| abstract |
Methods and devices for detecting and predicting parameter deviations and isolating failure modes in systems that are subject to failure. In a preferred embodiment, methods are provided for use with engines, including aircraft, automobile, and industrial combustion engines. However, numerous other applications are contemplated. Such engines may be described as having monitor points having current parameter values, where the monitor points may correspond to single physical sensors or to virtual or inferred monitor points having parameter values derived from multiple sensors. Acceptable ranges, limits, and values for each of the monitor point parameters may be provided for use with the present invention. Parameters lying outside of the acceptable ranges may be said to be in deviation. Ambiguity groups, including one or more failure modes or physical causes of the parameter deviations may also be provided. Parameter deviations, after optional filtering, may generate deviation signals which may be followed by analysis of the ambiguity groups to isolate the failure mode or modes causing the deviation. Courses of engine operation ameliorating the failure mode may be suggested. Methods are also provided for projecting current trends into the future to predict deviations and isolate failure modes early, prior to actual occurrence. One preferred use for the methods is early detection and isolation of faults in aircraft engines, leading to corrective action including early preventative maintenance. |