Gas Turbine Engine Capacity Facility

Investigators: Professor Tom Povey.

Students: Chris Hambidge, Tom Gammage

Sponsor: Rolls-Royce Turbines.

The Gas Turbine Engine Capacity FacilityA new semi-transient technique has been developed and demonstrated which will allow turbine designers to measure experimentally the effective throat area of an annular cascade of nozzle guide vanes under representative conditions of Mach number, and mainstream-to-cooling-flow mass flow rate ratio.

The technique offers order-of-magnitude savings over the continuously-running approach, because it removes the need for large and very costly compressor plant. It has been demonstrated by experiment, however, that the semi-transient technique delivers the same accuracy as can be achieved using a continuously-running facility.

The Gas Turbine Engine Capacity FacilityA high mass flow rate blow-down tunnel was commissioned in Oxford, and the effective throat area of a large civil engine experimentally determined. Results are presented and compared to CFD predictions.

During a test, approximately 1,000 kg of air (stored at room temperature in a high-pressure reservoir) is allowed to flow though the turbine cascade, achieving a peak mass flow rate of approximately 30 kg s-1. The design Mach number and coolant-to-mainstream pressure ratio are exactly matched at one point (in time) during the test. A test typically lasts 30 seconds. Unlike in a continuously-running facility, where total pressure and temperature are constant during an experiment, during a semi-transient test, pressure and temperature continually fall during the experiment.

The working section of the capacity facilityThe time-history of the mass flow rate at the turbine vane throat is calculated based on a mass flow rate measurement conducted some distance upstream of the throat, using a well-calibrated mass flow rate meter. The mass flow rates of air entering and leaving intervening plena must be calculated to a high degree of accuracy, and therefore the pressure and temperature must be measured in each of these plena. The increase in the overall measurement uncertainty associated with this unsteady correction was of the order of 0.06 per cent: this is negligible compared with the overall experimental uncertainty, which is approximately 0.9 per cent.

A thermocouple probe in the main flow path