Experimental Procedure The Test Facility

The tests were performed in the Rolls-Royce compressor test facility in Derby on a 34-inch diameter fan rig with an engine representative bypass ra­tio. The fan blade tested was a typical wide chord fan blade at 26-off with a design blade tip Mach no. of 1.4. The rotor path incorporated an array of 22 high-response, temperature compensated transducers for measuring the unsteady static pressure between leading edge and trailing edge. In addition, absolute pressure levels were measured for comparison using pneumatic pres­sure tappings fitted at the same axial locations, but offset circumferentially. The arrangement of the transducers and the pneumatic tappings is illustrated in Fig. 1. To obtain the density of measurement points required, the transducers were mounted in two parallel rows approximately aligned with the blade tip. The comparison between the Kulites and the pneumatic tappings was shown to give good agreement. One transducer close to the leading edge failed during testing, the results presented here are based on the remaining 21 transducers.

Data Acquisition

During operation of the facility, constant speed characteristics were mapped out by throttling the bypass, to the stability boundary. Overtip unsteady data were acquired during two stability events: a complete surge cycle at 105% speed and the recovery from rotating stall at 60% speed. The 105% speed data covered 0.18 seconds taken during a multiple surge cycle which lasted approximately 1.5 seconds. For the 60% speed the data acquisition system was set up to measure rotating stall for 1.5 seconds of an event which lasted 3.6 seconds. In both cases the unsteady pressure signals from the Kulites were recorded on a high speed digital data acquisition system at 500 kHz. The data acquisition system contained a signal conditioning unit which amplified the signal and provided anti-aliasing filters.

Data from three further rings of Kulites at varying axial locations were also acquired to give the overall performance of the rig during stall and surge cy­cles. The transducers were located approximately i/2 an axial chord upstream and downstream of the fan leading and trailing edge, and in the bypass duct downstream of the outlet guide vane. These data were sampled at a frequency of 2.5 kHz, therefore the bulk fhctuations produced by the stall and surge can be seen, but not those produced by the individual blade passing.

2. Results

Greitzer (Ref. 2) has previously defined the B parameter that defines whether a fan will operate in rotating stall or surge at the stability point. The B param­eter is equivalent to the ratio of the compressor pressure rise capability to the pressure rise required to induce mass fbw oscillations. Where,

B = {Py2/2)A pALuluhelm

pAL represents the mass of gas (density x duct area x duct length), u is the blade speed and whelm is the frequency of natural oscillations.

If B is low then a compressor does not have much pressure rise capability compared to that required to begin a surge. If B < Bcrit, then stall occurs, If B > Bcrit then surge occurs, where Bcrit is the boundary between stall and surge (where typically Bcrit =0.8).

The above equation shows that if only blade speed is changed then the fan is more likely to surge as speed is increased. The test experience for the fan discussed here shows rotating stall occurs between low speed and 100% with surge at 105% speed. The following sections present data for the overtip and rings of Kulites for a stall event and a surge event to show the nature of the two types of instability.