Diagnostic apparatus and diagnostic method

A diagnostic method for diagnosing a plurality of mechanical apparatuses including an industrial apparatus, the method being realized by a computer and comprising:

• a context information acquisition step of acquiring, from each of the plurality of mechanical apparatuses, context information corresponding to operation being performed among a plurality of pieces of context information set based on a type of operation of each of the plurality of mechanical apparatuses;
• a sensing information acquisition step of acquiring, as sensing information sensed by sensing units arranged in a target machine as a mechanical apparatus targeted for diagnosis and other mechanical apparatuses, a physical quantity of each of the target machine and the other mechanical apparatuses, the physical quantity changing according to operation of each of the target machine and the other mechanical apparatuses;
• a sensing information separation step of separating physical quantity information indicating the physical quantity of the target machine and physical quantity information indicating the physical quantities of the other mechanical apparatuses from each other based on the context information of each of the plurality of mechanical apparatuses acquired at the context information acquisition step and the sensing information of each of the target machine and the other mechanical apparatuses acquired at the sensing information acquisition step;
• a feature information extraction step of extracting feature information indicating a feature of the physical quantity information in an interval including a specific operation interval indicated by the context information of the target machine from the physical quantity information of the target machine separated at the sensing information separation step;
• and
• an anomaly determination step of determining, based on a model corresponding to the context information acquired at the context information acquisition step among one or more models corresponding to one or more pieces of context information and the feature information extracted at the feature information extraction step, whether or not the operation of the target machine is normal.