then, when the send button B2 is pressed, the probability of the failure mode is calculated again on the server 20 side, and the probability displayed in the failure mode state probability display column Cc2 of the terminal 10 is updated.

Note that the probabilities P (FM|I, MFM) calculated by the failure mode conditional probability calculation unit 24d (see FIG. 10) are the probabilities p5, p6, p7, and p8 shown in FIG. 15. That is, when the inspection items and lower-order failure modes reach the predetermined state stored in the temporary storage unit 23, the probability P (FM|I, MFM) that failure mode FM occurs is the probability p5, shown in FIG. They are p6, p7, and p8.

Further, for failure modes to which a warning flag is attached, a predetermined warning is displayed in the warning flag display column D2. Here, as the warning content of the warning flag, in addition to a predetermined highlighted display in the failure mode presentation field C2, information on the number of times the failure mode has occurred up to that point may be displayed.

In step S109 in FIG. 11, the maintenance person R determines whether additional repair is necessary. Additions are made based on criteria such as whether the probability of occurrence of a higher-order failure mode that causes lower-order failure modes that have been repaired is high, and whether a warning flag is attached to a higher-order failure mode. Maintenance person R determines whether or not repairs are necessary. If the maintenance person R determines that there is no need for additional repair (S109: No), the series of maintenance work ends (END).

Note that, at the end of the maintenance work, the final failure mode is input by the maintenance person R operating the terminal 10. This failure mode is stored in the temporary storage unit 23 of the server 20 and further saved in the failure history database 28. On the other hand, if the maintenance person R determines that additional repair is required in step S109 (S109: Yes), the process returns to step S105.

In this way, the processes of steps S102 to S109 are repeated as appropriate, and the maintenance personnel R inputs the results of inspection items and lower failure modes into the terminal 10. can be identified. Therefore, the maintenance person R can repair the device M quickly and appropriately.

<Effect>
According to the first embodiment, the causal relationships among the inspection items and failure modes of the device M are displayed in a tree shape on the terminal 10. Further, the probability P(FM|I,MFM) of failure mode FM occurring when the inspection items and lower-order failure modes of equipment M reach a predetermined state is displayed on the terminal 10 in a predetermined manner. As a result, when the maintenance person R appropriately inputs the results of inspection items and the results of lower-order failure modes into the terminal 10, it is possible to identify a higher-order failure mode that is close to the fundamental cause of the failure.

Furthermore, the information on the repaired failure mode input to the terminal 10 is transmitted to the server 20, and the probability of occurrence of the failure mode is recalculated based on the result. In this way, by reflecting the results of repaired failure modes, it is possible to improve the accuracy of estimating higher-order failure modes. Therefore, it is less likely that higher-level failure modes close to the root cause will be overlooked. As a result, the frequency with which failures occur again in the device M is reduced, so that the time and cost of repair operations during maintenance work can be reduced.

≪Second embodiment≫
In the second embodiment, the server 20A (see FIG. 16) extracts the causal relationship between the upper failure mode and the lower failure mode from the failure history database 28, and updates (changes) the causal relationship. This is different from one embodiment. Note that other points are the same as those in the first embodiment. Therefore, the parts that are different from the first embodiment will be explained, and the explanation of the overlapping parts will be omitted.

FIG. 16 is a functional block diagram showing a configuration for extracting the causal relationship between the upper failure mode and the lower failure mode in the server 20A of the failure diagnosis system according to the second embodiment.
Note that the server 20A shown in FIG. 16 has the configuration of the server 20 (see FIG. 1) described in the first embodiment, with the addition of a causal candidate extracting section 31 and a causal relationship changing section 32 (see FIG. 16). It is configured. Note that FIG. 16 illustrates the configuration related to changing the causal relationship of failure modes, and the remaining configurations are not illustrated.

As shown in FIG. 16, the server 20A of the failure diagnosis system includes a causal candidate extracting section 31, a causal relationship changing section 32, a failure history database 28, and a failure causal model 25.
The failure cause-and-effect model 25 has a configuration similar to that of the first embodiment (see FIG. 2), but a portion thereof is illustrated in FIG. 16. The failure history database 28 stores repair details for failure modes that have already been performed in association with predetermined data indicating the order in which they were performed.

The causal candidate extraction unit 31 has a function of extracting predetermined causal candidates (candidates for causal relationships between failure modes) from the failure history database 28 from among a plurality of failure modes. For example, regarding failure mode A and failure mode B, the number of times failure mode B has been repaired after failure mode A has been repaired is greater than or equal to a predetermined value, and the upper failure mode is failure mode B and the lower failure mode is Assume that the probability that the failure mode becomes failure mode A is not included in the failure causal model 25 (specifically, the lower failure mode-higher failure mode probability data 25f).

In such a case, even though failure mode B is actually the cause of failure mode A, there is a possibility that failure mode A is repaired before failure mode B is repaired. is high. Therefore, in order to change the causal relationship between failure modes A and B, the causal candidate extraction unit 31 extracts the combination of failure modes A and B described above as a failure candidate.

The causal relationship changing unit 32 has a function of changing the causal relationship between the plurality of failure modes in the failure causal model 25 based on the causal candidates extracted by the causal candidate extracting unit 31. For example, when the above-mentioned failure mode A is the highest failure mode data 25c, the causality changing unit 32 adds the data of the failure mode A to the lower failure mode data 25b, while adding the failure mode A from the higher failure mode data. Delete the contents of. Then, the causal relationship changing unit 32 sets failure mode B to be a higher level (cause) causal relationship with respect to failure mode A, and calculates the probability of occurrence of failure mode A when failure mode B occurs as lower failure mode - upper failure. It is stored in the inter-mode probability data 25f. Note that the operator of the failure diagnosis system may appropriately determine the value of the occurrence probability described above, or a predetermined value may be set in advance.

In this way, the causal candidate extracting unit 31 detects that the number of times the second failure mode is repaired after the first failure mode repair included in the failure causal model 25 is greater than or equal to a predetermined value, and , if the condition that a predetermined causal relationship in which the second failure mode is the cause and the first failure mode is the result does not exist in the failure causal model 25, this combination of predetermined causal relationships is designated as a causal candidate. , along with the order in which the repairs described above are performed. Then, the causal relationship changing unit 32 adds to the failure causal model 25 a predetermined causal relationship in which the second failure mode is the cause and the first failure mode is the result.

In addition, when the above-mentioned conditions are satisfied, the causal relationship changing unit 32 creates another causal relationship in the failure causal model in which the first failure mode is the cause and the second failure mode is the result. When it exists, the other causal relationship is deleted from the failure causal model 25. This makes it easier for the maintenance person R to repair the higher-order failure mode (second failure mode) that is closer to the root cause first, thereby reducing the time and cost required for repair.

Note that the predetermined value regarding the number of times repairs have been performed in the same order may be a fixed value, or may be changed as appropriate by the administrator of the failure diagnosis system. In addition, regarding the number of repairs for a combination of failure modes, the causal candidate extraction unit 31 may calculate the number of repairs from the entire past maintenance history, or may calculate the number of repairs within a predetermined period in the past. Good too.

<Effect>
According to the second embodiment, the causal candidate extracting unit 31 extracts the causal relationship between the upper failure mode and the lower failure mode from the failure history database 28, and the causal relationship changing unit 32 changes the failure causal model 25. . As a result, in the process of maintenance work for the device M, the maintenance person R can first repair a higher-order failure mode that is close to the root cause of the failure of the device M. Therefore, the time and cost required for maintenance work can be reduced compared to the case where repairs are performed on lower failure modes first.

≪Modification example≫
Although the failure diagnosis system 100 according to the present invention has been described above in each embodiment, the present invention is not limited to these descriptions, and various changes can be made.
For example, in each embodiment, a case has been described in which a Bayesian network is used in the failure diagnosis system 100 (see FIG. 1), but the present invention is not limited to this. In other words, it is only necessary to have a framework that can estimate the probability of a failure mode based on the information on the results of inspection items and the results of lower-order failure modes. techniques may be used. For example, when a neural network method is used, the failure causal model 25 is a well-known weight parameter in the neural network.

Further, in each embodiment, an example has been described in which the causal relationship between each inspection item and each failure mode is displayed in a tree shape on the output unit 13 of the terminal 10 (see FIG. 14), but the present invention is not limited to this. That is, if the format is such that the maintenance person R can grasp the causal relationships between inspection items and failure modes, and between lower-level failure modes and higher-level failure modes, for example, the output section 13 of the terminal 10 , each causal relationship may be displayed in a predetermined list format.

Furthermore, in each of the embodiments, the server 20 (see FIG. 1) has been described as having a failure mode probability estimating unit 24, a cause-and-effect relationship extracting unit 27, etc.; There may be.
Furthermore, in each of the embodiments, the failure diagnosis system 100 (see FIG. 1) has been described as having a configuration including the terminal 10 and the server 20. There may be. For example, as described above, the terminal display information generation processing section 29 may be included in the terminal 10 instead of the server 20.
Further, the processing executed by the server 20 may be executed as a predetermined program of a computer. The above program can be provided via a communication line, or can be written on a recording medium such as a CD-ROM and distributed.

Further, each embodiment is described in detail to explain the present invention in an easy-to-understand manner, and the present invention is not necessarily limited to having all the configurations described. Furthermore, it is possible to add, delete, or replace some of the configurations of the embodiments with other configurations. Further, the mechanisms and configurations described above are those considered necessary for explanation, and not all mechanisms and configurations are necessarily shown in the product.

100 Failure Diagnosis System 10 Terminal 20, 20A Server 11 Input Unit 12 Communication Unit 13 Output Unit 21 Communication Unit 22 Input Reading Unit 23 Temporary Storage Unit 24 Failure Mode Probability Estimation Unit (Failure Mode Occurrence Index Value Estimation Unit)
25 Failure causal model 26 Estimation result output unit 27 Causal relationship extraction unit 28 Failure history database 29 Terminal display information generation processing unit 31 Causal candidate extraction unit 32 Causal relationship changing unit S102 Step (first process)
S103 Step (second process)
S104 Step (third process)
S107 Step (4th process)
S108 Step (fifth process)
M Equipment R Maintenance person

Claims (8)

a communication unit that performs predetermined communication with a terminal operated by a maintenance person who performs maintenance work on the equipment;
A failure cause-and-effect model that includes causal relationships between a plurality of failure modes that are events that can cause failures of the device, and a causal relationship between a plurality of inspection items regarding the device and the plurality of failure modes, and a failure mode occurrence index value for calculating a predetermined index value indicating the likelihood of occurrence of each of the plurality of failure modes based on the results of the plurality of inspection items received from the terminal via the communication unit; Estimating section;
a terminal display information generation processing unit that generates information for displaying on the terminal by associating the estimation result of the failure mode occurrence index value estimation unit with the failure causal model;
The failure mode occurrence index value estimating unit is configured to display the index values of each of the plurality of failure modes based on the results of the inspection items on the terminal in association with the failure causal model, and then display the index values for each of the plurality of failure modes based on the results of the inspection items. When information on a repaired failure mode, which is a predetermined failure mode in which recalculating the index value for each of the failure modes;
The terminal display information generation processing unit associates the index values of each of the plurality of failure modes with the failure causal model and generates again information to be displayed on the terminal. . The terminal display information generation processing unit displays the causal relationships among the plurality of failure modes and the causal relationships between the plurality of failure modes and the results of the plurality of inspection items in a tree shape on the terminal. The fault diagnosis system according to claim 1, further comprising: generating information for causing the fault diagnosis to occur. The failure cause-and-effect model has the plurality of inspection items and the plurality of failure modes as nodes, and the causal relationships between the plurality of failure modes and the plurality of failure modes and the plurality of inspection items. The fault diagnosis system according to claim 1, wherein the fault diagnosis system is configured as a Bayesian network having the causal relationships between them as links. comprising a failure history database containing information on the history of failure repairs of the equipment;
In the failure history database, the number of repairs performed in the past is greater than or equal to a predetermined value, and a predetermined failure mode that has not yet been repaired for another higher-order failure mode linked in the failure causal model as its own cause. If a failure mode exists, the terminal display information generation processing unit generates information for causing the terminal to display a warning to the effect that repair of the other failure mode at a higher level should be performed. The fault diagnosis system according to claim 1. a failure history database containing information on the history of failure repairs of the equipment;
a causal candidate extraction unit that extracts a predetermined causal candidate from the failure history database from among the plurality of failure modes;
a causal relationship changing unit that changes the causal relationship between the plurality of failure modes in the failure causal model based on the causal candidate extracted by the causal candidate extracting unit,
The cause-and-effect candidate extraction unit is configured to determine whether the number of times a second failure mode repair is performed after a first failure mode repair included in the failure cause-effect model is equal to or greater than a predetermined value, and the second failure mode If a condition is established that a predetermined causal relationship in which the failure mode is the cause and the first failure mode is the result does not exist in the failure causal model, the combination of the predetermined causal relationships is used as the causal candidate. Extracted along with the order of carrying out the repair,
The causal relationship changing unit adds the predetermined causal relationship in which the second failure mode is a cause and the first failure mode is a result to the failure causal model. fault diagnosis system. The causal relationship changing unit is configured to determine, when the condition is satisfied, that another causal relationship exists in the failure causal model in which the first failure mode is the cause and the second failure mode is the result. 6. The fault diagnosis system according to claim 5, further comprising: deleting said other causal relationship from said fault causal model when said fault causality model exists. A first process of performing predetermined communication with a terminal operated by a maintenance person who performs maintenance work on the equipment and receiving information including the results of a plurality of inspection items regarding the equipment;
a failure cause-and-effect model including a causal relationship between a plurality of failure modes that are events that can cause a failure of the device, and a causal relationship between a plurality of the inspection items and a plurality of the failure modes regarding the device; , a second process of calculating a predetermined index value indicating the likelihood of occurrence of each of the plurality of failure modes based on the results of the plurality of inspection items received from the terminal in the first process;
a third process of associating the estimation result in the second process with the failure causal model and generating information for displaying on the terminal;
After the index values of each of the plurality of failure modes based on the results of the inspection items are displayed on the terminal in association with the failure causal model, the failure mode is a predetermined failure mode in which repair is performed by the maintenance person. a fourth process of recalculating the index value of each of the plurality of failure modes based on the failure causal model and the information of the repaired failure mode when information on the repaired failure mode is received from the terminal; ,
A fifth process of regenerating information for displaying on the terminal in association with the failure causal model, the index values of each of the plurality of failure modes being sequentially executed. Method. A program for causing a computer to execute the failure diagnosis method according to claim 7.

Images