3 Day Open Conference

A Probabilistic In-Flight Thrust Estimation Process
Tuesday 18th November 2008
15.30 - 16.00
Auditorium 3

Jet engines installed thrust are estimated in-flight by deterministic processes that starting from measured parameters as pressures and temperatures, calculates, by many different ways, the intermediate turbomachine parameters up to the exhaust nozzle pressure and temperature from which, and the previously determined model nozzle coefficients, the engine thrust and air mass flow are calculated. These methods are industry standards well documented on SAE Reports AIR 1703A and AIR 5450. However, they demand deep cooperation between engine and airframe manufacturer, and are frequently subject of disagreement between parties and subject of contractual penalties. A new approach has been recently proposed on Paper SAE Brazil 2007-01-2542, the more stochastic approach, which in fact estimates the engine fan and core pressures and temperatures from initial engine air mass flow and gross thrust values. Using the Output-Error Method the values of gross thrust and air mass flow are iteratively updated by a modified Newton-Raphson algorithm minimizing the error between the fan and core calculated and measured air pressures, temperatures, and the engine fuel flow. The new technique solves in fact, by optimization, the backward formulation of the in-flight thrust determination problem. The advantages of the new technique over the traditional one is that it has stochastic properties allowing to process the noisy flight test data samples without previous data averaging in the time interval. The above referenced paper (SAE 2007-01-2542) presented the new technique and its application to a mixed flow turbofan engine. Present paper demonstrates the application of the new technique to a separate-stream turbofan showing that the results of the new and the traditional technique are very close. Although the technique has been developed and validated for real engine flight test data, the results here presented where derived from engine Deck (cycle model) data transformed to sampled data by application of noise whose characteristics have been extracted from real flight test data.

About the Speaker(s):

Joao Carlos Hoff – SFTE Member
Organization: Embraer (Brazil)
Av. Faria Lima 2170. São Jose dos Campos – SP. Brazil