CPC G05B 23/0254 (2013.01) [G06F 21/552 (2013.01); G06F 21/57 (2013.01); G06N 5/04 (2013.01); G06N 20/00 (2019.01); G06F 2221/034 (2013.01)] | 15 Claims |
1. A method, comprising:
obtaining state data of an industrial automation system, wherein the state data describes a current operating state of a respective actuator of the industrial automation system and is recorded in one or more log files;
obtaining sensor data describing an environmental influence of the industrial automation system, wherein the sensor data quantifies a physical measurement;
obtaining control data for one or a plurality of actuators of the industrial automation system which bring about the environmental influence, wherein the control data describes a manner and/or an intensity of the environmental influence of the respective actuator; and
carrying out a comparison between the state data, the sensor data, and the control data using a model to output a result signal indicative of a probability of an impairment to an integrity of the industrial automation system, wherein the model is empirically determined by means of techniques of machine learning;
wherein the comparison takes account of a deviation of the environmental influence from a reference obtained on a basis of the model.
|