Failure time estimation device, machine learning device, and failure time estimation method

1. A failure time estimation device comprising:

a memory configured to store a machine-learned model obtained by performing machine learning using teaching data associating printer information with a failure time of a printer, the teaching data including at least one of an operation history of the printer, state information indicating a current state of the printer, and a print result image indicating a print result; and

a controller configured to obtain the printer information and estimate the failure time of the printer using the obtained printer information and the machine-learned model,

wherein the printer information includes maintenance failure information indicated by a difference between a number of maintenance executions and a number of resolving discharge failures after maintenance execution.

2. The failure time estimation device according to claim 1 , wherein

the print result image is an image indicating a print result of a predetermined pattern image,

the operation history is history of operation traced back from printing of the pattern image, and

the state information is information indicating a state of the printer at printing time of the pattern image.

3. The failure time estimation device according to claim 2 , wherein

the pattern image includes at least one of a pattern formed by a line and a pattern including a specified-size area filled with a specified color.

4. The failure time estimation device according to claim 1 , wherein

the print result image is captured by a sensor disposed on a carriage having thereon a print head including a nozzle discharging ink onto a print medium.

5. The failure time estimation device according to claim 1 , wherein

when a difference between a current time and the estimated failure time becomes less than a threshold value, the controller gives notification related to the failure time.

6. The failure time estimation device according to claim 1 , wherein

the printer information includes execution history of maintenance including at least one of flushing for discharging ink from a print head to resolve a discharge failure of ink and wiping for wiping a nozzle face of a print head.

7. The failure time estimation device according to claim 1 , wherein

the printer information includes at least one of a continuous operation time of the printer, a temperature history of the printer, and an ambient temperature history of the printer.

8. A machine learning device comprising:

a controller configured to obtain teaching data associating printer information with failure time of a printer, and to perform machine learning on a model receiving input of the printer information and outputting the failure time based on the teaching data, the teaching data comprising at least one of operation history of the printer, state information indicating a current state of the printer, and a print result image indication a print result, wherein the printer information includes maintenance failure information indicated by a difference between a number of maintenance executions and a number of resolving discharge failures after maintenance execution.

9. A failure time estimation method comprising:

storing a machine-learned model obtained by performing machine learning using teaching data associating printer information with failure time of a printer, the teaching data including at least one of operation history of the printer, state information indicating a current state of the printer, and a print result image indicating a print result; and

obtaining the printer information and determining the failure time of the printer using the obtained printer information and the machine-learned model,

wherein the printer information includes maintenance failure information indicated by a difference between a number of maintenance executions and a number of resolving discharge failures after maintenance execution.