Abnormality sign detection system

Abstract

The present invention detects an anomaly perceived by an equipment inspector, and detects signs of the perceived anomaly before an abnormality actually occurs. Provided is an abnormality sign detection system that detects signs of abnormalities in to-be-inspected equipment, and that includes a server and a terminal device which is capable of communicating with the server. The server includes: a storage unit that records an action table including a standard inspection action table; a terminal operation detection unit that detects an operation of the terminal device and converts the operation into an inspection action performed on the terminal device; a degree-of-deviation detection unit that, on the basis of the action table, calculates the degree of deviation between the inspection action performed on the terminal device and a standard inspection action; and a degree-of-deviation output unit that, if the total sum of the degrees of deviation is equal to or greater than a threshold, outputs a signal corresponding to "a present sign of abnormality". Said system detects a perceived anomaly even if an inspector does not report the perceived anomaly.

Description

Abnormal Sign Detection System

[0001] The present invention relates to a system for detecting signs of abnormalities before they occur when performing maintenance inspections of equipment.

[0002] In large-scale factories and the like, those related to manufacturing such as electricity, water, and raw materials are prepared as in-factory infrastructure such as power generation facilities, water supply facilities, and material storage facilities. These facilities are complexly related and exist as a group of facilities. Therefore, they cannot always be operated in a constant state, and the state of each facility changes due to various causes. However, there is a certain range (so-called "green zone") that can be judged as safe and stable in the parameters related to the operation of each individual facility. Therefore, it is necessary to operate while constantly checking that the state of each facility is within the green zone in order to maintain the group of facilities.

[0003] For example, if even one parameter of one facility in the group of facilities deviates from the green zone, that facility is in an abnormal state, and the influence spreads to the entire group of facilities. Therefore, in order to operate the group of facilities normally, an immediate response is required. However, in such a group of facilities, individual facilities are often related to each other, and when one facility becomes abnormal, a large-scale restoration such as stopping the entire group of facilities is required for restoration.

[0004] Therefore, if it is possible to know the signs leading to an abnormal situation before the abnormal situation occurs, it is also possible to adjust while operating the equipment, which can be useful information. However, in reality, a large part of the detection of such signs relies on the five senses of the inspector.

[0005] Regarding such a demand, in Patent Document 1, a method of creating a database by summarizing five-sense information such as the sound generated by equipment is disclosed.

[0006] Japanese Patent Application Laid-Open No. 2004-173,108

[0007] Patent Document 1 focuses on the five senses of humans and attempts to build a database for a system that can detect equipment malfunctions at an early stage. Certainly, information based on the five senses of inspectors is valuable, and utilizing it is considered a reasonable approach.

[0008] However, as is human nature, when it comes to recording data, we tend to seek some kind of justification or past performance. In other words, we tend to think that an unfounded premonition of an anomaly will only cause confusion. Therefore, a feeling of unease without a clear reason, such as "something's strange" or "it doesn't quite make sense," is information based on the five senses that is less likely to surface.

[0009] In view of the above-mentioned problems, the inventors of this invention conceived this idea in an attempt to identify any discomfort experienced by inspectors through their inspection procedures.

[0010] More specifically, the abnormality sign detection system according to the present invention is an abnormality sign detection system for detecting signs of abnormality in equipment under inspection, comprising a server and a terminal device capable of communicating with the server, wherein the server comprises: a storage unit that records an action table including a standard inspection action table; a terminal operation detection unit that detects operations of the terminal device and converts them into inspection actions performed on the terminal device; a deviation degree detection unit that determines the degree of deviation between the inspection actions performed on the terminal device and standard inspection actions based on the action table; and a deviation degree output unit that outputs a signal corresponding to "abnormality sign present" when the sum of the deviation degrees exceeds a threshold.

[0011] This invention extracts the "unease" that lurks within the inspector from the inspector's actions, eliminating the need for the inspector to rationalize their own unease. Therefore, even without verbally communicating it to other inspectors, other inspectors and equipment operators can share that unease, which has the effect of increasing everyone's attention to "signs of abnormality."

[0012] This allows equipment operators to be prepared for the possibility of malfunctions occurring, even in equipment that is currently functioning normally. Furthermore, it enables inspectors to perform more detailed inspections and maintenance than usual.

[0013] This diagram illustrates the operation of the abnormality detection system according to the present invention. This diagram shows the configuration of the server of the abnormality detection system. This is a flowchart showing the main processing flow of the abnormality detection system. This is a flowchart showing the processing flow of the equipment inspection process (step S110) within the main flow. This is a flowchart showing the processing flow of the non-standard inspection action processing process (step S206) within the equipment inspection process (step S110). This diagram illustrates the contents of the action table (standard inspection action table and non-standard inspection action table). This is a flowchart showing the processing flow of the data retrieval processing process (step S240) within the non-standard inspection action processing process (step S206). This is a flowchart showing the processing flow of the data creation processing process (step S242) within the non-standard inspection action processing process (step S206). This is a flowchart showing the processing flow of the chat retrieval processing process (step S244) within the non-standard inspection action processing process (step S206). This is a flowchart showing the processing flow of the prompt call processing step (step S246) within the non-standard inspection action processing step (step S206). This is a flowchart showing the processing flow of the specific condition processing step (step S248) within the non-standard inspection action processing step (step S206). This is a flowchart showing the processing flow of the time processing step (step S210) within the equipment inspection process (step S110). This is a flowchart showing the processing flow of the deviation degree processing step (step S212) within the equipment inspection process (step S110).

[0014] The abnormality detection system according to the present invention will be described below with reference to drawings. The following description illustrates one embodiment of the present invention, and the present invention is not limited to the following description. The following description may be modified without departing from the spirit of the present invention.

[0015] The abnormality indicator detection system 1 according to the present invention (hereinafter sometimes abbreviated as "this system 1") pre-defines standard inspection actions and other actions (non-standard inspection actions) for each piece of equipment being inspected, and assigns points to the degree of deviation between standard inspection actions and other actions. Whenever an inspector feels something is amiss during an inspection and performs an action other than a standard inspection action, the deviation score is accumulated, and if it exceeds a certain value, a signal corresponding to "sign of abnormality" is output. Examples of non-standard inspection actions include reviewing past data or reconfirming the operating status of surrounding equipment. Furthermore, "sign of abnormality" can be said to be when the inspector has a feeling that the operating parameters of the equipment will fall out of the green zone in the near future and that an operating abnormality is likely to occur.

[0016] In other words, this invention assumes that even when it is not possible to make a judgment based on clear evidence that an abnormality is about to occur in the equipment, actions taken by the inspector other than standard inspection procedures, such as reviewing past data or re-checking the operating status of surrounding equipment, reflect the inspector's sense of unease.

[0017] For equipment where the degree of deviation exceeds a certain level and is judged to have "signs of abnormality," even if there are no abnormalities at present, the equipment operator can take measures before an abnormality occurs, such as conducting more detailed inspections or replacing parts.

[0018] System 1's primary purpose is not to detect abnormalities, but rather to detect signs of abnormalities before an abnormality is detected (when an abnormality occurs). Therefore, System 1 may operate independently, or it may be integrated into existing inspection support systems that detect abnormalities.

[0019] Figure 1 shows the usage state of the abnormality detection system 1 according to the present invention. This system 1 consists of a server 20 and a terminal device 12 carried by the inspector CM. The equipment to be inspected An (n is a natural number) may be a single unit or multiple units. Figure 1 shows the equipment from equipment A1 to equipment An. The inspector CM inspects each equipment An while making rounds to the equipment An periodically (or irregularly).

[0020] Figure 2 shows the basic configuration of the server 20. The server 20 includes at least a terminal position detection unit 22, a terminal operation detection unit 24, a deviation degree detection unit 26, a deviation degree output unit 28, and a storage unit 30. It may also have an input / output device 32 for giving instructions to the server 20 or displaying output from the server 20.

[0021] The server 20 and the terminal device 12 are connected to each other via means such as wireless communication. By acquiring an external signal Sin and outputting data Dot, the server 20 can communicate with public networks such as the internet (indicated as "Net" in Figure 2) and the terminal device 12. The input / output device 32 can give instructions to the server 20 and display signals from the server 20.

[0022] Furthermore, the server 20 can acquire location information of the terminal device 12. Since the server 20 also has prior knowledge of the location of each piece of equipment An under inspection, it can also know the positional relationship between the inspector CM holding the terminal device 12 and each piece of equipment An under inspection. It can also know how the terminal device 12 was operated.

[0023] The server 20 can provide various information in response to the operations performed by the inspector CM on the terminal device 12. Furthermore, it can determine how the inspector CM compared that information on the terminal device 12.

[0024] The terminal position detection unit 22 acquires the position information Sp of the terminal device 12 and can determine its positional relationship with the equipment An under inspection. The method for acquiring the position information Sp of the terminal device 12 is not particularly limited. A GPS may be attached to the terminal device 12 to acquire its position information Sp. Alternatively, the position of the terminal device 12 may be determined by a camera installed in the facility or by a local wireless network. By knowing the position of the terminal device 12, it is possible to know whether the inspector CM has entered the inspectable area C_ari (see Figure 1) of the equipment An under inspection. Here, the inspectable area C_ari is the area around the equipment in which the inspector MC moves around when performing standard inspection actions. In other words, if the inspector MC is in the inspectable area C_ari during the inspection, it is one of the grounds for considering that standard inspection actions are being performed.

[0025] The terminal operation detection unit 24 can detect operations performed by the inspector CM on the terminal device 12. The terminal device 12 sends the operations performed by the inspector CM as an operation signal D_op to the server 20. The terminal operation detection unit 24 converts the operation signal D_op into an action signal Tm representing the inspector CM's actions and outputs it. The action signal Tm is sent to the deviation detection unit 26, which will be described later.

[0026] The deviation detection unit 26 detects how much the actions taken by the inspector CM while inspecting the equipment An deviate from standard inspection actions. This is determined based on the standard inspection action table HA and the non-standard inspection action table NA, which are stored in the memory unit 30, described later. The detected deviation value is expressed as the deviation degree PCX. The deviation degree PCX is accumulated and expressed as the total deviation degree PC.

[0027] The deviation degree output unit 28 outputs a signal Se corresponding to "signs of abnormality" when the total deviation degree PC exceeds a certain threshold (referred to as the "deviation degree threshold Th"). The output destination is preferably the input / output device 32 or the terminal device 12. However, signal output to other destinations is not excluded.

[0028] Furthermore, the deviation output unit 28 may output a signal Se corresponding to "signs of abnormality" based on various criteria other than when the total deviation PC exceeds the deviation threshold Th. For example, if multiple inspectors CM perform similar non-standard inspection actions on a specific piece of equipment An, or if the name of a specific piece of equipment An or a specific word frequently appears in a chat with other inspectors CM through the terminal device 12, the unit may output a signal Se indicating "signs of abnormality."

[0029] As described above, the memory unit 30 stores the standard inspection action table HA and the non-standard inspection action table NA. The standard inspection action table HA and the non-standard inspection action table NA are tables that define the standard and non-standard inspection actions for the equipment An under inspection, respectively. These are collectively called the action table CT. The action table CT may include not only the standard inspection action table HA and the non-standard inspection action table NA, but also tables that define matters related to these tables.

[0030] Furthermore, the memory unit 30 may also store data such as records of inspection actions for each inspector CM regarding each piece of equipment An under inspection, past inspection results for each piece of equipment An under inspection, and maintenance records. In addition, the action sheet CT may differ for each inspector CM. This is because even with the same actions, a novice and an experienced person are thought to have different sensibilities regarding "signs of abnormality."

[0031] <Main Routine> The following describes the processing flow of the abnormality detection system 1. Figure 3 shows the main routine of this system 1. The inspector CM carries a terminal device 12 and inspects the equipment An being inspected while patrolling. When this system 1 starts (step S100), a determination of the completion of the inspection (step S102) is made. Here, the main criterion for this determination is whether or not there is any equipment An that has not been inspected.

[0032] Furthermore, System 1 can be stopped if an actual malfunction occurs. This is because if a clear malfunction occurs, there is no longer a need to detect "signs of malfunction." However, if there are "signs of malfunction" in equipment other than the equipment that malfunctioned, these may lead to an actual malfunction, and there is value in continuing to operate System 1. Therefore, even if an actual malfunction occurs, System 1 may be kept running.

[0033] If it is determined that this system 1 should be stopped (Y branch in step S102), this system 1 is stopped (step S104). On the other hand, if it is determined that this system 1 should be continued (N branch in step S102), the process moves to the terminal device 12 position confirmation process (step S106). In this process, the server 20 confirms the current position of the terminal device 12 using the terminal position detection unit 22 and determines whether the inspector CM has entered the inspection of the equipment An. As described above, this determination may be made based on whether the terminal device 12 has entered the inspectionable area C_ari of the equipment An.

[0034] If the inspector CM is not in the inspectable area C_ari of the equipment An to be inspected (N branch in step S106), the process is repeated and the system waits. If the inspector CM is in the inspectable area C_ari of the equipment An to be inspected and it is determined that the system has entered the inspection state (Y branch in step S106), the process moves to the next equipment inspection update process (step S108).

[0035] In the equipment inspection update process (step S108), the server 20 identifies the equipment An to be inspected by the inspector CM, and prepares an action sheet CT for that equipment, which includes a standard inspection action sheet HA and a non-standard inspection action sheet NA. Then the process moves on to the equipment inspection process (step S110).

[0036] The equipment inspection process (step S110) involves performing an inspection on the equipment An under inspection based on the standard inspection action table HA. If a non-standard inspection action deviates from the standard inspection action, the deviation degree PCX is calculated and accumulated each time to calculate the total deviation degree PC. If the total deviation degree PC exceeds the deviation threshold Th, a signal Se corresponding to "sign of abnormality" is output. Details of the equipment inspection process (step S110) are explained in Figure 4.

[0037] When server 20 completes the equipment inspection process (step S110), it returns to the inspection completion determination process (step S102), and either stops or continues system 1, completing one cycle of system 1's processing. This completes the inspection process for one piece of equipment An under inspection.

[0038] <Processing in the Equipment Inspection Process> The equipment inspection process (step S110) shown in Figure 3 will be explained in detail below. Figure 4 shows the flow of the equipment inspection process (step S110). When the equipment inspection process (step S110) starts, initial setup (step S200) is performed. As part of the initial setup, a menu necessary for inspecting the equipment An to be inspected is prepared and sent from the server 20 to the terminal device 12. Here, the menu includes at least a list of inspection items. It may be part of the standard inspection action sheet HA. Next, the inspection start time T0 is recorded. At least these processes are included in the initial setup process.

[0039] Once the initial setup (step S200) is complete, the server 20 waits for communication from the terminal device 12 (step S202). The inspector CM performs inspection actions for the inspection items of the equipment An under inspection according to the menu displayed on the terminal device 12. Here, inspection actions include making inputs according to the inspection items of the equipment An under inspection. Specifically, this includes actions such as the inspector CM inputting the values ​​shown on each instrument of the equipment An under inspection into the menu of the terminal device 12, or specifying an item (an input other than a numerical value, such as Yes or No) from the pull-down menu displayed on the menu of the terminal device 12.

[0040] When an inspection action is performed on terminal device 12, terminal device 12 converts the inspection action into data and sends it to the server 20 as an operation signal D_op to terminal operation detection unit 24 in server 20. Server 20 knows that communication has occurred when it receives the operation signal D_op (step S202).

[0041] The operation signal D_op is converted by the terminal operation detection unit 24 into an action signal Tm that indicates what action was taken. Therefore, the server 20 can know all the actions that the inspector CM took with respect to the terminal device 12.

[0042] When the server 20 detects that there is input (Y branch in step S202), it determines, based on the action signal Tm, whether the inspection action was a standard inspection action (labeled "Standard?" in Figure 4) or not (step S204). A standard inspection action is an inspection action (input action) listed in the standard inspection action table HA. For example, if the inspection requires inputting a value shown on a specific instrument of the equipment An being inspected, the standard inspection action would be "input a numerical value into that item."

[0043] On the other hand, if, when entering a numerical value for that item, an inspection action other than the standard inspection action is taken, such as retrieving past data or communicating via chat with the previous inspector (CM), that inspection action will be judged as a non-standard inspection action.

[0044] When it is determined that the inspection action of the inspector CM is a non-standard inspection action (N-branch in step S204), the process proceeds to the non-standard inspection action processing step (step S206). The details of the non-standard inspection action processing step (step S206) will be described in detail later. The non-standard inspection action processing step is a step for evaluating the degree to which the inspection action of the inspector CM deviates from the standard inspection action. In this step, the degree of deviation is evaluated as the deviation degree PCX, and the cumulative process is performed.

[0045] When the non-standard inspection action processing step (step S206) ends, the process returns to the determination of the presence or absence of input again (step S202). In addition, if a specific non-standard inspection action occurs, it is also possible to output "sign of abnormality present". In that sense, it can be said that the non-standard inspection action processing step (step S206) is a part of the deviation degree output unit 28 (see FIG. 2).

[0046] When the inspection action by the inspector CM is a standard inspection action (Y-branch in step S204), it is confirmed whether all the inspection items to be input have been completed (step S208). If not all the inspection items have been input (N-branch in step S208), the process proceeds to the determination of the presence or absence of input again (step S202). On the other hand, when the input of all the inspection items has been completed (Y-branch in step S208), the process proceeds to the time processing step (step S210).

[0047] The time processing step (step S210) examines how much time the inspector CM spent on inputting the inspection items for the inspected equipment An, and evaluates it as the deviation degree PCX. It is defined that the inspection of the inspected equipment An takes approximately this much time when a standard inspection action is performed. Therefore, if it takes more time than that, there is a high possibility that the inspector CM felt some discomfort. Therefore, the amount of time exceeding the assumed time required for inspection is evaluated as the degree of deviation from the standard inspection action. The details of this process will also be described in detail later.

[0048] When the time processing step (step S210) ends, the process moves to the deviation degree processing step (step S212). The deviation degree processing step (step S212) is a step of calculating the total deviation degree PC due to the inspection actions performed by the inspector CM on the inspected equipment An. This is implemented by adding the deviation degree PCX from the standard inspection actions due to non-standard inspection actions and time overrun.

[0049] Also, in the deviation degree processing step (step S212), when the deviation degree PCX exceeds a certain value, a signal Se corresponding to "there is a sign of abnormality" is output. That is, this step can also be said to be a part of the deviation degree output unit 28, similar to the non-standard inspection action processing step (step S206). When the deviation degree processing step (step S212) ends, the process returns to the main routine (step S214).

[0050] Next, referring to FIG. 5, the non-standard inspection action processing step (step S206) will be described. The non-standard inspection action processing step (step S206) is a step of evaluating the degree to which the inspection actions performed by the inspector CM on the terminal device 12 deviate from the standard inspection actions. Since the server 20 has a non-standard inspection action table NA determined in advance for each inspected equipment An, the inspection actions of the inspector CM obtained through the terminal operation detection unit 24 are compared with the non-standard inspection action table NA to evaluate the deviation degree PCX.

[0051] FIG. 6 shows an example of the standard inspection action table HA and the non-standard inspection action table NA. This is the case of the inspected equipment Ak. The inspected equipment Ak is scheduled to input three parameters of pressure value, flow rate, and temperature as inspection input items. The input time considered necessary for each input is 20 seconds. That is, for this inspected equipment Ak, it takes a standard 1 minute to complete the input of all items.

[0052] On the other hand, in the non-standard inspection action table NA, inspection actions that are not standard inspection actions such as past data call, data call of other equipment, past inspector name call, chat call with other inspectors, manual call of this inspected equipment (included in prompt call), weather information call, creation of correlation diagram between items, creation of relationship between items and arbitrary parameters, etc. are assumed.

[0053] Furthermore, in the case of a chat call, the chat recipient is a specific person (here, a specific speaker: inspector CM4), and in prompt call processing, which is a chat or system inquiry, there are specific keywords whose occurrence frequency requires attention (here, a specific word: if the word "vortex" appears five or more times, a deviation degree PCX is assigned).

[0054] Furthermore, it is possible to include assigning a deviation score (PCX) when an inspector performs the same inspection action as a specific inspector CM (in this case, inspector CM6). For example, if both the current inspector CM and inspector CM6 are highly skilled, then the fact that two highly skilled inspectors perform the same action would be perceived as a "sign of abnormality." This is represented as "specific action" in Figure 6.

[0055] The inspection process is divided into two parts: requesting data from the server 20, or creating correlation diagrams such as graphs using the terminal device 12 itself. These are called "request" and "creation," respectively. For example, requesting past data means requesting the server 20 for the historical trends (chronological data) of each item of the equipment Ak being inspected. Creating inter-item correlation diagrams, on the other hand, is the act of creating various correlation diagrams between the various input items of the equipment Ak being inspected.

[0056] Furthermore, there are inspection actions such as counting specific words in chat calls and prompt calls, as well as specific conditions under which a deviation score (PCX) is assigned. In Figure 6, these specific conditions are categorized as "Other."

[0057] To briefly explain each inspection action, retrieving past data involves requesting past data for the input items of the equipment Ak being inspected from the server 20. Retrieving data for other equipment involves requesting data for equipment other than the equipment Ak being inspected from the server 20.

[0058] A chat call with another inspector means requesting server 20 to exchange information via chat with another inspector. Server 20 will respond by establishing a chat connection with the other inspector CM.

[0059] A prompt call is a request to the server 20 itself for information regarding the equipment Ak under inspection. For example, this could be a request for an explanation of the manual for equipment Ak. It could also be called a call for an explanation regarding equipment Ak. A weather information call is a request to the server 20 to provide weather information for a required period. This could be past weather records or future weather forecasts. It may also be possible to request the server 20 to retrieve other information from the network.

[0060] Creating an inter-item correlation diagram involves creating a correlation diagram on the terminal device 12 between data retrieved from the server 20. Creating a relationship between an item and an arbitrary parameter means creating a relationship between each item of the equipment Ak under inspection and information that can be obtained from the internet, etc.

[0061] Each of these inspection behaviors (non-standard inspection behaviors) is assigned a predetermined deviation score (PCX). The basis for this scoring is that behaviors that seem less related to the input items receive higher scores. In other words, behaviors that seem less related to the standard inspection behaviors receive higher scores. This is because such behaviors are considered to represent "discrepancies."

[0062] For example, when inputting the pressure value or flow rate of equipment Ak under inspection, weather information usually does not have an influence. However, it is conceivable that the inspector CM might feel something is amiss based on an intuition or hunch regarding the weather information and the equipment Ak under inspection, leading them to investigate the weather information. If the inspector CM is highly skilled as an inspector, they may also judge that this feeling of unease is likely to lead to an actual abnormality. This system 1 evaluates such feelings of unease using a score called the deviation degree PCX and extracts "signs of an abnormality" by accumulating these scores.

[0063] However, the non-standard inspection action sheet NA can be updated at any time, and the deviation degree PCX may be changed as appropriate during operation. Furthermore, the non-standard inspection action sheet NA may be changed for each inspector CM. For example, it is likely that the ability to detect "signs of abnormality" will differ depending on whether a highly experienced inspector CM calls up weather information or a novice inspector CM does so, even if the inspection action is the same.

[0064] As a way to configure the non-standard inspection action table NA, the server 20 may output a signal Se corresponding to "signs of abnormality" for that inspection action only when certain conditions are met. For example, if a deviation degree PCX greater than the deviation degree threshold Th is set for an inspection action in which two specific inspectors CM1 and CM2 retrieve the same data for the same inspected equipment Ak, then a signal Se corresponding to "signs of abnormality" can be output for that inspection action alone.

[0065] Refer to Figure 5 again. Once the process moves to the non-standard inspection action processing step (step S206), it is determined whether the inspection action performed by the inspector CM on the terminal device 12 is a data retrieval (step S230), data creation (step S232), chat call (step S234), prompt call (step S236), or specific condition (step S238).

[0066] If the inspection action performed by the inspector CM was a data retrieval (Y branch in step S230), the data retrieval processing step (step S240) is performed. The data that the terminal device 12 can request in the data retrieval processing step includes data from the server 20 and the internet via the server 20, as exemplified above. Here, the type of data that the inspector CM viewed is evaluated.

[0067] Here, "evaluating" refers to determining the degree of deviation PCX from the inspection action. The same applies to the non-standard inspection action (step S206) described below, including subroutines. Once the data retrieval processing step (step S240) is completed, the process returns to the equipment inspection routine (see Figure 4).

[0068] If the inspection action performed by the inspector CM was data creation (Y branch in step S232), the data creation processing step (step S242) is performed. In the data creation processing step, as illustrated above, the operation on the terminal device 12 is acquired to evaluate what kind of data combination relationships the inspector CM created. Once the data creation processing step (step S242) is completed, the routine of the equipment inspection process (see Figure 4) is returned.

[0069] If the inspection action performed by the inspector CM is a chat call (Y branch in step S234), the chat call processing step (step S244) is executed. The chat call processing step (step S244) is a chat with another inspector CM. It may also be a chat with an expert who is not on-site. Server 20 secures a communication circuit, accesses the requested person, and starts the chat. In the chat call processing step (step S244), the frequency of specific characters that appear during the chat is mainly evaluated. Once the chat call processing step (step S244) is completed, the routine of the equipment inspection process (see Figure 4) is returned.

[0070] If the inspection action performed by the inspector CM is a prompt call (Y branch in step S236), the server 20 becomes the consultant for the inspector CM. In this process, the manual for the equipment Ak being inspected or the relevant page of the manual is presented or read aloud. In the prompt call processing step (step S246), the frequency of specific strings is also evaluated. Once the prompt call processing step (step S246) is completed, the process returns to the equipment inspection routine (see Figure 4).

[0071] If the inspection action performed by the inspector CM meets a specific condition (Y branch in step S238), the process moves to the specific condition processing step (step S248). In the specific condition processing step (step S248), inspection actions other than data retrieval processing (step S240), data creation processing (step S242), chat retrieval processing (step S244), and prompt retrieval processing (step S246) are evaluated. Once the specific condition processing step (step S248) is completed, the process returns to the equipment inspection routine (see Figure 4).

[0072] Figure 7 shows the flow of the data retrieval process (step S240 in Figure 5). When the process moves to the data retrieval process (step S240), the server 20 selects the data requested by the terminal device 12 and sends it to the terminal device 12 (step S260).

[0073] Next, the deviation degree PCX is determined for this inspection action (step S262). This can be said to be an evaluation of this inspection action. Specifically, this inspection action is found in the non-standard inspection action table NA for the equipment Ak currently being inspected, and the corresponding deviation degree PCX is set as the deviation degree PCX for this inspection action.

[0074] Then, the deviation PCX is accumulated in the total deviation PC (step S264). The total deviation PC may be reset and initialized (returned to zero) for each inspector CM, each inspection of the equipment Ak being inspected, or at regular intervals. However, the total deviation PC will not be reset until the inspection of the equipment Ak currently being inspected is completed. Once the process is finished, the process returns to the non-standard inspection action processing step (step S206 in Figure 5) (step S266).

[0075] Figure 8 shows the flow of the data creation process (step S242 in Figure 5). Once the process moves to the data creation process (step S242), the server 20 acquires the operation of the terminal device 12 via the terminal operation detection unit 24 (step S270), and learns what kind of data correlations the inspector CM has created on the terminal device 12.

[0076] Next, the deviation degree PCX is determined for this inspection action (step S272). This can be said to be an evaluation of this inspection action. Specifically, this inspection action is found in the non-standard inspection action table NA for the equipment Ak currently being inspected, and the corresponding deviation degree PCX is set as the deviation degree PCX for this inspection action.

[0077] Then, the deviation PCX is accumulated in the total deviation PC (step S274). The total deviation PC may be reset and initialized (returned to zero) for each inspector CM, each inspection of the equipment Ak being inspected, or at regular intervals. Once the processing is complete, the process returns to the non-standard inspection action processing step (step S206 in Figure 5) (step S276).

[0078] Figure 9 shows the flow of the chat call processing step (step S244 in Figure 5). When processing moves to the chat call processing step (step S244), the server 20 searches for the person specified by the inspector CM and opens a line (step S280). Next, the operation of the terminal device 12 is acquired via the terminal operation detection unit 24 and the number of occurrences of the specific keyword is counted (step S282). Alternatively, the number of occurrences of the specific keyword may be counted from the data flowing over the line opened by the server 20.

[0079] Next, the deviation degree PCX is determined based on the frequency of occurrence of specific keywords (step S284). This can also be considered an evaluation of this inspection action. If the chat is with a specific person listed in the non-standard inspection action table NA, the deviation degree PCX is also calculated for that. Specifically, the deviation degree PCX for the frequency of occurrence of specific keywords is calculated from the non-standard inspection action table NA of the equipment Ak currently being inspected, and the corresponding deviation degree PCX is used as the deviation degree PCX for this inspection action. If the chat is with a specific person listed in the non-standard inspection action table NA, the deviation degree PCX is also calculated for that.

[0080] Then, the deviation PCX is accumulated in the total deviation PC (step S286). The total deviation PC may be reset and initialized (returned to zero) for each inspector CM, each inspection of the equipment Ak being inspected, or at regular intervals. Once the processing is complete, the process returns to the non-standard inspection action processing step (step S206 in Figure 5) (step S288).

[0081] Figure 10 shows the flow of the prompt call processing step (step S246 in Figure 5). When processing moves to the prompt call processing step (step S246), the server 20 starts responding to the inspector CM (step S290). Next, the operation of the terminal device 12 is acquired via the terminal operation detection unit 24, and the number of occurrences of the specific keyword is counted (step S292).

[0082] Next, the deviation degree PCX is determined based on the number of occurrences of specific keywords listed in the non-standard inspection action table NA (step S294). This can also be considered an evaluation of this inspection action. Specifically, the deviation degree PCX is obtained from the non-standard inspection action table NA of the equipment Ak currently being inspected, based on the number of occurrences of specific keywords, and the corresponding deviation degree PCX is set as the deviation degree PCX for this inspection action.

[0083] Then, the deviation PCX is accumulated in the total deviation PC (step S296). Note that the total deviation PC may be reset and initialized (returned to zero) for each inspector CM, each inspection of the equipment Ak being inspected, or at regular intervals. Once the processing is complete, the process returns to the non-standard inspection action processing step (step S206 in Figure 5) (step S298).

[0084] Figure 11 shows the flow of the specific condition processing (step S248 in Figure 5). The specific condition processing assigns a deviation degree PCX to certain inspection actions other than the data retrieval process, data creation process, chat processing process, and prompt processing process, and is therefore not limited to Figure 11. Figure 11 is just one example.

[0085] When the process moves to the specific condition processing step (step S248), the server 20 acquires the operation of the terminal device 12 via the terminal operation detection unit 24 (step S300). Next, it determines whether or not the operation is identical to a specific inspection action performed by a specific inspector CM (step S302).

[0086] Then, if they are the same (Y branch in step S302), the deviation degree PCX is determined from the non-standard inspection action table NA (step S304) and added to the total deviation degree PC (step S306). Then, the process returns to the non-standard inspection action processing step (step S206 in Figure 5) (step S308). If they are not the same (N branch in step S302), nothing is done and the process returns to the non-standard inspection action processing step (step S206 in Figure 5) (step S308).

[0087] In this type of processing, if a particular inspector (CM) is considered highly skilled, the fact that another inspector (CM) noticed something unusual from the same perspective will be considered a "sign of abnormality." Furthermore, if the inspector (CM) currently inspecting the equipment is considered highly skilled, the degree of the "sign of abnormality" is considered to be even higher.

[0088] Refer again to the equipment inspection process (step S110) in Figure 4. As explained above, when the inspector CM performs an inspection action on the equipment Ak to be inspected, for inspection actions that are not standard inspection actions, a deviation degree PCX is determined by the non-standard inspection action processing step (step S206), and the total deviation degree PC is obtained by accumulating these deviations. In other words, if there was a non-standard inspection action when the item input was completed in step S208, the total deviation degree PC has been calculated. System 1 then executes the time processing step (step S210).

[0089] Even without performing non-standard inspection actions, the inspector (CM) may sense something is amiss. In such cases, they input information into each inspection item using standard inspection actions. However, they may also stop or walk around the equipment to consider the cause of the discomfort. Or, they may sense that the air is different than usual, even if it's not a foul odor. Therefore, in this system 1, if the time required to inspect the equipment An significantly exceeds the scheduled time, a time processing step (step S210) is provided to assign a deviation degree PCX.

[0090] Refer to Figure 12. When the process moves to the time processing step (step S210), the server 20 determines, via the terminal position detection unit 22, whether the terminal device 12 has moved away from the inspectable area C_ari of the equipment An under inspection (step S320). If it has not moved (N branch in step S320), it waits. If it has moved (Y branch in step S320), it calculates the elapsed time Ts from the inspection start time T0 (step S322).

[0091] Then, the deviation degree PCX is determined based on this elapsed time Ts (step S324). For example, if a longer time has elapsed than the time required for the standard inspection action, the deviation degree PCX is assigned at regular intervals. For example, in the case of the equipment Ak to be inspected shown in Figure 6, if the standard inspection action is followed, inputting into the input items will be completed in 1 minute. Therefore, 20 minutes is considered a sufficiently long time compared to 1 minute, and the non-standard inspection action table NA is defined so that the deviation degree PCX is assigned every 20 minutes.

[0092] Figure 6 defines that if the excess time exceeds 20 minutes, 5 points are awarded for the deviation PCX, and if it exceeds 40 minutes, 7 points are awarded.

[0093] The total deviation is then calculated and added to the PC (step S326), and the process returns to the equipment inspection process shown in Figure 4 (step S110) (step S328).

[0094] Refer again to the equipment inspection process (step S110) in Figure 4. The total deviation PC accumulated in the non-standard inspection action processing process (step S206) is further accumulated with the deviation PCX for excess time in the time processing process (step S210). Finally, the deviation processing (step S212) is executed.

[0095] Figure 13 shows the flow of the deviation degree processing process (step S212). Once the process moves to the deviation degree processing process (step S212), the total deviation degree PC and the deviation degree threshold Th are compared (step S340). If the total deviation degree PC is not greater than the deviation degree threshold Th (N branch in step S340), the process is returned to the equipment inspection process (step S110) (step S344).

[0096] If the total deviation PC is greater than the deviation threshold Th (Y branch in step S340), a signal Se corresponding to "signs of abnormality" is output (step S342). The output destination is not particularly limited, but the input / output device 32 or terminal device 12 of the server 20 is preferred.

[0097] As described above, the abnormality indicator detection system 1 according to the present invention detects signs of equipment malfunction before an actual malfunction occurs.

[0098] The abnormality warning system according to the present invention is suitably usable when there is equipment that requires periodic inspection. In particular, when multiple pieces of equipment are interconnected, it is useful as it can detect signs of an abnormality before an actual abnormality is detected.

[0099] 1 Anomaly Detection System 1 This System 12 Terminal Device 20 Server 22 Terminal Location Detection Unit 24 Terminal Operation Detection Unit 26 Deviation Degree Detection Unit 28 Deviation Degree Output Unit 30 Storage Unit 32 Input / Output Device NA Non-Standard Inspection Action Table HA Standard Inspection Action Table Tm Action Signal D_op Operation Signal C_ari Inspectable Area CM Inspector Sp Location Information An Equipment Under Inspection Ak Equipment Under Inspection Sin Signal Dot Data CT Action Table PCX Deviation Degree PC Total Deviation Degree Th Deviation Degree Threshold Se Signal T0 Corresponding to "Anomaly Presence" Inspection Start Time Ts Elapsed Time

Claims

1. An abnormality sign detection system for detecting signs of abnormality in equipment under inspection, comprising a server and a terminal device capable of communicating with the server, wherein the server includes a storage unit that records an action table including a standard inspection action table, a terminal operation detection unit that detects operations of the terminal device and converts them into inspection actions performed on the terminal device, a deviation degree detection unit that determines the degree of deviation between the inspection action performed on the terminal device and the standard inspection action based on the action table, and a deviation degree output unit that outputs a signal corresponding to "abnormality sign present" when the sum of the deviation degrees exceeds a threshold.

2. The abnormality sign detection system according to claim 1, wherein the server further has a terminal position detection unit for detecting the position of the terminal device, and the deviation degree detection unit also determines the deviation degree, which corresponds to the time during which the terminal device was in the inspectable area of ​​the equipment under inspection that exceeds the time specified in the standard inspection action schedule.

3. The abnormality sign detection system according to claim 1 or 2, wherein the action schedule includes a non-standard inspection action schedule, and the non-standard inspection action schedule includes an action to retrieve sequential data of inspection items of the equipment being inspected.

4. The abnormality detection system according to claim 1 or 2, wherein the non-standard inspection action sheet includes an action to call up an explanation regarding the equipment being inspected.

5. The abnormal sign detection system according to claim 1 or 2, wherein the degree of deviation detection unit varies depending on the skill level of the inspector.

Images