Failure models for embedded analytics and diagnostic/prognostic reasoning
AssigneeSiemens Aktiengesellschaft
InventorsJustinian Rosca, Steffen Lamparter
Life of this patent
Estimating the term…
Abstract
A computer-implemented method for detecting faults and events related to a system includes receiving sensor data from a plurality of sensors associated with the system. A hierarchical failure model of the system is constructed using (i) the sensor data, (ii) fault detector data, (iii) prior knowledge about system variables and states, and (iii) one or more statistical descriptions of the system. The failure model comprises a plurality of diagnostic variables related to the system and their relationships. Probabilistic reasoning is performed for diagnostic or prognostic purposes on the system using the failure model to derive knowledge related to potential or actual system failures.
Claims17
A computer-implemented method for detecting faults and events related to an industrial system, the method comprising:
- receiving or retrieving sensor data from a plurality of sensors associated with the industrial system;
- constructing a hierarchical failure model of the industrial system using (i) the sensor data, (ii) fault data received from a plurality of fault detectors associated with the industrial system, and (iii) one or more statistical models of the industrial system, wherein the failure model comprises a plurality of diagnostic variables related to the industrial system and their relationships;
- and
- performing probabilistic reasoning for diagnostic or prognostic analytics on the industrial system using propagation of uncertainties in the failure model to derive knowledge related to potential or actual industrial system failures;
- wherein the derived knowledge related to potential or actually system failures comprises a list of industrial system components ranked according to likelihood of causing the potential or actually industrial system failures.
A system for detecting faults and events related to an industrial system, the system comprising:
- at least one processor;
- and
- a non-transitory, computer-readable storage medium in operable communication with the at least one processor, wherein the computer-readable storage medium contains one or more programming instructions that, when executed, cause the at least one processor to:
- receive sensor data from a plurality of sensors associated with the industrial system;
- receive fault data from a plurality of fault detectors associated with the industrial system;
- receive one or more statistical models of the industrial system;
- construct a hierarchical failure model of the industrial system using the sensor data, the fault data and one or more statistical models of the industrial system, wherein the failure model comprises a plurality of diagnostic variables related to the industrial system;
- and
- perform probabilistic reasoning for diagnostic or prognostic analytics on the industrial system using propagation of uncertainties in the failure model to derive knowledge related to potential or actual industrial system failures.
Dependencies are inferred from the claim text.