JP4795136B2 - Abnormality prediction apparatus, image forming apparatus, and maintenance system - Google Patents

Description

The present invention, abnormality predicting apparatus that performs prediction of the abnormal state of the image forming apparatus, an image forming apparatus and a maintenance system.

  Conventionally, in various image forming apparatuses on the market, if an abnormality of the apparatus such as a failure occurs, the apparatus cannot be used until the part is replaced or cleaned depending on the content of the apparatus. Inconvenience may occur. In particular, an electrophotographic image forming apparatus has a relatively complicated configuration and a large number of parts. Therefore, if maintenance of various parts is not performed regularly, abnormalities are likely to occur suddenly. Become.

Therefore, various methods for predicting the abnormality of the apparatus have been proposed, and the methods can be roughly divided into two.
One is a method based on time series analysis in which the state information acquired from the apparatus is observed every moment, the state of the apparatus at that time is estimated, and the future state is estimated from the time series change. This is based on the idea of extrapolating changes from the past to the present into the future. As a specific example, the time series change of the count value in which the cumulative operating time of various parts and devices in the apparatus such as the photoconductor and the developing apparatus is sequentially counted by a counter is as shown in FIG. When increasing with periodic fluctuations in period units, this time series change is approximated to the waveform, and the approximated waveform is extended toward the future (extrapolation) with the current point as the base point (extrapolation). The day to reach (multiple areas) can be predicted. Such prediction is accurate for relatively short-term predictions including the present, but is not accurate for long-term predictions.

  The other is a pattern recognition method that captures the correspondence between the state information of the device at a certain point in time and the state at a future point in time as a pattern. As a method based on this pattern recognition, multivariate analysis such as MTS method can be cited. This method is said to be able to cope with a longer-term prediction than the method based on the above time series analysis.

  Patent Document 1 proposes a method for predicting the occurrence of an abnormality in an image forming apparatus using the MTS method. That is, a method of acquiring a plurality of types of status information related to the status of the image forming apparatus, calculating an index value from the acquired plurality of types of status information, and predicting the occurrence of an abnormality in the image forming apparatus from the index value. . In this method, first, set data including a plurality of types of state information relating to an image forming apparatus is obtained from a normal state image forming apparatus or a normal state testing machine having the same specifications as the image forming apparatus. Then, a large number of group data are collected to construct a normal group data group (normal index information). When calculating the index value, a plurality of types of state information are acquired from the image forming apparatus. Then, a distance (index value) indicating what kind of relative positional relationship is present in the multi-dimensional space of the normal group data group constructed in advance for the state information. When an abnormality such as a failure is likely to occur away from the normal state, the correlation in the multidimensional space between the multiple types of state information and the normal set data group is disturbed, and the origin in the multidimensional space (normal The “distance” from the average state), that is, the index value increases. On the other hand, when the image forming apparatus is in a normal state, the “distance” from the origin (average of normal states), that is, the index value becomes small. Therefore, the normality of the image forming apparatus can be determined based on the index value. Therefore, in Patent Document 1, it is possible to detect a minor abnormality before a failure of the image forming apparatus is detected and predict the occurrence of the failure in advance. Based on the detection of anomalies, it is possible to reduce the downtime of the image forming apparatus by ordering parts in advance or by requesting a service person along with ordering the parts if they cannot replace the parts themselves. Can do. In addition, by exchanging only the components whose life is approaching so as to cause a slight abnormality, it is possible to avoid the high cost due to exchanging components that can still withstand use.

JP 2005-017874 A

In Patent Document 1, when calculating an index value, a plurality of types of state information regarding the image forming apparatus are acquired from a sensor or the like, and the index value is calculated based on the acquired information. Some state information has large variations and noise, so it was not possible to calculate a high-precision index value due to this state information variation and noise, and it was impossible to predict abnormalities with high accuracy. . Moreover, since patent document 1 is calculating an index value only using the instantaneous value (value acquired when calculating an index value) which does not have the information regarding a tendency (trend), it is highly accurate prediction of abnormality. I could not do it.
Therefore, the present inventors obtain a plurality of types of status information regarding the image forming apparatus for a plurality of days and calculate an index value such as a moving average or a change amount of the status information from the status information for a plurality of days. We considered extracting feature values and calculating index values by further adding these feature values. By adding the feature amount in this way, for example, information (trend) related to the temporal change of the state information such as whether the type of state information is in the descending process or the ascending process can be captured. As a result, the index value can be calculated in consideration of the temporal change of the state information, and factors such as variation and noise can be removed when calculating the index value, and the index value is more accurate than that of Patent Document 1. Can be calculated. Thereby, it is possible to predict an abnormality with high accuracy.

  However, if maintenance work such as replacement of parts and sensors in the image forming apparatus and cleaning of parts and equipment is performed based on the prediction of abnormality, the characteristics of the parts and equipment change before and after maintenance of the apparatus. For this reason, if maintenance is performed on a sensor that acquires status information from a part that has undergone maintenance or the sensor that detects the status itself, the status information before maintenance may differ greatly from the status information after maintenance. is there. Therefore, the feature quantity extracted using the state information before and after the maintenance is performed is not accurate. In other words, the extracted feature amount does not capture the temporal change of the state information. As a result, since the index value calculated based on this feature amount is not accurate, abnormality cannot be predicted correctly. For this reason, it is conceivable that after the maintenance of the part is performed, the abnormality determination is not performed until the state information for a plurality of days for the part is obtained. However, it is impossible to predict an abnormality until the state information for a plurality of days is complete, and even if an abnormality occurs while the state information for a plurality of days is complete, the image forming apparatus cannot be maintained. There was a problem.

  In the operation of the abnormality prediction system, when there is a sign of change in information acquired from equipment or parts before the occurrence of an abnormality such as a failure, measures such as parts replacement, cleaning, and adjustment are performed. At the time of carrying out, the device is not yet in an abnormal state. Therefore, even if maintenance personnel perform maintenance according to the notification information of the abnormality prediction system, a visible change as an operating state does not appear before and after maintenance. Therefore, whether or not the maintenance work has been effective cannot be confirmed unless the abnormality is determined and the maintenance is correctly performed. However, as described above, if the abnormality determination is not performed until the state information for a plurality of days is collected, there is a problem that the effect of the maintenance work cannot be confirmed.

The present invention has been made in view of the above problems, and the object of the present invention is to obtain even if the predetermined number of state information necessary for extracting a feature amount used for calculating an index value has not been acquired. , abnormal predicting abnormality prediction system that can ensure correct image forming apparatus is to provide an image forming apparatus and a maintenance system.

Information In order to achieve the above object, the invention according to claim 1, for obtaining the abnormality prediction system for performing prediction of the abnormal state of the image forming apparatus, a mutually different kinds of status information on the image forming apparatus at a predetermined timing An acquisition means, a storage means for storing the plurality of types of state information in time series, and a date of acquisition a predetermined number of times from the time of acquisition of the latest plurality of types of state information stored in the storage means. Based on a plurality of types of state information acquired up to that point in time, feature amount extraction means for extracting a feature amount that captures the temporal change of the state information for each type of state information, and feature amounts of the plurality of types of state information and the index value calculating means for calculating an index value indicating the state of the image forming apparatus on the basis of a first abnormality predicting means based on the index value to predict the occurrence of abnormality of the image forming apparatus, wherein Comprising a second abnormality predicting means for predicting the occurrence of abnormality in the image forming apparatus on the basis of the latest state information stored in憶means and, after performing maintenance for the image forming apparatus, said acquisition means while retrieving a predetermined number of times the plurality of types of state information, wherein the abnormality of the image forming apparatus is predicted by the first abnormality prediction means, after performing the maintenance for the image forming apparatus, said acquisition means Until the plurality of types of state information are acquired a predetermined number of times, the second abnormality predicting unit is configured to predict the occurrence of an abnormality in the image forming apparatus .
According to a second aspect of the present invention, in the abnormality prediction apparatus according to the first aspect, when the first abnormality prediction unit predicts the occurrence of an abnormality in the image forming apparatus , the second abnormality prediction unit also detects an abnormality in the image forming apparatus . The present invention is characterized in that the occurrence is predicted .
Further, the invention of claim 3, in the abnormality prediction system for performing prediction of the abnormal state of the image forming apparatus, an information obtaining unit configured to obtain mutually different types of status information about the image forming apparatus at a predetermined timing, the plurality The storage means for storing the state information of the types in time series, and acquired from the time when the latest state information of the plurality of types stored in the storage means is acquired by a predetermined number of acquisitions to the past. Based on a plurality of types of state information, feature amount extraction means for extracting a feature amount that captures temporal changes in the state information for each type of state information in the previous period, and the image formation based on the feature amounts of the plurality of types of state information and the index value calculating means for calculating an index value indicating the state of the apparatus, based on the index value, and an abnormality predicting means for predicting the occurrence of abnormality in the image forming apparatus, the image forming apparatus From the time maintenance is performed until the acquisition unit acquires the plurality of types of state information a predetermined number of times, the acquisition unit acquires the plurality of types of state information in advance. Is.
Further, the invention of claim 4, in the abnormality prediction system for performing prediction of the abnormal state of the image forming apparatus, an information obtaining unit configured to obtain mutually different types of status information about the image forming apparatus at a predetermined timing, the plurality The storage means for storing the state information of the types in time series, and acquired from the time when the latest state information of the plurality of types stored in the storage means is acquired by a predetermined number of acquisitions to the past. Based on a plurality of types of state information, feature amount extraction means for extracting a feature amount that captures a change over time of the state information for each type of state information, and the image formation based on the feature amount of the plurality of types of state information and the index value calculating means for calculating an index value indicating the state of the apparatus, based on the index value, and an abnormality predicting means for predicting the occurrence of abnormality in the image forming apparatus, the image forming apparatus After performing maintenance with, until the acquisition unit acquires a predetermined number of times the plurality of types of state information, the index value calculating means, based on the state information about the site maintenance is performed of the image forming apparatus A provisional index value is calculated based on feature quantities of a plurality of types of state information excluding extracted feature quantities, and the abnormality prediction unit predicts occurrence of an abnormality in the image forming apparatus based on the provisional index values. It is configured as described above.
According to a fifth aspect of the present invention, in the abnormality prediction device according to the third or fourth aspect, a second determination is made as to abnormality of the image forming apparatus based on the latest plurality of types of state information stored in the storage means. An abnormality determination means is provided.
According to a sixth aspect of the present invention, in the abnormality prediction device according to any one of the first to fifth aspects, a notification means for notifying a result of the abnormality prediction of the image forming apparatus is provided.
The invention of claim 7, in the abnormality prediction system according to claim 6, wherein the notification means, when the abnormality of the image forming apparatus Ru is expected, to notify the maintenance is necessary portion of the image forming apparatus It is a feature.
According to an eighth aspect of the present invention, in the abnormality prediction apparatus according to any one of the first to seventh aspects, the index value calculation means calculates an index value using normal index information that is an index of a normal state of the image forming apparatus. An abnormality prediction apparatus characterized by:
According to a ninth aspect of the present invention, there is provided a recording medium conveying means for conveying a recording medium, a visible image forming means for forming a visible image on a recording medium conveyed by the recording medium conveying means, and the whole or a part of the apparatus. In an image forming apparatus comprising an abnormality prediction unit that predicts an abnormality, any one of claims 1 to 8 is used as the abnormality prediction unit.
The invention of claim 10 is provided with an abnormality predicting means for predicting an abnormality of the image forming apparatus, based on a determination result of the abnormality prediction means, in the maintenance system for performing maintenance of the image forming apparatus, the abnormality prediction means As described above, any one of claims 1 to 8 is used.

According to the first aspect of the present invention, the abnormality is normally predicted based on the temporal change data of the state information by performing the abnormality determination by the first abnormality determination means, so that highly accurate prediction can be performed. In addition, while the first abnormality determination unit cannot perform correct abnormality determination after the maintenance of the image forming apparatus until the acquisition unit acquires a predetermined number of the plurality of types of state information, the second abnormality determination unit To determine whether the image forming apparatus is abnormal. Since the second abnormality determination unit makes the determination based on the latest state information stored in the storage unit, even if the predetermined number of state information is not acquired as in the first abnormality determination unit, the image forming apparatus It is possible to determine the abnormality. Therefore, even when the first abnormality determination unit cannot perform the correct abnormality determination of the image forming apparatus , the second abnormality determination unit can determine the abnormality of the image forming apparatus . Accordingly, it is possible to perform maintenance of the image forming apparatus until the acquisition unit acquires a plurality of types of state information a predetermined number of times, and immediately check whether the maintenance of the image forming apparatus has been performed correctly. Can do.

According to the third aspect of the present invention, the timing at which the acquisition unit acquires a plurality of types of status information after the image forming apparatus is maintained until the acquisition unit acquires a predetermined number of the plurality of types of status information. Therefore, a predetermined number of types of state information can be stored in the storage means at an early stage. Accordingly, it is possible to extract feature amounts for each of the plurality of types of state information based on a predetermined number of state information at an early stage, and to calculate index values based on the extracted feature amounts of the plurality of types of state information. Therefore, it is possible to determine the subsequent abnormality of the image forming apparatus based on the index value at an early stage, and to shorten the period during which the abnormality determination cannot be performed. Thereby, it can be confirmed at an early stage whether or not the maintenance of the image forming apparatus is correctly performed.

Further, according to the invention of claim 5, the state for the image forming apparatus after performing maintenance, until obtaining means obtains a predetermined number multiple types of status information, regarding site maintenance has been performed in the image forming apparatus A provisional index value from which information is removed is calculated, and abnormality is determined based on the provisional index value. Since the status information related to the parts where the image forming apparatus has been maintained varies greatly before and after the maintenance, it is possible to calculate a highly accurate index value by excluding this status information, and to detect abnormalities with high accuracy. Judgment can be made. Thus, after performing the maintenance for an image forming apparatus, it is possible to acquire means until obtaining a predetermined number multiple types of status information, it determines an abnormality of at least maintenance performed the image forming apparatus other than the site . Even if there is an abnormality in a part other than the part where the maintenance of the image forming apparatus is performed before the acquisition unit acquires a predetermined number of types of state information, the part can be maintained.

A first embodiment in which the present invention is applied to an electrophotographic copying machine (hereinafter simply referred to as a copying machine) as an image forming apparatus will be described below.
First, the basic configuration of the copier according to the first embodiment will be described. FIG. 1 is a schematic configuration diagram showing the copying machine. The copier includes an image forming unit including a printer unit 100 and a paper feeding unit 200, a scanner unit 300, and a document conveying unit 400. The scanner unit 300 is mounted on the printer unit 100, and a document transport unit 400 including an automatic document transport device (ADF) is mounted on the scanner unit 300.

  The scanner unit 300 reads the image information of the document placed on the contact glass 32 by the reading sensor 36 and sends the read image information to a control unit (not shown). Based on the image information received from the scanner unit 300, the control unit controls lasers and LEDs (not shown) disposed in the exposure device 21 of the printer unit 100 to provide four drum-shaped photoconductors 40K, Y, and M. , C is irradiated with laser writing light L. By this irradiation, an electrostatic latent image is formed on the surface of the photoreceptors 40K, Y, M, and C, and this latent image is developed into a toner image through a predetermined development process. Note that the subscripts K, Y, M, and C added after the reference numerals indicate specifications for black, yellow, magenta, and cyan.

  In addition to the exposure device 21, the printer unit 100 includes primary transfer rollers 62K, Y, M, and C, a secondary transfer device 22, a fixing device 25, a paper discharge device, a toner supply device (not shown), a toner supply device, and the like. Yes.

  The paper feeding unit 200 includes an automatic paper feeding unit disposed below the printer unit 100 and a manual feeding unit disposed on a side surface of the printer unit 100. The automatic paper feed unit separates the two paper feed cassettes 44 arranged in multiple stages in the paper bank 43, the paper feed roller 42 that feeds transfer paper as a recording medium from the paper feed cassette, and the fed transfer paper. A separation roller 45 and the like are sent to the paper feed path 46. Further, it also includes a transport roller 47 that transports the transfer paper to the paper feed path 48 of the printer unit 100. On the other hand, the manual feed section includes a manual feed tray 51 and a separation roller 52 that separates transfer sheets on the manual feed tray 51 one by one toward the manual feed path 53.

  A registration roller pair 49 is disposed near the end of the paper feed path 48 of the printer unit 100. The registration roller pair 49 is formed between the intermediate transfer belt 10 serving as an intermediate transfer body and the secondary transfer device 22 at a predetermined timing after receiving the transfer paper sent from the paper feed cassette 44 or the manual feed tray 51. To the secondary transfer nip.

  In this copying machine, an operator sets a document on the document table 30 of the document transport unit 400 when copying a color image. Alternatively, after the document conveying unit 400 is opened and a document is set on the contact glass 32 of the scanner unit 300, the document conveying unit 400 is closed and the document is pressed. Then, a start switch (not shown) is pressed. Then, when an original is set on the original conveying unit 400, after the original is conveyed onto the contact glass 32, immediately after the original is set on the contact glass 32, the scanner unit 300 is driven. Start. Then, the first traveling body 33 and the second traveling body 34 travel, and the light emitted from the light source of the first traveling body 33 is reflected by the document surface and then travels toward the second traveling body 34. Further, after being reflected by the mirror of the second traveling body 34, it reaches the reading sensor 36 via the imaging lens 35 and is read as image information.

  When the image information is read in this way, the printer unit 100 rotates the other two support rollers while rotating one of the support rollers 14, 15, and 16 with a drive motor (not shown). Then, the intermediate transfer belt 10 stretched around these rollers is moved endlessly. Further, laser writing as described above and a development process described later are performed. Then, while rotating the photoconductors 40K, Y, M, and C, monochrome images of black, yellow, magenta, and cyan are formed on them. These are four-color superposed toners that are sequentially superposed and electrostatically transferred at the primary transfer nips for K, Y, M, and C where the photoreceptors 40K, Y, M, and C and the intermediate transfer belt 10 abut. Become a statue. Toner images are formed on the photoreceptors 40K, 40Y, 40M, and 40C.

  On the other hand, the paper feed unit 200 operates any one of the three paper feed rollers to feed the transfer paper having a size corresponding to the image information, and feeds the transfer paper to the paper feed path of the printer unit 100. Lead to 48. After the transfer paper that has entered the paper feed path 48 is sandwiched between the registration roller pair 49 and temporarily stopped, the intermediate transfer belt 10 and the secondary transfer roller 23 of the secondary transfer device 22 come into contact with each other at the appropriate timing. To the secondary transfer nip, which is a part. Then, in the secondary transfer nip, the four-color superimposed toner image on the intermediate transfer belt 10 and the transfer paper are brought into close contact in synchronization. Then, the four-color superimposed toner image is secondarily transferred onto the transfer paper due to the influence of the transfer electric field formed at the nip, the nip pressure, etc., and becomes a full color image combined with the white color of the paper.

  The transfer paper that has passed through the secondary transfer nip is fed into the fixing device 25 by the endless movement of the transport belt 24 of the secondary transfer device 22. Then, after the full color image is fixed by the action of the pressure applied by the pressure roller 27 of the fixing device 25 and the heating by the heating belt, the paper discharge tray 57 provided on the side surface of the printer unit 100 via the discharge roller 56. Discharged to the top.

  FIG. 2 is an enlarged configuration diagram illustrating the printer unit 100. The printer unit 100 includes a belt unit, four process units 18K, Y, M, and C that form toner images of respective colors, a secondary transfer device 22, a belt cleaning device 17, a fixing device 25, and the like.

  The belt unit moves the intermediate transfer belt 10 stretched around a plurality of rollers endlessly while contacting the photoreceptors 40K, Y, M, and C. In the primary transfer nip for K, Y, M, and C where the photoreceptors 40K, Y, M, and C are brought into contact with the intermediate transfer belt 10, the intermediate transfer belt 10 is driven by the primary transfer rollers 62K, Y, M, and C. Is pressed from the back side toward the photoconductors 40K, Y, M, and C. A primary transfer bias is applied to the primary transfer rollers 62K, Y, M, and C by a power source (not shown). As a result, a primary transfer electric field for electrostatically moving the toner images on the photoreceptors 40K, Y, M, and C toward the intermediate transfer belt 10 is formed in the primary transfer nips for K, Y, M, and C. Has been. Between each primary transfer roller 62K, Y, M, and C, a conductive roller 74 that is in contact with the back surface of the intermediate transfer belt 10 is disposed. In these conductive rollers 74, the primary transfer bias applied to the primary transfer rollers 62K, Y, M, and C is applied to adjacent process units via the intermediate resistance base layer 11 on the back side of the intermediate transfer belt 10. It prevents the flow.

  The process unit (18K, Y, M, C) supports the photosensitive member (40K, Y, M, C) and several other devices as a single unit on a common support. The unit 100 is detachable. Taking the black process unit 18K as an example, this has a developing unit 61K as developing means for developing the electrostatic latent image formed on the surface of the photosensitive member 40K into a black toner image in addition to the photosensitive member 40K. ing. Further, it also has a photoreceptor cleaning device 63K that cleans transfer residual toner adhering to the surface of the photoreceptor 40K after passing through the primary transfer nip. Further, there are a static elimination device (not shown) that neutralizes the surface of the photoreceptor 40K after cleaning, a charging device (not shown) that uniformly charges the surface of the photoreceptor 40K after static elimination, and the like. The process units 18Y, M, and C for other colors have almost the same configuration except that the color of the toner to be handled is different. The copying machine has a so-called tandem configuration in which these four process units 18K, Y, M, and C are arranged so as to face the intermediate transfer belt 10 along the endless movement direction.

  FIG. 3 is a partially enlarged view showing a part of the tandem unit 20 including the four process units 18K, Y, M, and C. Since the four process units 18K, Y, M, and C have substantially the same configuration except that the colors of the toners to be used are different, K, Y, M, and C attached to the respective reference numerals in FIG. The subscript is omitted. As shown in the figure, the process unit 18 includes a charging device 60 as a charging unit, a developing device 61, a primary transfer roller 62 as a primary transfer unit, a photoconductor cleaning device 63, a charge eliminating device around the photoconductor 40. A device 64 or the like is provided.

  As the photosensitive member 40, a drum-shaped member is used in which a photosensitive organic layer is applied to a base tube made of aluminum or the like to form a photosensitive layer. However, an endless belt may be used. In addition, as the charging device 60, a charging device to which a charging roller to which a charging bias is applied is rotated while being in contact with the photoreceptor 40 is used. A scorotron charger or the like that performs a non-contact charging process on the photoreceptor 40 may be used.

  The developing device 61 develops a latent image using a two-component developer containing a magnetic carrier and a nonmagnetic toner. An agitating unit 66 that conveys the two-component developer accommodated in the inside while stirring and supplies the developer sleeve 65 to the developing sleeve 65; , C, a developing unit 67 for transferring to C.

  The stirring unit 66 is provided at a position lower than the developing unit 67, and includes two screws 68 arranged in parallel to each other, a partition plate provided between the screws, and a toner provided on the bottom surface of the developing case 70. It has a density sensor 71 and the like.

  The developing unit 67 includes a developing sleeve 65 that faces the photosensitive member 40 through the opening of the developing case 70, a magnet roller 72 that is non-rotatably provided inside the developing sleeve 65, a doctor blade 73 that approaches the developing sleeve 65, and the like. ing. The distance at the closest portion between the doctor blade 73 and the developing sleeve 65 is set to about 500 [μm]. The developing sleeve 65 has a non-magnetic rotatable sleeve shape. In addition, the magnet roller 72 that is prevented from being rotated around the developing sleeve 65 has, for example, five magnetic poles N1, S1, N2, S2, and S3 in the rotation direction of the developing sleeve 65 from the position of the doctor blade 73. ing. Each of these magnetic poles applies a magnetic force to the two-component developer on the sleeve at a predetermined position in the rotation direction. As a result, the two-component developer sent from the stirring unit 66 is attracted and carried on the surface of the developing sleeve 65, and a magnetic brush is formed along the magnetic field lines on the sleeve surface.

  The magnetic brush is regulated to an appropriate layer thickness when passing through the position facing the doctor blade 73 as the developing sleeve 65 rotates, and then conveyed to the developing area facing the photoreceptor 40. Then, due to the potential difference between the developing bias applied to the developing sleeve 65 and the electrostatic latent image on the photoreceptor 40, it is transferred onto the electrostatic latent image and contributes to development. Further, as the developing sleeve 65 rotates, the developing sleeve 65 returns to the developing portion 67 again, and after being separated from the sleeve surface due to the influence of the repulsive magnetic field between the magnetic poles of the magnet roller 72, the developing sleeve 65 is returned to the stirring portion 66. In the stirring unit 66, an appropriate amount of toner is supplied to the two-component developer based on the detection result by the toner density sensor 71. The developing device 61 may employ a one-component developer that does not include a magnetic carrier, instead of the one that uses a two-component developer.

  As the photoconductor cleaning device 63, a system in which a cleaning blade 75 made of polyurethane rubber is pressed against the photoconductor 40 is used, but another system may be used. In order to improve the cleaning property, in this example, a cleaning device 63 having a contact conductive fur brush 76 whose outer peripheral surface is in contact with the photoreceptor 40 is rotatable in the direction of the arrow in the figure. A metal electric field roller 77 for applying a bias to the fur brush 76 is provided so as to be rotatable in the direction of the arrow in the figure, and the tip of the scraper 78 is pressed against the electric field roller 77. The toner removed from the electric field roller 77 by the scraper 78 falls on the collection screw 79 and is collected.

  The photoconductor cleaning device 63 having such a configuration removes residual toner on the photoconductor 40 with a fur brush 76 that rotates in the counter direction with respect to the photoconductor 40. The toner adhering to the fur brush 76 is removed by the electric field roller 77 to which a bias that rotates in contact with the fur brush 76 in the counter direction is applied. The toner adhering to the electric field roller 77 is cleaned by the scraper 78. The toner recovered by the photoconductor cleaning device 63 is brought to one side of the photoconductor cleaning device 63 by the recovery screw 79 and returned to the developing device 61 by the toner recycling device 80 for reuse.

  The static eliminator 64 includes a static elimination lamp or the like, and removes the surface potential of the photoreceptor 40 by irradiating light. The surface of the photoreceptor 40 thus neutralized is uniformly charged by the charging device 60 and then subjected to optical writing processing.

  A secondary transfer device 22 is provided below the belt unit in the figure. In the secondary transfer device 22, a secondary transfer belt 24 is stretched between two rollers 23 and moved endlessly. One of the two rollers 23 is a secondary transfer roller to which a secondary transfer bias is applied by a power source (not shown), and the intermediate transfer belt 10 and the secondary transfer belt 24 are between the roller 16 of the belt unit. Is sandwiched. Thus, a secondary transfer nip is formed in which both belts move in the same direction at the contact portion while contacting. On the transfer paper fed from the registration roller pair 49 to the secondary transfer nip, the four-color superimposed toner image on the intermediate transfer belt 10 is secondarily transferred collectively under the influence of the secondary transfer electric field and the nip pressure, so that full color is obtained. An image is formed. The transfer sheet that has passed through the secondary transfer nip is separated from the intermediate transfer belt 10 and is conveyed to the fixing device 25 along with the endless movement of the belt while being held on the surface of the secondary transfer belt 24. Note that secondary transfer may be performed by a transfer charger or the like instead of the secondary transfer roller.

  The surface of the intermediate transfer belt 10 that has passed through the secondary transfer nip approaches a support position by the support roller 15. Here, the intermediate transfer belt 10 is sandwiched between a belt cleaning device 17 that contacts the front surface (loop outer surface) and a support roller 15 that contacts the back surface. Then, after the transfer residual toner adhering to the front surface is removed by the belt cleaning device 17, it sequentially enters the primary transfer nip for K, Y, M, and C, and the next four-color toner The images are superimposed.

  The belt cleaning device 17 has two fur brushes 90 and 91. By rotating a plurality of raised brushes in contact with the intermediate transfer belt 10 in a counter direction with respect to the flocking direction, the transfer residual toner on the belt is mechanically scraped off. In addition, a cleaning bias is applied by a power source (not shown) to electrostatically attract and recover the scraped transfer residual toner.

  With respect to the fur brushes 90 and 91, the metal rollers 92 and 93 are rotating in the forward or reverse direction while being in contact with each other. Among these metal rollers 92 and 93, a negative polarity voltage is applied to the metal roller 92 located on the upstream side in the rotation direction of the intermediate transfer belt 10 by the power source 94. Further, a positive polarity voltage is applied to the metal roller 93 located on the downstream side by the power source 95. The tips of the blades 96 and 97 are in contact with the metal rollers 92 and 93, respectively. In such a configuration, the upstream fur brush 90 first cleans the surface of the intermediate transfer belt 10 with the endless movement of the intermediate transfer belt 10 in the direction of the arrow in the drawing. At this time, for example, when −400 [V] is applied to the fur brush 90 while −700 [V] is applied to the metal roller 92, first, the positive polarity toner on the intermediate transfer belt 10 is moved to the fur brush 90 side. Electrostatic transfer to The toner transferred to the fur brush side is further transferred from the fur brush 90 to the metal roller 92 due to a potential difference, and is scraped off by the blade 96.

  As described above, the toner on the intermediate transfer belt 10 is removed by the fur brush 90, but a lot of toner still remains on the intermediate transfer belt 10. These toners are negatively charged by a negative polarity bias applied to the fur brush 90. This is considered to be charged by charge injection or discharge. Next, by using the fur brush 91 on the downstream side to apply a positive polarity bias and perform cleaning, the toner can be removed. The removed toner is transferred from the fur brush 91 to the metal roller 93 due to a potential difference and scraped off by the blade 97. The toner scraped off by the blades 96 and 97 is collected in a tank (not shown).

  Most of the toner is removed from the surface of the intermediate transfer belt 10 after being cleaned by the fur brush 91, but a little toner is still left. The toner remaining on the intermediate transfer belt 10 is charged with a positive polarity by a positive polarity bias applied to the fur brush 91 as described above. Then, it is transferred to the photoconductors 40K, Y, M, and C by the transfer electric field applied at the primary transfer position, and is collected by the photoconductor cleaning device 63.

  In general, the registration roller pair 49 is often used while being grounded. However, a bias can be applied to remove the paper dust of the transfer paper P.

  Under the secondary transfer device 22 and the fixing device 25, a transfer paper reversing device 28 (see FIG. 1) is provided so as to extend in parallel with the tandem portion 20 described above. As a result, the transfer paper that has undergone the image fixing process on one side is switched by the switching claw to the transfer paper path to the transfer paper reversing device side, where it is reversed and enters the secondary transfer transfer nip again. Then, after the secondary transfer process and the fixing process of the image are performed on the other side, the sheet is discharged onto a discharge tray.

FIG. 4 is a functional block diagram of a control unit that controls the copying machine. The control unit 1 actually performs overall control of the entire copying machine, but in the drawing, only functions relating to an abnormality prediction and maintenance system of the image forming apparatus are displayed.
As shown in the figure, the control unit 1 is a control unit that controls the entire copying machine, and includes a calculation unit 1a including a CPU and a storage unit 1c including a RAM and a ROM. The storage means 1c stores status information of a plurality of types of copying machines in time series. Further, the control unit 1 serves as an index value calculation unit that calculates an index value for predicting an abnormal state such as a failure of the image forming apparatus based on a plurality of types of state information of the copying machine stored in the storage unit 1c. It has the function of The control unit 1 also has a function as an abnormality determination unit that determines an abnormal state of the image forming apparatus based on the index value and the state information of the copying machine.
Further, the control unit 1 acquires information from the various sensors 2, the operation display unit 3 and the like on the state of the components of the copying machine and the phenomenon occurring inside, and stores the information in the storage unit 1c. An information acquisition unit 1e serving as acquisition means is provided. The operation display unit 3 includes a display unit configured by a liquid crystal display or the like that displays character information or the like, an operation unit that receives input information from an operator using a numeric key or the like, and sends the input information to the control unit 1.
The control unit 1 also has a function as a notification unit that displays the determination result of the abnormal state of the image forming apparatus on the operation display unit 3 to notify the user. Further, when the determination result shows that there is an abnormality in the image forming apparatus or when an abnormality is predicted, the control unit 1 specifies a device or a part where the occurrence of the abnormality is predicted (or an abnormality has occurred). Also, it has a function as a notification means for displaying on the operation display unit and notifying the user.

  The copying machine status information acquired by the information acquisition unit 1e, which is an information acquisition unit, includes sensing data, control parameter data, input data, image reading data, and the like, and examples include the following.

(A) Sensing information The sensing information includes information related to driving, various characteristics of recording media, developer characteristics, photoreceptor characteristics, various process conditions of electrophotography, environmental conditions, various characteristics of recorded materials, etc. Conceivable. The outline of the sensing information is as follows.

(A-1) Driving information / Rotation speed of the photosensitive drum is detected by an encoder, the current value of the driving motor is read, and the temperature of the driving motor is read.
Similarly, the drive state of a cylindrical or belt-like rotating part such as a fixing roller, a paper transport roller, or a drive roller is detected.
-Detect sound generated by driving with a microphone installed inside or outside the device.

(A-2) Paper transport status ・ Uses a transmission or reflection type optical sensor or contact type sensor to detect the position of the leading and trailing edges of the transported paper and detect that a paper jam has occurred. The deviation of the passage timing of the leading and trailing edges of the paper and the fluctuation in the direction perpendicular to the feeding direction are read.
Similarly, the paper moving speed is obtained based on the detection timing between a plurality of sensors.
The slip between the paper supply roller and the paper during paper supply is obtained by comparing the measured value of the rotation speed of the roller with the amount of movement of the paper.

(A-3) Various characteristics of recording media such as paper This information greatly affects the image quality and stability of sheet conveyance. There are the following methods for acquiring information on this paper type.
-The thickness of the paper should be equal to the amount of movement of the member pushed up when the paper is sandwiched between two rollers and the relative positional displacement of the rollers is detected by an optical sensor or the paper enters. Find by detecting.
-The surface roughness of the paper is such that a guide or the like is brought into contact with the surface of the paper before transfer, and vibrations or sliding noises caused by the contact are detected.
-For the gloss of paper, a light beam with a specified opening angle is incident at a specified incident angle, and a light beam with a specified opening angle reflected in the specular reflection direction is measured by a sensor.
The paper rigidity is obtained by detecting the amount of deformation (curvature) of the pressed paper.
The judgment as to whether or not the paper is recycled is made by irradiating the paper with ultraviolet rays and detecting its transmittance.
The determination as to whether the paper is a backing paper is performed by irradiating light from a linear light source such as an LED array and detecting the light reflected from the transfer surface with a solid-state image sensor such as a CCD.
Whether or not the sheet is for OHP is determined by irradiating the paper with light and detecting regular reflection light having a different angle from the transmitted light.
• The amount of water contained in the paper is determined by measuring the absorption of infrared or μ-wave light.
-The amount of curl is detected by an optical sensor, contact sensor, etc.
-The electrical resistance of the paper is measured directly by contacting a pair of electrodes (such as paper feed rollers) with the recording paper, or by measuring the surface potential of the photoconductor or intermediate transfer body after paper transfer, and recording from that value. Estimate the resistance of the paper.

(A-4) Developer characteristics The characteristics of the developer (toner carrier) in the apparatus affect the basic function of the electrophotographic process. Therefore, it becomes an important factor for the operation and output of the system. Obtaining developer information is extremely important. Examples of the developer characteristics include the following items.
-For toner, list charge amount and distribution, fluidity / cohesion / bulk density, electrical resistance, amount of external additive, consumption or remaining amount, fluidity, toner concentration (mixing ratio of toner and carrier) Can do.
-As for carriers, magnetic properties, coat film thickness, spent amount, etc. can be mentioned.

It is usually difficult to detect such items alone in the image forming apparatus. Therefore, it is detected as a comprehensive characteristic of the developer. The overall characteristics of this developer can be measured, for example, as follows.
A test latent image is formed on the photoconductor, developed under predetermined development conditions, and the reflection density (light reflectance) of the formed toner image is measured.
A pair of electrodes is provided in the developing device, and the relationship between applied voltage and current is measured (resistance, dielectric constant, etc.).
-Install a coil in the developing device and measure the voltage-current characteristics (inductance).
-A level sensor is provided in the developing device to detect the developer capacity. The level sensor includes an optical type and a capacitance type.

(A-5) Photoreceptor characteristics Like the developer characteristics, the photoreceptor characteristics are closely related to the function of the electrophotographic process. Information on the photoconductor characteristics includes photoconductor film thickness, surface characteristics (friction coefficient, unevenness), surface potential (before and after each process), surface energy, scattered light, temperature, color, surface position (flare), linear velocity Potential decay rate, resistance / capacitance, surface moisture content, and the like. Among these, the following information can be detected in the image forming apparatus.
・ Detect the current flowing from the charging member to the photoconductor, and simultaneously compare the change in capacitance with the change in film thickness with the voltage-current characteristics with respect to the voltage applied to the charging member and the preset dielectric thickness of the photoconductor. Thus, the film thickness is obtained.
・ Surface potential and temperature can be obtained by a conventionally known sensor.
-The linear velocity is detected by an encoder attached to the rotating shaft of the photoconductor.
-Scattered light from the surface of the photoreceptor is detected by an optical sensor.

(A-6) Electrophotographic process state As is well known, electrophotographic toner image formation is performed by uniformly charging the photoreceptor, forming a latent image (image exposure) using laser light, etc., and charged toner (colored particles). Development by transfer, transfer of toner image onto transfer material (in the case of color, overlay on the recording medium that is the intermediate transfer body or final transfer material, or over development on the photoreceptor during development), toner on the recording medium This is done in the order of image fixing. Various information at each of these stages greatly affects the output of images and other systems. Obtaining these is important in evaluating the stability of the system. Specific examples of the information acquisition of the electrophotographic process state include the following.
The charging potential and the exposure portion potential are detected by a conventionally known surface potential sensor.
The gap between the charging member and the photosensitive member in non-contact charging is detected by measuring the amount of light that has passed through the gap.
・ Electromagnetic waves from electrification are captured by a broadband antenna.
・ Sound generated by charging ・ Exposure intensity ・ Exposure light wavelength

Further, as a method for acquiring various states of the toner image, the following can be mentioned.
The pile height (the height of the toner image) is obtained by measuring the depth from the vertical direction with a displacement sensor and the light shielding length from the horizontal direction with a linear sensor of parallel light.
The toner charge amount is measured by a potential sensor that measures the potential of the electrostatic latent image on the solid part and the potential when the latent image is developed, and the ratio to the adhesion amount converted from the reflection density sensor at the same location. Ask for.
The dot fluctuation or dust is detected by detecting the dot pattern image with an infrared light area sensor on the photosensitive member and with an area sensor having a wavelength corresponding to each color on the intermediate transfer member, and performing appropriate processing.
The offset amount (after fixing) is obtained by reading the corresponding locations on the recording paper and the fixing roller with optical sensors and comparing them.
An optical sensor is installed after the transfer process (on the PD and on the belt), and the transfer remaining amount is determined based on the amount of reflected light from the transfer residual pattern after the transfer of the specific pattern.
-Detect color unevenness during overlay with a full-color sensor that detects the recording paper after fixing.

(A-7) The characteristics, image density, and color of the formed toner image are optically detected (either reflected light or transmitted light may be used. The projection wavelength is selected depending on the color). In order to obtain density and single color information, it may be on a photoconductor or an intermediate transfer body, but it is necessary on paper to measure a color combination such as color unevenness.
For gradation, the reflection density of the toner image formed on the photosensitive member or the toner image transferred to the transfer member is detected by an optical sensor for each gradation level.
Sharpness is obtained by reading an image in which a line repetition pattern is developed or transferred using a monocular sensor having a small spot diameter or a high-resolution line sensor.
The graininess (roughness) is obtained by reading a halftone image and calculating a noise component by the same method as the sharpness detection.
The registration skew is obtained from the difference between the registration roller ON timing and the detection timing of both sensors by providing optical sensors at both ends in the main scanning direction after registration.
Color misregistration is detected by a monocular small-diameter spot sensor or high-resolution line sensor at the edge portion of the superimposed image on the intermediate transfer member or recording paper.
Banding (density unevenness in the feed direction) measures density unevenness in the sub-scanning direction with a small-diameter spot sensor or high-resolution line sensor on the recording paper, and measures the signal amount of a specific frequency.
Glossiness (unevenness) is provided so that a recording paper on which a uniform image is formed is detected by a regular reflection optical sensor.
・ Fog is a method of reading the image background with an optical sensor that detects a relatively wide area on the photoconductor, intermediate transfer member, or recording paper, or image information for each area of the background with a high-resolution area sensor. And the number of toner particles contained in the image is counted.

(A-8) The physical characteristics, image flow, and fading of the printed matter of the image forming apparatus are detected on the photosensitive member, the intermediate transfer member, or the recording paper by the area sensor, and the acquired image information is displayed. Judge by image processing.
・ Chile is obtained by taking an image on recording paper with a high-resolution line sensor or area sensor and calculating the amount of toner scattered around the pattern area.
The trailing edge blank and the solid cross blank are detected by a high resolution line sensor on the photosensitive member, the intermediate transfer member, or the recording paper.
・ Curls, undulations, and folds are detected by a displacement sensor. In order to detect a fold, it is effective to install a sensor near the both ends of the recording paper.
-For the dirt and scratches on the edge surface, the area sensor is used to capture and analyze the edge surface when paper discharge has accumulated to some extent by the area sensor installed vertically on the paper discharge tray.

(A-9) For detecting environmental conditions and temperature, the thermocouple method that extracts thermoelectromotive force generated at the contact point between dissimilar metals or between metal and semiconductor as a signal, the resistivity of metal or semiconductor changes with temperature Resistivity change element using this, pyroelectric element that generates a potential on the surface due to bias in the arrangement of charges in the crystal due to the rise in temperature in certain crystals, and magnetic characteristics depending on temperature A thermomagnetic effect element or the like that detects a change in temperature can be employed.
・ For humidity detection, there are optical measurement methods that measure the light absorption of H ₂ O or OH groups, humidity sensors that measure changes in the electrical resistance of materials due to the adsorption of water vapor, etc. Detection is performed by measuring a change in electric resistance of the oxide semiconductor accompanying gas adsorption.
Although there are optical measurement methods and the like for detection of airflow (direction, flow velocity, gas type), an air bridge type flow sensor that can be made smaller in consideration of mounting on a system is particularly useful.
・ Pressure-sensitive materials are used to detect atmospheric pressure and pressure, and mechanical displacement of the membrane is measured. A similar method is used for vibration detection.

(B) Control Parameter Information Since the operation of the image forming apparatus is determined by the control unit, it is effective to directly use the input / output parameters of the control unit.

(B-1) Image Forming Parameters Direct parameters output by the control unit through image processing for image formation include the following examples.
・ Setting values of process conditions by the control unit, such as charging potential, development bias value, fixing temperature setting value, etc. ・ Similarly, setting values of various image processing parameters such as halftone processing and color correction. Various parameters to be set, such as paper transport timing, execution time of preparation mode before image formation, etc.

(B-2) User operation history-Frequency of various operations selected by the user, such as the number of colors, number of sheets, and image quality instructions-Frequency of the paper size selected by the user

(B-3) Power consumption / Total power consumption or its distribution, change amount (differentiation), cumulative value (integration) for the whole period or specific period unit (1 day, 1 week, 1 month, etc.)

(B-4) Consumables consumption information-Total amount or specific period unit (1 day, 1 week, 1 month, etc.) toner, photoconductor, paper usage or distribution, change (differentiation), cumulative value ( Integration)

(B-5) Failure occurrence information ・ Frequency or distribution of failure occurrence (by type), change amount (differentiation), cumulative value (integration) of whole period or specific period unit (1 day, 1 week, 1 month, etc.)

(C) Input image information The following information can be acquired from image information sent directly as data from the host computer or image information obtained after image processing is performed by reading a document image from a scanner.
The cumulative number of colored pixels is obtained by counting image data for each GRB signal for each pixel.
The original image can be separated into characters, halftone dots, photographs, and backgrounds by the method described in, for example, Japanese Patent No. 2621879, and the ratio of the character part, halftone part, etc. can be obtained. Similarly, the ratio of color characters can be obtained.
The toner consumption distribution in the main scanning direction can be obtained by counting the cumulative value of the colored pixels for each area divided in the main scanning direction.
The image size is obtained from the distribution of colored pixels in the image size signal or image data generated by the control unit.
-Character type (size, font) is obtained from character attribute data.

The information acquisition unit 1e acquires the above-described copying machine status information once a day at a predetermined timing, calculates a difference or ratio between the initial value data and the acquired status information, and uses the acquired status information as general-purpose data. To. Thus, by using general-purpose data, it is possible to eliminate the difference in state information due to individual differences between parts, and the same algorithm can be applied. In this way, the state information that is the general-purpose data is stored in the storage means in a predetermined number m (m days) in time series. When new state information is acquired, the oldest state information among the state information stored in the storage unit 1c is discarded, and the acquired new state information is stored. In addition, the present invention is not limited to this, and status information may be acquired a plurality (n) times a day, and feature values such as an average value and a change amount may be stored in the storage unit as daily status information. Further, the status information acquired a plurality of times a day may be stored in the storage means as the status information for the day as it is. In this case, the time series data is m × n.
The copying machine status information stored in the storage means need not be stored as daily status information. For example, the status information is acquired every 4 hours, and m status information is stored in the storage means. You may make it memorize | store in a series.

  From the various information as described above, the above-described index value D is calculated, and based on the index value, the possibility of occurrence of an abnormality such as a failure is determined, and the occurrence of an abnormality such as a failure is predicted. Basically, if the magnitude of the index value D calculated from a plurality of types of information is larger than a predetermined threshold as described above, the possibility of occurrence of an abnormality is high. This threshold is generally determined by prior experiments.

  The index value D represents a scale by which the correlation between the incorporated information is deviated from the normal state. Since it is determined that the deviation from the normal state is larger as the index value is larger, it is possible to foresee the possibility of occurrence of abnormality even when the mechanism of abnormality is unknown.

  The index value is calculated using a multivariate analysis method such as a principal component analysis, an MTS method, or an MTA method. Hereinafter, calculation of an index value using the MTS method will be described as an example. Details of the MTS method are explained in detail in “Technology Development Publication Committee Chairman Genichi Taguchi, published by the Japanese Standards Association,” published by the Japanese Standards Association. is there. That is, first, set data composed of a plurality of types of information is acquired from an object to be examined in a normal state or the same specification object having the same specification. Then, a large number of sets of data are collected to construct an inverse matrix that becomes standard data (normal index information). Thereafter, when it is desired to examine the normality of the subject to be examined, set data is acquired from the subject to be examined. Then, the Mahalanobis distance indicating the relative positional relationship in the multi-dimensional space based on the standard data (normal index information) that has been pre-constructed is obtained for this set data, and the subject to be examined is based on the result. The amount of normality is measured.

  In this copying machine, an inverse matrix based on the MTS method is stored in the storage means as standard data. This inverse matrix predicts the state of the device from various kinds of feature information such as moving averages and change amounts calculated from a plurality of types of state information of the tester acquired in advance and a plurality (m) of state information. An item that is valid for the item is selected and constructed based on the selected item. That is, for certain types of state information, the moving average and the amount of change are used as items for calculating the inverse matrix, and for certain types of state information, only the moving average is used as the item for calculating the inverse matrix. It is used.

  The index value is calculated as follows. First, feature quantities such as a moving average and a change amount (differentiation) are extracted for each type of status information of the copying machine from the status information for a plurality of (m) days stored in the storage unit 1c. As the index value, the Mahalanobis distance as the index value is obtained based on the extracted feature quantity and the above-described inverse matrix.

  FIG. 5 is a graph showing a relationship between an index value (for example, Mahalanobis distance) and an elapsed time since the last maintenance. Immediately after maintenance, the index value does not show a very high value, but then gradually increases as the usage time of the apparatus increases. Then, after reaching a predetermined threshold value, it continues to rise, and eventually an abnormality occurs. Since the threshold is set to a value lower than the index value at the time of occurrence of an abnormality, the occurrence of abnormality is predicted in advance when the index value reaches the threshold, and the occurrence of the abnormality is predicted in advance to the user. I can tell you. Therefore, the control unit 1 of the copying machine transmits abnormality occurrence information when the index value reaches a threshold value.

Next, features of the present embodiment will be described.
FIG. 6 shows time-series fluctuation data of drive current of an optical sensor which is a sensor for detecting copier status information such as paper transport status. As shown in the figure, it can be seen that the drive current values are significantly different before and after the replacement or cleaning of the optical sensor. For this reason, when feature values are extracted using data of optical sensors before and after replacement or cleaning, the trend of trends in the state of parts and devices detected by optical sensors It cannot be captured correctly with features. As a result, the index value calculated based on the feature amount is not accurate, and the state of the image forming apparatus cannot be predicted correctly. Therefore, until the data for a plurality of (m) days are acquired, the index value cannot be calculated and the abnormal state cannot be predicted.
Therefore, in the present embodiment, a second abnormality prediction mode serving as a second abnormality prediction unit is provided, and while the data for a plurality of (m) days are acquired, the second abnormality prediction mode allows the image forming apparatus to Anomalies are predicted.

FIG. 7 is a control flow when switching to the second abnormality prediction mode.
As shown in the figure, first, the control unit checks whether maintenance such as replacement of parts or cleaning has been performed (S1). Whether or not the maintenance has been performed is detected, for example, by inputting the component information maintained by the service person to the image forming apparatus when the service person maintains the parts in the image forming apparatus. Also, a detection means for detecting whether or not there is a component in the apparatus is provided. For example, when the detection result changes from a state where there is no component to a certain state, it is determined that the component in the apparatus has been replaced. Then, when the maintenance of the part is performed (YES in S1), the state information in which the acquired data greatly varies due to the maintenance of the part is specified (S2). When maintenance is performed, a plurality of types of state information may be acquired, and state information that varies greatly may be searched by examining previous data and acquired data. Further, state information that fluctuates in advance with respect to the parts to be maintained is checked, and the parts and the state information are associated and stored in the storage means. When the parts are maintained, the parts information maintained from the storage means is updated. You may make it identify the status information which fluctuates based on it. Then, the state information that varies greatly depending on the specified component maintenance is deleted from the storage means. Next, it is checked whether or not the maintenance is cleaning (S3), and in the case of replacement (NO in S3), status information is acquired and initial value data is updated (S4). This is because the characteristic value is different for each part. On the other hand, in the case of cleaning (YES in S3), since the characteristic value does not change, the initial value is not updated. Next, the control unit changes from the normal abnormality prediction mode to the second abnormality prediction mode (S5).
On the other hand, when maintenance such as replacement of parts or cleaning is not performed (NO in S1), it is checked whether or not the second abnormality prediction mode is set (S6). If it is not the second abnormality prediction mode (NO in S6), the abnormality of the image forming apparatus is predicted in the normal abnormality prediction mode. On the other hand, in the case of the second abnormality prediction mode (YES in S6), it is checked whether or not there are all m (= integer) pieces of state information stored in the storage means (S7). If there is no plural (m) pieces of state information among any one of a plurality of types of state information (NO in S7), the normal abnormality prediction mode cannot be executed, so that the image forming apparatus remains in the second abnormality prediction mode. Predict abnormalities. On the other hand, when there are m (= integer) pieces of state information stored in the storage means (YES in S8), the normal abnormality prediction mode can be executed, so change to the normal abnormality prediction mode ( S8) An abnormality of the image forming apparatus is predicted in a normal abnormality prediction mode.

FIG. 8 is a control flow in a normal abnormality prediction mode.
When a predetermined time comes, such as when a predetermined time has elapsed or when a predetermined number of sheets have been printed, the normal or more prediction mode is started, and the state information is read from the storage means (S11). Is extracted (S12). Then, an index value (for example, Mahalanobis distance) is calculated from the extracted feature amount (S13) and compared with a threshold value (S14). If the index value is greater than or equal to the threshold value (YES in S14), the occurrence of an abnormality of the apparatus is predicted, so the part where the occurrence of the abnormality is predicted is identified (S15) and notified to the user (S16). If the index value is equal to or less than the threshold value (NO in S14), it is predicted that no abnormality of the apparatus will occur, so the index value is notified to the user as “predicted state value after ◯ day” (S17).

  Next, the second abnormality prediction mode will be described in detail based on Examples 1 to 3.

[Example 1]
The second abnormality prediction mode according to the first embodiment predicts an abnormality for each state information individually acquired from each sensor or the like while data for a plurality of (m) days is acquired. For example, the second abnormality prediction means predicts whether the value is abnormal from the value of the state information itself, such as predicting an abnormality when the output value of the encoder that detects the driving speed of the photoreceptor exceeds a certain threshold. Yes (hereinafter, anomaly detection type).
FIG. 9 is a control flow of the second abnormality prediction mode in the first embodiment.
First, when a predetermined time comes, such as when a predetermined time has elapsed or when a predetermined number of sheets have been printed, the control unit 1 checks whether the abnormality prediction mode is the normal abnormality prediction mode or the second abnormality prediction mode, and the second In the case of the abnormality prediction mode, the second abnormality prediction mode is started.
When the second abnormality prediction mode starts, first, the control unit 1 reads a plurality of types of latest state information stored in the storage unit (S21). Next, the control unit 1 reads a threshold value corresponding to each state information from the storage unit (S22), and checks whether or not the latest state information exceeds the threshold value (S23). When there is state information that exceeds the threshold (YES in S23), it is predicted that there is an abnormality in the apparatus, and the user is notified by displaying it on the display unit (S24). At this time, an abnormal component may be identified from the state information exceeding the threshold, and the predicted component information may be displayed on the display unit. On the other hand, if none of the state information exceeds the threshold value (NO in S23), since no abnormal state has occurred in the apparatus, the mode is ended as it is.

  As a result, due to maintenance such as replacement and cleaning of the parts, the state and characteristics of the parts greatly fluctuate before and after the maintenance of the parts. From the feature values extracted from the data before and after the parts maintenance, the state of the parts in time series Therefore, it is possible to simply predict an abnormality in the second abnormality prediction mode even while the abnormality cannot be predicted based on the feature amount of the state information. Therefore, even if an abnormality occurs in the device until the necessary state information is stored so that changes in the state of the components in time series can be accurately grasped from the extracted feature amount, an appropriate device Maintenance work can be performed. In addition, it is possible to immediately determine whether the maintenance work of the apparatus such as replacement or cleaning is appropriate.

  In the second abnormality prediction mode, whether or not the value is an abnormal value is predicted from the value of the state information itself. For example, an index value (for example, Mahalanobis distance) is calculated using only the current state information, May be determined. In this case, an inverse matrix (normal index information) constructed based on time-series data of a plurality of types of state information of the test machine acquired in advance by the test machine is stored in the storage means. Then, in the second abnormality prediction mode, the inverse constructed based on the time series data of the plurality of types of state information from the inverse matrix constructed based on the time series data of the feature amounts of the plurality of types of state information. Switch to the matrix. Then, a temporary index value (for example, Mahalanobis distance) is calculated from the latest plural types of state information stored in the storage means and an inverse matrix constructed based on the time series data of the plural types of state information. Then, abnormality determination is performed based on the provisional index value. Even in this case, it is appropriate even if an abnormality of the device occurs until the necessary state information is stored because it is possible to accurately grasp changes in the state of parts in time series from the extracted feature amount. It is possible to carry out maintenance work on various devices.

[Example 2]
Next, the second abnormality prediction mode of Example 2 will be described.
In the second embodiment, a plurality of (m) pieces of state information are acquired in a short time with respect to state information that has greatly changed before and after maintenance due to the maintenance of the parts, and the changed state information is acquired for a plurality of (m) days. Until this time, the normal abnormality prediction mode is provisionally operated with the dummy state information acquired in a short time.
FIG. 10 is a control flow of the control unit when the second abnormality prediction mode is executed in the second embodiment.
When the mode is changed to the second abnormality prediction mode, the information acquisition timing of the predetermined state information that has greatly changed before and after replacement or cleaning due to maintenance of parts is changed (S31). That is, in the normal time, the acquisition cycle is changed to be very short, for example, every 10 minutes, for example, at a timing acquired about once a day. When a plurality of (m) pieces of identified state information are acquired in time series in the storage unit (YES in S32), the state information acquisition timing is returned to the normal timing (S33). As a result, a plurality of (m) pieces of predetermined state information that greatly fluctuates before and after replacement or cleaning due to parts maintenance are stored in a short time, and a plurality of types of state information stored in the storage means are all stored in a plurality. (M) It can be in a certain state. Therefore, it can be operated in the normal abnormality prediction mode at an early stage. Note that the feature quantity extracted from multiple (m) pieces of state information acquired in a short time does not accurately capture changes in the state of parts, devices, etc., but the parts are subject to maintenance such as replacement or cleaning. Since the state is restored to the normal state, between the feature amount extracted from the state information correctly acquired for a plurality of (m) days and the feature amount extracted from the m pieces of state information temporarily acquired in a short time There is no big difference. Therefore, even if the index value is calculated using the feature amount extracted from the state information provisionally acquired in a short time, the prediction accuracy does not significantly decrease.
Then, the state information acquired at the normal acquisition timing is sequentially replaced with the oldest temporary state information, and after m days, all the temporary state information disappears, and the state information acquired at the normal acquisition timing Will be replaced.
As described above, since the abnormality of the apparatus can be predicted in the normal abnormality prediction mode in which the index value is calculated at an early stage, highly accurate prediction can be performed.

[Example 3]
Next, the second abnormality prediction mode of Example 3 will be described.
In the third embodiment, an index value is calculated by excluding predetermined state information that greatly fluctuates before and after maintenance such as replacement of parts or cleaning, and an abnormality of the apparatus is predicted.
FIG. 11 is a control flow of the control unit when the second abnormality prediction mode is executed in the third embodiment.
First, when the control unit 1 is changed to the second abnormal mode, the maintenance part is identified (S41), and the provisional arithmetic expression (for example, temporarily used when this part is maintained is determined from the identified maintenance part (for example, (Inverse matrix) is specified (S42) and read from the storage means. This provisional arithmetic expression (for example, inverse matrix) is a provisional arithmetic expression derived from normal set data excluding predetermined state information that varies greatly before and after maintenance due to the maintenance of this part.
When a predetermined timing comes, such as when a predetermined time has elapsed or when a predetermined number of sheets have been printed (YES in S43), specific state information excluding predetermined state information that varies greatly before and after maintenance due to maintenance of this part is stored. The information is read from the means (S44), and the feature amount of the state information is extracted (S45). Then, a provisional index value (for example, Mahalanobis distance) is calculated from the extracted feature quantity and provisional arithmetic expression (inverse matrix) (S46), and compared with a threshold (S47). If the provisional index value is greater than or equal to the threshold value (YES in S47), the occurrence of an abnormality of the apparatus is predicted, so the part where the occurrence of the abnormality is predicted is identified (S48) and notified to the user (S49). If the provisional index value is equal to or less than the threshold value (NO in S47), it is predicted that no abnormality will occur in the apparatus, so the provisional index value as shown in FIG. 12 is notified to the user (S51).
As a result, it is possible to predict an abnormality in a device or a part other than the part that has undergone maintenance such as replacement or cleaning.

  It should be noted that abnormality detection type abnormality determination for predicting whether or not there is an abnormality from the value of the state information shown in the first embodiment may always be performed. In the abnormality prediction mode that is normally performed, the feature value is extracted and the index value is calculated based on the time series change of each state information. Although it is high, the prediction accuracy is reduced for sudden occurrences of abnormalities. This is because feature amounts are extracted, and such sudden fluctuations are processed as noise and variations. On the other hand, since the abnormality detection type abnormality prediction predicts an abnormality with respect to the value acquired at that time, it is possible to satisfactorily detect an abnormality that occurs suddenly. Thus, by predicting the abnormality of the apparatus by using the abnormality detection type and the normal abnormality prediction mode in combination, both long-term abnormality prediction and sudden abnormality prediction can be performed. When such an abnormality determination is performed, as shown in FIG. 13, the result of the abnormality detection type prediction is displayed as “current state information”, and the result of the abnormality prediction is calculated as “ It is displayed as “status information”.

  In the above description, the image forming apparatus is provided with the abnormality determining device as abnormality determining means, and abnormality determination is performed inside the image forming apparatus. However, as shown in FIG. 14, the image forming apparatus and the abnormality determining device are provided. A maintenance system may be constructed using the information management apparatus. In this maintenance system, an information management apparatus is installed in, for example, a service company that maintains an image forming apparatus. The image forming apparatus acquires a plurality of types of status information at a predetermined timing by a status information acquisition unit configured by sensors of the image forming apparatus and transmits the acquired status information to the information management apparatus via a communication line. The information management apparatus stores a plurality of types of status information acquired from the image forming apparatus in a status information storage unit in time series. When the predetermined timing comes, the index value calculation unit extracts feature amounts for each state information, calculates the index value, and the state determination unit predicts the state of the image forming apparatus based on the index value. When a predetermined part needs to be replaced, the information processing unit of the information management apparatus notifies a serviceman that the part needs to be replaced. In addition, when a part is exchanged, the image forming apparatus identifies the part exchanged by the part information acquisition unit, and uses the communication line from the information processing unit to identify the exchanged part information. Send to. The information management device changes the abnormality prediction mode based on the received component information.

  Moreover, in the normal abnormality prediction mode, abnormality prediction using the MTS method is performed. However, the present invention is not limited to this, and in general, an abnormality due to pattern recognition that can cope with a long-term prediction than the method based on time series analysis. Any prediction method may be used.

  As described above, according to the present embodiment, as shown in the first embodiment, any part of the state information is maintained, and the state information changes before and after the part maintenance due to the part maintenance. When the change with time cannot be captured, the abnormality determination is performed in the second abnormality prediction mode as the second abnormality determination means. In the second abnormality determination mode, for example, if the output value of the encoder that detects the driving speed of the photosensitive member exceeds a certain threshold value, the abnormality is predicted based on only the value of the current state information, such as predicting an abnormality and notifying the user. It is to predict whether or not. Therefore, in the second abnormality prediction mode, an abnormality can be predicted even if the storage means does not have a plurality of state information. As a result, while a plurality of data is stored in the storage unit, it is possible to tentatively predict an abnormality of the apparatus by making a prediction in the second abnormality prediction mode. Therefore, an abnormality of the apparatus can be predicted even when a plurality of data is stored in the storage means, and even if an abnormality of the apparatus occurs while a plurality of data is stored in the storage means, it can be appropriately dealt with. Further, if the maintained work is not correct, an abnormality is predicted while a plurality of data is stored in the storage means, so it is possible to grasp whether the maintained work has been correctly performed.

  In addition, as shown in the second embodiment, when a part is maintained, at least the state information acquisition timing at which the value largely fluctuates before and after the maintenance due to the maintenance of the part is extremely shortened, and a plurality of states are obtained in a short time. I try to get information. Thereby, a plurality of state information is stored in the storage means, and a temporary feature amount can be extracted. Therefore, the normal abnormality prediction mode can be provisionally operated at an early stage, and a highly accurate abnormality prediction can be performed at an early stage.

  Further, as shown in the third embodiment, when a part is maintained, an index value is calculated by excluding the state information until a plurality of pieces of state information whose values greatly fluctuate before and after the maintenance are accumulated. , Provisional abnormality prediction is implemented. Thus, it is possible to accurately predict an abnormality in a device or a part other than a part that has undergone maintenance such as replacement or cleaning until a plurality of state information whose values greatly fluctuate before and after maintenance.

  Also, abnormality prediction that predicts whether or not an abnormal value is based only on the current state information is performed in parallel with normal abnormality prediction that extracts feature values from a plurality of state information and calculates an index value based on the feature values. You may do it. Thereby, it is possible to perform both of an abnormality that occurs suddenly and a long-term abnormality.

  Further, the image forming apparatus according to the present embodiment may be provided with a notification unit that displays a prediction result of the occurrence of abnormality predicted from the time change data of the index value D on an operation display unit such as a display to notify the user. . By displaying the prediction result of the occurrence of the abnormality in this way, the user can know the predicted occurrence of the abnormality without determining the state change of the image forming apparatus from the time change of the index value D. Maintenance before occurrence becomes possible. Also, in the image forming apparatus of the present embodiment, an abnormality occurrence prediction result predicted from the time change data of the index value D is transmitted to an external device via a communication line such as a dedicated line, a public line, the Internet, or a local area network. You may provide the communication means to transmit to. In this case, the occurrence of abnormality in a plurality of image forming apparatuses can be predicted centrally at a monitoring center or the like.

  Moreover, you may provide the alerting | reporting means which identifies the components which generate | occur | produce abnormality, displays on operation display parts, such as a display, and alert | reports to a user. Thereby, since the predicted abnormality occurrence part can be known, the apparatus can be more reliably maintained.

  In addition, an index value (for example, Mahalanobis distance D) is calculated using an inverse matrix A of a group of data sets made up of a plurality of state information acquired from a normal state device, which is normal index information, and used for abnormality determination. In such a configuration, unlike a conventional abnormality determination device that determines an abnormality simply by comparing a threshold value with acquired data, it is possible to predict the occurrence of an abnormality whose cause is not clearly specified using the MTS method.

1 is a schematic configuration diagram illustrating the copier that is an image forming apparatus that can be a test target of an abnormality determination apparatus to which the present invention is applied. FIG. 2 is a schematic configuration diagram illustrating a printer unit of the copier. The elements on larger scale which show the tandem part of the copier. FIG. 2 is a block diagram showing a part of an electric circuit of the copier. 6 is a graph showing an example of a relationship between an index value D for a general abnormality in the copier and an elapsed time (operation time). 6 is a graph showing an example of the relationship between drive current and elapsed time of an optical sensor that is a sensor that detects status information of the copier. The control flow figure when switching to the 2nd abnormality prediction mode is performed. The control flowchart of a normal abnormality prediction mode. The control flowchart of the 2nd abnormality prediction mode of Example 1. FIG. FIG. 10 is a control flow diagram of a second abnormality prediction mode according to the second embodiment. The control flowchart of the 2nd abnormality prediction mode of Example 3. FIG. FIG. 10 is a diagram illustrating an example displayed on the operation display unit of the copier after execution of the second abnormality prediction mode according to the third embodiment. The figure which shows an example displayed on the operation display part, after performing two abnormality prediction modes in parallel. 2 is a diagram showing an example of a maintenance system that performs maintenance of the copier. FIG. The figure which shows an example of the abnormality prediction by a time series analysis.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Control part 2 Various sensors 3 Operation display part 100 Copying machine

Claims (10)

In the abnormality prediction device that predicts the abnormal state of the image forming apparatus ,
Information acquisition means for acquiring a plurality of different types of status information about the image forming apparatus at a predetermined timing ;
Storage means for storing the plurality of types of state information in time series ;
For each type of status information, based on a plurality of types of status information acquired from the time when the latest plurality of types of status information stored in the storage unit are acquired to a time point that is traced back a predetermined number of times. A feature amount extraction means for extracting feature amounts that capture changes over time in state information;
Index value calculation means for calculating an index value indicating the state of the image forming apparatus based on feature quantities of the plurality of types of state information;
First abnormality prediction means for predicting the occurrence of an abnormality in the image forming apparatus based on the index value;
Second abnormality prediction means for predicting the occurrence of an abnormality of the image forming apparatus based on the latest state information stored in the storage means,
When the acquisition unit acquires the plurality of types of status information for a predetermined number of times after performing maintenance on the image forming apparatus , the first abnormality prediction unit predicts the occurrence of an abnormality in the image forming apparatus , Between the maintenance of the image forming apparatus and the acquisition unit acquiring the plurality of types of state information a predetermined number of times, the second abnormality prediction unit predicts the occurrence of an abnormality in the image forming apparatus. An abnormality prediction apparatus characterized by comprising. In the abnormality prediction device according to claim 1,
When predicting the occurrence of abnormality in the image forming apparatus in the first abnormality prediction means, abnormality predicting apparatus characterized by being configured to predict the abnormality of the image forming apparatus in the second abnormality prediction means. In the abnormality prediction device that predicts the abnormal state of the image forming apparatus ,
Information acquisition means for acquiring a plurality of different types of status information about the image forming apparatus at a predetermined timing ;
Storage means for storing the plurality of types of state information in time series ;
Based on the plurality of types of state information acquired from the time when the latest plurality of types of state information stored in the storage means are acquired to a time point that is a predetermined number of acquisitions, the type of the previous state information A feature amount extraction means for extracting a feature amount that captures a change in state information over time,
Index value calculation means for calculating an index value indicating the state of the image forming apparatus based on feature quantities of the plurality of types of state information;
An abnormality prediction means for predicting the occurrence of an abnormality in the image forming apparatus based on the index value,
The timing for acquiring the plurality of types of status information of the acquisition unit is advanced from when maintenance is performed on the image forming apparatus until the acquisition unit acquires the plurality of types of status information a predetermined number of times. An abnormality determination device characterized by the above. In the abnormality prediction device that predicts the abnormal state of the image forming apparatus ,
Information acquisition means for acquiring a plurality of different types of status information about the image forming apparatus at a predetermined timing ;
Storage means for storing the plurality of types of state information in time series ;
The type of the state information based on the plurality of types of state information acquired from the time when the latest plurality of types of state information stored in the storage unit are acquired to a time point that is traced back by the predetermined number of acquisition times. A feature amount extraction means for extracting a feature amount that captures a change in state information over time,
Index value calculation means for calculating an index value indicating the state of the image forming apparatus based on feature quantities of the plurality of types of state information;
An abnormality prediction means for predicting the occurrence of an abnormality in the image forming apparatus based on the index value,
State after performing maintenance for the image forming apparatus, until the acquisition unit acquires a predetermined number of times the plurality of types of state information, the index value calculating means, regarding site maintenance is performed of the image forming apparatus The provisional index value is calculated based on the feature amounts of the plurality of types of state information excluding the feature amount extracted based on the information , and the abnormality prediction unit is configured to detect an abnormality of the image forming apparatus based on the provisional index value . An abnormality prediction apparatus characterized by being configured to predict occurrence . In the abnormality prediction device according to claim 3 or 4,
An abnormality prediction apparatus comprising: a second abnormality determination unit that determines an abnormality of the image forming apparatus based on the latest plurality of types of state information stored in the storage unit. In the abnormality prediction device according to any one of claims 1 to 5,
An abnormality determination apparatus comprising an informing means for informing a result of abnormality prediction of the image forming apparatus. The abnormality prediction device according to claim 6,
The notification means, wherein when the abnormality of the image forming apparatus Ru is predicted, the abnormality determination apparatus characterized by notifying the maintenance is necessary portion of the image forming apparatus. In the abnormality prediction apparatus in any one of Claims 1 thru | or 7,
The abnormality prediction apparatus, wherein the index value calculation means calculates an index value using normal index information that is an index of a normal state of the image forming apparatus. A recording medium conveying means for conveying the recording medium, a visible image forming means for forming a visible image on the recording medium conveyed by the recording medium conveying means, and an abnormality predicting means for predicting an abnormality of the entire apparatus or a part thereof. In an image forming apparatus comprising:
9. An image forming apparatus according to claim 1, wherein the abnormality prediction means is any one of claims 1 to 8. In a maintenance system that includes an abnormality prediction unit that predicts an abnormality of the image forming apparatus , and that performs maintenance of the image forming apparatus based on a determination result of the abnormality prediction unit,
A maintenance system according to any one of claims 1 to 8, wherein said abnormality prediction means is used.

Images