System and method for aircraft fault detection

A method, comprises:

• receiving, by an aircraft controller, information captured by at least one aircraft data logging device, wherein the information comprises a collection of time-series flight data, from at least one aircraft subsystem, for at least one flight leg of an aircraft flight;
• grouping, by the aircraft controller, the collections of time-series flight data from the at least one aircraft subsystem into a plurality of test states, the plurality of test states being determined from a plurality of training states and one or more of the test states being different from other test states in the plurality of test states;
• generating, by the aircraft controller, at least one test transition matrix based on the plurality of test states and determine anomalous behavior of the at least one aircraft subsystem based on the at least one test transition matrix;
• and
• forecasting, by the aircraft controller, a fault within the at least one aircraft subsystem based on the anomalous behavior of the at least one aircraft subsystem determined from the at least one test transition matrix.

A method, comprising:

• capturing, with at least one aircraft data logging device, a collection of time-series flight data, from at least one aircraft subsystem, for at least one flight leg of an aircraft flight;
• and
• forecasting, with an aircraft controller, faults within the at least one aircraft subsystem based on anomalous behavior of the at least one aircraft subsystem determined from at least one test transition matrix,
• wherein the at least one test transition matrix is generated with the aircraft controller by
• grouping the collections of time-series flight data from the at least one aircraft subsystem into a plurality of test states, the plurality of test states being determined from a plurality of training states where one or more of the test states is different from other test states in the plurality of test states,
• for more than one time point in the collections of time-series flight data, identifying a current state of a subject matter expert identified test variable of the collections of time-series flight data and counting a frequency of the subject matter expert identified test variable transitions from the current state to another state or to a same state at a next sequential time point, and
• normalizing the frequency so that one or more entry in a respective cell of a respective test transition matrix is a probability number that lies between zero and one.