CPC B25J 9/163 (2013.01) [G05B 23/024 (2013.01); G06F 16/2474 (2019.01); G06N 5/04 (2013.01); G06N 20/00 (2019.01); G05B 2219/50391 (2013.01)] | 6 Claims |
1. A learning data confirmation support device for confirming contamination of inappropriate data caused by a factor including at least a different setting value or a different operation from a normal one when learning data including only normal data are acquired in advance, in order to detect an anomaly of an industrial machine using machine learning, the learning data confirmation support device comprising:
a data acquisition unit that acquires measurement data including time-series data representing at least one of a predetermined state quantity or control quantity relating to control when the industrial machine is made to execute a control program for a health check by performing a predetermined operation according to a preset schedule and under a preset condition;
a display control unit that aligns a plurality of pieces of the time-series data acquired by the data acquisition unit in a direction of a time axis, and superimposes a same parameter of measurement data acquired at different times from the plurality of pieces of data of the time-series data as separate waveforms in a graph that is displayed, thereby confirming the contamination of the inappropriate data by visually listing the learning data in the form of the graph and comparing the separate waveforms;
a data storage unit that stores the time-series data acquired by the data acquisition unit; and
a data selection unit that excludes, from the data storage unit, time series data determined as inappropriate data as learning data from the plurality of pieces of time-series data displayed in the graph by the display control unit,
wherein the inappropriate data as learning data is determined from any waveform among the separate waveforms that do not overlap on a line, and the data selection unit excludes the measurement data that corresponds to the waveforms that do not overlap on the line as the inappropriate data as the learning data.
|