JP2011237576A - Failure-related item setting device, program and failure prediction system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To propose a technology to effectively select a device status item suitable for use for failure prediction of device.SOLUTION: A failure-related item extractor 12 takes the following steps in order: a step to obtain device status data included in maintenance data of image formation device (normal-system machine) in the status of nonoccurrence of failure while operating and those included in the maintenance data of image formation device (abnormal-system machine) in the status of occurrence of failure while operating, from a maintenance data storage unit 15; a step to compare a group of normal-system device status data consisting of more than one device status datum relating to the normal-system machine to a group of abnormal-system device status data consisting of more than one device status datum relating to the abnormal-system machine; a step to identify a device status item strongly linked to failure occurrence among many items about the device status; a step to generate failure-related item information in which the identified device status item is set as the device status item used for failure prediction; and a step to store the generated information in a failure-related item storage unit 16.

Description

  The present invention relates to a failure-related item setting device, a program, and a failure prediction system.

Various inventions have been proposed for failure prediction relating to devices to be monitored.
For example, after a plurality of image forming apparatuses are manufactured for a plurality of image forming apparatuses of the same model including the image forming apparatus whose state is to be determined, a plurality of sets of data related to the state of the image forming apparatus are acquired. In order to determine an index value calculation formula for calculating an index value for determining a state of a reference data group including all of a plurality of sets of data acquired during an operation test performed during the operation test of the plurality of image forming apparatuses. An invention used as an initial reference data group has been proposed (see Patent Document 1).

JP 2005-227518 A

  An object of the present invention is to propose a technique for effectively selecting an item of a device state suitable for use in device failure prediction.

  The present invention described in claim 1 is a first acquisition means for acquiring information on a device state detected by a device in which a failure has occurred during operation, and a device detected by a device in which no failure has occurred during operation. Based on the second acquisition means for acquiring state information, the apparatus state information acquired by the first acquisition means and the apparatus state information acquired by the second acquisition means, the state of the apparatus related to the failure A failure-related item setting device comprising: a specifying unit that specifies an item; and a setting unit that sets the item of the device state specified by the specifying unit as a device state item used for failure prediction. is there.

  According to a second aspect of the present invention, in the failure-related item setting device according to the first aspect, the first acquisition unit and the second acquisition unit are the same model and have an operating time close to satisfying a standard. Information on the device state detected by the device having the information is acquired as a comparison target.

  According to a third aspect of the present invention, in the failure-related item setting device according to the first aspect, the device is an image forming device, and the first acquisition unit and the second acquisition unit are of the same model. The apparatus status information detected by the apparatus having the number of image forming sheets close to satisfying the standard is acquired as a comparison target.

  The present invention described in claim 4 is a first acquisition function for acquiring information on a device state detected by a device in which a failure has occurred during operation, and a device in which a failure has not occurred during operation. A second acquisition function for acquiring information on the device state obtained, a device state information acquired by the first acquisition function, and a device state information acquired by the second acquisition function. It is a program for realizing a specific function for specifying an apparatus state item and a setting function for setting an apparatus state item extracted by the extraction function as an apparatus state item used for failure prediction.

  The present invention according to claim 5 is a first acquisition means for acquiring information on a device state detected by a device in which a failure has occurred during operation, and a device detected by a device in which no failure has occurred during operation. Based on the second acquisition means for acquiring state information, the apparatus state information acquired by the first acquisition means and the apparatus state information acquired by the second acquisition means, the state of the apparatus related to the failure Specifying means for specifying an item; setting means for setting an item of the device state specified by the specifying means as an item of an apparatus state used for failure prediction; and an apparatus for an item of the device state set by the setting means A failure prediction system comprising: a prediction unit that performs failure prediction on a device that is a target of failure prediction using state information.

  According to the present invention described in claims 1, 4, and 5, it is possible to effectively select an item of a device state suitable for use in device failure prediction as compared with a case where the present invention is not applied.

  According to the second and third aspects of the present invention, the information on the device state effective for selecting the item of the device state suitable for use in predicting the failure of the device can be efficiently obtained compared with the case where the present invention is not applied. Obtainable.

It is a figure which illustrates the structure of the failure prediction system which concerns on one Embodiment of this invention. It is a figure which illustrates the item of the apparatus state which concerns on one Embodiment of this invention. It is a figure which illustrates the structure of the failure related item extraction part which concerns on one Embodiment of this invention. It is a figure which illustrates the processing flow by the failure related item extraction part which concerns on one Embodiment of this invention. It is a figure which shows an example of the processing flow by the distance calculation part and failure determination part which concern on one Embodiment of this invention. It is a figure explaining an example of the process which estimates the remaining days until the failure which concerns on one Embodiment of this invention. It is a figure which shows the other example of the processing flow by the distance calculation part and failure determination part which concern on one Embodiment of this invention. It is a figure which illustrates the main hardware constitutions of the computer which operate | moves as a failure prediction apparatus which concerns on one Embodiment of this invention.

An embodiment of the present invention will be described with reference to the drawings.
FIG. 1 illustrates the configuration of a failure prediction system according to an embodiment of the present invention.
The failure prediction system of this example is installed in a market (the environment of a customer such as a customer) and is operating (used), and a failure prediction that performs failure prediction for the market installation machine M. Device S.

  In this example, a plurality of image forming apparatuses that can communicate with the failure prediction apparatus S wirelessly or by wire are used as the market setting machine M. However, the present invention is not limited to this, and the failure prediction apparatus S For example, devices that do not have a plurality of systems (for example, two systems of an operation system and a standby system), and are installed assuming normal operation Devices that are required to respond promptly in the event of a failure, such as devices that have been used, are targeted.

  In the image forming apparatus of this example, the apparatus state is monitored for a plurality of predetermined items to generate apparatus state data, which is included in the maintenance data and transmitted to the failure prediction apparatus S. As the item of the apparatus state relating to the image forming apparatus, as exemplified in FIG. 2, for example, the energization time for each operation mode, the operation time of main parts such as the heat fixing apparatus, and the environmental measurement values such as temperature and humidity There are various items such as the number of transported sheets, the distribution of the number of continuous printed pages in color or black and white, the number of prints by color / monochrome, paper size, paper type, double-sided print type, and the like.

  In the image forming apparatus of this example, maintenance data including apparatus status data for each item illustrated in FIG. 2 is transmitted to the failure prediction apparatus S at predetermined time intervals (for example, every day). However, the maintenance data may be transmitted to the failure prediction apparatus S at another timing. For example, the maintenance data may be transmitted in response to a request from the failure prediction apparatus S.

  In the image forming apparatus of this example, the maintenance data transmitted to the failure prediction apparatus S identifies the device status detection date and time (device status data creation date and time), device ID information for identifying the image forming device, and its model. The model ID information is included, and when the occurrence of a failure requiring maintenance work is detected in the image forming apparatus, the failure occurrence data indicating the occurrence of the failure is included in the maintenance data and transmitted promptly. I have to. In this example, maintenance data including failure occurrence data is transmitted only once after the occurrence of a failure is detected, but transmission may be repeated until the failure is removed by maintenance work or the like.

  The failure prediction device S of this example is a device that performs failure prediction related to an image forming device based on maintenance data obtained from a plurality of image forming devices (market installation machines M). The maintenance data acquisition unit 11, failure related item extraction Unit 12, distance calculation unit 13, failure determination unit 14, maintenance data storage unit 15, and failure related item storage unit 16.

  The maintenance data acquisition unit 11 acquires maintenance data from each of the plurality of image forming apparatuses and stores the maintenance data in the maintenance data storage unit 15. As described above, the maintenance data of this example includes the detection date and time, device ID information (and model ID information), and device status data, and the image forming apparatus corresponding to the device ID information is at the time of detection date and time. You can see what the device status was. Further, if the maintenance data includes failure occurrence data, it can be understood that a failure has occurred in the image forming apparatus.

  The failure-related item extraction unit 12 receives from the maintenance data storage unit 15 the device status data included in the maintenance data of the image forming apparatus (normal machine) in which no failure has occurred during operation, and the failure during operation. The apparatus status data included in the maintenance data of the image forming apparatus (abnormal machine) in the generated state is obtained, and the normal system apparatus status data group including one or more apparatus status data related to the normal system machine and the abnormal system machine are acquired. By comparing the abnormal system state data group composed of one or more pieces of the device state data, an item of the device state (hereinafter referred to as a failure related item) having a strong relationship with the occurrence of failure among various items regarding the device state. ) Is specified, and failure related item information in which the specified failure related item is set as an item of the device state used for failure prediction is generated and stored in the failure related item storage unit 16. In this example, the failure-related item is re-specified every predetermined period (for example, January), and the failure-related item information in the failure-related item storage unit 16 is updated.

  The distance calculation unit 13 includes the distance L from the normal system device state data group to the abnormal system device state data group and the normal system for the fault related items set in the fault related item information stored in the fault related item storage unit 16. A distance X from the apparatus state data group to the apparatus state data (the latest one) for the image forming apparatus targeted for failure prediction is calculated.

The failure determination unit 14 performs failure prediction on the image forming apparatus that is the target of failure prediction based on the distance L and the distance X calculated by the distance calculation unit 13.
In the failure prediction device S of this example, the result of failure prediction by the failure determination unit 14 is displayed on the display device provided in the failure prediction device S and presented to the operator. Other output methods may be used, for example, a method of printing out on a medium such as paper, or a predetermined transmission destination (such as an administrator or maintenance person of an image forming apparatus targeted for failure prediction). For example, a method of sending an e-mail to can be used.

Next, the failure related item extraction unit 12 will be described in more detail.
The failure-related item extraction unit 12 of this example includes a grouping unit 21, an average value calculation unit 22, a difference ratio extraction unit 23, and a difference ratio upper item extraction unit 24, as illustrated in FIG. Yes.

  The grouping unit 21 reads maintenance data for a plurality of image forming apparatuses to be compared from the maintenance data storage unit 15, and converts the device status data included in the read maintenance data into a normal system status data group and an abnormal system device. Classification (grouping) into state data groups.

In this example, when a failure occurs in the image forming apparatus, the maintenance data transmitted from the image forming apparatus includes the failure occurrence data. Therefore, depending on whether the maintenance data includes the failure occurrence data. It is determined whether or not there is a failure in the image forming apparatus.
Note that the presence or absence of a failure may be determined by another method. For example, the image forming apparatus includes a configuration in which a failure flag indicating the presence or absence of failure is included in the maintenance data and is included in the maintenance data. The presence / absence of a failure may be determined based on the value of the failure flag.

  Further, in this example, for a plurality of image forming apparatuses that are the same model and have an operation time close to satisfying the standard (same operation time), the respective apparatus state data are classified as comparison targets. The operating time includes the time elapsed since the device was installed, the time when the device was activated (the time when the device was in the ON state), the normal mode when the device was activated (an image that is an original function of the device) The time in which the forming function can be executed) can be used.

Note that the apparatus status data of a plurality of image forming apparatuses having other relationships may be used as comparison targets. For example, the number of image forming sheets that are the same model and meet the standard (the number of image forming sheets of the same degree) For a plurality of image forming apparatuses, the apparatus state data may be classified as comparison targets. As the number of images to be formed, the total number of prints, the number of prints for each print format such as color / monochrome, paper size, paper type, double-sided, and the like can be used.
Here, the model and operating time (or the number of images formed) of the image forming apparatus can be specified by the model ID information in the maintenance data and the value of the corresponding apparatus status data.

  Also, in this example, first, maintenance data (maintenance data including failure occurrence data) of the abnormal machine is collected, and then the same operating time (or the number of images formed, etc.) with the same model as the abnormal machine. Maintenance data (maintenance data that does not include failure occurrence data) is collected for the normal system machines that are included, and the device status data included in these is used as a comparison target, but the device status data to be compared is collected using another method. You may do it. For example, collection conditions related to the model and operating time (or the number of images to be formed) are set, maintenance data matching the collection conditions is collected, and the device status data included in the collected maintenance data is broken in the maintenance data. According to the presence / absence of the generated data, the data may be assigned to a normal system or an abnormal system, and these may be used as comparison targets.

  The average value calculation unit 22 calculates the average value of the device state data for each item of the device state for each of the normal device state data group and the abnormal device state data group obtained by the grouping unit 21. The average value for the normal system state data group is Ni, and the average value for the abnormal system state data group is ANi.

  Based on the average value Ni for the normal system state data group and the average value ANi for the abnormal system state data group calculated by the average value calculation unit 22 for each device state item, the difference ratio extraction unit 23 The difference ratio (ANi-Ni) / Ni is calculated for each state item. The difference ratio is an index of the degree of abnormality, and indicates the ratio (how far away) the abnormal data is separated from the normal data. The larger the value, the greater the degree of abnormality (different from normal) Become. That is, it is estimated that the larger the difference ratio, the stronger the relationship with the occurrence of a failure.

Based on the difference ratio calculated for each item of the apparatus state by the difference ratio extraction unit 23, the difference ratio upper item extraction unit 24 extracts the item of the apparatus state as a failure-related item in descending order of the difference ratio, and extracts the extracted items. Failure-related item information in which the failure-related item is set as an item of the device state used for failure prediction is generated and stored in the failure-related item storage unit 16.
The extraction of failure-related items includes various methods such as a method of extracting a predetermined number (or ratio) of items from a larger difference ratio and a method of extracting items having a difference ratio equal to or greater than a predetermined threshold. It can be performed by the method of.

FIG. 4 illustrates a processing flow by the failure-related item extraction unit 12.
First, the grouping unit 21 reads maintenance data for a plurality of image forming apparatuses to be compared from the maintenance data storage unit 15, and the apparatus status data included in the read plurality of maintenance data is abnormal with the normal system apparatus status data group. It classifies into a system device state data group (step S11). Thereafter, the average value calculation unit 22 calculates the average value Ni for the normal system state data group and the average value ANi for the abnormal system state data group for each device state item (step S12), and extracts the difference ratio. The unit 23 calculates the difference ratio (ANi-Ni) / Ni for each item of the device state (step S13), and the difference ratio higher-level item extraction unit 24 sets the items of the device state as failure related items in descending order of the difference ratio. Extraction is performed (step S14), and failure-related item information in which the extracted failure-related item is set as a device state item used for failure prediction is generated and stored in the failure-related item storage unit 16.

  As described above, in this example, the maintenance data of the image forming apparatus that is actually operating in the market is used, and the failure-related items are identified by classifying and comparing the normal machine and the abnormal machine. Failure related items are identified without using unnecessary data as much as possible. For this reason, compared to a configuration in which a failure-related item is specified using a large amount of data at the time of an operation test before actually operating as in the conventional method, the processing time for specifying the failure-related item is shortened. Further, the possibility that an item unsuitable for failure prediction is specified as a failure-related item is reduced, and failure-related items can be effectively selected.

Note that the method for specifying the failure-related item as described above is merely an example, and the failure-related item may be specified by another method.
In this example, the failure-related item is specified based on the apparatus status data for the same model image forming apparatus. However, in consideration of the influence of the version difference in the same model, further image formation of the same version is performed. The device status data about the device may be used, or another model that is assumed to have the same tendency although it is a different model is associated in advance, and the unit of the other model is summarized. You may make it use apparatus state data.
Also, the failure type may be considered. In this case, for example, information indicating the failure type is included in the failure occurrence data, and for each failure type, the failure related to the failure of that type is included. What is necessary is just to specify an item.

In this example, when a failure occurs in the image forming apparatus, the failure occurrence data is included in the maintenance data transmitted from the image forming apparatus to the failure prediction apparatus S. However, a failure occurs separately from the maintenance data. The data may be transmitted to the failure prediction apparatus S. In this case, for example, the failure occurrence data is dealt with by adding the model ID information and the failure occurrence date and time to the failure occurrence data. The correspondence with the maintenance data (failure occurrence data) at the time may be specified.
In this example, the image forming apparatus itself detects the occurrence of a failure and automatically generates failure occurrence data. However, this is requested by a user who has a sense of incongruity in the operation of the image forming apparatus or the user. Failure occurrence data may be generated in response to an operation input by a maintenance person who has performed maintenance work.

FIG. 5 shows an example of a processing flow by the distance calculation unit 13 and the failure determination unit 14.
First, for the failure related items set in the failure related item information stored in the failure related item storage unit 16, the distance calculation unit 13 uses the average vector for the normal system state data group and the abnormal system state data group. Are calculated (step S21), and the Euclidean distance L between these average vectors is calculated (step S22).
Further, the distance calculation unit 13 calculates the Euclidean distance X between the corresponding apparatus state data in the image forming apparatus that is a target of failure prediction and the average vector for the normal system state data group (step S23).
Thereafter, the failure determination unit 14 determines whether or not the condition of X / L <1 is satisfied based on the distance L and the distance X calculated by the distance calculation unit 13 (step S24). Performs prediction processing (for example, calculating the number of remaining days until failure) according to the value of X / L (step S25), and if the condition is not satisfied, it is determined that a failure has occurred (step S25). S26).

  The processes in steps S21 and S22 do not have to be performed every time a failure is predicted. For example, when a failure-related item is specified by the failure-related item extraction unit 12, the process L is performed to calculate the distance L, and the calculation is performed. The information indicating the average vector for the distance L and the normal system state data group used for the calculation is stored in the failure related item storage unit 16 in association with the failure related item information, and at the time of failure prediction. You may make it read from the failure related item memory | storage part 16. FIG.

Here, as the prediction process according to the value of X / L, for example, the value of X / L is calculated a plurality of times (the latest apparatus state data in the image forming apparatus that is the target of failure prediction, and one or more previous ones) And calculating the remaining number of days for which X / L ≧ 1 by linear approximation or the like.
An example of a process for predicting the number of remaining days until failure will be described with reference to FIG. FIG. 6 is an example in which a date when X / L exceeds 1 is obtained by extrapolating from the calculated value of X / L and used as a failure prediction date. In FIG. The value is “0.4”, the value of X / L calculated for June 1 is “0.7”, and the failure prediction date estimated based on these values is used until the failure. The remaining number of days is predicted.

FIG. 7 shows another example of the processing flow by the distance calculation unit 13 and the failure determination unit 14.
First, the distance calculation unit 13 is normal for the failure-related items set in the failure-related item information stored in the failure-related item storage unit 16 based on the normal system state data group and the abnormal system state data group. A set of abnormal state data of the system machines is used as an extracted data vector to calculate a covariance matrix M from the vectors of all normal machines (step S31), and an average vector A of all abnormal systems is calculated (step S32). Based on the covariance matrix M and the average vector A of abnormal machines, the Mahalanobis distance L from the vector of all normal machines to the average vector A of all abnormal machines is calculated (step S33).
Further, the distance calculation unit 13 calculates the Mahalanobis distance X from the vectors of all normal machines based on the corresponding apparatus state data vector and the covariance matrix M in the image forming apparatus targeted for failure prediction (step) S34).
Thereafter, the failure determination unit 14 determines whether or not the condition X / L <1 is satisfied based on the distance L and the distance X calculated by the distance calculation unit 13 (step S35). Performs prediction processing (for example, calculating the number of remaining days until failure) according to the value of X / L (step S36), and determines that the failure has occurred if the condition is not satisfied (step S36). S37).

Note that the processing in steps S31 to S33 does not have to be performed every time a failure is predicted. For example, when a failure-related item is specified by the failure-related item extraction unit 12, the processing is performed to calculate the distance L, and the calculation is performed. The information indicating the distance L and the covariance matrix M used for the calculation is stored in the failure-related item storage unit 16 in association with the failure-related item information, and the failure-related item storage unit 16 at the time of failure prediction. You may make it read from.
Moreover, as a prediction process according to the value of X / L, for example, a method of calculating the number of remaining days where X / L ≧ 1 by calculating the value of X / L by a plurality of times by linear approximation or the like is used. It can be carried out.

  Here, when the prediction process based on the processing flow illustrated in FIG. 5 is compared with the prediction process based on the processing flow illustrated in FIG. 7, the processing time tends to be shorter in the processing flow illustrated in FIG. Yes, the prediction accuracy tends to be higher in the processing flow illustrated in FIG.

FIG. 8 illustrates the main hardware configuration of a computer that operates as the failure prediction apparatus S of this example.
In this example, a CPU 31 that performs various arithmetic processes, a RAM 32 that is a work area of the CPU 31, a main storage device such as a ROM 33 that records basic control programs, an auxiliary that stores programs and various data according to an embodiment of the present invention. Storage device (for example, magnetic disk such as HDD, rewritable nonvolatile memory such as flash memory, etc.) 34, display device for displaying and outputting various information, operation buttons used for input operation by operator, touch panel, etc. The computer includes hardware resources such as an input / output I / F 35 that is an interface with the input device and a communication I / F 36 that is an interface for performing wired and wireless communication with other devices.
Then, the program according to the embodiment of the present invention is read from the auxiliary storage device 34 and the like, loaded into the RAM 32, and executed by the CPU 31, thereby causing the function of the failure prediction apparatus S according to the embodiment of the present invention to be executed by a computer. Realized above.

In the failure prediction apparatus S of this example, each functional unit is provided in one computer. However, each functional unit may be provided in a distributed manner in a plurality of computers.
In addition, the program according to the embodiment of the present invention may be a computer according to the present example, for example, in a format read from an external storage medium such as a CD-ROM storing the program or a format received via a communication line. Set to
Moreover, it is not restricted to the aspect which implement | achieves each function part by software configuration like this example, You may make it implement | achieve each function part with a dedicated hardware module.

M: Machine installed on the market, S: Failure prediction device,
11: Maintenance data acquisition unit, 12: Failure related item extraction unit, 13: Distance calculation unit, 14: Failure determination unit, 15: Maintenance data storage unit, 16: Failure related item storage unit, 21: Grouping unit, 22: Average Value calculation unit, 23: difference ratio extraction unit, 24: difference ratio upper item extraction unit

Claims (5)

First acquisition means for acquiring information on a device state detected by a device in which a failure has occurred during operation;
Second acquisition means for acquiring information on a device state detected by a device in which no failure has occurred during operation;
A specifying unit for specifying an item of the device state related to the failure based on the information on the device state acquired by the first acquiring unit and the information on the device state acquired by the second acquiring unit;
Setting means for setting the device status item specified by the specifying means to the device status item used for failure prediction;
A failure-related item setting device. The first acquisition unit and the second acquisition unit acquire, as a comparison target, information on a device state detected by a device that is the same model and has an operation time close to satisfying a standard,
The failure related item setting device according to claim 1. The apparatus is an image forming apparatus,
The first acquisition unit and the second acquisition unit acquire, as a comparison target, information on an apparatus state detected by an apparatus having the same model and having an image forming number close to satisfying a reference.
The failure related item setting device according to claim 1. On the computer,
A first acquisition function for acquiring information on a device state detected by a device in which a failure has occurred during operation;
A second acquisition function for acquiring information on a device state detected by a device in which no failure has occurred during operation;
A specifying function for identifying an item of the device state related to the failure based on the information on the device state acquired by the first acquisition function and the information on the device state acquired by the second acquisition function;
A setting function for setting the device status item extracted by the extraction function as a device status item used for failure prediction;
A program to realize First acquisition means for acquiring information on a device state detected by a device in which a failure has occurred during operation;
Second acquisition means for acquiring information on a device state detected by a device in which no failure has occurred during operation;
A specifying unit for specifying an item of the device state related to the failure based on the information on the device state acquired by the first acquiring unit and the information on the device state acquired by the second acquiring unit;
Setting means for setting the device status item specified by the specifying means to the device status item used for failure prediction;
Prediction means for performing failure prediction for a device to be subject to failure prediction using information on the device state for the item of device state set by the setting means;
A failure prediction system comprising:

Images