Abstract: A method for establishing an alert or an alert limit for monitoring a mechanical device includes identifying faults that can be recognized by their effect on measured trends and determining a set of available sensors for monitoring the trends. Representative data are acquired and processed to obtain a statistical characterization of the trends. In addition, a parameter variance is determined for at least one alert test in accordance with the statistical characterization. The method further includes analyzing Type I and Type II error risks and associated financial penalties to adjust parameters associated with the alert test and programming an automated alert system to utilize the adjusted parameters to identify measured trends of a mechanical device that are to be brought to the attention of a human analyst.

Abstract: A method and system for evaluating whether faults detected during the testing of a gas turbine engine are related to the performance problems of the engine or to some other abnormality unrelated to engine performance. One performance parameter of the engine is evaluated under one performance condition to generate a first set of current engine data that is then compared to a first set of prior engine data to determine if there is an abnormality. If one is detected, this performance parameter is then evaluated under at least two different performance conditions to generate a second set of current engine data that is compared to a second set of prior engine data to determine if there is an abnormality. If one is detected, the abnormality is evaluated to see if the fault is unrelated to the performance of the engine.

Abstract: A method and system for evaluating whether faults detected during the testing of a gas turbine engine are related to the performance problems of the engine or to some other abnormality unrelated to engine performance. One performance parameter of the engine is evaluated under one performance condition to generate a first set of current engine data that is then compared to a first set of prior engine data to determine if there is an abnormality. If one is detected, this performance parameter is then evaluated under at least two different performance conditions to generate a second set of current engine data that is compared to a second set of prior engine data to determine if there is an abnormality. If one is detected, the abnormality is evaluated to see if the fault is unrelated to the performance of the engine.