JP2017016382A - Information processing system, and failure prediction model adoption determination method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an information processing system capable of adopting a profitable failure prediction model by performing a trail calculation of cost during operation.SOLUTION: The information processing system having one or more information processing devices includes cost trial calculation means for applying failure history information of one or more electronic apparatuses in which a failure occurs to a failure prediction model obtained by associating a sign detection method of failures that occur in the electronic apparatuses with prevention measures against the failures that occur in the electronic apparatuses to perform a trail calculation of cost of the failure prediction model, and adoption determination means for determining the adoption of a profitable failure prediction model on the basis of a result of the trial calculation.SELECTED DRAWING: Figure 1

Description

  The present invention relates to an information processing system, a failure prediction model adoption determination method, and a program.

  In recent years, electronic devices such as copiers and printers have undergone measures necessary to detect failure signs and prevent failures using failure prediction models. The failure prediction model consists of a method for detecting an electronic device that is likely to fail (failure prediction logic) and a method for optimally preventing a failure of an electronic device that has detected a failure sign using the failure prediction logic (prevention treatment). ).

  For example, each image forming apparatus transmits state information representing its own state to the management apparatus, and when the management apparatus receives the state information, the contents of the state information are analyzed, and an image is selectively transmitted to each terminal apparatus. 2. Description of the Related Art An image forming apparatus management system that transmits information related to maintenance or repair of a forming apparatus is conventionally known (see, for example, Patent Document 1).

  For example, the development of failure prediction models that respond to various phenomena is ongoing. The developed failure prediction model is employed, for example, for detecting a failure sign based on information collected from an electronic device and for notifying a customer engineer (CE) of an optimal treatment method for preventing the failure.

  However, since the accuracy of the failure prediction logic is not necessarily 100%, there is a problem that if all of the developed failure prediction models are adopted, it may not be profitable due to excessive prevention measures.

  An object of one embodiment of the present invention is to provide an information processing system capable of adopting a failure prediction model that can be profitable by estimating costs during operation.

  In order to achieve the above object, claim 1 of the present application is an information processing system having one or more information processing apparatuses, a method for detecting a sign of failure occurring in an electronic device, and a failure occurring in the electronic device. A cost estimation means for estimating the cost of the failure prediction model by applying failure history information of one or more electronic devices in which the failure has occurred to a failure prediction model associated with On the basis of the result of the above, the adoption determining means for determining the adoption of the profitable failure prediction model.

  According to one embodiment of the present invention, it is possible to adopt and determine a failure prediction model that can be calculated by estimating the cost during operation.

It is a system configuration figure of an example of an information processing system concerning this embodiment. It is a hardware block diagram of an example of the computer which concerns on this embodiment. FIG. 2 is a hardware configuration diagram of an example of a multifunction peripheral according to the present embodiment. It is a processing block diagram of an example of an information processor concerning this embodiment. It is a processing block diagram of an example of a failure prediction model development part. It is a figure explaining an example of the process which searches the data pattern of the status information which has generate | occur | produced in common before failure occurrence. It is a figure explaining an example of the precision of failure prediction logic. It is a figure explaining an example of the parameter | index showing the precision of failure prediction logic. It is explanatory drawing of an example of the prevention process determined when the precision of failure prediction logic is high. It is explanatory drawing of an example of the prevention action determined when the precision of failure prediction logic is not high. It is a figure explaining an example of a failure prediction model. It is a formula for calculating the total service cost for a failure. It is a figure which shows four patterns divided by the result (predictive medium / missing) of the failure sign detection by failure prediction logic, and the possibility / impossibility of plus one correspondence. It is a figure explaining the trial calculation of the service cost effect at the time of selecting the prevention measure by plus one. It is a figure which shows an example of the trial calculation result of the service cost effect at the time of selecting the prevention measure by plus one. It is a figure which shows an example of the fluctuation | variation of the electric current output value of the sensor A. FIG. It is a figure which shows an example of the trial calculation result of the service cost effect in the case of emergency visit. 6 is a diagram illustrating an example of fluctuations in a current output value of a sensor B and development Y toner density. It is a response flow of an example by CE in the case of an emergency visit. It is a figure which shows an example of the cause of failure (error B). It is a classification diagram of an example of a failure (error B). It is a figure which shows an example of the fluctuation | variation of a sensor light quantity.

Next, embodiments of the present invention will be described in detail.
[First Embodiment]
<System configuration>
FIG. 1 is a system configuration diagram of an example of an information processing system according to the present embodiment. In the information processing system 1, an information processing apparatus 10, a plurality of multifunction machines 12, and a terminal apparatus 14 are connected via a network N1 such as the Internet.

  The multifunction machine 12 is an example of an electronic device. The multifunction machine 12 performs image forming processing such as scanning, printing (output), and FAX. The electronic apparatus according to this embodiment includes various electronic apparatuses that perform maintenance using a failure prediction model, in addition to image forming apparatuses such as multifunction peripherals, scanners, printers, facsimiles, projectors, and electronic blackboards.

  The information processing apparatus 10 acquires a plurality of variables related to the multifunction machine 12 (for example, status information such as sensor current output values) via the network N1. Further, the information processing apparatus 10 may read a plurality of variables related to the multifunction machine 12 recorded on a recording medium such as a USB memory by the customer engineer (hereinafter referred to as CE) or the like. The information processing apparatus 10 performs a cost simulation (trial calculation) when the failure prediction model is used, confirms whether it can be profitable as described later, and adopts a failure prediction model that can be profitable.

  The terminal device 14 is a device operated by a user who performs maintenance such as CE or a user who instructs the CE to perform maintenance. The terminal device 14 is, for example, a PC (Personal Computer), a tablet terminal, a smartphone, a mobile phone, a PDA (Personal Digital Assistance), or the like. The terminal device 14 performs a result of failure sign detection using the failure prediction model adopted by the information processing device 10 and notification of preventive measures for the CE.

  The network N1 of the information processing system 1 in FIG. 1 may be a wired communication network or a wireless communication network. The information processing system 1 in FIG. 1 is an example of a system configuration. For example, the information processing apparatus 10 may be configured to be distributed among a plurality of computers.

<Hardware configuration>
"Computer"
The information processing apparatus 10 and the terminal apparatus 14 are realized by, for example, a computer having a hardware configuration shown in FIG. FIG. 2 is a hardware configuration diagram of an example of a computer according to the present embodiment.

  A computer 500 in FIG. 2 includes an input device 501, a display device 502, an external I / F 503, a RAM 504, a ROM 505, a CPU 506, a communication I / F 507, an HDD 508, and the like. Note that the input device 501 and the display device 502 may be connected and used when necessary.

  The input device 501 includes a keyboard, a mouse, a touch panel, and the like, and is used by a user to input each operation signal. The display device 502 includes a display and the like, and displays a processing result by the computer 500.

  A communication I / F 507 is an interface for connecting the computer 500 to various networks. Thereby, the computer 500 can perform data communication via the communication I / F 507.

  The HDD 508 is an example of a nonvolatile storage device that stores programs and data. The stored programs and data include an OS, which is basic software for controlling the entire computer 500, and application software (hereinafter simply referred to as an application) that provides various functions on the OS. The computer 500 may use a drive device (for example, a solid state drive: SSD) that uses a flash memory as a storage medium instead of the HDD 508.

  The external I / F 503 is an interface with an external device. The external device includes a recording medium 503a. Accordingly, the computer 500 can read and / or write the recording medium 503a via the external I / F 503. Examples of the recording medium 503a include a flexible disk, a CD, a DVD, an SD memory card, and a USB memory.

  The ROM 505 is an example of a nonvolatile semiconductor memory (storage device) that can retain programs and data even when the power is turned off. The ROM 505 stores programs and data such as BIOS, OS settings, and network settings that are executed when the computer 500 is started up. The RAM 504 is an example of a volatile semiconductor memory (storage device) that temporarily stores programs and data.

  The CPU 506 is an arithmetic device that realizes control and functions of the entire computer 500 by reading a program and data from a storage device such as the ROM 505 and the HDD 508 onto the RAM 504 and executing processing.

  The information processing apparatus 10 and the terminal apparatus 14 can realize various processes as will be described later by using a computer 500 having a hardware configuration shown in FIG.

<Machine>
1 is realized by a computer having a hardware configuration as shown in FIG. 3, for example. FIG. 3 is a hardware configuration diagram of an example of a multifunction peripheral according to the present embodiment. 3 includes a controller 601, an operation panel 602, an external I / F 603, a communication I / F 604, a printer 605, a scanner 606, and the like.

  The controller 601 includes a CPU 611, a RAM 612, a ROM 613, an NVRAM 614, an HDD 615, and the like. The ROM 613 stores various programs and data. The RAM 612 temporarily stores programs and data. The NVRAM 614 stores setting information, for example. The HDD 615 stores various programs and data.

  The CPU 611 reads out programs, data, setting information, and the like from the ROM 613, the NVRAM 614, the HDD 615, and the like onto the RAM 612 and executes processing, thereby realizing the control and functions of the entire multifunction machine 12.

  The operation panel 602 includes an input unit that receives input from the user and a display unit that performs display. An external I / F 603 is an interface with an external device. The external device includes a recording medium 603a. Thereby, the multifunction machine 12 can read and / or write the recording medium 603a via the external I / F 603. Examples of the recording medium 603a include an IC card, a flexible disk, a CD, a DVD, an SD memory card, and a USB memory.

  The communication I / F 604 is an interface for connecting the multifunction machine 12 to the network N1. Thereby, the multifunction machine 12 can perform data communication via the communication I / F 604. A printer 605 is a printing device for printing print data on paper. A scanner 606 is a reading device for reading image data (electronic data) from a document.

<Software configuration>
《Information processing device》
The information processing apparatus 10 according to the present embodiment is realized by a processing block as illustrated in FIG. 4, for example. FIG. 4 is a processing block diagram of an example of the information processing apparatus according to the present embodiment.

  The information processing apparatus 10 executes the program to obtain a state information acquisition unit 21, a failure data acquisition unit 22, a failure prediction model development unit 23, a cost estimation unit 24, an adoption determination unit 25, a state information storage unit 31, and a failure data storage. The unit 32 and the failure prediction model storage unit 33 are realized.

  The state information acquisition unit 21 acquires state information that is a plurality of variables related to the multifunction machine 12 and stores the state information in the state information storage unit 31. The state information is a current output value of a sensor provided in the multifunction machine 12, information (for example, a counter value or the number of times of use) regarding the degree of wear of consumable parts.

  The failure data acquisition unit 22 acquires failure data of the multifunction machine 12 and stores the failure data in the failure data storage unit 32. The failure data includes state information at the time of failure (current output value of the sensor, information on the degree of wear of consumable parts, etc.).

  The failure prediction model development unit 23 uses the state information stored in the state information storage unit 31 and the failure data stored in the failure data storage unit 32 to develop a failure prediction logic based on statistical analysis and determine preventive measures. Develop. The failure prediction model development unit 23 may automatically develop a failure prediction model, or may support development of a failure prediction model by a data analyst. The failure prediction model development unit 23 stores the developed failure prediction model in the failure prediction model storage unit 33.

  The cost trial calculation unit 24 uses the status information in the status information storage unit 31 and the fault data in the fault data storage unit 32 to perform trial calculation of the service cost effect when the developed fault prediction model is used.

  For example, the cost estimation unit 24 calculates a service cost that increases when the developed failure prediction model is operated and a service cost that is reduced when the developed failure prediction model is operated.

  The adoption deciding unit 25 adopts a profitable failure prediction model based on the trial calculation result of the service cost effect when the developed failure prediction model is used. For example, if the service cost that is reduced when the developed failure prediction model is operated is larger than the service cost that is increased when the developed failure prediction model is operated, the adoption deciding unit 25 can make a profitable failure prediction model. Adopt as. Note that the adoption determination unit 25 may determine whether to adopt the failure prediction model for each site where the CE is located.

  It should be noted that various criteria for determining whether or not the failure prediction model can be profitable can be considered, such as those that can reduce service costs and those that can reduce service costs by a predetermined percentage.

  The failure prediction model development unit 23 in FIG. 4 is realized by, for example, processing blocks as shown in FIG. FIG. 5 is a processing block diagram of an example of the failure prediction model development unit. The failure prediction model development unit 23 includes a data pattern search unit 41, a failure prediction logic accuracy index calculation unit 42, and a preventive action determination unit 43.

  The data pattern search unit 41 uses the state information in the state information storage unit 31 and the failure data in the failure data storage unit 32 to search for a data pattern of state information that occurs in common before the failure occurs.

  The failure prediction logic accuracy index calculation unit 42 calculates the accuracy of failure prediction logic (the magnitude of failure occurrence probability after data pattern generation). For example, the failure prediction logic accuracy index calculation unit 42 according to the present embodiment calculates two indexes of hit rate and cover rate, which will be described later, as indexes indicating the accuracy of the failure prediction logic. The preventive action determining unit 43 determines the preventive action as described later based on the accuracy of the failure prediction logic.

<Details of processing>
Below, the detail of the process of the information processing system 1 which concerns on this embodiment is demonstrated.

<< Search for data pattern >>
The data pattern search unit 41 uses the state information in the state information storage unit 31 and the failure data in the failure data storage unit 32, as shown in FIG. Search for data patterns.

  FIG. 6 is a diagram for explaining an example of a process for searching for a data pattern of state information that occurs in common before the failure occurs. The data pattern search unit 41 is generated before the occurrence of the failure A from the failure history of the device (multifunction device 12) represented by the state information in the state information storage unit 31 and the failure data in the failure data storage unit 32. Search for the current data pattern. The relationship between the failure A and the data pattern that occurs in common before the occurrence of the failure A is the failure prediction logic. By detecting the occurrence of the searched data pattern in this way, the information processing apparatus 10 can detect the sign of the failure A.

<< Calculation of accuracy of failure prediction logic >>
FIG. 7 is a diagram for explaining an example of the accuracy of the failure prediction logic. Here, it is assumed that a sign of failure is detected by the occurrence of the data pattern searched by the data pattern search unit 41 and the preventive measure is performed.

  FIG. 7A shows a state in which a failure actually occurs after the data pattern is generated. As described above, if a failure actually occurs after the data pattern is generated, the previously performed preventive measure can realize the prevention of the failure.

  FIG. 7 (2) shows a state in which no failure actually occurred after the data pattern was generated. As described above, if a failure does not actually occur after the data pattern is generated, the preventive measures that are performed are excessive measures.

  Therefore, the data pattern search unit 41 searches for data patterns that are commonly generated before the occurrence of the failure in the case of FIG. 7 (1) and less in the case of FIG. 7 (2).

  FIG. 8 is a diagram for explaining an example of an index representing the accuracy of the failure prediction logic. In FIG. 8, the hit rate and the coverage rate are shown as indices indicating the accuracy of the failure prediction logic. The failure prediction logic is an expression for detecting a data pattern that occurs in common before the failure occurs.

  The result of the failure sign detection by the failure prediction logic is divided into the following three cases, for example, during the following predictive operation, the fact that it is missed and cannot be found. Predictive is a case where a failure sign is detected by the failure prediction logic and a failure actually occurs. The idling is a case where a failure sign has been detected by the failure prediction logic and no failure has actually occurred. “Not found” is a case where a failure has not actually been detected by the failure prediction logic and a failure has actually occurred.

  The accuracy of the failure prediction logic is represented by two indicators of the coverage rate and hit rate shown in FIG. Coverage ratio indicates the ratio of the number of cases in predictive to the number of cases (predictive medium + not found). That is, the coverage is an index indicating the rate of failure that can be detected by the data pattern. The hit rate indicates the ratio of the number of predictive cases to the number of (predictive medium + missing) cases. In other words, the hit rate is an index that indicates the ratio of a sign of a failure that actually occurs, and indicates the probability that a failure will occur after the data pattern is generated.

《Determination of prevention measures》
The preventive action determination unit 43 determines the preventive action as shown in FIG. 9 or 10 according to the accuracy of the failure prediction logic indicated by the hit rate. FIG. 9 is an explanatory diagram of an example of a preventive measure determined when the accuracy of the failure prediction logic is high. FIG. 10 is an explanatory diagram of an example of a preventive measure that is determined when the accuracy of the failure prediction logic is not high.

  As shown in FIG. 9, when the accuracy of the failure prediction logic is high, the preventive action determining unit 43 selects a preventive action by an emergency visit. As described above, since the failure occurs almost certainly after the data pattern is generated, the prevention action determining unit 43 selects the prevention action to be performed before the failure occurs due to the emergency arrangement.

  In addition, when the accuracy of the failure prediction logic is not high, the prevention process determining unit 43 then selects a prevention process by work (plus one). When an increase in the number of cases in which the preventive action ends up being missed is anticipated, the preventive action determining unit 43 reduces the cost increase due to the missed swing by performing the preventive action with a plus one during regular work. At the same time, measures to prevent failures are implemented.

《Failure prediction model》
FIG. 11 is a diagram illustrating an example of a failure prediction model. In the failure prediction model of FIG. 11, failure prediction logic and preventive measures are associated with each other. The failure prediction logic shows how to detect a device that is likely to fail. In addition, the preventive measure presents a method of an optimum measure for preventing a failure of an electronic device in which a failure sign is detected by the failure prediction logic.

  The failure prediction model development unit 23 calculates the accuracy of the failure prediction logic, selects the prevention measures as shown in FIG. 9 or 10 according to the accuracy of the failure prediction logic, and associates the selected prevention measures with the failure prediction logic. A failure prediction model is being developed.

<Cost estimation>
The cost trial calculation unit 24 calculates a service cost, for example, as shown in FIG. FIG. 12 is a diagram for explaining an example of a service cost calculation formula. FIG. 12 is a formula for calculating the total service cost for a failure.

  As shown in FIG. 12, the formula for calculating the total service cost for a failure is calculated by multiplying the number of failures that require CE arrangement by the service cost for maintenance of the failure.

  Further, the service cost for maintenance of the failure includes the labor cost of the CE and the parts cost. As shown in FIG. 12, the labor cost of CE is calculated by multiplying the time required for one visit, 1 + re-visit rate, and CE unit price. The part cost is the total part cost for maintenance of the failure. The time required for one visit is represented by travel time and work time. The travel time varies depending on the distance from the location where the CE is located to the location where the multifunction machine 12 is installed.

  For example, the cost trial calculation unit 24 performs the trial calculation of the service cost effect when the preventive measure by plus one is selected as follows. FIG. 13 is divided into four patterns depending on the result of failure sign detection by the failure prediction logic (predictive medium / missing) and possible / impossible of plus one correspondence.

  Pattern (1) in FIG. 13 is a case where the result of failure sign detection by the failure prediction logic is predictive, and a plus one response is possible. Further, pattern (2) in FIG. 13 is a case where the result of failure sign detection by the failure prediction logic is predictive, and plus one correspondence is impossible.

  Pattern (3) in FIG. 13 is a case where the result of failure sign detection by the failure prediction logic is idle, and a plus one response is possible. Pattern (4) in FIG. 13 is a case where the result of failure sign detection by the failure prediction logic is idle, and plus one correspondence is not possible.

  FIG. 14 is a diagram for explaining a trial calculation of the service cost effect when the preventive measure by plus one is selected. FIG. 14 (1) classifies the patterns (1) to (4) of FIG. 13 into the predictive state shown in FIG. In FIG. 14 (1), the numbers of patterns (1) and (2) are classified in a predictive manner, and the numbers of patterns (3) and (4) are classified in an empty manner.

  Here, in order to explain the trial calculation of the service cost effect when the plus one prevention measure is selected, the number of patterns (1) and (3) that can be treated as plus one is narrowed down as shown in FIG. 14 (2). . In FIG. 14 (2), the number of patterns (1) is classified predictively, and the number of patterns (3) is classified empty.

  In the case of FIG. 14 (2), the cost trial calculation unit 24 calculates the service cost generated due to the plus one correspondence by the following equation (1).

Increasing service cost = A x B x C yen (1)
A case: number of patterns (1) + (3) B minutes: work time for plus one C yen: unit price for CE Further, the cost calculation unit 24 is reduced by work that is unnecessary due to plus one. The service cost is calculated by the following equation (2).

Reduced service cost = D cases x E minutes x C yen (2)
D cases: Number of patterns (1) E minutes: Travel time and work time that can be reduced by Plus One correspondence The service cost effect when selecting the preventive measures by Plus One is based on the service cost calculated by Equation (2) Trial calculation is performed by subtracting the service cost calculated by Equation (1). For example, if the service cost calculated by subtracting the service cost calculated by the formula (1) from the service cost calculated by the formula (2) is positive, it is adopted as a profitable failure prediction model.

  FIG. 15 is a diagram illustrating an example of a trial calculation result of the service cost effect in the case where the plus one prevention measure is selected. In FIG. 15, 50 cases are classified in a predictive manner, 82 cases are classified in an idle manner, and 110 cases are classified without being found. The failure prediction logic of FIG. 15 is a state in which a failure (error A) is likely to occur when the current output value of the sensor A exceeds a predetermined value. Note that the accuracy of the failure prediction logic assumes that a failure with error A occurs within 30 days with a probability of 38%. In addition, the plus one correspondence is to clean the sensor A with water to prevent a failure caused by the sensor A contamination.

  In the example of FIG. 15, the hit rate is 38% and the cover rate is 31%. For example, in the example of FIG. 15, even if it is calculated that half of the predictive “25 cases” can be dealt with plus one, the service cost reduced by the plus one deal is 82 cases of plus one. The service cost increases due to

  FIG. 16 is a diagram illustrating an example of fluctuations in the current output value of the sensor A. As shown in FIG. 16, a failure (error A) has occurred after the current output value of sensor A has become equal to or greater than a predetermined value (predictive detection). Thus, before the failure (error A) occurs, the failure sensor A can be cleaned by wiping with a plus one to prevent an emergency visit due to the occurrence of the failure (error A).

  In addition, the cost trial calculation unit 24 performs the trial calculation of the service cost effect in the case of an emergency visit as follows. When parts replacement is required in an emergency visit, the cost estimation unit 24 can estimate the service cost effect of reducing the part-free revisit rate by arranging parts in advance. If the service cost reduction by reducing the no-part revisit rate is greater than the service cost increase by skipping, the preventive measure determination unit 43 is adopted as a profitable failure prediction model.

  FIG. 17 is a diagram showing an example of a trial calculation result of service cost effect in the case of an emergency visit. In FIG. 17, 3 cases are classified in a predictive manner, 0 cases are classified in an empty manner, and 11 cases are classified without being found. The failure prediction logic of FIG. 17 is that the current value of the current output value of the sensor B increases and the current value of the developing Y toner concentration decreases, so that the toner concentration becomes light and a failure (error C) occurs.

  The accuracy of the failure prediction logic is such that a failure with error C occurs with a probability of 100%. In addition, it is assumed that the work instruction to the CE at the emergency visit is to replace the toner supply unit.

  In the example of FIG. 17, the hit rate is 100% and the cover rate is 21.4%. For example, in the example of FIG. 17, calculation is performed assuming that an emergency visit is possible for “3 cases” in the prediction. The failure of error C has a high revisit rate, and the service cost reduced by the decrease in the revisit rate is estimated as the service cost effect in the case of an emergency visit.

  FIG. 18 is a diagram illustrating an example of fluctuations in the current output value of the sensor B and the development Y toner density. As shown in FIG. 18, a failure (error C) occurs after the current output value of the sensor B becomes equal to or higher than a predetermined value and the developing Y toner density becomes equal to or lower than a predetermined value (predictive detection). By replacing the toner replenishment unit by an emergency visit before the failure (error C) occurs, it is possible to prevent an emergency visit due to the occurrence of the failure (error C).

  Moreover, the cost trial calculation part 24 can perform the trial calculation of the service cost effect in the case of an emergency visit as follows using a cause diagnosis model. The service cost reduced by the cause diagnosis model may be due to failure cause diagnosis time, recovery treatment time, and travel time due to revisit.

  FIG. 19 is an example of a response flow by the CE in the case of an emergency visit. In step S11, the CE moves for an emergency visit. In step S12, the CE confirms the status of the multifunction machine 12. In step S13, failure cause diagnosis is performed using a cause diagnosis model. In step S14, the CE performs a recovery process based on the failure cause diagnosis result.

  In step S15, the CE checks the operation of the multifunction machine 12. In step S16, the CE performs standard work such as cleaning. In step S17, the CE ends the work flow in FIG.

  By performing the cause diagnosis of the failure by the cause diagnosis model before the emergency visit and notifying the CE of the cause diagnosis result of the failure and necessary parts before the emergency visit, the failure cause diagnosis time of step S13 in the corresponding flow of FIG. And the recovery treatment time of step S14 can be shortened. Further, in the corresponding flow of FIG. 19, the travel time due to the no-part revisit can be eliminated.

  FIG. 20 is a diagram showing an example of the cause of failure (error B). From FIG. 20, it can be determined that the main cause of the failure (error B) is “dirt of sensor A”. Therefore, a cause diagnosis model of the failure (error B) due to the cause of occurrence “dirt of sensor A” was examined.

  For example, contamination of the sensor A is suspected when the predetermined sensor light amount is equal to or greater than a predetermined value. Therefore, “predetermined sensor light quantity is a predetermined value or more” is defined as a determination condition. The failure (error B) is classified as shown in FIG. 21 depending on whether it is caused by contamination of the sensor A and whether the determination condition is met.

  FIG. 21 is a classification diagram of an example of a failure (error B). In FIG. 21, 25 cases are classified in a predictive manner, and 43 cases are classified as failures that do not meet the determination condition due to contamination of sensor A. Four cases are classified as failures that did not occur. In FIG. 21, the hit rate is 86% and the cover rate is 37%. The work instruction to the CE is to wipe the sensor A with water.

  In the example of FIG. 21, the failure cause diagnosis time in step S13 and the recovery treatment time in step S14 can be shortened for 25 cases that are predictive out of 68 cases (error B) due to contamination of sensor A. . In addition, the travel time due to the re-visit without parts can be eliminated.

  FIG. 22 is a diagram illustrating an example of fluctuations in the sensor light amount. In FIG. 22, a failure (error B) occurs after the sensor light quantity becomes equal to or greater than a predetermined value (2385) of the determination condition. When the sensor A is wiped and cleaned by an emergency visit, the sensor light amount is reduced to a level equal to or lower than a predetermined value (2385) of the determination condition.

<Summary>
According to the present embodiment, it is possible to perform a trial calculation of service cost when the developed failure prediction model is used, and to adopt a failure prediction model that can be profitable. For this reason, according to the present embodiment, a failure prediction model that cannot be profitable is adopted, and a state in which a loss is generated as the prevention measures are performed can be avoided.

  The present invention is not limited to the specifically disclosed embodiments, and various modifications and changes can be made without departing from the scope of the claims. The multifunction machine 12 is an example of an electronic device described in the claims. The information processing apparatus 10 is an example of one or more information processing apparatuses. The failure data is an example of failure history information. The work by the emergency visit is an example of the visit work. The work by revisit is an example of revisit work.

DESCRIPTION OF SYMBOLS 1 Information processing system 10 Information processing apparatus 12 Multifunction machine 14 Terminal device 21 State information acquisition part 22 Failure data acquisition part 23 Failure prediction model development part 24 Cost estimation part 25 Employment determination part 31 State information storage part 32 Failure data storage part 33 Failure Prediction model storage unit 41 Data pattern search unit 42 Failure prediction logic accuracy index calculation unit 43 Precaution prevention measure determination unit 500 Computer 501 Input device 502 Display device 503 External I / F
503a Recording medium 504 RAM
505 ROM
506 CPU
507 Communication I / F
508 HDD
601 Controller 602 Operation panel 603 External I / F
603a Recording medium 604 Communication I / F
605 Printer 606 Scanner 611 CPU
612 RAM
613 ROM
614 NVRAM
615 HDD
B bus N1 network

JP 2004-37941 A

Claims (8)

An information processing system having one or more information processing devices,
Failure history information of one or more electronic devices in which the failure has occurred in a failure prediction model in which a method for detecting a sign of a failure occurring in an electronic device and a preventive measure for a failure occurring in the electronic device are associated with each other And a cost calculation means for calculating the cost of the failure prediction model,
Based on the result of the trial calculation, adoption determination means for determining the adoption of the profitable failure prediction model;
An information processing system. The cost estimation means calculates a cost that is increased by the operation of the failure prediction model and a cost that is reduced by the operation of the failure prediction model,
If the cost reduced by the operation of the failure prediction model is larger than the cost increased by the operation of the failure prediction model, the adoption determination means determines to adopt the failure prediction model that can be profitable. The information processing system according to claim 1. The cost estimation means can be profitable if the cost of the work that is unnecessary due to the work is larger than the cost that is caused by the work, then the cost that is caused by the work when the preventive action is a work-preventive measure. The information processing system according to claim 2, wherein adoption is determined as the failure prediction model. The cost estimation means is characterized in that a sign of a failure is detected by the sign detection method, and the cost of work corresponding to the number of cases in which the failure actually occurred is calculated as the cost of work that is unnecessary due to work. The information processing system according to claim 3. The cost calculation means determines the cost of the revisiting work that is unnecessary than the cost generated by arranging the parts in advance, when the preventive action is a precautionary action that requires replacement of parts by visiting work. The information processing system according to claim 2, wherein if it is larger, the adoption is determined as the profitable failure prediction model. 6. The cost estimation unit according to any one of claims 1 to 5, wherein the cost calculation means calculates the cost of the failure prediction model based on a labor cost generated by a time required for the preventive measure and a parts cost of the preventive measure. An information processing system according to claim 1. A failure prediction model adoption determination method executed in an information processing system having one or more information processing devices,
Failure history information of one or more electronic devices in which the failure has occurred in a failure prediction model in which a method for detecting a sign of a failure occurring in an electronic device and a preventive measure for a failure occurring in the electronic device are associated with each other And a cost estimation step for estimating the cost of the failure prediction model,
Based on the result of the trial calculation, an adoption determination step for determining the adoption of the profitable failure prediction model;
A failure prediction model adoption determination method having Information processing device
Failure history information of one or more electronic devices in which the failure has occurred in a failure prediction model in which a method for detecting a sign of a failure occurring in an electronic device and a preventive measure for a failure occurring in the electronic device are associated with each other A cost calculation means for calculating the cost of the failure prediction model,
Based on the result of the trial calculation, adoption determination means for determining the adoption of the failure prediction model that can be profitable,
Program to function as.

Images