JP2021077274A - Failure prediction system and failure prediction method - Google Patents

Abstract

To accurately perform failure prediction while reducing a load.SOLUTION: A failure prediction system for predicting a failure in an information processing apparatus has: a control unit 100 that acquires respective pieces of measurement data measured by a measurement unit that measures the state of devices in the information processing apparatus; a first operation unit 1001 that performs a first operation based on a predetermined algorithm by using the respective pieces of measurement data; and a second operation unit 110 that performs a second operation based on a predetermined algorithm according to a result of the first operation. The first operation unit 1001 primarily determines the presence or absence of a sign of a failure in the information processing apparatus on the basis of a result of the first operation. When the first operation unit 1001 determines that there is the sign of failure, the second operation unit 110 secondarily determines the presence or absence of a sign of a failure in the information processing apparatus on the basis of a result of the second operation.SELECTED DRAWING: Figure 1

Description

The present invention relates to a failure prediction technique.

Conventionally, in an image forming apparatus such as a printer or a multifunction device, a failure prediction system for predicting a failure of an apparatus is known by collecting and analyzing state data of the apparatus using sensor information in the apparatus. Since the amount of data such as sensor information required for prediction is large and the calculation load is heavy, failure prediction is required to be performed efficiently while suppressing the amount of data and the calculation load.

In the technique described in Patent Document 1, first, data is received from a plurality of devices at low frequency, a first failure prediction is performed, and an instruction is instructed to transmit data at high frequency to a device that determines an abnormality. put out. Next, data is frequently received from the instructed device to perform a second failure prediction. By doing so, the load is suppressed without significantly reducing the accuracy of failure prediction.

Japanese Unexamined Patent Publication No. 2016-146020 Japanese Unexamined Patent Publication No. 2018-18411

However, with the technique described in Patent Document 1, it is difficult to accurately predict a failure based on highly random data while suppressing the load. For example, if the numerical value accidentally fluctuates in the abnormal direction at the timing of the first failure prediction, the second failure prediction, which is normally unnecessary, is performed, so that the load increases despite the normal state. It ends up. On the other hand, if the numerical value accidentally fluctuates in the normal direction at the timing of the first failure prediction, the second failure prediction is not performed and the detection of the sign of the failure is delayed.

The present invention has been made in view of such a problem, and an object of the present invention is to accurately predict a failure while suppressing a load.

In one embodiment of the present invention, the failure prediction system for predicting a failure of the information processing device includes an acquisition means for acquiring each of the measurement data measured by the measurement means for measuring the state of each device of the information processing device, and the above-mentioned acquisition means. A first calculation means that performs a first calculation based on a predetermined algorithm using each of the measurement data, and a second calculation means that performs a second calculation based on the predetermined algorithm according to the result of the first calculation. The first calculation means has two calculation means, and the first calculation means primary determines the presence or absence of a sign of failure of the information processing apparatus based on the result of the first calculation, and the second calculation means is When it is determined by the first calculation means that there is a sign of failure, the presence or absence of a sign of failure of the information processing apparatus is secondarily determined based on the result of the second calculation.

According to the present invention, failure prediction can be performed with high accuracy while suppressing the load.

It is a figure which shows the configuration example of the failure prediction system in 1st Embodiment. It is a figure which shows the structural example of the measurement data in 1st Embodiment. It is a flowchart which shows the processing flow of the 1st operation in 1st Embodiment. It is a flowchart which shows the processing flow of the 2nd operation in 1st Embodiment. It is a figure which shows the example of the inverse matrix A in which the weight of each element is biased in 2nd Embodiment. It is a flowchart which shows the processing flow of the 1st operation in 2nd Embodiment. It is a flowchart which shows the processing flow of the 2nd operation in 2nd Embodiment. It is a figure which shows the configuration example of the failure prediction system in 3rd Embodiment. It is a hardware block diagram of the information processing apparatus which can carry out 1st to 5th Embodiment.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. The following embodiments do not limit the present invention. Moreover, not all combinations of features described in the present embodiment are essential to the present invention. The present invention also includes various modifications within the scope of the gist of the present invention. Further, the present invention may be configured by combining some of the following embodiments. Further, in the following, an image forming apparatus will be used as an example, but the present invention is not limited to this, and the present invention can be applied to a general information processing apparatus.

(First Embodiment)
FIG. 1 shows a configuration example of the failure prediction system according to the first embodiment. The failure prediction system includes an image forming apparatus 10, a calculation server 11, a management server 13, and a network 12 connecting them.

The calculation server 11 performs failure prediction based on a predetermined algorithm based on the data received from the image forming apparatus 10 connected via the network 12. The failure prediction performed by the calculation server 11 is set as the second calculation, and the calculation server 11 has a second calculation unit. The calculation server 11 notifies the management server 13 connected to the image forming apparatus 10 and the network 12 of the failure prediction information obtained by the second calculation, and urges the user and the service person to perform maintenance.

The management server 13 is installed in a service center or the like and remotely manages the image forming apparatus 10 connected via the network 12. The management server 13 acquires failure prediction information from the calculation server 11 and notifies the service person of the image forming apparatus having a high possibility of failure.

The image forming apparatus 10 has a plurality of devices that operate as a control unit 100, a developing unit 101, a fixing unit 102, a toner unit 103, and a paper feeding unit 104. The image forming apparatus 10 can be, for example, an electrophotographic laser beam printer.

The control unit 100 includes a first calculation unit 1001, an external communication unit 1002, and a measurement information storage unit 1003, and performs calculations related to overall control and failure prediction of the image forming apparatus 10. The control unit 100 is also referred to as a control device. The calculation related to the failure prediction performed by the first calculation unit 1001 of the image forming apparatus 10 is defined as the first calculation.

The control unit 100 controls the developing unit 101, the fixing unit 102, the toner unit 103, and the paper feeding unit 104, and carries out a series of image forming processes such as development, fixing, toner supply, and paper feeding. Further, the control unit 100 receives the information (measurement data) measured by the measurement unit in each unit, which will be described later, and holds it in the measurement information storage unit 1003. The control unit 100 primaryly determines the possibility of a failure sign of the image forming apparatus 10 by the first calculation using the measurement information stored in the measurement information storage unit 1003. The measurement information storage unit 1003 may be a dedicated storage area, or may be configured such that the control unit 100 allocates a part of a large-capacity DRAM, SRAM, or the like.

The external communication unit 1002 communicates with the calculation server 11 via the network 12, and transmits the result obtained by the first calculation to the calculation server 11.

The developing unit 101 carries out the developing step in the image forming process based on the instruction from the control unit 100. Further, the developing unit 101 has a motor speed and torque measuring unit 1010, measures the speed and torque of the photosensitive drum drive motor provided inside the developing unit 101 based on an instruction from the control unit 100, and collects the measurement data. It is sent to the control unit 100.

The fixing unit 102 carries out the fixing step in the image forming process. The fixing unit 102 has a temperature measuring unit 1020, measures the temperature of the fixing device in the fixing unit 102 based on an instruction from the control unit 100, and sends the measurement data to the control unit 100.

The toner unit 103 supplies toner onto a charged photosensitive drum in the developing process in the image forming process. Further, the toner unit 103 has a toner measuring unit 1030, measures the remaining amount of toner based on an instruction from the control unit 100, and sends the measurement data to the control unit 100.

The paper feed unit 104 feeds and conveys paper in the image forming process. The paper feed unit 104 includes a motor speed and torque measuring unit 1040 for measuring the motor speed and torque, a paper transport angle measuring unit 1041 for measuring the paper angle during paper transport, an acoustic measuring unit 1042 for measuring sound, and a paper type and paper. It has a paper type for measuring the thickness and a paper thickness measuring unit 1043. Each measurement unit performs measurement based on an instruction from the control unit 100, and sends the measurement data to the control unit 100.

The configuration of the failure prediction system and the image forming apparatus 10 and each measurement unit are not limited to the above. For example, it may have a measuring unit that measures a voltage, a current value, a motor rotation speed, and the like.

Further, the failure prediction system may include two or more image forming devices 10.

Subsequently, a failure prediction algorithm using the Mahalanobis distance disclosed in Patent Document 2 used in the first calculation and the second calculation will be described.

Mahalanobis distance measures the measurement data of multiple measurement items (motor speed, torque, voltage, current value, etc.) to be monitored at predetermined time intervals, and calculates the average and distribution of the measurement data for each measurement item. After standardization, it is calculated from the correlation of measurement data for each measurement item. Then, when the calculated Mahalanobis distance exceeds a predetermined threshold value, it can be determined that the image forming apparatus 10 has a sign of failure.

The failure prediction algorithm in this embodiment is not limited to this, and may be, for example, an algorithm using a neural network.

Hereinafter, the method of calculating the Mahalanobis distance will be described in detail.

<Preparation>
First, the normal state is defined from the accumulated measurement data of the normal state for a plurality of measurement items to be monitored. The definition of the normal state is performed before the failure prediction, and may be defined for each image forming apparatus, or may be defined for the same model of image forming apparatus connected to the network 12.

Specifically, the average μ and standard deviation S for each measurement item are calculated. Further, the correlation coefficient r for each measurement item is obtained and used as the correlation coefficient matrix R. The following equation (1) shows the correlation coefficient matrix R when the types of measurement items to be monitored are k.

Next, the inverse matrix A of the correlation coefficient matrix R is obtained. The following equation (2) shows the inverse matrix A of the correlation coefficient matrix R of the equation (1).

<Calculation of Mahalanobis distance>
Next, a method of calculating the Mahalanobis distance D ^ 2 from the measurement data of k measurement items when actually performing failure prediction after the preliminary preparation will be described.

FIG. 2 shows a configuration example of measurement data in this embodiment. It is assumed that the measurement items are, for example, the monitoring targets in the paper feed unit 104, and there are k types such as motor speed, torque, voltage, and current value. The average μ and the standard deviation S are numerical values in the measurement data in the normal state calculated in advance for each measurement item. The measurement data is measured values of k measurement items acquired at predetermined time intervals, for example, acquired once for each page printed by the image forming apparatus. The number of data sets (1 to n) in the figure corresponds to the number of acquisitions of measurement data. That is, in FIG. 2, the data set 1 shows the measurement value of the first time, and shows an example in which the measurement data for n times is acquired.

_{In order to obtain the Mahalanobis distance D i} ^ 2 of the i-th data set, first, the measurement data X is standardized for each measurement item using the following equation (3). That is, the standardized variable U for each measurement item is obtained.

_{Next, the Mahalanobis distance D i} ^ 2 is calculated by the following equation (4) using the inverse matrix A of the equation (2).

FIG. 3 shows a flowchart of the processing flow of the first calculation in the present embodiment. The series of processes shown in the flowchart is performed by the CPU of the information processing device expanding the control program stored in the ROM or HDD into the RAM and executing it. Alternatively, some or all the functions of the steps in the flowchart may be realized by hardware such as an ASIC or an electronic circuit. The symbol "S" in the description of the flowchart means a "step" in the flowchart. The same applies to other flowcharts. Hereinafter, the process related to the first calculation in the present embodiment will be described in detail with reference to FIG. Here, the first calculation performed from the start to the end of printing by the image forming apparatus will be described.

First, in S300, the control unit 100 determines whether or not a predetermined time interval has elapsed after printing is started. If the predetermined time interval has elapsed, the process proceeds to S301. On the other hand, if the predetermined time interval has not elapsed, the determination process is repeated. The predetermined time interval is such that after printing one page, after finishing one print job, after printing multiple pages, after finishing multiple print jobs, or multiple times during printing of one page, signs of failure are likely to occur in advance. It can be a fixed time interval.

In S301, the control unit 100 gives an instruction to acquire measurement data to each unit.

In S302, the control unit 100 confirms that the measurement data of each unit has been written in the measurement information storage unit 1003, and reads out the measurement data. That is, the control unit 100 acquires the measurement data.

In S303, the control unit 100 uses the measurement data read by the first calculation unit 1001 and the mean μ and standard deviation S for each measurement item calculated in advance to standardize variables for each measurement item using the equation (3). Calculate U.

In S304, the control unit 100 determines whether or not the standardized variable U for each measurement item calculated by the first calculation unit 1001 satisfies a predetermined condition. By doing so, the control unit 100 primarily determines whether or not there is a sign of failure. When the standardized variable U satisfies a predetermined condition, it is determined that there is a sign of failure. The predetermined condition can be, for example, one of the standardized variables U of k types of measurement items having measurement data exceeding 3σ, or a plurality of measurement data exceeding 2σ. If the predetermined conditions are met, it is judged that there is a sign of failure. By making a judgment using the standardized measurement data in this way, it is possible to quantitatively and easily measure how much the numerical value that stands out from the measurement data in the normal state is measured. However, the above determination is not limited to this. Further, for example, by focusing on a specific measurement item having a large influence among k types of measurement items, the determination accuracy by the first calculation can be improved. If it is determined by the above determination that the standardized variable U has a sign of failure (that is, a predetermined condition is satisfied), the process proceeds to S305, and if not, the process proceeds to S306.

In S305, the control unit 100 transmits to the calculation server 11 a failure prediction instruction by the second calculation and data indicating the calculated standardized variable U for each measurement item via the external communication unit 1002. ..

In S306, the control unit 100 determines whether or not printing is completed. If printing is not completed, the process returns to S300 and the process is repeated. On the other hand, when printing is completed, the process is terminated.

In the above-mentioned processing flow, the control unit 100 has been described so as to alternately perform the measurement data acquisition instruction and the determination of the possibility of failure sign, but the present invention is not limited to this. For example, the control unit 100 may periodically issue only the acquisition instruction of the measurement data of S301 after the start of printing, and may execute the processes of S302 to S305 after the end of printing.

FIG. 4 shows a flowchart of the processing flow of the second calculation in the present embodiment. Hereinafter, the process related to the second calculation in the present embodiment will be described in detail with reference to FIG. The calculation server 11 that executes the second calculation is in a standby state until it receives an instruction for failure prediction by the second calculation from any of the plurality of image forming devices 10 connected to the network 12 and starts processing.

In S401, the calculation server 11 receives from the image forming apparatus 10 the instruction of failure prediction by the second calculation and the data indicating the standardized variable U for each measurement item. That is, the calculation server 11 receives a part of the calculation result calculated in the process of the primary determination processing executed by the control unit 100 in the image forming apparatus 10 from the image forming apparatus 10.

In S402, the calculation server 11 calculates the Mahalanobis distance from the received standardized variable U and the inverse matrix A calculated in advance by the second calculation unit 110 using the equation (4).

In S403, the calculation server 11 determines whether or not the calculated Mahalanobis distance exceeds a predetermined threshold value by the second calculation unit 110. This is also called a secondary judgment. If the Mahalanobis distance exceeds a predetermined threshold value, it is determined that there is a sign of failure, and the process proceeds to S404. On the other hand, if not, the process proceeds to S405.

In S404, the calculation server 11 notifies the image forming apparatus 10 and the management server 13 via the network 12 by the external communication unit 1002 that the image forming apparatus has a sign of failure. The image forming apparatus 10 and the management server 13 that have received the notification urge the user and the service person to perform maintenance, respectively. For example, the image forming apparatus 10 and the management server 13 may display a message to that effect to the user or the service person.

In S405, the calculation server 11 notifies the image forming apparatus 10 via the network 12 by the external communication unit 1002 that there is no sign of failure. The image forming apparatus 10 that has received the notification may display a message to that effect to the user, or may simply receive the notification and do nothing.

Next, the calculation server 11 ends the process, returns to the standby state, and waits for a new failure prediction instruction from the image forming apparatus.

As described above, according to the present embodiment, the image forming apparatus first determines the sign of failure by the first calculation, and the calculation server 11 determines only the measurement data that is likely to indicate the sign of failure. On the other hand, the secondary determination can be performed by the second calculation. By doing so, the frequency of performing the second calculation having a heavy calculation load can be reduced, and the calculation load on the calculation server 11 can be suppressed. Further, since the calculation load of the calculation server 11 is suppressed, it is not necessary to reduce the transmission frequency of the measurement data. In addition, since it is always determined whether or not the second calculation can be performed, it is possible to accurately detect a sign of failure even with highly random data. Furthermore, since the failure prediction calculation algorithm is divided into two calculation processes and the second calculation is performed using at least a part of the first calculation result, the overall calculation amount is reduced and the load is suppressed. can do.

(Second embodiment)
In the first embodiment, a method of suppressing the overall calculation load by dividing the failure prediction calculation algorithm into two calculation processes has been described. On the other hand, in the second embodiment, a method of suppressing the calculation load of failure prediction by a method of performing the same calculation process twice without dividing the failure prediction calculation algorithm will be described.

Also in this embodiment, the failure prediction using the Mahalanobis distance as the calculation algorithm for the failure prediction will be described, but the failure prediction algorithm is not limited to this.

In some failure prediction algorithms including Mahalanobis distance and neural networks, the amount of calculation tends to increase synergistically according to the number of input measurement items. For example, when the number of measurement items is two, the formula (4) can calculate the Mahalanobis distance by multiply-accumulate four times, but when the number of measurement items is five, it is 25 times, and when the number of measurement items is eight. It is necessary to perform the product-sum operation 64 times. Therefore, the arithmetic processing based on the failure prediction algorithm using many measurement items becomes a heavy load.

On the other hand, since the Mahalanobis distance is calculated by the product-sum calculation of each element a of the inverse matrix A and the standardized variable U, it is the measurement data multiplied by the element having a large absolute value that greatly affects the determination result.

Therefore, in the present embodiment, as the first calculation, only the product-sum calculation of the measurement items having a large influence on the determination result is performed to calculate the Mahalanobis distance with a small number of input data, so that the calculation amount is suppressed and the accuracy is high. Make a primary judgment of the possibility of failure sign. Next, as a second calculation, the Mahalanobis distance is calculated using the remaining measurement data to make a secondary determination.

FIG. 5 shows an example of the inverse matrix A in which the weights of each element are biased in the present embodiment. In this example, for the sake of explanation, there are four types of measurement items for failure prediction, the absolute values of the three elements a13, a31, and a33 are relatively large, and the weight occupies more than 60% of the total.

FIG. 6 shows a flowchart of the processing flow of the first calculation in the present embodiment. Hereinafter, the process related to the first calculation in the present embodiment will be described in detail with reference to FIG.

Since the processing of S600 and S601 is the same as the processing of S300 and S301 in the first embodiment, the description thereof will be omitted.

In S602, the control unit 100 confirms that the measurement data of each unit is written in the measurement information storage unit 1003, and among them, some measurement data having a large weight (measurement items 1 and 3 in the example of FIG. 5). Is read.

In S603, the control unit 100 calculates the Mahalanobis distance from the measurement data read in S602 by the first calculation unit 1001 according to the equations (3) and (4). The standardized variable U for each measurement item, which is the calculation result of the equation (3), may be stored in the measurement information storage unit 1003 for use in the second calculation in the calculation server 11.

In S604, the control unit 100 determines whether or not the Mahalanobis distance calculated in S603 exceeds a predetermined threshold value by the first calculation unit 1001. That is, the primary determination is performed. If the Mahalanobis distance exceeds a predetermined threshold value, it is determined that there is a sign of failure, and the process proceeds to S605. On the other hand, if not, the process proceeds to S607.

In S605, the control unit 100 reads the remaining measurement data that was not read in S602 from the measurement information storage unit 1003.

In S606, the control unit 100 instructs the calculation server 11 to predict the failure by the second calculation, the standardized variable which is the intermediate result of the first calculation in S603, and the remaining measurement data read in S605. Is transmitted via the external communication unit 1002. By transmitting the standardized variable, it is possible to omit a part of the processing of the equation (3) in the second operation, but the measurement data read in S602 is read again and transmitted to the calculation server 11. You may.

In S607, the control unit 100 determines whether or not printing is completed. If printing is not completed, the process returns to S600 and the process is repeated. On the other hand, when printing is completed, the process is terminated.

In this embodiment as well, the control unit 100 may periodically issue only the acquisition instruction of the measurement data of S601 after the start of printing, and may execute the processes of S602 to S606 after the end of printing.

FIG. 7 shows a flowchart of the processing flow of the second calculation in the present embodiment. Hereinafter, the process related to the second calculation in the present embodiment will be described in detail with reference to FIG. 7.

The calculation server 11 that executes the second calculation receives an instruction for failure prediction by the second calculation from any of the plurality of image forming devices 10 connected to the network 12 and starts processing, the first embodiment. It is in a standby state as well as.

In S701, the calculation server 11 receives the failure prediction instruction by the second calculation, the standardized variable calculated by the first calculation, and the remaining measurement data from the image forming apparatus 10.

In S702, the calculation server 11 calculates the standardized variable U for each measurement item from the remaining measurement data received and the average μ and standard deviation S for each measurement item calculated in advance by the calculation shown in the equation (3). To do. The average μ and standard deviation S for each measurement item may be received from the image forming apparatus 10 or may be stored in advance in the calculation server 11.

In S703, the calculation server 11 calculates the Mahalanobis distance from the received standardized variable, the calculated standardized variable, and the preset inverse matrix A according to the equation (4).

Since the processes of S704 to S706 are the same as the processes of S403 to 405 in the first embodiment, the description thereof will be omitted.

As described above, according to the present embodiment, the first calculation is performed focusing on the measurement items having a large weight on the failure prediction, and the failure prediction including other calculation items is performed by the second calculation. By doing so, the overall computational load on failure prediction can be suppressed. This embodiment is effective when the correlation between some measurement items is very strong or when it is difficult to predict the possibility of failure sign from the intermediate data of the failure prediction algorithm.

(Third Embodiment)
In the first and second embodiments, the control unit 100 in the image forming apparatus 10 performs the first calculation, and the calculation server 11 connected via the network 12 performs the second calculation to predict the failure. A method of reducing such a calculation load has been described. On the other hand, in the third embodiment, a method will be described in which the development unit 101, the fixing unit 102, the toner unit 103, and the paper feed unit 104 each perform the first calculation, and the control unit 100 performs the second calculation. ..

FIG. 8 shows a configuration example of the failure prediction system according to the present embodiment. The failure prediction system includes an image forming apparatus 10, a management server 13, and a network 12 connecting them. In the image forming apparatus 10 of the present embodiment, each of the developing unit 101, the fixing unit 102, the toner unit 103, and the paper feeding unit 104 has first calculation units 1011, 1021, 1031, and 1044 that perform the first calculation. .. That is, each unit has a CPU (not shown), and the CPU functions as a first calculation unit that performs the first calculation. Further, in the present embodiment, the control unit 100 has a second calculation unit 110.

Each unit receives an instruction from the control unit 100, acquires measurement data at predetermined time intervals by each measurement unit, and performs the first operation shown in the first or second embodiment by each first calculation unit. Do. In addition, each unit primaryly determines the presence or absence of a sign of failure based on the result of the first calculation by each first calculation unit, and controls the judgment result together with the result of the first calculation or the measurement data. Send to 100. The result of the first calculation or the measurement data may be transmitted to the control unit 100 without each unit making a determination, and the control unit 100 may first determine whether or not there is a sign of failure.

When it is determined that there is a sign of failure based on the result of the first calculation, the control unit 100 receives the result of the first calculation or the measurement data together with a flag indicating the execution of the second calculation for the unit. Is written in the measurement information storage unit 1003.

On the other hand, when it is determined that there is no sign of failure, the control unit 100 writes a flag indicating that the second calculation is not performed to the unit in the measurement information storage unit 1003. When it is determined that there is no sign of failure, the control unit 100 does not have to write the result of the first calculation and the measurement data to the measurement information storage unit 1003.

Next, the control unit 100 refers to the flag of each unit written in the measurement information storage unit 1003, and determines whether or not the second calculation is performed. When the flag indicates the execution of the second operation, the control unit 100 reads the result of the first operation corresponding to the flag or the measurement data from the measurement information storage unit 1003, and in the first or second embodiment. Perform the second operation shown.

Next, when the control unit 100 determines that there is a sign of failure based on the result of the second calculation, the external communication unit 1002 notifies the management server 13 of the sign of failure via the network 12 and provides a service. Encourage the man to maintain. Further, the control unit 100 displays a notification that there is a sign of failure on a display unit (not shown) of the image forming apparatus 10, and urges the user to perform maintenance.

As described above, according to the present embodiment, it is possible to suppress the calculation load even when the failure prediction is performed only by the image forming apparatus without using an external server.

(Fourth Embodiment)
In the first and second embodiments, it is determined whether or not there is a sign of failure depending on whether or not the result of the first calculation satisfies a predetermined condition or exceeds a predetermined threshold value. In this embodiment, an embodiment in which a predetermined condition or a predetermined threshold value is changed depending on the state or operation mode of the image forming apparatus 10 will be described.

In the present embodiment, the determination result of the failure sign by the first first calculation is stored, and when the determination result is the failure sign, the above predetermined condition or the threshold value is set to the failure sign. Change to make it easier to judge. If it is determined that there is a sign of failure based on the result of the first calculation, but there is no sign of failure based on the result of the second calculation, there is no sign of failure, but the sign is likely to appear. there is a possibility. In such a case, in the present embodiment, it is possible to prevent the sign from being overlooked in the first calculation due to an accidental change in the measurement data, and to improve the accuracy of failure prediction.

In the present embodiment, by storing some determination conditions or threshold values in advance, it is possible to read and switch the determination conditions or threshold values according to the state and operation mode of the image forming apparatus 10.

Further, the image forming apparatus 10 has an operation mode in which a failure is likely to occur due to an abnormal state that is milder than that of ordinary paper, such as a deviation in transport speed, such as thin paper printing. In addition, there are operation modes that place a heavy load on each unit, such as thick paper printing and continuous printing that exceeds a predetermined number of pages. When printing in such an operation mode, the failure prediction may be performed by changing the determination condition or the threshold value as described above. Alternatively, when the unit or load used is different from the operation mode in which the previous failure prediction was performed, such as when changing monochrome or color or when changing the paper quality, the condition or threshold value is changed as described above to predict the failure. It may be carried out.

By doing so, the frequency of performing the second calculation increases, but the sign of failure can be detected with high accuracy.

(Fifth Embodiment)
In the first and second embodiments, measurement data is acquired according to a predetermined time interval and the first calculation is performed to determine whether or not there is a sign of failure. In this embodiment, an embodiment in which a predetermined time interval is changed depending on the state and the operation mode of the image forming apparatus 10 will be described.

In the present embodiment, the determination result of the failure sign by the first first calculation is stored, and when the determination result is a failure sign, the predetermined time interval is changed so as to be shortened. As described above, if it is determined that there is a sign of failure based on the result of the first calculation, but there is no sign of failure based on the result of the second calculation, there is no sign of failure, but a sign is issued. It may be in an easy situation. In such a case, in the present embodiment, by increasing the frequency of execution of the first calculation, it is possible to prevent accidental changes in the measurement data from overlooking a sign in the first calculation and improve the accuracy of failure prediction. it can.

Further, the image forming apparatus 10 has an operation mode in which a failure is likely to occur due to an abnormal state that is milder than that of ordinary paper, such as a deviation in transport speed, such as thin paper printing. In addition, there are operation modes that place a heavy load on each unit, such as thick paper printing and continuous printing that exceeds a predetermined number of pages. When printing in such an operation mode, the predetermined time interval may be shortened as described above to increase the frequency of performing the first calculation. Alternatively, when the unit or load used is different from the operation mode in which the previous failure prediction was performed, such as when changing monochrome or color or when changing the paper quality, the predetermined time interval is shortened as described above. The frequency of the operation of 1 may be increased.

By doing so, the frequency of performing the first calculation increases, but the sign of failure can be detected with high accuracy.

(Hardware configuration)
With reference to FIG. 9, the hardware configuration of the information processing apparatus 900 capable of carrying out the first to fifth embodiments described above will be described. The information processing device 900 includes a CPU 911, a ROM 912, a RAM 913, an auxiliary storage device 914, a display unit 915, an operation unit 916, a communication I / F 917, and a bus 918.

The CPU 911 can control the entire information processing apparatus 900 by using computer programs and data stored in the ROM 912 and the RAM 913. The information processing apparatus 900 may have one or more dedicated hardware different from the CPU 911, and the dedicated hardware may execute at least a part of the processing by the CPU 911. Examples of dedicated hardware include ASICs (application specific integrated circuits), FPGAs (field programmable gate arrays), and DSPs (digital signal processors). The ROM 912 stores programs and the like that do not require changes. The RAM 913 temporarily stores programs and data supplied from the auxiliary storage device 914, data supplied from the outside via the communication I / F 917, and the like. The auxiliary storage device 914 is composed of, for example, a hard disk drive or the like, and stores various data such as image data and audio data.

The display unit 915 is composed of, for example, a liquid crystal display, an LED, or the like, and displays a GUI (Graphical User Interface) for the user to operate the information processing device 900. The operation unit 916 is composed of, for example, a keyboard, a mouse, a joystick, a touch panel, or the like, and inputs various instructions to the CPU 911 in response to an operation by the user.

The communication I / F 917 is used for communication with an external device of the information processing device 900. For example, when the information processing device 900 is connected to an external device by wire, a communication cable is connected to the communication I / F 917. When the information processing device 900 has a function of wirelessly communicating with an external device, the communication I / F 917 includes an antenna. The bus 918 connects each part of the information processing device 900 to transmit information.

In FIG. 9, it is assumed that the display unit 915 and the operation unit 916 exist inside the information processing device 900, but at least one of the display unit 915 and the operation unit 916 exists as another device outside the information processing device 900. You may be doing it. In this case, the CPU 911 may operate as a display control unit that controls the display unit 915 and an operation control unit that controls the operation unit 916.

(Other embodiments)
The present invention supplies a program that realizes one or more functions of the above-described embodiment to a system or device via a network or storage medium, and one or more processors in the computer of the system or device reads and executes the program. It can also be realized by the processing to be performed. It can also be realized by a circuit (for example, ASIC) that realizes one or more functions.

10 Image forming device 11 Calculation server 12 Network 13 Management server 100 Overall control unit 101 Development unit 102 Fixing unit 103 Toner unit 104 Paper feed unit 110 Second calculation unit 1001 First calculation unit 1002 External communication unit 1003 Measurement information storage unit 1010 Motor Speed and torque measurement unit 1020 Temperature measurement unit 1030 Toner measurement unit 1040 Motor speed and torque measurement unit 1041 Paper transport angle measurement unit 1042 Acoustic measurement unit 1043 Paper type and thickness measurement unit

Claims (19)

It is a failure prediction system that predicts the failure of information processing equipment.
An acquisition means for acquiring each of the measurement data measured by the measurement means for measuring the state of each device of the information processing device, and
A first calculation means for performing a first calculation based on a predetermined algorithm using each of the measurement data, and a first calculation means.
It has a second calculation means for performing a second calculation based on the predetermined algorithm according to the result of the first calculation.
The first calculation means primarily determines whether or not there is a sign of failure of the information processing apparatus based on the result of the first calculation.
When the first calculation means determines that there is a sign of failure, the second calculation means secondarily determines whether or not there is a sign of failure of the information processing apparatus based on the result of the second calculation. A failure prediction system characterized by making a judgment. The second calculation means is characterized in that the second calculation is executed using at least a part of the data calculated in the process of the first calculation executed by the first calculation means. The failure prediction system according to claim 1. The first calculation means executes a part of the operations based on the predetermined algorithm as the first calculation.
The second calculation means uses the result of the first calculation as the second calculation for the remaining operations that the first calculation means has not executed among the operations based on the predetermined algorithm. The failure prediction system according to claim 1 or 2, wherein the failure prediction system is executed. The first calculation means executes the first calculation using a part of the measurement data acquired.
The failure prediction system according to claim 1 or 2, wherein the second calculation means executes the second calculation by using the remaining measurement data of the acquired measurement data. The information processing apparatus performs the first calculation by the first calculation means.
The second calculation by the second calculation means is characterized in that a server connected to the information processing device via a network receives the result of the first calculation from the information processing device. The failure prediction system according to any one of claims 1 to 4. The first calculation by the first calculation means is performed by each device of the information processing apparatus.
The failure prediction system according to any one of claims 1 to 4, wherein the second calculation by the second calculation means is performed by a control device of the information processing apparatus that controls each device. .. The failure prediction system according to any one of claims 1 to 6, wherein the first calculation means changes the determination condition of the primary determination. The failure prediction system according to claim 7, wherein the determination condition of the primary determination is changed according to the result of the previous primary determination that there is a sign of failure. The failure prediction system according to claim 7, wherein the determination conditions for the primary determination are changed according to the operation mode of the information processing apparatus. The failure prediction system according to claim 9, wherein the operation mode is changed when operating in an operation mode of continuous printing, thin paper printing, or thick paper printing exceeding a predetermined number of pages. The failure according to claim 7, wherein the determination condition of the primary determination is changed when the information processing apparatus operates in an operation mode different from the operation mode when the previous primary determination is performed. Prediction system. The failure prediction system according to any one of claims 1 to 11, wherein the first calculation means changes the frequency of executing the first calculation. The failure prediction system according to claim 12, wherein the frequency of executing the first calculation is changed according to the result of the previous primary determination that there is a sign of failure. The failure prediction system according to claim 12, wherein the frequency of executing the first calculation is changed according to an operation mode of the information processing apparatus. The failure prediction system according to claim 14, wherein the operation mode is an operation mode of continuous printing, thin paper printing, or thick paper printing exceeding a predetermined number of pages. The claim is characterized in that the frequency of executing the first calculation is changed when the information processing apparatus is operating in an operation mode different from the operation mode when the previous primary determination is performed. Item 14. The failure prediction system according to item 14. The failure prediction system according to any one of claims 1 to 16, wherein the information processing device is an image forming device. The failure prediction system according to any one of claims 1 to 17, wherein the predetermined algorithm is an algorithm for calculating the Mahalanobis distance. It is a failure prediction method that predicts the failure of information processing equipment.
An acquisition step of acquiring each of the measurement data measured by the measurement means for measuring the state of each device of the information processing device, and
A first calculation step of performing a first calculation based on a predetermined algorithm using each of the measurement data, and
A second calculation step of performing a second calculation based on the predetermined algorithm according to the result of the first calculation is included.
In the first calculation step, the presence or absence of a sign of failure of the information processing apparatus is primarily determined based on the result of the first calculation.
When it is primarily determined in the first calculation step that there is a sign of failure, in the second calculation step, the presence or absence of a sign of failure of the information processing apparatus is two based on the result of the second calculation. A failure prediction method characterized in that the next determination is made.

Images