US 11,669,068 B2
Diagnostic device, diagnostic method, and recording medium
Kyohei Suzuki, Tokyo (JP); and Tetsushi Ishida, Tokyo (JP)
Assigned to MITSUBISHI ELECTRIC CORPORATION, Tokyo (JP)
Appl. No. 17/907,974
Filed by Mitsubishi Electric Corporation, Tokyo (JP)
PCT Filed Jul. 10, 2020, PCT No. PCT/JP2020/027135
§ 371(c)(1), (2) Date Aug. 30, 2022,
PCT Pub. No. WO2022/009434, PCT Pub. Date Jan. 13, 2022.
Prior Publication US 2023/0011992 A1, Jan. 12, 2023
Int. Cl. G05B 19/4065 (2006.01)
CPC G05B 19/4065 (2013.01) [G05B 2219/50276 (2013.01)] 7 Claims
OG exemplary drawing
 
1. A diagnostic apparatus to diagnose existence of an abnormality in a tool for processing of a processing target by each of a plurality of working machines, the diagnostic apparatus comprising:
processing circuitry configured as an acquirer to acquire program identification information, tool information, and transition information at execution of processing by each of the plurality of working machines, the program identification information identifying a program executed in the working machine, the tool information indicating a type of the tool applied to the processing, the transition information indicating transition of load in the working machine from a start to an end of the processing executed by execution of the program;
the processing circuitry further configured as a diagnoser to diagnose whether an index value of the load is out of a predetermined range corresponding to a combination of the program identification information, the tool information, and machine information identifying the working machine that transmits the transition information, the index value being obtained from the transition information; and
the processing circuitry is further configured as an estimator to estimate, for at least one model of models that each are defined by the combination of the program identification information, the tool information, and the machine information, how many more times the tool of the type indicated by the tool information is to be used until the index value goes out of the range based on a history of the index value for the at least one model.