In this paper a filtered sequential regression parameter identification technique is developed which is feasible for on-board digital computer implementation on gas turbine engines. When applied to a test second order system, this filtered sequential technique is superior to a nonfiltered sequential technique for noisy data. The practicality of the technique is demonstrated by application to data generated from a twinspool gas turbine engine model. A simplified linear set point model was identified using the two spool speeds as state variables. The model obtained with this filtered sequential regression technique matched the system state response to within 2 percent accuracy.

This content is only available via PDF.
You do not currently have access to this content.