Menu

Turbo-shaft set-up

1. Introduction

The most critical and essential components are compressor, turbine and bearings: compressor and turbine because of their complex geometry and their impact on the cycles efficiency, the bearings because of the high required rotor speed.

To test the performance of these critical components without the overhead of the complete system, a special set-up is built, containing only compressor, turbine, diffuser and bearings (see figures 1 and 2). This set-up is called a turbo-shaft set-up, due to its similarity with a turbo of a car engine. Compressed air is used to drive the turbine which in its turn drives the compressor. While the final gas turbine has an annular design, the turbo-shaft set-up uses supply and exhaust tubes connected to volutes collecting and distributing the compressed air. This approach is necessary to connect both compressor and turbine to valves and sensors.

The exhaust of the compressor is connected to an adjustable valve which serves as a variable load. Turbine pressure is controlled manually with a pressure regulator from Festo. By varying the load (through the valve) and speed (via the turbine pressure), and measuring flow and pressure in the exhaust tube, compressor maps can be obtained.

Fig. 1. Turbo-shaft setup. Fig. 2. Compressor (left), turbine (right) and air bearing insert. Rotor diameter: 20 mm. Size compared to 1 euro coin.

2. Instrumentation

Figure 3 shows the complete set-up including pressure regulator, variable load, tubing, and various sensors to measure flow, pressure, rotor speed, angular rotor reference and rotor vibration. Speed and phase reference are measured on the compressor nose using a KD310 Fotonic Sensor from Mechanical Technology Inc., by means of a black reference mark. Turbine pressure is measured in the supply tube using a PMP 1400 pressure transducer from Druck with a range of 16 bar (relative pressure). Compressor pressure is measured in the exhaust tube using a pressure sensor from Keller Druckmesstechnik, type PR-21S/2.5bar/80549.3, with a pressure range of 0-2.5 bar (relative pressure). Both compressor and turbine flow are measured with rotameters.

Rotor vibration is measured with two home-made single-fiber optical probes, based on reflected light intensity. The probes are positioned on the 20 mm rims of compressor and turbine. A data acquisition system from National Instruments (PXI-6123 with 500 kHz simultaneous sampling rate) is used to monitor rotor vibration.


Fig. 3. Turbo-shaft set-up.

3. Balancing

The rotor has to be balanced in two planes to minimize both cylindrical and conical rotor vibrations. This balancing is very critical because the nominal rotor speed lies above the resonance frequencies of the suspension modes. Above these critical speeds, the shaft rotates around its center of gravity. Because radial bearing clearance is only 5 m, the imbalance should be only a fraction of this value, thus maximally 1-2 m. Also the conical imbalance has to be compensated very well as both thrust bearings have an axial clearance of only 8 m. Balancing is currently done in situ by adding small masses inside the cavity of the compressor nose and at the turbine nose.

4. Measured characteristics

Figure 4 shows measured maps for turbine and compressor. Speed was limited to 4000 Hz (240 000 rpm) due to rotordynamic instability. With newly developed aerodynamic bearings, speeds of 500 000 rpm and beyond are expected soon. The compressor generated a maximum pressure of 450 mbar at zero flow rate (not shown on the graph). Figure 5 shows an extrapolated compressor map for the nominal speed of 500 000 rpm. For this purpose, flow rate is considered proportional to speed, while pressure is considered proportional to the square of speed. At zero flow (not shown on the graph) an extrapolated pressure of 1.95 bar (relative) or a pressure ratio of 2.95 is obtained. This corresponds to the design value (pressure ratio 3) at nominal flow (20 kg/s), while the pressure ratio at zero flow should be higher. Thus the performance of the current prototype is below the theoretical expectations.

Fig. 4. Measured maps for compressor and turbine (relative pressures). Fig. 5. Compressor map extrapolated to nominal speed (500 000 rpm).

5. Future work

New aerodynamic bearings have been developed that remain stable at speeds up to 1 200 000 rpm. These have been successfully tested on a simplified set-up and will now be implemented on the turbo-shaft set-up.


Reply on content: Jan.Peirs at mech.kuleuven.be Copyright 2009 Katholieke Universiteit Leuven, Department of Mechanical Engineering