Applications of the Cycle Deck Models for Test Cell Performance Evaluation and Engine Simulation, 09-R9705Printer Friendly Version
Inclusive Dates: 04/02/07 Current
Background - The current practice in U.S. Air Force jet engine testing is to verify, correlate and compare engine test cells using "sweeper" jet engines and "gold-plate" engines. Also, the current practice of verifying engine test software prior to and during field installation is to perform software injection of interpolated values based on observed data, as well as manual signal checkout, and finally live engine runs. The use of a high-fidelity, repeatable dynamic engine signal simulator based on an engine cycle model would provide an alternate means of performing these tasks, with the potential of significantly reducing fuel, personnel, and other costs related to the shipment and use of the sweeper and gold engines. The typical engine cycle model is a computational software model of a gas turbine engine for use in quantifying theoretical performance at different engine stages based on gas flow path calculations.
Another common practice in engine test cells is validation of sensor data during engine testing. All engine performance and engine acceptance calculations are based on sensor data corrected to "standard day" readings. Therefore accurate and consistent sensor measurements are a necessity. This is most often accomplished by measurement averaging and sensor redundancy.
Approach - The approach used takes advantage of an existing engine cycle model for the T56 engine. The first objective will be to develop a cycle model-based engine simulator. Investigation into the theoretical engine model and actual engine performance will need to be accomplished to determine if a relationship can be established. If this can be accomplished, a practical and useable error band within which the simulator must operate will be developed. Statistical analysis of a collection of actual correlation factors for a given engine over multiple test cells will also be performed to quantify the effect correlation factors have on engine performance compared to the theoretical model.
For engine sensor validation, algorithms will be developed based on the cycle model for real-time validation of acquired data in a typical test cell software program.
Accomplishments - Development of a generic code for the cycle-deck model has been completed up to the simulator hardware interface. The software architecture will allow future engine models to be incorporated with minimal change to the overall simulator. The relationship between the cycle deck model and actual engine data has been established. Equations have been developed to relate cycle model parameters to engine sensor parameters for depot-type test cells, and will soon be completed for field-level test cells. The project team believes that the development of these equations has not been established in such a way before. The team has also discovered some potential errors in the way that engine efficiency is currently calculated in depot-type test cells. Time-permitting, this finding will be explored, as this would be helpful to potential clients. Analysis of the effect of correlation factors to comparison between theoretical and real engine parametric data is under way. Comparison of simulator output data to engine test data is near completion.