JP2012147049A - Failure prediction apparatus, image forming apparatus and program - Google Patents

Abstract

The present invention relates to failure prediction of an image forming apparatus, and accurately detects a change in an apparatus state that is a sign of failure, thereby increasing the accuracy of failure prediction.
In the image forming apparatus of the present example, based on image forming parameters that are items to be adjusted for each page (that is, for each image for each page) in order to maintain an appropriate image quality in the execution of the image forming process. The fluctuation tendency of the image formation parameter is calculated, and the image formation parameter at the target time point is compared with the fluctuation tendency with respect to the image formation parameter at the target time point when the image forming process is executed. On the other hand, it is determined whether an inflection point is reached, and it is predicted that a failure related to image formation will occur in the apparatus after the inflection point.
[Selection] Figure 1

Description

  The present invention relates to a failure prediction apparatus, an image forming apparatus, and a program.

  In an image forming apparatus having an image forming function for forming and outputting an image on a recording material such as paper, the use of the image forming function when a failure such as a paper jam or a transfer failure that impedes the operation of the image forming function occurs. Will be restricted, causing inconvenience to the user. Therefore, by predicting the occurrence of such a failure, it is possible to perform necessary measures such as replacement or repair of the relevant part immediately before the occurrence of the failure or after the occurrence of the failure. It is desired to reduce the time during which usage is limited.

So far, various inventions have been proposed for predicting a failure by monitoring the state of a device to be monitored, and some examples will be shown below.
For example, process values input from time to time from a plant are converted into time-series process data and stored in sequence, the predetermined inspection section is divided into sections 1 and 2, and the process data is statistically processed. An invention has been proposed in which a section division point at which the variance of the process value with respect to the first-order approximation equation is determined to be a point at which the state of the plant has changed has been proposed (see Patent Document 1).

  For example, in a control system such as an air conditioning system, control amounts and manipulated variables are collected sequentially and accumulated as time-series data. These time-series data are processed to generate processed data, and control amount processed data and manipulated variable are processed. An invention has been proposed in which correlation values are sequentially calculated for each combination of data, and the presence or absence of abnormality is determined based on equivalence regarding a series of correlation values calculated for each combination (see Patent Document 2).

JP 2002-278621 A JP 2003-208219 A

  The present invention relates to a failure prediction of an image forming apparatus, and an object thereof is to propose a technique for accurately detecting a change in an apparatus state that is a precursor of a failure and increasing the accuracy of failure prediction.

  The present invention described in claim 1 is a collection unit that collects image formation parameters in execution of image formation processing of an image forming apparatus that is a target of failure prediction, a storage unit that accumulates the collected image formation parameters, Extraction means for extracting an inflection point with respect to the tendency of the image forming parameter based on the tendency of fluctuation of the accumulated image forming parameter, and the failure based on the case of the failure of the image forming apparatus that has occurred in the past For the image forming parameter in which an inflection point has occurred before the occurrence of the above, based on the reference creation means for extracting a feature at the inflection point and creating a failure prediction reference, the variation tendency and the reference, A failure prediction apparatus comprising: prediction means for predicting a failure that occurs in the image forming apparatus that is a target of failure prediction after an inflection point.

  According to a second aspect of the present invention, in the first aspect of the present invention, the extracting means includes a direction of a convex shape of a curve that approximates the tendency in a time-series interval before a certain time point, and the or Compare the direction of the convex shape of the curve approximating the tendency in the time series interval including the time point when the convex shape direction is different, and extract the value of the image formation parameter at the certain time point as the inflection point This is a failure prediction apparatus characterized by the above.

  According to a third aspect of the present invention, in the present invention according to the first or second aspect, the present invention further comprises storage means for storing a parameter group in which a plurality of image forming parameters are combined and a failure type in association with each other. The means predicts that when an inflection point is extracted for all image forming parameters related to the parameter group, a failure of the type associated with the parameter group occurs. It is.

  According to a fourth aspect of the present invention, there is provided image forming means for forming an image on a recording material, collecting means for collecting image forming parameters in execution of image forming processing by the image forming means, and the collected image forming parameters. Storage means for accumulating, extraction means for extracting an inflection point with respect to the tendency of the image forming parameter based on the tendency of fluctuation of the accumulated image forming parameter, and other image forming apparatuses that have occurred in the past A reference acquisition means for acquiring a failure prediction reference, which is created in advance by extracting a feature at the inflection point for an image formation parameter in which the inflection point has occurred before the occurrence of the failure based on the case of the failure; An image forming apparatus comprising: prediction means for predicting a failure occurring in the apparatus after the inflection point on the basis of the variation tendency and the reference. .

  According to a fifth aspect of the present invention, in a computer, a collecting function for collecting image forming parameters in execution of an image forming process of an image forming apparatus that is a target of failure prediction, and a storage for storing the collected image forming parameters. Function, an extraction function for extracting an inflection point with respect to the tendency of the image forming parameter based on the tendency of fluctuation of the accumulated image forming parameter, and a case of a failure of another image forming apparatus that has occurred in the past. Based on the image acquisition parameter in which an inflection point has occurred before the occurrence of the failure, a reference acquisition function for acquiring a failure prediction reference, which is created in advance by extracting features at the inflection point, A prediction function for predicting a failure occurring in the image forming apparatus that is the target of the failure prediction after the inflection point based on the tendency and the reference. A gram.

  According to the first, fourth, and fifth aspects of the present invention, it is possible to accurately detect a change in the apparatus state that is a precursor of a failure, and to improve the accuracy of failure prediction.

  According to the second aspect of the present invention, it is possible to easily detect a change in the apparatus state that is a sign of failure.

  According to the third aspect of the present invention, the accuracy of failure prediction can be increased.

FIG. 2 is a diagram illustrating functional blocks of an image forming apparatus according to an embodiment of the invention. It is a figure which illustrates transition of the section used as the calculation object of the regression curve concerning the time series data of the image formation parameter concerning one embodiment of the present invention. It is a figure which illustrates the extraction flow of the inflection point which concerns on the time series data of the image formation parameter which concerns on one Embodiment of this invention. It is a figure which illustrates the time series data of the image formation parameter which concerns on one Embodiment of this invention, and transition of the secondary term coefficient of the regression curve group which concerns on the said time series data. It is a figure which illustrates the regression curve group which concerns on the time series data of the image formation parameter which concerns on one Embodiment of this invention. It is a figure which illustrates the correspondence table of the combination of the image formation parameter which concerns on one Embodiment of this invention, and the classification of a failure. It is a figure which illustrates the correspondence table of the combination of the image formation parameter which concerns on one Embodiment of this invention, and the classification of a failure. It is a figure which illustrates the relationship between the image formation parameter which concerns on one Embodiment of this invention, and the classification of failure. It is a figure which illustrates the functional block which concerns on the extended example of the failure prediction determination part which concerns on one Embodiment of this invention. It is a figure explaining the determination process by the probability inference model determination part which concerns on one Embodiment of this invention. It is a figure explaining the determination process by the fault tree determination part which concerns on one Embodiment of this invention. It is a figure explaining the selection process by the determination selection part which concerns on one Embodiment of this invention. 1 is a diagram illustrating the structure of an image forming apparatus according to an embodiment of the invention. 1 is a diagram illustrating a hardware configuration of a computer according to an image forming apparatus according to an embodiment of the present disclosure.

An embodiment of the present invention will be described with reference to the drawings.
First, an image forming apparatus according to an embodiment of the present invention will be described.
The image forming apparatus is an apparatus having an image forming function for forming and outputting an image on a recording material such as paper. Examples of the image forming apparatus include a printer (printing apparatus), a copier (copying apparatus), a facsimile machine, and the like, as well as a multi-function machine having multiple functions such as printing, copying, and facsimile.

FIG. 13 illustrates the structure of the image forming unit in the image forming apparatus of this example.
The image forming apparatus of this example is an intermediate transfer method generally called a tandem type, and as a representative functional unit, a plurality of image forming units 1Y, 1M, 1C, and the like that form toner images of respective color components by an electrophotographic method. 1K, a primary transfer unit 10 that sequentially transfers (primary transfer) each color component toner image formed by each of the image forming units 1Y, 1M, 1C, and 1K to the intermediate transfer belt 15, and is transferred onto the intermediate transfer belt 15. A secondary transfer unit 20 that collectively transfers (secondary transfer) the superimposed toner image onto a sheet P (an example of a recording material), and a fixing unit 34 that fixes the secondary transferred image onto the sheet P are provided. .
Further, the image forming apparatus of this example includes a control unit 40 that controls the operation of each unit, and a user interface (UI) 41 for receiving information presented to the user and instructions from the user.

In this example, each of the image forming units 1Y, 1M, 1C, and 1K includes a photosensitive drum 11 (11Y, 11M, 11C, and 11K) that rotates in the arrow A direction. Further, around each of the photosensitive drums 11, a charger 12 that charges the photosensitive drum 11, an exposure unit 13 that writes an exposure beam Bm on the photosensitive drum 11 to write an electrostatic latent image, and each color component toner And a developing device 14 that visualizes the electrostatic latent image on the photosensitive drum 11 with toner, and each color component toner image formed on the photosensitive drum 11 is transferred to the intermediate transfer belt 15 by the primary transfer unit 10. Various types of electrophotographic devices such as a primary transfer roll 16 for transferring to the drum and a drum cleaner 17 for removing residual toner on the photosensitive drum 11 are sequentially arranged.
These image forming units 1Y, 1M, 1C, and 1K are arranged substantially linearly in the order of yellow (Y), magenta (M), cyan (C), and black (K) from the upstream side of the intermediate transfer belt 15. The intermediate transfer belt 15 can be contacted and separated.

  Further, the image forming apparatus of the present example includes, as a paper transport system, a paper feed mechanism unit 31 that performs a paper feed operation of taking out the paper P from the paper storage unit and feeding it to the secondary transfer unit 20, and the secondary transfer unit 20. A conveyance belt 32 that conveys the passed paper P to the fixing device 34 side, a fixing inlet guide 33 that guides the paper P to the inlet of the fixing device 34, and a discharge that guides the paper P discharged from the fixing device 34. A paper guide 35 and a paper discharge roll 36 for discharging the paper P guided by the paper discharge guide 58 to the outside of the apparatus are provided.

  That is, the paper P fed from the paper storage unit to the secondary transfer unit 20 by the paper feed mechanism unit 31 is electrostatically transferred to the toner image on the intermediate transfer belt 15 by the secondary transfer unit 20. The paper is transported to the transport belt 32 while being peeled from the intermediate transfer belt 15. Then, the toner is conveyed by the conveyance belt 32 to the fixing device 34 through the fixing inlet guide 33 in accordance with the operation speed of the fixing device 34. The unfixed toner image on the paper P conveyed to the fixing device 34 is fixed on the paper P by receiving a fixing process in which heat and pressure are applied by the fixing device 34. Thereafter, the paper P on which the fixed image is formed is conveyed to a paper discharge container (not shown) provided outside the apparatus via a paper discharge guide 35 and a paper discharge roll 36.

  Here, the image forming apparatus according to the present example is based on the image forming parameter that is an item to be adjusted for each page (that is, for each image for each page) in order to maintain an appropriate image quality in the execution of the image forming process. The fluctuation tendency of the image formation parameter is calculated, and the image formation parameter at the target time point is compared with the fluctuation tendency with respect to the image formation parameter at the target time point when the image forming process is executed. It also operates as a failure prediction device that determines whether or not an inflection point is detected and predicts that a failure related to image formation will occur in the device after the inflection point.

  That is, the image forming parameters are sequentially determined according to the characteristics of the image forming process (the type of paper used as the recording material, the coverage (the image forming density with respect to the recording material), the number of images formed, etc.) and the presence or absence of disturbance at that time. For example, in the conventional method in which failure prediction is performed from the viewpoint of whether the image formation parameters are within the normal range, for example, the variation in the image formation parameters is due to the characteristics of image formation processing, disturbances, etc. However, in the image forming apparatus of this example, a new technique for effectively detecting fluctuations in image forming parameters caused by the state changes of the components related to image formation is used. The accuracy of failure prediction is improved.

  As a method for determining whether the image formation parameter at the target time point is an inflection point with respect to the change tendency, for example, the target time point with respect to an Nth order (N ≧ 3) regression curve that approximates the change tendency is used. A method of calculating the distance (degree of divergence) of the image forming parameter and determining that the image forming parameter at the target time point is an inflection point when the distance is equal to or greater than a predetermined threshold, or a time before the target time point Compare the direction of each convex shape (second-order coefficient) between the secondary regression curve that approximates the fluctuation trend in the series section and the secondary regression curve that approximates the fluctuation trend in the time series section including the target time point. For example, there is a method of determining that the image forming parameter at the target time point is an inflection point when they are different. Below, the case where an inflection point is discriminate | determined by the latter method is demonstrated to an example.

  In addition, the target time point may be a time point when a new image forming process is executed, or may be any time point between the time when failure prediction was performed in the past and the current time point. The occurrence of a failure after the target time point related to the inflection point can be predicted.

FIG. 1 illustrates functional units related to failure prediction in the image forming apparatus of this example.
The image forming apparatus of the present example includes a parameter collection unit 51, a parameter statistical processing unit 52, a time series data storage unit 53, a sequential interval linear regression calculation unit 54, an inflection point extraction as functional units related to failure prediction as described above. Unit 55, failure prediction determination unit 56, failure prediction determination information update unit 57, and the like.

  Here, when the image forming apparatus of this example receives an execution instruction for image forming processing related to one page or a plurality of pages, the image forming apparatus performs page-by-page (ie, page-by-page image) in order to maintain the quality of the image formed on the sheet. The image formation parameters are adjusted to perform image formation, and the values of each image formation parameter are measured at any time during the image formation operation by the sensors provided at each part controlled based on the image formation parameters. I can do it. Hereinafter, a series of image forming processes (image forming processes in units of execution instructions) performed based on execution instructions are referred to as jobs.

The parameter collection unit 51 collects image formation parameters adjusted during job execution for each page together with the measurement time. The collection is performed collectively for each job.
Examples of the image forming parameters to be collected include a photosensitive member potential (a potential of the photosensitive drum 11), a photosensitive member charging current (a charging current of the photosensitive drum 11), a semiconductor laser light amount (a laser light amount of the exposure device 13), and a developing device. Density (density of toner image formed by developing unit 14), primary transfer current (transfer current in primary transfer unit 10), secondary transfer current (transfer current in secondary transfer unit 20), fixing unit heat roll temperature (fixing) Temperature of the heat roll that heats the container 34), procompat density (density of the patch image used for process control such as toner image alignment and density control), and the like. In this example, an actual measurement value measured at a part corresponding to the image formation parameter is used as the image formation parameter. For example, a target value for controlling each part or a difference between the actual measurement value and the target value is used. Other types of values may be used. Further, for example, environmental information such as temperature and humidity inside the apparatus may be added to the monitoring item and collected for each page together with the image forming parameter group.

  The parameter statistical processing unit 52 performs statistical processing based on the measured values of the image formation parameters collected for each page by the parameter collection unit 51 for each job, and calculates a statistical value for each image formation parameter for each job. As the statistical processing, for example, an average of measurement values for each image forming parameter, calculation of a standard deviation, or the like is performed.

  The time-series data storage unit 53 adds the statistical value for each image formation parameter calculated by the parameter statistical processing unit 52 with date / time information (for example, job start time and end time) about the job. Store (accumulate) as time-series data, arranged in sequence.

The sequential interval linear regression calculation unit 54 performs time-series data of image forming parameters stored in the time-series data storage unit 53 at a predetermined timing (for example, at the end of a certain number of jobs, every predetermined time, etc.). (Statistics and date / time information of image forming parameters) are read out, and for each image forming parameter, a curve approximating a transition of statistical values in a time-series direction section including statistical values of a predetermined number of data, one data at a time Calculation is performed for each moved section. In this example, the regression curve group is calculated by performing linear regression of quadratic regression for each section. Note that, as in this example, instead of a configuration in which a regression curve is calculated for each section moved by one data unit, a regression curve may be calculated for each section moved by two or more data units.
FIG. 2 exemplifies the time series data of the image formation parameter “ACGridVoltage3” and the transition of the section that is the calculation target of the regression curve related to the time series data, and the calculation target section is moved by one data unit. It shows how to make it conceptual.

The inflection point extraction unit 55 extracts (detects) inflection points related to the time series data of the image forming parameters based on the regression curve group calculated by the sequential interval linear regression calculation unit 54.
FIG. 3 illustrates an inflection point extraction flow related to time-series data of image forming parameters. In this example, inflection point extraction processing is performed for each image forming parameter as follows. First, a coefficient of a quadratic term (hereinafter referred to as a quadratic term coefficient) of each regression curve is extracted (step S11), and a change in the polarity (sign plus / minus) of the quadratic term coefficient is detected (step S12), 2 When a change in the polarity of the second-order coefficient is detected, the time when the polarity of the second-order coefficient changes is output as an inflection point (step S13).
That is, in this example, the inflection point is detected while switching each time point between the time point when the failure prediction was performed in the past and the current time point as the target time point.

In this example, for example, as illustrated in FIG. 5, when curved lines L1 and L2 that are convex downward and curved lines L3 and L4 that are convex upward are obtained in chronological order, the convex shape is downward. Since the second-order term coefficients of the curves L1 and L2 are negative and the convex L3 and L42-order terms are positive in the convex shape, it is specified that the polarities of the second-order terms have changed in the curves L2 and L3. The time point is detected as an inflection point.
In this way, in this example, the regression curve group is a quadratic curve, so the regression curve calculated while moving the section is either an upward convex curve or a downward convex curve when the trend of the original data changes. The inflection point extraction unit 55 identifies the time when the polarity of the quadratic coefficient of each regression curve has changed (the time when the sign has changed from positive to negative or from negative to positive) and detects it as an inflection point. To do.

Here, the inflection points relating to the time series data of the image forming parameters will be described with reference to FIG.
FIG. 4A illustrates time-series data for “ACGridVoltage3” which is an example of an image forming parameter. As described with reference to FIG. 2, the regression curve group is calculated for the time series data shown in FIG. 2 while moving the section in units of one data, and the quadratic coefficient of each regression curve is obtained. As a result, as illustrated in FIG. 4B, it is possible to obtain the transition of the quadratic coefficient of the regression curve group related to the time series data of the image forming parameters.
In the example of FIG. 4B, among the inflection points at which the switching of the sign of the secondary term coefficient has occurred, the sign of the secondary term coefficient is switched from negative to positive, and the fluctuation amount of the secondary term coefficient is a threshold value within a certain period. (In this example, 0.1) or more, an inflection point is marked with a circle, and a failure of image quality trouble occurs several days after the inflection point.

  Therefore, in the failure prediction apparatus that operates on the image forming apparatus of the present example, the image forming parameter in which an inflection point occurs before the occurrence of the failure for each type of failure based on the case of the failure that occurred in the past. The characteristics at the inflection point (for example, how to switch the sign of the quadratic coefficient (the direction of the convex shape of the regression curve), the amount of fluctuation of the quadratic coefficient at the inflection point, etc.) are extracted in advance as a failure prediction criterion. By preparing, as will be described later, when an inflection point having the same tendency as the feature is detected, a type of failure corresponding to the image formation parameter related to the inflection point occurs. Predict that.

  When the inflection point related to the time series data of the image formation parameters is extracted by the inflection point extraction unit 55, the failure prediction determination unit 56 prepares the types of image formation parameters (and A prediction process for the type of failure corresponding to the image formation parameter related to the inflection point, with reference to the correspondence table between the combination) and the type of failure targeted for failure prediction based on the time series data, and the prediction result Is output to a display or the like of the image forming apparatus to notify the user. Instead of outputting the prediction result, a display indicating that an inflection point has been detected or a graph as illustrated in FIG. 4B may be output to a display or the like.

The failure prediction determination unit 56 of this example performs failure prediction processing using the correspondence table illustrated in FIG. 6 as an example.
FIG. 6A illustrates a correspondence table between parameter groups that are combinations of image forming parameters and failure types. In this example, when an inflection point is detected for time series data of all image forming parameters related to a certain parameter group, the probability that a type of failure corresponding to the parameter group will occur is set. The probability in the correspondence table is calculated in advance based on past cases.
FIG. 6B illustrates a combination of image forming parameters related to the parameter group in the correspondence table of FIG. In this example, a plurality of image forming parameters related to the part are grouped as a parameter group in a unit of the part related to image formation.
For example, when an inflection point is detected for the time series data of all the image forming parameters related to the parameter group 1, based on the occurrence probability of the failure A and the failure C associated with the parameter group 1, the failure A Will occur with a probability of 70% and failure C is predicted to occur with a probability of 10%. Further, in this state, when an inflection point is further detected for the time series data of all the image forming parameters related to the parameter group 3, the occurrence probability of the failure A associated with the parameter group 3 is added. The failure A is predicted to occur with a probability of 90%. In addition, if the period from the detection of an inflection point to the occurrence of a failure is set based on past cases, the time when the failure occurs can also be predicted.

In addition, the failure prediction determination unit 56 of this example performs failure prediction processing using the correspondence table illustrated in FIG. 7 as another example.
FIG. 7A illustrates a correspondence table between parameter groups that are combinations of image forming parameters and failure types. In this example, when an inflection point is detected for time series data of all image forming parameters related to a certain parameter group, the probability that a type of failure corresponding to the parameter group will occur is set. The probability in the correspondence table is calculated in advance based on past cases.
FIG. 7B illustrates a combination of image forming parameters related to the parameter group in the correspondence table of FIG. In this example, a plurality of image forming parameters related to the failure are collected as a parameter group in units of failure types.
For example, when an inflection point is detected for the time series data of all image forming parameters related to the parameter group 5, the failure A is 90% based on the occurrence probability of the failure A associated with the parameter group 5. It is expected to occur with a probability of. In addition, if the period from the detection of an inflection point to the occurrence of a failure is set based on past cases, the time when the failure occurs can also be predicted.

The correspondence tables illustrated in FIGS. 6 and 7 are generated based on the relationship between the image forming parameters illustrated in FIG. 8 and the failure types.
In FIG. 8, “Up” indicates that the value of the image forming parameter is increasing, and “Down” indicates that the value of the image forming parameter is decreasing. For example, when an inflection point of “photoconductor potential parameter 1” is detected, the occurrence of failure A is predicted if the measured value tends to increase, and the occurrence of failure B is predicted if the measured value tends to decrease. Is set. From another point of view, for example, when a failure C occurs, an inflection point of “photoreceptor potential parameter 2” is detected and the measured value tends to increase, and “toner density parameter 1”. It is set that the inflection point is detected and the measured value tends to decrease.

  The failure prediction determination information update unit 57 acquires (receives) the latest version of the correspondence table generated based on the latest failure case from an external database or update server, and uses the failure prediction determination unit 56 Update the table. The correspondence table is updated at a predetermined timing (for example, every week). In the case of a configuration that performs threshold determination, probability inference model determination, fault tree determination, and the like, which will be described later, information used for these determinations is also updated. As a result, failure prediction based on the latest failure case can be performed, and the accuracy of failure prediction is improved.

Next, an extended example of the failure prediction determination unit 56 will be described.
FIG. 9 illustrates functional blocks according to an extension example of the failure prediction determination unit 56.
The failure prediction determination means 56 of this example is a plurality of determination units 61 to 63 as means for determining whether or not the detected inflection point leads to failure, and selection of determination results by the determination units 61 to 63 And a determination selection unit 64 that performs processing.

The threshold value determination unit 61 sets a threshold value for the variation amount of the quadratic term coefficient at the inflection point for the quadratic term coefficient of the regression curve, and determines that the inflection point leads to failure when the threshold value is exceeded.
The probabilistic inference model determination unit 62 uses a failure model built in advance in a probability network (for example, a Bayesian network), and inputs probabilistic inference by inputting detailed device information such as measured values of image formation parameters at inflection points. Determine whether the point is an inflection point leading to a failure.
The fault tree determination unit 63 uses a model in which causal relationships such as failure locations and device states are preliminarily constructed in a tree shape, and determines whether the inflection point leads to a failure by tracing the tree.
In addition, these determination parts 61-63 are examples, and you may provide the determination part which performs determination by another method.

The determination selection unit 64 selects only one of the determination results by the determination units 61 to 63, or determines the final determination result by majority decision of each determination result. For example, instead of the majority decision, the determination result by each of the determination units 61 to 63 may be weighted, and the final determination result may be determined in consideration of the weight.
As described above, the failure prediction accuracy is improved by performing the failure prediction determination by the optimal determination method according to the state of the device, the model, and the type of failure using a plurality of determination methods. .

The determination process by the probabilistic inference model determination unit 62 will be described with reference to FIG.
FIG. 10A is an example of a causal network diagram of a probabilistic inference model based on a Bayesian network. For example, factor 1 or factor 4 may occur due to the occurrence of factor 10, and the occurrence of factor 1 may occur. This indicates that failure A or failure C may occur. FIG. 10B is an example of a conditional probability table for each node in the network. For example, when factor 1 is Yes (inflection point is detected), factor 3 is also Yes (inflection point is detected). In this case, the probability of occurrence of the failure A is 99%, while when the factor 3 is No (the inflection point is not detected), the occurrence probability of the failure A is 80%.

  Thus, in this example, each node of the factors 1 to 4, 10 to 12 is linked to each of the other causal factors or the nodes of the faults A to D, and the conditional probability table The probabilities between the nodes calculated along the lines are propagated to the network to infer the probabilities of the final nodes (failures A to D). That is, for example, when an inflection point is detected with an image forming parameter corresponding to factor 1, data is set as evidence information in factor 1 and the network is propagated with probability according to a conditional probability table. The probability of occurrence is inferred, and if the probability estimation result is equal to or higher than a predetermined threshold, it is output that an inflection point that leads to the occurrence of the failure is detected on that day.

The determination process by the fault tree determination unit 63 will be described with reference to FIG.
FIG. 11 shows an example of a fault tree used in the fault tree determination unit 63. In FIG. 11, factors 1 to 5 are basic parameters, high-order factors 1 to 3 are high-order parameters related to basic parameters (factors 1 to 5), and the factors and high-order factors are AND (logical product) or OR, respectively. The tree leading to the event of the highest fault A is configured by combining them with (logical sum). In this example, an event on the fault tree (factors 1 to 5) corresponding to the image forming parameter in which the inflection point is detected is traced to the top, and the occurrence probability of failure A, which is the highest event, is equal to or higher than a specified threshold value. Then, it outputs that the inflection point leading to the occurrence of the failure A is detected on that day.

The selection process by the determination selection part 64 is demonstrated with reference to FIG.
FIG. 12 shows an example of a selection table used in the determination selection unit 64. For each type of failure, each of the determination units 61 to 63 is selected from “O”, “−”, “Δ”, and a numerical value. Is set. “◯” indicates a determination unit that is selected as a determination result, “−” indicates a determination unit that is not selected as a determination result, “Δ” indicates a determination unit that is a majority decision candidate, and a numerical value depends on each determination unit This represents a weighting coefficient for the determination result. According to FIG. 12, for failure A, the determination result of the threshold determination unit 61 is selected. For failure B, the majority of the determination results of the three determination units 61 to 63 is selected. For C, it indicates that the determination result of the fault tree determination unit 63 is selected. For the failure D, the determination results of the determination units 61 to 63 are multiplied by a weighting coefficient, and the total value is a predetermined threshold value. If it is above, it determines with having detected the inflection point leading to generation | occurrence | production of a failure.
In this example, the method of using the determination results by the determination units 61 to 63 is set for each type of failure. For example, the type (model) or installation location of the image forming apparatus is classified and the classification is performed. You may make it set for every.

  In the above description, the example in which the image forming apparatus itself operates as a failure prediction apparatus that predicts the occurrence of a failure in the own apparatus has been described. However, the image forming apparatus and the failure prediction apparatus are configured separately. May be. That is, for example, a failure prediction apparatus connected to one or more image forming apparatuses so as to be able to communicate wirelessly or by wire receives the measured values of each image forming parameter from each image forming apparatus and accumulates them as time series data. A failure prediction system that predicts the occurrence of a failure for each image forming apparatus based on the time series data may be used.

FIG. 14 illustrates a hardware configuration of a computer that operates as the failure prediction apparatus of this example.
In this example, a main memory such as a CPU (Central Processing Unit) 81 that performs various arithmetic processing, a RAM (Random Access Memory) 82 that is a work area of the CPU 81, and a ROM (Read Only Memory) 83 that records basic control programs. Device, auxiliary storage device such as HDD (Hard Disk Drive) 84 for storing programs and various data according to an embodiment of the present invention, display device for displaying and outputting various information, and operations used for input operations by the operator Consists of a computer having hardware resources such as an input / output I / F 85 that is an interface with an input device such as a button or a touch panel, and a communication I / F 86 that is an interface for performing wired or wireless communication with other devices. Has been.
Then, the program according to the embodiment of the present invention is read from the auxiliary storage device 84 or the like, loaded into the RAM 82, and executed by the CPU 81, thereby realizing each functional unit related to the above-described failure prediction device on the computer. ing.

Note that the program according to the embodiment of the present invention is based on, for example, a computer according to this example depending on a format read from an external storage medium such as a CD-ROM storing the program or a format received via a communication network. Set to
Moreover, it is not restricted to the aspect which implement | achieves each function part by software configuration like this example, You may make it implement | achieve each function part with a dedicated hardware module.

51: Parameter collection unit 52: Parameter statistical processing unit 53: Time series data storage unit 54: Sequential interval linear regression calculation unit 55: Inflection point extraction unit 56: Failure prediction determination unit 57: Failure prediction determination Information update unit 61: threshold determination unit 62: probability inference model determination unit 63: fault tree determination unit 64: determination selection unit

Claims (5)

A collecting means for collecting image forming parameters in execution of image forming processing of an image forming apparatus to be a target of failure prediction;
Storage means for storing the collected image forming parameters;
Extracting means for extracting an inflection point for the tendency of the image forming parameter based on the tendency of fluctuation of the accumulated image forming parameter;
Criteria for extracting a feature at an inflection point and creating a failure prediction criterion for an image formation parameter in which an inflection point occurred before the occurrence of the failure, based on a case of an image forming apparatus failure that occurred in the past Creating means;
Predicting means for predicting a failure occurring in the image forming apparatus that is a target of the failure prediction after the inflection point, based on the tendency of the change and the reference;
A failure prediction apparatus comprising: The extraction means includes a convex shape direction of the curve approximating the tendency in a time series interval before a certain time point, and a convex shape direction of the curve approximating the tendency in a time series interval including the certain time point. When the direction of the convex shape is different, the value of the image forming parameter at a certain point is extracted as an inflection point.
The failure prediction apparatus according to claim 1. A storage means for storing a parameter group combining a plurality of image forming parameters and a failure type in association with each other,
The predicting means predicts that when an inflection point is extracted for all image forming parameters related to the parameter group, a type of failure associated with the parameter group occurs;
The failure prediction apparatus according to claim 1, wherein the apparatus is a failure prediction apparatus. Image forming means for forming an image on a recording material;
Collecting means for collecting image forming parameters in execution of image forming processing by the image forming means;
Storage means for storing the collected image forming parameters;
Extracting means for extracting an inflection point for the tendency of the image forming parameter based on the tendency of fluctuation of the accumulated image forming parameter;
Based on the cases of other image forming device failures that occurred in the past, for image forming parameters in which an inflection point occurred before the occurrence of the failure, a failure that was created in advance by extracting features at the inflection point A reference acquisition means for acquiring a prediction reference;
Predicting means for predicting a failure occurring in the device after the inflection point based on the tendency of variation and the reference;
An image forming apparatus comprising: On the computer,
A collection function for collecting image formation parameters in the execution of image formation processing of an image forming apparatus that is a target of failure prediction;
A storage function for storing the collected image forming parameters;
An extraction function for extracting an inflection point for the tendency of the image forming parameter based on the tendency of fluctuation of the accumulated image forming parameter;
Based on the cases of other image forming device failures that occurred in the past, for image forming parameters in which an inflection point occurred before the occurrence of the failure, a failure that was created in advance by extracting features at the inflection point A reference acquisition function for acquiring a prediction reference;
A prediction function that predicts a failure that occurs in the image forming apparatus that is the target of the failure prediction after the inflection point, based on the variation tendency and the reference;
A program to realize

Images