Sensor data processing device, abnormality detection processing method, and program

Abstract

To realize an abnormality detection capable of adapting to changes in the environment without the need to re-learn the normal model.SOLUTION: A sensor data processing device includes one or more processors and one or more memory resources. The processor acquires time series sensor data from a sensor attached to a machine via a sensor signal transmission material, extracts a feature amount of the time-series sensor data, determines whether the working space is in a state before an environmental change or after an environmental change based on the feature amount, corrects the feature amount using correction data corresponding to a state before the environmental change when the processor determines that the working space is in a state after an environmental change, and diagnoses the presence or absence of an abnormality in the machine based on the uncorrected feature amounts or the corrected feature amounts, depending on the result of determining whether the work space is in a state before the environmental change or after the environmental change.SELECTED DRAWING: Figure 1

Images