Malfunction handling assistance system and malfunction handling assistance program

Abstract

To properly cope with a malfunction.SOLUTION: A malfunction handling assistance system includes an assistance computer 3 configured to: store, as past information, a past monitoring panel image obtained by photographing an alarm state of a monitoring panel 101 that displays various alarms when a malfunction occurred in a facility in the past, a past device image obtained by photographing a situation of a device related to the malfunction, and a handling method performed for the malfunction; determine a situation of which device is to be confirmed based on the image obtained by photographing the alarm state of the monitoring panel 101 when the malfunction in the assistance target occurred and based on the past information; and determine a handling method for the malfunction in the assistance target, based on the image obtained by photographing the situation of the determined device and the past information.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a failure response support system and a failure response support program that support responses when a failure occurs in a facility. [Background technology]

[0002] For example, if a malfunction occurs at a hydroelectric power plant, the person in charge of managing the site will look at the monitoring panel that displays various alarms and estimate which equipment or facility may be malfunctioning or malfunctioning, in other words, which equipment's status should be checked.The person in charge will then go to the site, inspect the suspected equipment, confirm the nature of the malfunction, consider and decide on the necessary measures and responses, and implement them with the approval of their superiors.

[0003] However, this required a great deal of time and effort, as the person in charge had to look at the monitoring panel, estimate which equipment's status should be checked, and then go to the site to determine the necessary response.In addition, there was a risk that the estimations and judgments of the person in charge would vary, preventing appropriate and prompt response.

[0004] Meanwhile, a plant monitoring support device is known that allows operators to understand the causes of accidents without having to memorize the patterns of alarms that occur when an accident occurs (see, for example, Patent Document 1). This device photographs a common monitoring panel that has alarm displays that show alarms and indicators that show process quantities, processes the photographed image, and displays the image on a small screen with information related to the alarm added. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Laid-Open No. 2000-270430 Summary of the Invention [Problem to be solved by the invention]

[0006] However, the device described in Patent Document 1 only displays information related to alarms based on the displays and instructions on a common monitoring panel installed in the plant control room on a small screen, and is unable to provide information based on the status of on-site equipment. Therefore, it is not possible to take appropriate action against a malfunction based solely on the information displayed on the small screen.

[0007] SUMMARY OF THE INVENTION It is therefore an object of the present invention to provide a failure response support system and a failure response support program that enable appropriate responses to failures. [Means for solving the problem]

[0008] In order to solve the above problem, the invention of claim 1 is a failure response support system characterized by comprising: a past information storage means for storing, as past information, past monitoring panel images of the alarm status of a monitoring panel that displays various alarms when a failure occurred in a facility in the past, past equipment images of the status of equipment installed in the facility that is related to the failure, and the response method taken in response to the failure; a first photographing means for photographing the alarm status of the monitoring panel; a second photographing means for photographing the status of the equipment installed in the facility; a first support means for determining which equipment's status should be checked based on the image photographed by the first photographing means when a failure of the support target occurs and the past information; and a second support means for determining the response method for the failure of the support target based on the image photographed by the second photographing means of the status of the equipment determined by the first support means and the past information.

[0009] The invention of claim 2 is characterized in that, in the failure response support system described in claim 1, the first support means uses a first support learning model that has been machine-learned based on past performance data so that when an image captured by the first photographing means is input, the device whose status should be checked is output.

[0010] The invention of claim 3 is characterized in that, in the failure response support system described in claim 1, the second support means uses a second support learning model that has been machine-learned based on past performance data so that when an image captured by the second photographing means is input, a response method for the failure of the support target is output.

[0011] The invention of claim 4 is characterized in that, in the failure response support system described in claim 1, the past information storage means stores a report on the failure as the past information, and is equipped with a report creation means that creates a report on the failure of the support target based on the date and time when the failure of the support target occurred, images taken by the first and second photographing means, the response method determined by the second support means, and the past information.

[0012] The invention of claim 5 is a failure response support program that causes a computer to function as a past information storage means that stores, as past information, past monitoring panel images of the alarm status of a monitoring panel that displays various alarms when a failure occurred in a facility in the past, past equipment images of the status of equipment installed in the facility and related to the failure, and the response method taken in response to the failure, a first support means that determines which equipment installed in the facility should have its status checked based on an image of the alarm status of the monitoring panel when a failure to be supported occurs and the past information, and a second support means that determines how to respond to the failure to be supported based on an image of the status of the equipment determined by the first support means and the past information.

[0013] The invention of claim 6 is characterized in that, in the failure response support program described in claim 5, the first support means uses a first support learning model that has been machine-learned based on past performance data so that when an image of an alarm state of the monitoring panel is input, the equipment whose status should be checked is output.

[0014] The invention of claim 7 is characterized in that, in the failure response support program described in claim 5, the second support means uses a second support learning model that has been machine-learned based on past performance data so that when an image of the status of the equipment is input, a response method for the failure of the support target is output.

[0015] The invention of claim 8 is characterized in that, in the failure response support program of claim 5, the past information storage means stores a report on the failure as the past information, and causes the computer to function as a report creation means that creates a report on the failure of the support target based on the date and time when the failure of the support target occurred, an image of the alarm status of the monitoring panel and an image of the status of the equipment, the response method determined by the second support means, and the past information. [Effects of the Invention]

[0016] According to the inventions of claims 1 and 5, when a failure occurs, taking a picture of the alarm status on the monitoring panel will determine which equipment's status should be checked, and by checking the status of that equipment, it becomes possible to take an appropriate response based on the status of the equipment at the site. Furthermore, taking a picture of the status of that equipment will determine how to respond to the failure, and by responding in accordance with this response method, it becomes possible to take an appropriate and prompt response based on the status of the equipment at the site.

[0017] According to the inventions of claims 2 and 6, the device whose status should be checked is output using the machine-learned first support learning model, making it possible to check the status of a more appropriate device. As a result, it becomes possible to respond to failures more appropriately and quickly.

[0018] According to the inventions of claims 3 and 7, a countermeasure against a fault is output using the second support learning model that has been machine-learned, making it possible to acquire a more appropriate countermeasure. As a result, it becomes possible to respond to the fault more appropriately and quickly.

[0019] According to the inventions described in claims 4 and 8, reports on failures are automatically created based on images of the alarm status on the monitoring panel and images of the equipment status, thereby reducing the time and effort required to create reports and making it possible to create uniform reports. [Brief explanation of the drawings]

[0020] [Figure 1] 1 is a schematic configuration diagram showing a failure response support system according to an embodiment of the present invention; [Figure 2] 2 is a schematic block diagram showing the configuration of a support computer of the failure response support system of FIG. 1. FIG. [Figure 3] 3A and 3B are diagrams showing examples of a past monitoring panel image and a past device image in the past information stored in the past information database of the support computer in FIG. 2. [Figure 4] 3 is a functional block diagram showing a schematic configuration of a first support learning model of the support computer of FIG. 2. FIG. [Figure 5] 3 is a functional block diagram showing a schematic configuration of a second support learning model of the support computer of FIG. 2. FIG. DETAILED DESCRIPTION OF THE INVENTION

[0021] The present invention will be described below based on the illustrated embodiments.

[0022] Fig. 1 is a schematic diagram showing a fault response support system 1 according to an embodiment of the present invention. This fault response support system 1 is a system that supports responses when a fault or abnormality occurs in a facility, and in this embodiment, a case will be described in which the facility is a hydroelectric power plant 100. This hydroelectric power plant 100 is equipped with a variety of devices and facilities, and the devices and facilities include not only machines but also instruments, piping, containers, and other items that can be used to grasp a wide range of on-site conditions.

[0023] A monitoring panel 101 that displays various alarms and abnormalities is installed in the monitoring room of the hydroelectric power plant 100. The configuration and display method of this monitoring panel 101 may be any, but for example, it may be provided with multiple alarm lamps that indicate the content of the alarm, or multiple meters that indicate whether the temperature of each piece of equipment has exceeded a predetermined temperature.

[0024] The failure response support system 1 mainly comprises a smartphone (first image capturing means, second image capturing means) 2 and a support computer 3, which are connected to each other so as to be able to communicate freely.

[0025] The smartphone 2 is a multi-function mobile phone (with a camera) owned and used by the manager M in charge of managing the site, and mainly photographs the alarm status of the monitoring panel 101 and the status of the equipment installed in the hydroelectric power plant 100, and transmits the images to the support computer 3. In other words, when a failure occurs in the hydroelectric power plant 100 and an alarm is displayed on the monitoring panel 101, the smartphone 2 photographs the alarm status (display status) of the monitoring panel 101 and transmits a first support request to the support computer 3, including the image of the target monitoring panel and the identification information of the smartphone 2.

[0026] Furthermore, as will be described later, when receiving information about a device whose status should be checked from the support computer 3, the smartphone 2 takes a picture of the status of the device and transmits a second support request including the image of the target device and identification information of the smartphone 2 to the support computer 3. Furthermore, the smartphone 2 is capable of transmitting a third support request for assistance in creating a report to the support computer 3. On the other hand, when receiving first support information, second support information, and a report as will be described later from the support computer 3, the smartphone 2 displays this information on the display.

[0027] Here, such a series of photographing and image transmission may be performed according to an application installed on the smartphone 2. In this manner, in this embodiment, the smartphone 2 serves as both the first photographing means and the second photographing means, but these may be configured as separate cameras. Also, the first photographing means that photographs the alarm state of the monitoring panel 101 may be configured as a stationary camera installed in a monitoring room.

[0028] The support computer 3 is a computer server that provides information to support troubleshooting, and in this embodiment is installed in the monitoring room of the hydroelectric power plant 100. As shown in Fig. 2, the support computer 3 mainly comprises an input unit 31, a display unit 32, a communication unit 33, a memory unit 34, a support task (first support means, second support means) 35, a report creation task (report creation means) 36, a learning task 37, and a central processing unit 38 that controls these. Note that the support computer 3 may be configured from multiple computer servers.

[0029] The input unit 41 is an interface for inputting various types of information and commands, and specifically, inputs commands to start the support task 35 and the report creation task 36, and inputs past information, which will be described later. The display unit 42 is a display that displays various types of data and information, and specifically, displays the output results of the support task 35 and the report creation task 36, and displays past information. The communication unit 43 is an interface for communicating with the outside via the Internet network, a telephone communication network, etc., and specifically, receives images from the smartphone 2, and transmits to the smartphone 2 information about devices whose situations should be checked and how to respond.

[0030] The storage unit 34 mainly includes a past information database (past information storage means) 341, a first support learning model 342, and a second support learning model 343. Here, the past information database 341 will be explained, and the learning models 342 and 343 will be described later.

[0031] The past information database 341 is a database that stores, as past information, past monitoring panel images of the alarm state of the monitoring panel 101 when a failure occurred in the hydroelectric power plant 100 in the past, past equipment images of the status of the equipment related to the failure (related to the content and cause of the failure), the response method taken in response to the failure, reports on the failure, etc. That is, in this embodiment, for each pattern of past monitoring panel images (pattern of alarm state) in failures that occurred in the hydroelectric power plant 100 in the past, the past equipment images (which may include identification information of the equipment), response methods, reports, etc. are stored in association with each other.

[0032] Here, if there are multiple past device images for one past monitoring panel image pattern, that is, if there are multiple devices (or their situations) related to the failure, response methods, reports, etc. are stored for each past device image. By storing such past information, as will be described later, it becomes possible to identify past device images based on the target monitoring panel image and past monitoring panel images, and to identify and acquire response methods and reports based on the target device image and past device images.

[0033] Here, the past monitoring panel image and the target monitoring panel image are, for example, images that show the lighting status of the alarm lamps on the monitoring panel 101 and the indication status of the meters, as shown in Fig. 3(a). The past equipment image and the target equipment image are, for example, images that show the measurement values ​​and alarm contents (messages) of the meters 102, the presence or absence of leakage from the pipes, deformation or abnormalities of the equipment, as shown in Fig. 3(b). The response method is the measures taken in response to the failure and the procedures for doing so, and the report contains the date and time when the failure occurred, the contents of the failure, the past monitoring panel image and past equipment image, the response method, etc.

[0034] The support task 35 is a program task that, when a new failure occurs in the hydroelectric power plant 100 (when an alarm is displayed on the monitoring panel 101), identifies the equipment whose status should be checked and determines how to respond to the failure. In this embodiment, the support task 35 is activated when a first support request including the identification information of the smartphone 2 and an image of the target monitoring panel (an image of the monitoring panel 101 when the failure to be supported occurred) is received from the smartphone 2, or when a second support request including the identification information of the smartphone 2 and an image of the target equipment is received. Alternatively, the support task 35 may be activated by inputting a start command by specifying the image of the target monitoring panel or the image of the target equipment at the input unit 41.

[0035] First, as a first support, when a first support request is received, it determines which equipment's status should be checked based on the target monitoring panel image and the past information in the past information database 341. That is, it searches and acquires a past monitoring panel image in an alarm state that is the same as or similar to the alarm state of the target monitoring panel image from the past information database 341. It then acquires a past equipment image associated with the acquired past monitoring panel image, and determines the equipment in this past equipment image as the equipment whose status should be checked.

[0036] At this time, if there are multiple past device images associated with the past monitoring panel image, the device whose status should be checked is identified from each past device image. Also, if there are multiple past monitoring panel images with the same or similar alarm status, the device whose status should be checked is identified from the past device images associated with each past monitoring panel image. Then, first support information including the identified device and the past device image is sent to the smartphone 2.

[0037] Next, when a second support request is received as second support, a response method (current response method) for the failure of the support target is determined based on the target device image and past information in the past information database 341. That is, a past device image of the device status that is the same as or similar to the device status in the target device image is searched and acquired from the past information database 341. Then, the response method associated with the acquired past device image is acquired to determine the current response method.

[0038] In this case, first, the past device image (including single or multiple) acquired in the first support is compared with only the target device image, and if they are identical or similar, the response method associated with that past device image is determined as the current response method. Next, if they are not identical or similar, the past device image acquired in the first support and other past device images are also compared with the target device image, and the current response method is determined based on the response method associated with the identical or similar past device image. At this time, the target monitoring panel image and target device image are compared comprehensively with the past monitoring panel image and past device image to determine the current response method. Then, second support information including the current response method determined in this way is sent to the smartphone 2.

[0039] Here, in the first support, when a target monitoring panel image is input, a first support learning model 342 that has been machine-learned based on past performance data is used to output the device whose status should be checked. This first support learning model 342 is created by learning task 37.

[0040] That is, as shown in Fig. 4, the learning task 37 uses past performance data recorded and accumulated in a first support performance database 344 to create a first support learning model 342 using a known machine learning algorithm such as a neural network. This first support performance database 344 is a database in which performance data is recorded and accumulated, including equipment whose status should be checked, determined by experts and skilled personnel in response to failures at the hydroelectric power plant 100, based on target monitoring panel images as input information. The past performance data includes data created based on the actual target monitoring panel images and the equipment whose status should be checked, determined actually by experts and skilled personnel, as well as data created through prior training, etc.

[0041] This learning task 37 uses machine learning and deep learning that utilizes a neural network to create a neural network based on the performance data recorded in the first support performance database 344, with, for example, the target monitoring panel image as the input layer, the device whose status should be checked as the output layer, and the analysis process from the input layer to the output layer as the intermediate layer. Then, the learning task 37 uses the performance data of the first support learning model 342 as learning data to learn various parameters in the intermediate layer. In other words, the learning task 37 learns various parameters in the intermediate layer so that the device whose status should be checked is appropriately output based on the target monitoring panel image.

[0042] Similarly, in the second support, when an image of the target device is input, a second support learning model 343 that has been machine-learned based on past performance data is used to output a method of dealing with a malfunction of the support target. This second support learning model 343 is created by learning task 37.

[0043] That is, as shown in Fig. 5, the learning task 37 uses past performance data recorded and accumulated in a second support performance database 345 to create a second support learning model 343 using a known machine learning algorithm such as a neural network. This second support performance database 345 is a database in which performance data including response methods determined by experts and skilled workers in responding to failures at the hydroelectric power plant 100 based on target equipment images as input information is recorded and accumulated. Note that the past performance data includes data created based on actual target equipment images and response methods actually determined by experts and skilled workers, as well as data created through pre-training, etc.

[0044] This learning task 37 uses machine learning and deep learning using a neural network to create a neural network based on the performance data recorded in the second support performance database 345, with, for example, a target device image as an input layer, a response method as an output layer, and analysis processing from the input layer to the output layer as an intermediate layer. The learning task 37 then uses the performance data of the second support learning model 343 as learning data to learn various parameters in the intermediate layer. In other words, the learning task 37 learns various parameters in the intermediate layer so that an appropriate response method is output based on the target device image. Note that the learning task 37 may be divided into a learning task for the first support learning model 342 and a learning task for the second support learning model 343.

[0045] The report creation task 36 is a program task for creating a report for a failure of a support target. In this embodiment, the report creation task 36 is started when a third support request (including the identification information of the smartphone 2) for support in creating a report is received from the smartphone 2. In response to this, the report creation task 36 may be started by inputting a start command specifying an image of the target monitoring panel or an image of the target device.

[0046] When activated, a report (draft report) for the failure of the support target is created based on the date and time when the failure of the support target occurred (date and time when the first support request was received from smartphone 2), the target monitoring panel image and target device image received from smartphone 2 in support task 35, the response method determined in support task 35, and reports stored in past information database 341. Specifically, the report is created by filling in and inputting the date and time, image, response method, etc. into a pre-stored report form.

[0047] Furthermore, the user fills in and inputs the necessary information by referring to reports stored in the past information database 341 that relate to the equipment and the response method for which the situation identified in the support task 35 should be checked. Alternatively, the user may create a report by inputting or overwriting the date and time of the failure to be supported, images, the response method, etc. into the related report. The report created in this way is then sent to the smartphone 2.

[0048] Next, the operation and behavior of the failure response support system 1 configured as above, and the failure response support method using the failure response support system 1 will be described.

[0049] When a new fault occurs in the hydroelectric power plant 100 and an alarm (for example, a generator major fault trip) is displayed on the monitoring panel 101, the manager M first takes a photo of the alarm state on the monitoring panel 101 with the smartphone 2 and sends a first support request including the image of the target monitoring panel to the support computer 3. In response to this, the support task 35 is started in the support computer 3, which determines the equipment whose status should be checked as described above, and sends first support information including this equipment to the smartphone 2.

[0050] Next, the manager M checks the condition of the equipment (for example, a decrease in the flow rate of cooling water for the water turbine bearings), takes a photo with the smartphone 2, and sends a second support request including an image of the target equipment to the support computer 3. In response to this, the support task 35 is started in the support computer 3, a response method is determined as described above, and second support information including the response method is sent to the smartphone 2.

[0051] Thereafter, the manager M implements the response / measures according to this response method and sends a third support request to the support computer 3. In response to this, the report creation task 36 is started in the support computer 3, and a report is created as described above and sent to the smartphone 2. The manager M then adds to and corrects this report as necessary to create the final report.

[0052] In this way, according to this failure response support system 1, when a failure occurs, taking a picture of the alarm status on the monitoring panel 101 will determine which device's status should be checked, and by checking the status of that device, it becomes possible to take an appropriate response based on the status of the on-site device. Furthermore, taking a picture of the status of that device will determine how to respond to the failure, and by responding in accordance with this response method, it becomes possible to take an appropriate and prompt response based on the status of the on-site device.

[0053] In this case, the device whose status should be checked is output using the machine-learned first support learning model 342, making it possible to check the status of a more appropriate device. As a result, it becomes possible to respond to failures more appropriately and quickly.

[0054] Similarly, a more appropriate response method can be acquired by using the machine-learned second support learning model 343 to output a response method to a failure, which results in a more appropriate and prompt response to the failure.

[0055] Furthermore, reports on failures are automatically created based on images of the alarm status on the monitoring panel 101 and images of the equipment status, thereby reducing the time and effort required to create reports and enabling the creation of uniform quality reports.

[0056] Although the embodiments of the present invention have been described in detail above, the specific configuration is not limited to these embodiments, and the present invention also includes design changes within the scope of the present invention. For example, in the above embodiment, the facility is a hydroelectric power plant 100, but the present invention can also be applied to other facilities such as factories and public halls. Furthermore, although one support task 35 serves as both a first support means and a second support means, these may each be configured as separate task programs.

[0057] On the other hand, the above-described failure response support system 1 and support computer 3 may be configured by installing the following failure response support program on a general-purpose computer (including a smartphone).

[0058] That is, the computer includes a past information storage means (past information database 341) that stores, as past information, past monitoring panel images of the alarm status of the monitoring panel 101 that displays various alarms when a failure occurred in the facility in the past, past equipment images of the status of equipment installed in the facility and related to the failure, the response method taken in response to the failure, and a report on the failure, a first support means (support task 35) that determines which equipment installed in the facility should have its status checked based on the image of the alarm status of the monitoring panel 101 taken when the failure to be supported occurred and the past information, and a second support means (support task 35) that determines how to respond to the failure to be supported based on the image of the status of the equipment determined by the first support means and the past information. This is a failure response support program that functions as a report creation means (report creation task 36) that creates a report on the failure of the support target based on the date and time the failure of the support target occurred, images of the alarm state of the monitoring panel 101 and images of the status of the equipment, the response method determined by the second support means, and past information. The first support means uses a first support learning model 342 that has been machine-learned based on past performance data so that when an image of the alarm state of the monitoring panel 101 is input, the equipment whose status should be checked is output, and the second support means uses a second support learning model 343 that has been machine-learned based on past performance data so that when an image of the status of the equipment is input, the response method for the failure of the support target is output. [Explanation of symbols]

[0059] 1. Failure response support system 2. Smartphone (first and second imaging means) 3. Support Computer 341 Past information database (past information storage means) 342 First Assisted Learning Model 343 Second Supportive Learning Model 35 Support Task (First Support Means, Second Support Means) 36 Report writing task (report writing means) 100 Hydroelectric Power Plants (Facilities) 101 Monitoring board 102 Instruments (equipment) M Administrator

Claims

1. past information storage means for storing, as past information, past monitoring panel images in which alarm states of a monitoring panel that displays various alarms when a failure has occurred in the facility in the past, past equipment images in which the status of equipment disposed in the facility and related to the failure has been photographed, and the countermeasures taken in response to the failure; a first photographing means for photographing an alarm state of the monitoring panel; a second photographing means for photographing the status of the equipment installed in the facility; a first support means for determining which device's status should be checked based on an image taken by the first image taking means when a failure occurs in the support target and the past information; a second support means for determining a method of dealing with the failure of the support target based on an image of the state of the equipment determined by the first support means, taken by the second photographing means, and the past information; A failure response support system comprising:

2. the first support means uses a first support learning model that has been machine-learned based on past performance data so that, when an image captured by the first imaging means is input, a device whose status should be checked is output; 2. The failure response support system according to claim 1.

3. the second support means uses a second support learning model that has been machine-learned based on past performance data so that, when an image captured by the second imaging means is input, a response method for the failure of the support target is output.

2. The failure response support system according to claim 1.

4. the past information storage means stores a report on the failure as the past information, a report creation means for creating a report on the failure of the support target based on the date and time when the failure of the support target occurred, the images taken by the first and second photographing means, the response method determined by the second support means, and the past information; 2. The failure response support system according to claim 1.

5. Computer, past information storage means for storing, as past information, past monitoring panel images in which alarm states of a monitoring panel that displays various alarms when a failure has occurred in the facility in the past, past equipment images in which the status of equipment disposed in the facility and related to the failure has been photographed, and the countermeasures taken in response to the failure; a first support means for determining which equipment in the facility should be checked for its status based on an image of the alarm state of the monitoring panel taken when a failure of the support target occurs and the past information; a second support means for determining a method of dealing with the failure of the support target based on an image of the state of the device determined by the first support means and the past information; A failure response support program characterized by functioning as a

6. The first support means uses a first support learning model that has been machine-learned based on past performance data so that, when an image of an alarm state of the monitoring panel is input, a device whose status should be checked is output.

6. The failure response support program according to claim 5.

7. the second support means uses a second support learning model that has been machine-learned based on past performance data so that, when an image of the status of the device is input, a response method for the failure of the support target is output.

6. The failure response support program according to claim 5.

8. the past information storage means stores a report on the failure as the past information, Computer, a report creation means for creating a report on the failure of the support target based on the date and time when the failure of the support target occurred, an image of the alarm state of the monitoring panel and an image of the status of the equipment, the response method determined by the second support means, and the past information; 6. The failure support program according to claim 5, wherein the program functions as:

Images