JPH06124019A - Image forming device provided with self-repairing function - Google Patents

Abstract

PURPOSE:To shorten the time required for repairing a fault so that a self- repairing system is made a practical machine incorporating type system. CONSTITUTION:The diagnosis of the fault is performed by comparing a value obtained by making each sensor value a fuzzy qualitative value basing on a fuzzy membership function with a virtual case previously stored in a virtual case holding part 12. When the fault is diagnosed to exist, the repair of the fault is performed. Repairing operation is performed so that a target area is previously set and an area to which the device belongs is transited toward the target area. In the case where a contradiction is found in the transition of the area by the time where the repair is completed, the repair is performed from a different point of view by resetting a transition target area. Therefore, an image forming device having a width in the repair is obained.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image forming apparatus having a self-repairing function, and more particularly, it converts a detected value obtained from the image forming apparatus into a qualitative value and stores the converted qualitative value in advance. The present invention relates to an image forming apparatus capable of performing self-diagnosis as to whether or not a failure has occurred in the apparatus by comparing with a virtual case that has occurred, and self-repairing the failure that has occurred.

[0002]

2. Description of the Related Art In the field of development of precision machinery and industrial machinery, experts who have recently used artificial intelligence (so-called AI) technology in order to save labor in repair work and lengthen automatic operation. The research of the system is actively done. Some expert systems self-diagnose whether a failure has occurred in the device and also self-repair the failure that has occurred.

However, the conventional expert system (automatic adjustment system or failure diagnosis system) basically only operates the corresponding actuator based on the output of a certain sensor, and therefore, as a self-healing machine. It wasn't perfect. Therefore, the applicant of the present application has found a machine control method using diagnosis / repair inference on a target model based on qualitative physics, and using such a machine control method, a new self-diagnosis and self-diagnosis for an image forming apparatus are performed. He invented a restoration system and filed a patent application (for example, see Japanese Patent Laid-Open No. 4-130459).

The self-diagnosis and self-repair system for the image forming apparatus according to this prior application has the following features. That is, (1) The detection value of the sensor provided in the image forming apparatus is converted into a qualitative value and used for control. (2) The structure and characteristics of the image forming apparatus are qualitatively expressed using a causal relationship network (parameter model) of parameters representing the characteristics of the image forming apparatus.

(3) The sensor value converted into a qualitative value is applied to a parameter model to perform a qualitative simulation for fault diagnosis and fault repair inference. In other words, a qualitative model-based system (Qualitat
ive Model Based System (QMS) for failure diagnosis and repair.

According to the self-diagnosis and self-repair system according to the prior application of the applicant of the present invention having such characteristics, even if a failure such as a structural change occurs in the image forming apparatus, it is flexible. Is available. This is because it is possible to dynamically change the control point and control loop of the target machine by using the qualitative simulation.

[0007]

However, in order to operate the self-diagnosis and self-repair system according to the above-mentioned prior application as a practical machine-incorporated system, the system size should be reduced with respect to the failure diagnosis. It is necessary to improve the execution speed for failure repair. As one method therefor, the inventor of the present application and the applicant previously proposed a control device for performing failure diagnosis using virtual case knowledge and filed a patent application (Japanese Patent Application No. 4-66439- (See 7 cases of Japanese Patent Application No. 4-66445). When such virtual case knowledge, that is, the virtual case is used, it is not necessary to apply the qualitative value obtained from the image forming apparatus to the parameter model and perform the qualitative simulation at the time of failure diagnosis. Can be separated, and it is only necessary to provide the image forming apparatus with a failure diagnosis unit that stores virtual cases. Therefore, it is possible to reduce the system capacity for failure diagnosis mounted on the image forming apparatus.

The invention of the present inventor and the invention of the applicant's earlier application have such characteristics, but now, the inventor of the present application further expands the fault diagnosis method using the virtual case to create a virtual The present invention provides an image forming apparatus capable of favorably performing not only failure diagnosis using a case but also failure recovery plan. That is, a main object of the present invention is to store a virtual case obtained in advance by qualitative simulation in a memory, perform accurate failure diagnosis using the virtual case, and derive a failure repair plan to perform failure repair. An object of the present invention is to provide an image forming apparatus having a self-diagnosis and self-repair system which is automatically built and is practical.

A more specific object of the present invention is to divide the qualitative space that can be taken by the device into a plurality of regions so that the device can be properly repaired in accordance with a failure that has occurred in the device. If there is a contradiction, it is determined from which area the area transition should be made so that the device state belongs, and whether or not there is a contradiction when the area transition occurs. Another object of the present invention is to provide an image forming apparatus in which the transition target area is reset and repair is performed from a different viewpoint to improve self-repair efficiency.

[0010]

The invention according to claim 1 is
An image forming apparatus having a self-repairing function, which is provided in a predetermined portion of the apparatus, has a sensor unit for detecting a predetermined physical quantity or a change in the physical quantity, and has a predetermined portion of the apparatus to perform a predetermined operation. Actuator means, a conversion means for converting a detection value of the sensor means into a fuzzy qualitative value using a fuzzy membership function, and a qualitative value represents a possible state of the apparatus when a failure symptom occurs in the apparatus. The current state of the device by comparing the case storage means in which the plurality of virtual cases stored, the fuzzy qualitative value converted in the conversion means, and the plurality of virtual cases stored in the case storage means The qualitative space that can be taken by the diagnostic means and the device is divided into a plurality of areas, and the actuator is used as operation information for restoration corresponding to each area. In the operation information storage unit that stores the operation content of the step and the region that can be transitioned by the operation, the target region setting unit that sets the repair target in any one of the divided regions, and the diagnosis unit. It is determined which one of the plurality of divided regions the present state of the diagnosed device belongs to, the operation information of the belonging region is read from the operation information storage unit, and the actuator unit is read according to the read operation information. Is provided in the repair executing means for performing repair so that the area to which the apparatus state belongs moves toward the area set by the target area setting means, and the state of the apparatus is provided by executing the repair. When it transits to belong to another area,
Confirmation means for confirming whether or not the transitioned area is a consistent area, and resetting means for resetting the target area set by the area setting means when the transitioned area is inconsistent by the confirmation means, It is characterized by including.

[0011]

According to the present invention, the present state of the device is represented by a fuzzy qualitative value, which is represented by the qualitative value and compared with a plurality of prestored virtual cases. A plurality of virtual cases are obtained in advance by qualitative simulation and stored. Therefore, the fuzzy qualitative value and the virtual case need only be read out from the case storage means to compare the fuzzy qualitative value and the virtual case, and the current state of the apparatus can be diagnosed in a short time.

Further, as a result of the diagnosis, if the device is out of order, the operation information for repair is read from the operation information storage means. In the operation information storage means, the qualitative space that the device can take is divided into a plurality of areas, and operation information corresponding to each area is obtained in advance and stored. The operation information includes the operation content of the actuator and a region that can be transitioned by the operation. Therefore, it is possible to determine which divided region the current state of the device belongs to and obtain which region to transit to next as the operation information. Therefore, the repair can be efficiently performed based on the obtained operation information.

In the repair operation, the target area is set in advance,
A repair operation is performed such that the state to which the device belongs transitions towards its target area. If the area transitions to a theoretically impossible area while transitioning toward the target area, it is possible that the sensitivity of the sensor means, etc. may decrease, so confirm that such a situation has not occurred. . In addition, when a contradiction occurs in the above-mentioned region transition during restoration, the set target region is not valid, and the target region is reset. This allows the approach to restoration to be restarted from a different perspective.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A small electrophotographic copying machine capable of self-diagnosis and self-repair will be described below as an embodiment with reference to the drawings. FIG. 1 is a machine configuration diagram of a small-sized electrophotographic copying machine to which the present invention is applied, and is a diagram schematically showing only a portion related to the present invention. In FIG. 1, 1
Is a photosensitive drum, 2 is a main charger, 3 is a halogen lamp for illuminating an original, 4 is a developing device, and 5 is a transfer / separation charger.

The main charger 2 is connected to a main charger controller 2C for changing the discharge voltage of the main charger. Further, a halogen light amount controller 3C for controlling the light amount of the halogen lamp 3 is connected to the halogen lamp 3. further,
The transfer / separation charger 5 includes a transfer charger controller 5 for controlling the discharge voltage of the charger 5, that is, the transfer voltage between the photosensitive drum 1 and the copy sheet.
C is connected.

In an electrophotographic copying machine, it is of utmost importance whether or not the obtained copy image is beautifully finished (normal). Therefore, in this embodiment, it is automatically detected whether the obtained copy image is normal, has image fog, or has a light image, and the obtained copy has image fog or has a light image fog. Describes an example of an electrophotographic copying machine capable of diagnosing the cause of such a symptom, that is, a failure and self-repairing the failure.

In this embodiment, for example, four sensors are provided. That is, the light amount sensor X for measuring the amount of light that exposes the photosensitive drum 1, in other words, the amount of light of the halogen lamp 3, and the surface potential sensor Vs for measuring the surface potential of the photosensitive drum 1 after exposure. , A toner density sensor Ds for detecting the toner density on the photosensitive drum 1, and a copy density sensor Os.

The copy density sensor Os is for detecting the density of the copy image formed by this electrophotographic copying machine. Detection output Os of copy density sensor Os
Based on the above, it is determined whether the electrophotographic copying machine is normal, image fogging occurs as a failure symptom, or whether the image is light. That is, as shown in FIG. 3, when the detection output Os of the copy density sensor Os is, for example, less than 2.5 (V), it is determined that the copy density is low. When the detected output Os is 2.5 (V) or more and less than 2.9 (V), it is determined that the copy density is normal. Further, when the detected output Os is 2.9 (V) or more, it is determined that the image has fog in the copy. This determination is made by a diagnosis / repair reasoning unit 11 (described later) which is provided in the electrophotographic copying machine and is composed of, for example, a microcomputer.

FIG. 2 is a functional block diagram of the small-sized electrophotographic copying machine shown in FIG. 1, and shows only the portion related to the present invention. In FIG. 2, blocks with rounded corners represent functions by so-called hardware, and blocks with sharp corners represent functions by so-called software (program processing executed in a computer). Note that the division of functions by hardware and functions by software is an example, and the functions of software may be realized by hardware.

The correspondence relationship between the functional blocks of FIG. 2 and the mechanical configuration of FIG. 1 is as follows. That is, the sensor of FIG. 2 includes the light quantity sensor X, the surface potential sensor Vs, and
A toner density sensor Ds and a copy density sensor Os are included. The actuator controller of FIG. 2 includes the main charger controller 2C, the halogen light amount controller 3C and the transfer charger controller 5C of FIG. The actuator of FIG.
The main charger 2, the halogen lamp 3 and the transfer / separation charger 5 of FIG. 1 are included.

In FIG. 2, the software section incorporated in the electrophotographic copying machine, that is, the functional block by software includes, for example, five functional blocks. That is, the diagnosis / repair reasoning unit (reasoning engine) 1
1, virtual case holder 12, membership function generator 1
3, a pseudo failure generation unit 14, and a work script table 15.

The virtual case holding unit 12 stores in advance virtual cases generated by a virtual case compiler 16 provided in an external computer or the like. Here, the virtual case is a virtual state case obtained as a result of performing behavior simulation based on qualitative physics in advance for all possible failures and all actuator operations in the electrophotographic copying machine.

The virtual case is obtained in advance by the virtual case compiler 16, and the obtained virtual case is held by the virtual case holding unit 12.
The virtual case compiler 16 can be separated from the software part incorporated in the electrophotographic copying machine. In other words, by storing virtual cases,
The failure diagnosis unit that must be incorporated in the electrophotographic copying machine and the simulation unit (virtual case compiler 16) that does not have to be incorporated in the electrophotographic copying machine can be separated, and the system installed in this electrophotographic copying machine can be separated. A reduction in capacity can be realized.

The virtual case generation in the virtual case compiler 16 may be performed using the qualitative simulation described in, for example, the applicant's prior application (Japanese Patent Laid-Open No. 4-130459). Just in case, I will briefly explain the method of qualitative simulation. When this electrophotographic copying machine is viewed from a physical point of view as a combination of a plurality of elements and the behavior and attributes of each element and the connection relationship between each element are qualitatively expressed using parameters, for example, FIG. The parameter model shown in is obtained. In addition,
The parameter model shown in FIG. 4 is a simplified model in which only the parameters relating to the copy density parameter Os are taken out.

In the parameter model shown in FIG. 4, H
l is the light amount parameter of the halogen lamp 3, D is the optical density parameter of the document, X is the light amount parameter for exposing the photosensitive drum 1, and β is the sensitivity parameter of the photosensitive drum 1.
Vn is the surface potential parameter of the photosensitive drum 1 after the main charge, Vs is the surface potential parameter of the photosensitive drum 1 after exposure, Vb is the development bias parameter, γ ₀ is the toner sensitivity parameter, and Ds is the drum sensitivity parameter. An image density (toner density) parameter, Vt is a transfer voltage parameter, and ζ is a paper sensitivity parameter. Of these parameters, D, β, γ _0, and ζ are unlikely to change, and thus can be regarded as fixed values. Therefore, the reason why the copy density parameter Os changes is H
It can be inferred that this is due to a change in any one of l, Vn, Vb, and Vt. When these four parameters Hl, Vn, Vb, or Vt change and Os changes, the change must be three sense target parameters H, Vs, or Ds (enclosed with circles in FIG. 4).
(However, only when Vt is changed,
Neither X, Vs nor Ds changes. ).

In the qualitative simulation for generating the virtual case, the change of the copy density parameter Os and the four parameters Hl, Vn, Vb and Vt when the change of Os occurs on the parameter model shown in FIG. The sense target parameters X, Vs and Ds are simulated to simulate the sense target parameters X, Vs.
The states of s and Ds are grasped and the states of four causal parameters Hl, Vn, Vb and Vt are grasped.

Next, how to generate a virtual case based on the above qualitative simulation will be specifically described with reference to the parameter model of FIG. It is assumed that the copy density becomes abnormal and image fogging occurs, for example. That is, O
Suppose s goes high (+). In the parameter model of FIG. 4, the cause of Os becoming high (+) is Hl: low (−), Vn: high (+), Vb: low (−), or Vt: high (+) by qualitative simulation. It can be seen that it is due to changes or a combination of those changes. Then, these parameters Hl, Vn,
The changes in Vb and Vt affect the sense target parameters X, Vs, and Ds on the parameter model. This is because if the parameter changes to such an extent that it does not affect, Os will not change as a result.

Therefore, the combination of how each parameter has changed is the sense target parameters X, Vs, D.
change s. The result of simulating how the sense target parameters X, Vs, and Ds change based on the parameter change is a virtual case, and Table 1 shows the virtual case.

[0029]

[Table 1]

Table 1 shows eight virtual cases a, b, c, d, e, f, g and h when image fog occurs. The numerical value of "1.0" added to each of the three sensor parameter states in Table 1 indicates the grade in the membership function of the fuzzy theory described later. The advantages of introducing the membership function of this fuzzy theory will be described later.

The virtual cases illustrated in Table 1 above are generated by qualitative simulation in advance in the virtual case compiler 16 for each failure symptom, and as described above, these are stored in the virtual case holding unit 12. . The diagnosis / repair inference unit 11 compares the detection output Os of the copy density sensor Os with the failure determination reference value shown in FIG. 3 as described above, and determines whether the image density is normal. If it is determined that the image density is abnormal, for example, if image fogging is determined, the cause of the failure symptom “image fogging” is estimated. That is, the failure diagnosis is performed. In the fault diagnosis, three sensors X, Vs,
The detection value given from Ds is converted into a qualitative value in the diagnosis / repair inference unit 11, and the converted qualitative value is compared with the sensor grades listed in the failure symptom “image fog” shown in Table 1 to find the highest degree of agreement. High virtual cases are retrieved.

By the way, when converting the detection values of the three sensors X, Vs and Ds into qualitative values, it is necessary to define a boundary mark (landmark) on the qualitative amount space. Boundary landmarks must be defined correctly.If this boundary landmark is not defined correctly, qualitative quantification of detected values will be performed based on incorrect boundary landmarks, and subsequent failure diagnosis etc. I can't do it right.

Therefore, in this embodiment, a method of introducing the pseudo fault method (Imitation Fault method: IF method) into the determination of the landmark and dynamically determining the landmark by fuzzyizing the qualitative value is a method. Has been adopted. First, the membership function used when making a qualitative value will be described. This membership function is stored in the membership function generation unit 13 shown in FIG. 2, and the membership function is updated by the pseudo failure generation unit 14 as described later.

FIG. 5 shows the membership function generating section 13.
The toner concentration sensor Ds, the surface potential sensor Vs, and the light amount sensor X used at the time of the "fogging of image" as illustrated in FIG.
A fuzzy membership function for qualitatively converting the detected value of is stored. As is well known, the fuzzy membership function is a function that defines the degree (grade) of a certain element belonging to a certain set in the fuzzy theory.

In the diagnosis / restoration inference section 11, the detection output of the copy density sensor Os is compared with the failure determination reference value shown in FIG. 3, and when Os is 2.9 (V) or more, the image is fogged on the copy. Is determined to have occurred. When it is determined that the image fog has occurred, the detected values of the toner density sensor Ds, the surface potential sensor Vs, and the light amount sensor X at that time are stored in the membership function generating unit 13 and the fuzzy membership shown in FIG. It is qualitatively based on a function.

For example, if the detected value of the toner density sensor Ds is less than 2.33 (V) in quantitative value, Ds (N: 1.0,
Qualitative value is set to +: 0.0). Toner concentration sensor Ds
When the detected quantitative value of Ds is 2.82 (V), for example, Ds
It is qualitatively converted to (N: 0.7, +: 0.3). Also,
When the detected quantitative value of the toner concentration sensor Ds is 3.97 (V) or higher, the qualitative value is set to Ds (N: 0.0, +: 1.0).

Similarly, the detected quantitative value of the surface potential sensor Vs and the detected quantitative value of the light quantity sensor X are also qualitatively converted into values based on the fuzzy membership function of Vs and the fuzzy membership function of X shown in FIG. 5, respectively. It Next, how to generate this fuzzy membership function will be described. Generally, in order to convert the quantitative value detected by the sensor into a qualitative value, it is necessary to define a boundary mark on the quantity space as described above. However, it is not easy to determine the boundary mark as a static mark in consideration of the change of the state accompanying the use of the electrophotographic copying machine and the limit of the measurement accuracy of the sensor. If the boundary mark is determined as static and there is an error in the determination, it means that the qualitative quantification of the sensor value, which is a prerequisite for the fault diagnosis and fault repair control according to this embodiment, cannot be performed accurately. In the failure diagnosis and the failure repair, there is a great possibility that the misdiagnosis and the incorrect repair are performed.

Therefore, in this embodiment, the boundary mark is defined for each failure symptom, and the boundary mark is defined by using the membership function of the fuzzy theory. If the detected quantitative value of the sensor is made into a qualitative value by using the membership function according to the failure symptom, it is possible to flexibly and appropriately cope with the reading error of the sensor and the fluctuation of the sensor output due to the change of the usage environment.

If a fuzzy theory membership function is introduced when converting the detected quantitative value of the sensor into a qualitative value, the correspondence between the measured quantitative value and the qualitative value, which depends on the measurement accuracy of the sensor and changes in the operating environment, etc. It is possible to flexibly deal with the problem relating to marking, and it becomes difficult for an error to occur when the sensor value is qualitatively converted. A specific target is determined by a method called a pseudo-fault method (IF method) that fuzzy qualitative values to dynamically determine the target. The IF method forces the electrophotographic copying machine to fail by operating the actuator at the initial stage before shipping the electrophotographic copying machine, after repairing the failure, or at any timing based on manual input. This is a method of dynamically determining the boundary mark by using sensor information before the occurrence and when the failure occurs.

If the IF method is used, it is possible to dynamically determine the boundary mark on the quantity space necessary for converting the quantitative value detected by the sensor into a qualitative value for each electrophotographic copying machine to be controlled. It is possible to accurately define the boundary mark that is the basis of qualitative value conversion for each device. In addition, if the IF method is used, the boundary mark defined when the device is in the initial state can be corrected, for example, each time the failure repair is completed. The upper boundary mark can always be updated to the optimum value.

The IF method is used in the pseudo fault generating section 14 shown in FIG.
Executed by. Next, regarding the processing contents of the IF method,
This will be specifically described with reference to FIG. First, the diagnostic / repair inference unit 11 reads the detection values of the toner concentration sensor Ds, the surface potential sensor Vs, and the light amount sensor X (step S1). The detection value of each sensor read at this time is, for example, Ds: 2.33 (V), Vs: 0.44.
(V), X: 3.87 (v).

Next, the pseudo failure generating section 14 operates the halogen light amount controller C3 to forcibly reduce the light amount of the halogen lamp 3 (step S2). Then, each time the light quantity of the halogen lamp 3 is reduced by a small amount, the electrophotographic copying machine is made to perform a copy operation, and the copy density obtained at that time is detected by the copy density sensor Os.
The detected value is read by the diagnosis / repair reasoning unit 11 (step S3). The detection value of the copy density sensor Os is
In the light of the failure determination reference value shown in FIG.
When the value of has reached the reference value at which image fogging occurs, the process of lowering the light amount of the halogen lamp 3 is stopped (step S4).

Then, the toner density sensor D when the light quantity of the halogen lamp 3 is lowered until the image fogging occurs
The detection values of s, the surface potential sensor Vs, and the light amount sensor X are read (step S5). The read detection values are, for example, Ds: 3.97 (V), Vs: 0.76.
(V) and X: 3.24 (v). The detected values of Ds, Vs, and X read in step S1 and D at the time when the image fog read in step S5 occurs
The detected values of s, Vs, and X are the membership function generation unit 1
3, the membership function at the time of image fogging is generated. That is, the value detected in step S1 is the boundary marker at the normal time, and the value read in step S5 is the boundary marker at the start of the image fog generation. As a fuzzy membership function at the time of the image fog, as shown in FIG. What is shown is obtained (step S6).

When the IF method is used, for example, the boundary mark of the electrophotographic copying machine can be dynamically updated every time a failure is repaired.
It is possible to define a boundary mark suitable for the electrophotographic copying machine after restoration. In addition, even if the electrophotographic copying machine changes over time due to long-term use, the boundary marker can be dynamically updated in accordance with the change over time, so it is always detected by each sensor based on the optimum boundary marker. Qualitative value conversion of quantitative values can be performed.

By the way, in the fault diagnosis based on the virtual case, the result is given as the coincidence order with respect to the virtual case. For example, when the failure symptom “image fogging” occurs, the eight virtual cases a, b, and
Ranking is performed on c, d, e, f, g, and h.
Eight virtual cases ah are three sensor parameters D
It can be expressed in a three-dimensional quantity space of grades of s, Vs, and X. This three-dimensional quantity space is shown in FIG. In FIG. 7, eight vertices of the cube represent the states of the eight virtual cases shown in Table 1, respectively. As described above, the result of the fault diagnosis based on the virtual case is given as the coincidence order for the eight virtual cases. Therefore, the state space that the electrophotographic copying machine can take is divided into regions according to the coincidence order for each virtual case, that is, the case order. It is possible.

Table 2 shows a list of areas and case orders when the state space that the electrophotographic copying machine can take in the failure symptom "image fog" is divided into 96 areas.

[0047]

[Table 2]

The order of cases in each area represents the function development state of the electrophotographic copying machine. That is, the functional state of the electrophotographic copying machine obtained as a result of the failure diagnosis is closest to which virtual case, which virtual case is next, which virtual case is the next, and so on. You can see if you are moving away. Therefore, by using this area information in the restoration plan, the restoration target state can be set according to the function development state of the electrophotographic copying machine. Then, in the repair, since the function of interest has only to be expressed, the function of interest has been expressed, but it is possible to examine whether it is possible to compromise without expressing other functions. That is, it is possible to realize a function trade-off regarding repair.

The list of divided areas and case rankings shown in Table 2 is also stored in advance in the virtual case holding unit 12 of FIG. Similarly, for failure symptoms other than image fogging, the regions are preliminarily segmented and the case rankings are obtained and stored. The number of region divisions is determined by the number of sense target parameters, and further determined by what rank of the matching order of the failure diagnosis result for each virtual case is considered effective.
Since the order of the virtual cases considered to be effective directly affects whether or not the restoration is successful, it is necessary to determine in advance according to the sensor accuracy of the system or the like before the restoration. In this embodiment, this validity order will be referred to as "diagnosis resolution". For example, when the target machine is equipped with n sensors and the matching order for virtual cases is regarded as all valid, the maximum area division number N is limited to two qualitative values to be considered at the same time (2 One qualitative value means, for example, that qualitative value is + / N or − / N), and is given by the following equation (1).

[0050]

[Equation 1]

Specifically, in this embodiment, three sensors D are used.
s, Vs, X, and the qualitative amount space of the three sensors is limited to two qualitative values (+ / N or − / N), the maximum division number is 96, as shown in Table 2. , Area A-1 to A-96 can be divided. The division into 96 regions in this region division will be described more specifically.

Considering the cube of the three-dimensional quantity space shown in FIG. 7, eight virtual cases are located at each vertex of the cube. Therefore, the first virtual case is selected from any of the virtual cases located at the eight vertices of the cube, and thus there are eight cases. For example, it is assumed that the virtual case a is selected as the first virtual case. Then, the second virtual case includes three virtual cases b, which are located at vertices close to the first virtual case a.
Since it will be selected from either d or e, there are three ways to select. Furthermore, the third virtual case is selected from the remaining two unselected virtual cases b, d, and e that may have been selected as the second virtual case described above. Further, the 4th virtual case is a non-selected one of the above 3 virtual cases b, d, and e, or a virtual case c.
Chosen from.

Then, when the first virtual case is determined to be a, for example, a virtual case g of the vertex opposite to the virtual case a.
Is determined as the eighth virtual case, and similarly, if the second through fourth virtual cases are determined, the seventh through fifth virtual cases are also determined. Therefore, in the end, it can be divided into 96 areas of 8 × 3 × 2 × 2 = 96. Next, the work scripts stored in the work script table 15 (FIG. 2) will be described.

Between the areas divided by the case rank, there are cases where transition is possible based on the property of the target machine and cases where transition is not possible. This transition rule can be derived from the causal relationship of the parameters of the electrophotographic copying machine which is the target machine and the configuration of the actuator, and can be used in the restoration plan of the electrophotographic copying machine. FIG. 8 is a schematic diagram planarly representing the three-dimensional quantity space shown in FIG. 7 as a two-dimensional quantity space, and is an illustrative view showing the relationship between the above-described transition rule and the path in the restoration process. Referring to FIG. 8, assuming that the current state area of the electrophotographic copying machine is A-1 and the goal area to be restored is A-19,
It suffices to search for a path from the current state area A-1 to the goal area A-19, and operate the actuator while monitoring the matching order of the virtual cases so that the state transits through the path. Knowledge representing the relationship between the contents of this actuator operation and the area change is the work script. Actual repair is performed by using this work script as the minimum unit operation.

Tables 3, 4, 5, and 6 show examples of work scripts.

[0056]

[Table 3]

[0057]

[Table 4]

[0058]

[Table 5]

[0059]

[Table 6]

As shown in Tables 3 to 6, the work script is set for each area and represents a certain operation and an area which can be transitioned by the operation. For example, the work script of the area A-1 shown in Table 3 describes that, as an operation, if the halogen light amount (parameter Hl) is increased, the operation script may transit to the area A-4. Further, when the halogen light amount (parameter Hl) is lowered, the grade of the light amount parameter X is maximized or the grade of the toner density parameter Ds is maximized.

Further, as an operation, when the main charging voltage (parameter Vn) is increased, it is possible to transit to the area A-3 or the parameter Ds may be maximized. ) Is lowered,
If the developing bias voltage (parameter Vb) can be changed to the area A-4, the area can be changed to A-2, and the developing bias voltage (parameter Vb) can be changed.
It is described in the work script that the parameter Ds may become maximum when b) is lowered.

Further, the work script may describe not only the type of operation, for example, increasing the halogen light amount (parameter Hl), but also the operation amount. It is preferable that the operation amount describes the operation amount that causes the region transition by one operation when the region transition described later is performed. In the repair process described with reference to FIG. 8, first, in the area A-1, for example, the halogen light amount (parameter Hl) is increased based on the work script A-1 shown in Table 3, and as a result, the state area A -4, the operation based on the work script A-4 shown in Table 4 is performed.

As described above, the work script table 1
5 (see FIG. 2) stores a corresponding work script for each divided area. The work script stored in the work script table 15 is created by simulation based on the parameter model shown in FIG. Work script table 15
If the work script is stored in advance by storing the work script, it is not necessary to generate the work script based on the parameter model each time a failure occurs in the electrophotographic copying machine, and the electrophotographic copying machine can be downsized. . Further, since it is not necessary to generate a work script for each failure repair, it is possible to shorten the failure repair time.

FIG. 9 is a flow chart showing the procedure of control performed in the machine built-in software section shown in FIG. 2, particularly in the diagnosis / repair reasoning section 11. Next, FIG.
The failure diagnosis and failure repair processing in this electrophotographic copying machine will be described along the flow of. The flowchart shown in FIG. 9 also includes those that are not automatically executed based on a predetermined program but are manually processed. Specifically, the resolution setting process shown in step S11, the goal space setting process shown in step S13, the determination as to whether the goal space shown in step S22 can be traded off, and the goal space resetting process shown in step S23. Is done in person.

When the process is started, the "diagnosis resolution" is set (step S11). As described above, the setting of the resolution of diagnosis is to determine the matching effective order of virtual cases which is the basis for dividing the state space into regions. In this embodiment, when the eight virtual cases a to h are given the matching ranks, the ranks of the third to sixth virtual cases are not considered. Therefore, the state space is divided into 24 regions as shown in Table 7 based on the first, second, seventh, and eighth virtual cases of the order of the eight virtual cases. To be done.

[0066]

[Table 7]

In Table 7, virtual cases marked with * are cases in which the order is not considered. The setting of the resolution may be determined in accordance with the sensor accuracy of the system as described above. In this embodiment, as a result of setting the resolution as described above, the state space is divided into 24 regions. Within the range of 96, the number of area divisions can be increased or decreased depending on how to set the resolution.

When the resolution is set, then a failure determination is made (step S12). The failure determination is made by reading the detected value of the copy density sensor Os by the diagnosis / repair inference unit 11 and comparing the read copy density Os with the failure determination reference value shown in FIG. The failure determination process of step S12 is automatically performed by the diagnosis / repair reasoning unit 11 in this embodiment, but may be performed manually.

When the failure determination is performed manually, the copy density sensor Os need not be provided. In the manual process, a serviceman or the like may judge that the copy causes image fogging, for example, by looking at the copy output from the electrophotographic copying machine. In this case, the fact that image fogging has occurred as a failure symptom is input to the apparatus. The failure symptom may be input, for example, by changing the electrophotographic copying machine to a failure diagnosis mode and inputting it by a ten-key pad or the like usually provided in the electrophotographic copying machine.

In step S12, a failure symptom is determined, and if it is determined that "image fog" has occurred as a failure symptom, then a goal area is set (step S13). The setting of the goal area is to set any one of, for example, 24 areas divided by the setting of the resolution in step S11 as the final target area of the transition. This goal area is set manually.

Then, failure diagnosis is performed (step S14). FIG. 10 shows a detailed subroutine of the failure diagnosis processing in step S14. Next, the failure diagnosis processing will be specifically described with reference to FIG. First,
In the diagnosis / repair reasoning section 11, the toner concentration sensor D
The detection values of s, the surface potential sensor Vs, and the light amount sensor X are read (step S31). The detected quantitative value of each sensor read now is Ds: 4.00 (V), V, respectively.
It is assumed that s: 0.625 (V) and X: 3.418 (V). The read sensor quantitative values are applied to the fuzzy membership function (FIG. 5) at the time of image fogging stored in the membership function generation unit 13 and converted into a fuzzy qualitative value (step S32).

In this example, Ds: 4.00, Vs:
0.625 and X: 3.418 are applied to the fuzzy membership function of FIG. 5, respectively, and Ds: +1.
00, Vs: +0.58, and X: -0.72 are obtained (step S32). Next, it is judged whether or not the obtained fuzzy qualitative value exists in the neutral space (step S33).

Here, the fact that the fuzzy qualitative value exists in the neutral space means that, as shown in FIG. 11, two or more fuzzy qualitative value grades out of the three fuzzy qualitative values Ds, Vs, and X. , The case where it exists in a predetermined intermediate area (neutral space). When the fuzzy qualitative value exists in the neutral space, it is difficult to accurately obtain the failure diagnosis result. More specifically, when the fuzzy qualitative value exists in the neutral space, it means that the sensor value is included in a region where many regions are concentrated in the three-dimensional quantity space. Therefore, when a sensor value enters a place where a large number of such regions are concentrated, the region to which the sensor value belongs varies depending on the error in the sensor detection value and the sensor accuracy. Therefore, it is difficult to correctly determine the area to which the device state belongs, and it may not be possible to accurately perform the subsequent repair operation. Therefore, in this embodiment, the actuator is operated (step S34 in FIG. 10) so that the fuzzy qualitative value does not exist in the neutral space.

FIG. 12 shows a diagram in which the neutral space is visualized in a three-dimensional quantity space. In the three-dimensional quantity space shown in FIG. 13, 3 shown by the one-dot chain line located at the center of the cube
The neutral space is inside the three-dimensional cross space. Focusing only on the magnitude relationship of the grades of the fuzzy qualitativeized sensor values, there are six patterns of the sensor values existing in the neutral space, that is, patterns 1 to 6 shown in FIG. Further, the condition that the sensor value exists in the neutral space is that two or more sensor values exist in the neutral space width, and therefore, there are three possible states for each pattern in FIG.
Therefore, there are 6 × 3 = 18 combinations of sensor values existing in the neutral space. In this example,
In the actuator operation (step S34) in FIG. 10, processing is performed so as to get out of the neutral space earliest.

An example of actuator operation in step S34 of FIG. 10 is as follows. For example, if the grade of the sensor value is in the form of pattern 1 in FIG. 13,
And it is Ds and V that belong to the neutral space width.
If it is s, the actuator operation is Vs.
Lower (main charger output voltage).

When the sensor value grade is the pattern 1 of FIG. 13 and X and Vs belong to the neutral space width, Hl (halogen lamp light amount) is lowered as the actuator operation. Further, it is Ds, Vs, and X that the sensor value grade is the pattern 1 of FIG. 13 and belongs to the neutral space width.
And, if the actuator operation is
Lower the (halogen lamp light intensity).

The operation of these actuators was determined based on the parameter model shown in FIG. Returning to FIG. 10, in step S33, when the fuzzy qualitative value does not exist in the neutral space, or when the actuator operation in step S34 is performed, the sensor value is read again and converted into the fuzzy qualitative value (steps S31 and S32). ), If the fuzzy qualitative values do not exist in the neutral space, the degree of coincidence C between the obtained fuzzy qualitative value and the virtual case is calculated (step S35). The calculation of the degree of coincidence C is performed as follows.

First, the qualitative value Ds obtained in step S32: +1.00, Vs: +0.58, X: -0.72
Is located at a position p indicated by a black circle in the three-dimensional quantity space of FIG. Therefore, next, the distance D from the point p to each of the virtual cases a to h (the vertex of the cube in the three-dimensional quantity space of FIG. 7)
(Da to Dh) is calculated, each calculated distance D is normalized, and the matching degree C is calculated. The degree of coincidence C is calculated by the following equations (2), (3) and (4).

[0079]

## EQU2 ## C = 1-D / √n (2) (where n is the number of sense parameters: n = 3 in this case)

[0080]

## EQU3 ## C = 1-√ {C (p1) ² + C (p2) ² + ... + C (pn) ² } / √n (3)

[0081]

## EQU00004 ## C (pn) = Gm (qn) -Gs (qn) (4) (where C is the degree of agreement of the entire model, pn is a measurable variable, and C (pn) is the degree of agreement with the variable pn. , Qn: variable p
n can be a qualitative value, Gm (qn): grade of qualitative value qn in failure model, Gs (qn): grade of qualitative value qn in measured value) As a result of calculation, the qualitative value obtained in step S32 and eight The degree of coincidence C (Ca to Ch) with the virtual cases a to h is calculated. The result of the calculation is Ca: 0.709 Cb: 0.628 Cc: 0.313 Cd: 0.353 Ce: 0.519 Cf: 0.466 Cg: 0.214 Ch: 0.248. Therefore, the current state area is determined by arranging these in the order of the highest degree of coincidence (step S36). In the region determination, the resolution is set as described with reference to FIG. 9. Therefore, specifically, the one having the highest degree of coincidence, the second highest degree of coincidence, the second lowest degree of coincidence, and The area is determined based on the four matching degrees of the smallest matching degree.

In the case of this embodiment, the matching order is "ab.
******* hg ", the area is determined to be A-1. Returning to FIG. 9, when the above-described failure diagnosis is performed and the area is determined in step S14, the work script corresponding to the area is read from the work script table 15 (see FIG. 2). In the case of this embodiment, the area A-
The work script A-1 (see Table 3) corresponding to 1 is read (step S15).

Then, the read work script A-1
The repair operation is performed based on the above (step S16). Next, the specific processing of the repair operation performed in step S16 will be described based on the subroutine of FIG. Referring to FIG. 14, first, step S15 of FIG.
All transitionable areas are derived from the work script A-1 (see Table 3) read in step 3. In this case, the work script A1 has areas A-2 and A- as areas to which transition can be made.
3 and A-4 are described. Therefore, these three areas A-2, A-3, A-4 and the area A-1 to which the user currently belongs are derived (step S41).

Next, after setting the counter N to N = 1 (step S42), the proximity of the transitionable area to the goal area is provisionally determined from the rank N in the goal area, that is, the virtual case of the first order (step S42). S43). A specific example will be described with reference to Table 7 and FIG. Now, the goal area is the area A-19, and the case ranking is A-19: gc **.
** ea, so the virtual case of rank 1 in this area A-19 is g.

On the other hand, the transitionable areas described in the work script A-1 and the case ranks of the areas are as follows. A-2: ab **** fg A-3: ae **** cg A-4: ba **** gh Further, the currently belonging area is A-1: ab **** hg. There is a possibility that the area A-1 to which it currently belongs does not change and remains in the area A-1 as it is.
Therefore, the transitionable areas are A-2, A-3, A-
4 and A-1 to which he currently belongs.

By the way, if the order of the virtual case g in these four transitionable areas is detected, it is A-2: 8th position A-3: 8th position A-4: 7th position A-1: 8th position. . At this time, in a region where the virtual case of rank 1 can transit in the target region, the higher the rank, the closer to the goal region. Therefore, in this case, area A-4
Is the area closest to the goal area, and is determined to be the area to be transitioned.

Next, in step S44, it is judged whether or not the order of closeness of all the transitionable areas has been determined. In the case of the above specific example, it is determined that the area A-4 is the closest, but since the order of the proximity of the areas A-1, A-2, and A-3 to the goal area is not determined, In the goal area, the proximity of the transitionable area to the goal area is provisionally determined based on the virtual case of rank N = 2, that is, the virtual case c in this specific example (step S43).

As described above, the proximity of all transitionable areas to the goal area is determined. In addition, in the case where there are a plurality of transitionable areas, at the stage when it is determined that any one of the transitionable areas is closest to the goal area, that is, in the above-described specific example, the virtual case ranked first in the goal area. The determination of the proximity of the transitionable area to the goal area may be finished when it is determined that the area A-4 is closest to the goal area based on g.

After that, in step S46, the closest transitionable area is set as the area Z (in the above-described specific example, the area Z is the area A-4), and the area to which the current area belongs is changed to the area Z. The actuator operation is derived from the work script A-1 (step S47). Then, the actuator operation is executed (step S4).
8).

In the actuator operation in step S48, if the operation amount of the actuator is not described in the work script, a predetermined unit amount of operation is performed. If not only the type of actuator that is the operation content to be operated in the work script as described above,
When the operation amount is also described, the described operation amount is operated. It is preferable that this operation amount is an operation amount that can cause a region transition by one operation.

The above description has been made by taking the resolution shown in Table 7, that is, the case where the divided region is 24 as an example. However, as a more general example, an example when the resolution is 96 will be described below. In order to make the explanation easy to understand, it is assumed that the goal area: A-40 (gfhcebda), the area to which the player currently belongs: A-2 (abedfchg), and the area that can transition from A-2: A-1 (abdeccfh).
g) A-5 (aedbfhcg) A-7 (bacfdegh) A-8 (bafcedgh) A-2 itself.

Next, the virtual case of rank 1 is examined in the goal area. In this case, it is virtual case g. Therefore, the order of the virtual case g in the above-mentioned transitionable area is examined. The results are: A-1: 8th position A-5: 8th position A-7: 7th position A-8: 7th position A-2 (self): 8th position. At this time, the rank of the virtual case g is A-2 (sel
If it is above f), it can be said that it is a transitionable area that approaches the goal area, and if it is below it, it can be said that it is a transitionable area that moves away from the goal area. Therefore, in the above case,
It is clear that A-1 and A-5 are unknown, and that A-7 and A-8 are closer to the goal space than the present.

Therefore, the tentative determination of the closeness of the transitionable area to the goal space at this point is as follows. A-7, A-8 <A-1, A-5, A-2 (self) Next, a virtual case of rank 2 is examined in the goal area A-40. In this case, it is a virtual case f.

Then, the order of the virtual cases f in the transitionable area is examined as follows. A-1: 6th place A-5: 5th place A-7: 4th place A-8: 3rd place A-2 (self): 5th place The ranking at this time is used for the tentative determination of the closeness to the goal space described above. Considering this, it can be seen that among the transitionable areas, the area closest to the goal is A-8, the area next to the goal is A-7, and the area A-1 is an area away from the goal area. Therefore, the tentative determination of the proximity of the transitionable region to the goal space will be detailed as follows.

A-8 <A-7 <A-5, A-2 (self) <A-1 Next, a virtual case of rank 3 is examined in the goal area A-40. In this case, it is a virtual case h. Therefore, the order of the virtual case h in the transitionable area is examined. The results are as follows.

A-1: 7th place A-5: 6th place A-7: 8th place A-8: 8th place A-2 (self): 7th place At this point, the order of the distance to the goal area is known. The regions not shown are A-5 and A-2. Therefore, from the order of the virtual case h, it can be seen that A-5 is an area closer to the goal area than A-2. Therefore, the final decision on the proximity of the transitionable area to the goal area is as follows.

A-8 <A-7 <A-5 <A-2 (self) <A-1 The above general processing is performed in step S42 shown in FIG.
This is performed in the routine of S43 → S44 → S45 → S43 → S44. Also, an image diagram of this processing is shown in FIG.
Shown in. In FIG. 15, it can be seen that the transitionable area around the area to which the user currently belongs is derived and the proximity to the goal area is ranked.

Returning to FIG. 9 again, the explanation of the overall control processing is continued. After the repair operation is performed by the above-described actuator operation in step S16, a failure diagnosis is performed (step S17). This failure diagnosis is performed in step S14.
The content is the same as that described in. That is, the failure diagnosis is performed according to the subroutine shown in FIG.

Thereafter, a failure judgment is made (step S1).
8). The failure determination is performed by comparing the output of the copy density sensor Os with the value shown in FIG. 3, in the same way as in step S12. Then, as a result of the failure determination, when the failure symptom does not occur, that is, when the output value of the copy density sensor Os is within the normal image range of the failure determination reference value (FIG. 3), the repair processing is considered successful and the processing ends. .

On the other hand, if it is determined as a result of the failure determination that the failure symptom still occurs, whether the area determined by the failure diagnosis performed in step S17 has changed from the area determined in step S14. It is determined whether or not (step S19). If the belonging area has not changed, for example, in the case of the above-mentioned specific example, if the belonging area remains the area A-1 and has not changed, then the operation area of the actuator related to the repair operation is changed. It is determined (step S20). Then, if the operation area of the actuator is within the operation limit, the operation returns to step S16 again, and the repair operation is performed.

This is illustrated diagrammatically in the above-mentioned FIG. 8, and in FIG. 8, one arrow indicates the amount of change changed by one repair operation. The actuator involved in restoration is operated multiple times within its operating area,
It can be seen that the area may be changed by a plurality of operations. In step S20, when the actuator is out of the operation limit, further operation of the actuator is impossible, and in that case, it is determined whether or not the goal area can be traded off (step S2).
2). This goal area trade-off is the step S
This means that the goal area set in 13, for example, the area A-19 in the above-described specific example is changed to another area, for example, the area A-18. If the goal area cannot be traded off, this repair operation fails and the processing ends.

In step S19, it is determined whether or not the area has changed, and if the area has changed, it is determined whether or not the change is inconsistent with the transition rule (step S21). Whether or not the transition rule is inconsistent is determined based on whether or not the area has changed to the transitionable area described in the work script. That is, as a result of performing the repair operation described in a certain work script, if the transition is made to a different area instead of the transitionable area described in the work script, it is determined that the transition rule is inconsistent. If there is no contradiction, the process returns to step S14 and the failure diagnosis is performed and the process is repeated.

If it is determined in step S21 that the transition is made to a region inconsistent with the transition rule, it is considered that the diagnostic resolution has changed due to, for example, a decrease in sensor sensitivity. It is dangerous to continue working. Therefore, in that case, the work may be forcibly interrupted and ended. Alternatively, as shown in this flowchart, it is determined whether or not the goal area can be traded off (step S22), and if the tradeoff is possible, the goal area is reset (step S2).
3) The process from the failure diagnosis in step S14 may be repeated.

Further, step S in FIG. 9 described above.
In the failure diagnosis of 14 or S17, the processing as shown in FIG. 16 may be performed to improve the diagnosis accuracy. Figure 1
Describing with reference to the flowchart of FIG. 6, first, a failure diagnosis is performed (step S51), a work script is searched (step S52), and a repair operation described in the work script is executed (step S53).

Then, it is judged whether or not there is an area transition inconsistent with the transition rule as a result of executing the repair operation described in the work script (step S54). Then, it is judged whether or not all the repair operations of the work script have been executed (step S55), and if not all of them, the repair operations of the remaining work scripts are executed (step S5).
3).

In this way, when all the repair operations described in the work script are executed and the area transition described in the work script occurs, it is determined that the failure diagnosis has succeeded. If a region transition inconsistent with the transition rule occurs, it is determined that the failure diagnosis has failed. For example, if the current affiliation area is A-1 as a result of the failure diagnosis, each operation should transition to A-4, A-2, A-3. When all of these transitions occur, it is possible to confirm that the diagnosis result determined based on the sensor output, that is, the failure diagnosis that the current belonging area is A-1 is correct.

If it is determined that the diagnosis has failed in the processing shown in FIG. 16, the sensor output is incorrect, and it can be determined that the diagnosis is erroneous due to, for example, a decrease in sensor sensitivity. In the above description, it is stated that the process shown in FIG. 16 can be performed in the failure diagnosis in step S14 or S17 of the flowchart in FIG. 9, but the process shown in FIG. 16 is the process in the flowchart shown in FIG. Independently of, the failure diagnosis may be performed at an arbitrary timing to confirm that the sensor output is correct.

The electrophotographic copying machine having the self-diagnosis and self-repair functions according to this embodiment described above uses the C ++ language on MS-DOS as a system including a microcomputer or the like which constitutes the machine-embedded software section. The system capacity was about 280 KB and could be specifically configured. Also, using this system, the actual fault diagnosis and fault repair processing, that is, FIG.
It was confirmed by experiments that the time required for restoration varies depending on the state of the target at the time of failure occurrence, but is completed within a few minutes when the processing of the flowchart shown in is performed.

The actuator operation amount in the work script may be a fixed value or may be variable depending on the result of the operation. If the actuator operation amount is variable, the efficiency of the repair process can be improved. Next, another embodiment of the present invention will be described. In the above-described embodiment, the virtual case holding unit 12 (see FIG. 2) stores virtual cases for each failure symptom, and the divided areas for each failure symptom are defined. Then, it has been explained that the work script, which is knowledge representing the relationship between the actuator operation content and the area change at the time of restoration, is set in the work script table 15 (see FIG. 2).

However, without providing the work script table 15, not only the order of cases in the area but also the repair method in the area may be described for each of the divided areas, as shown in Table 8.

[0111]

[Table 8]

In Table 8, Hl, Vn in the repair method described , Vb represent the manipulated variable for each actuator parameter to be manipulated, and + in front of the manipulated variable means to raise the actuator,
-Means to lower the actuator. When Table 8 is stored in advance, the following processing of expansion plan 1 or expansion plan 2 can be performed.

That is, as the development plan 1, as shown in Table 8, the repair method is described in advance in each area. Therefore, if the belonging area is determined when a failure occurs, it is described in the belonging area. The repair method can be performed at the rate described in the repair method until the failure symptom is corrected. In other words, it is possible to perform the repair from the area to which the failure belongs at a stretch toward the goal area without treating the repair process of how the area transits.

As the expansion plan 2, the repair process can be performed as follows. First, the repair method described in the area to which the failure belongs is performed, and the repair is performed until the area transition occurs. Then, when a region transition occurs, the repair method described in the transitioned region is performed to cause the region transition. In this way, the region can be transitioned to the goal region while handling the restoration process.

FIG. 17 shows an image diagram of the above-mentioned expansion plan 1 and expansion plan 2. The development plan 1 and the development plan 2 are
Although this is an example of performing a simpler repair process than the first described example, the following example can be created as an example of performing an even simpler repair process. Figure 4
When a virtual case is generated based on the parameter model shown in (1), the following limitation is added. That is, for example, when the copy density parameter Os shows an abnormality, it is limited that the cause is always due to the change of a single parameter, and the parameter change does not affect the sense target parameter (parameter surrounded by circle in FIG. 4) without fail. I'm giving it. Then, if virtual cases are generated under the condition of, for example, the virtual cases in the failure symptom “image cast” are not eight as shown in Table 1, and the virtual cases a, e, f, and It becomes four of g.

[0116]

[Table 9]

Therefore, a restoration method is described in advance for each of these four virtual cases. Then, the repair method described in the virtual case having the highest degree of coincidence is executed according to the failure diagnosis result. Then, each time repair is executed, failure diagnosis is performed, the virtual case with the highest degree of coincidence is detected, and the repair method described in the virtual case is executed. According to such an embodiment, it is possible to provide a device that can perform processing even if the capacity of the virtual case storage unit is small.

The present invention is not specifically limited to the embodiments described above, but various concrete examples can be considered based on the scope of the claims.

[0119]

According to the present invention, when a failure symptom occurs in the image forming apparatus, the failure which is the cause can be self-diagnosed in a short time based on a virtual case. Further, the self-diagnosed failure can be repaired so that the device state moves to a desired transition area according to the operation content of the actuator, which is the operation information stored in advance, and the area that can be transitioned by the operation. Further, when a region transition occurs, it is confirmed that there is no contradiction in the region transition, so that the self-repair can be performed while confirming that the self-repair is correctly performed. Therefore, it has the ability to self-diagnose and self-repair,
In addition, it is possible to provide an image forming apparatus having a practical self-repairing function, which is compact in size for self-diagnosis and self-repairing and has a relatively high execution speed.

Further, if there is a contradiction in the area transition before the transition to the preset target area, the set target area is not appropriate, and therefore a different target area is reset and It is possible to provide an image forming apparatus that can perform restoration processing from different viewpoints and that has a wide range of restoration functions.

[Brief description of drawings]

FIG. 1 is a mechanical configuration diagram of a small electrophotographic copying machine according to an embodiment of the present invention.

FIG. 2 is a functional block diagram of the small electrophotographic copying machine shown in FIG.

FIG. 3 is a diagram showing an example of a failure determination reference value in the small electrophotographic copying machine.

FIG. 4 is a diagram of a simplified parameter model of the small electrophotographic copying machine shown in FIG.

FIG. 5 is a diagram showing an example of fuzzy membership functions of Ds, Vs, and X used at the time of image fogging.

FIG. 6 is a flowchart showing an example of processing of a pseudo failure method (IF method).

FIG. 7 is an image diagram showing the distance from the current state position p of the electrophotographic copying machine to each of the virtual cases a to h in the failure symptom “image fog”.

FIG. 8 is an illustrative view showing a relationship between a transition rule and a path in a repair process.

FIG. 9 is a flowchart showing the overall procedure of control performed in the diagnosis / restoration inference unit in the electrophotographic copying machine according to this embodiment.

FIG. 10 is a flowchart showing a specific content of failure diagnosis processing.

FIG. 11 is a diagram for explaining a neutral space in a fuzzy qualitative value.

FIG. 12 is a diagram visualizing a neutral space in a fuzzy qualitative value in a three-dimensional quantity space.

FIG. 13 is a diagram showing a presence pattern of each sensor value existing in a neutral space in a fuzzy qualitative value.

FIG. 14 is a flowchart showing a specific processing content of a repair operation.

FIG. 15 is an image diagram showing a process for determining the proximity of a transitionable area to a goal area.

FIG. 16 is a flowchart showing processing contents for improving diagnosis accuracy.

FIG. 17 is a diagram showing an image of a deployment plan 1 and a deployment plan 2 according to another embodiment of the present invention.

[Explanation of symbols]

 1 Photoconductor Drum 2 Main Charger 3 Halogen Lamp 4 Developing Device 5 Transfer / Separation Charger 2C Main Charger Controller 3C Halogen Light Level Controller 5C Transfer Charger Controller 11 Diagnostic / Repair Inference Unit 12 Virtual Case Holding Unit 13 Membership Function Generation Unit 14 Pseudo-fault Generation unit 15 Work script table 16 Virtual case compiler X Light intensity sensor Vs Surface potential sensor Ds Toner density sensor Os Copy density sensor

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Tetsuo Tomiyama 1035 Hanazonocho, Chiba City, Chiba Prefecture 2-203 (2) Inventor ▲ Yoshi ▼ Hiroyuki Kawa 804 Yonbancho, Yonbancho, Chiyoda-ku, Tokyo 804

Claims (1)

[Claims] 1. A sensor means for detecting a predetermined physical quantity or a change in a physical quantity, which is provided at a predetermined portion of the apparatus, an actuator means for performing a predetermined operation, which is provided at a predetermined portion of the apparatus, and the sensor. A conversion means for converting the detected value of the means into a fuzzy qualitative value using the fuzzy membership function, and multiple virtual cases in which the possible states of the equipment when the equipment has a failure symptom are represented by qualitative values. A stored case storage means, a fuzzy qualitative value converted by the conversion means, and a plurality of virtual cases stored in the case storage means, to diagnose a current state of the apparatus; The qualitative amount space that can be taken is divided into a plurality of areas, and the operation content of the actuator means and the operation thereof as operation information for restoration corresponding to each area. Therefore, an operation information storage unit that stores a transitionable region, a target region setting unit that sets a repair target in any one of the divided regions, and a current state of the device diagnosed by the diagnosis unit. Determines which one of the divided areas belongs to,
By reading the operation information of the area to which it belongs from the operation information storage means and operating the actuator means in accordance with the read operation information, the area to which the device state belongs shifts toward the area set by the target area setting means. Repair execution means for carrying out repair, and provided in the repair execution means, when the status of the device is changed to belong to another area by execution of the repair, it is confirmed whether or not the transitioned area is a consistent area. An image forming having a self-repairing function, which comprises: a confirming unit; and a resetting unit that resets the target region set by the region setting unit when the region transitioned by the confirming unit is inconsistent. apparatus.