When the US Marine Corps asked Rolls-Royce to improve the performance of fan blades for its Pegasus engine for the USMC Harrier (see figure above), it set a test challenge that involved some sophisticated vibration modeling and trials, even including the reactivation of a mothballed test facility.
The Rolls-Royce Pegasus engine powers the Harrier aircraft operated by the USMC as well as the Italian, Indian and Spanish navies. In 2012, the USMC awarded Rolls-Royce a contract under the Component Improvement Program (CIP) to develop a modified fan blade to improve the robustness of the engine.
Rolls-Royce created a number of finite element (FE) stress models to understand the dynamic behavior of the blade and establish the root cause of some of the problems that had been experienced in initial operations. The fan blade, as shown in the figure on the right, is manufactured from titanium alloy and rotates at in excess of 8,000rpm. It is retained within the engine by a feature commonly referred to as a dovetail.
The blade features a ‘snubber’ at around two-thirds height, which is intended to control the flap and torsional response of the blade. After the analysis was completed, it was established that issues could be caused by high stresses within the blade, which were often exacerbated by damage from material drawn into the engine during operation. From completed analysis, it was determined that the dynamic mode most likely to cause issues was a part speed stall flutter mode, which gives rise to high stress levels at several locations on the blade.
The importance of this mode was further underlined by an analysis of the operating regime of the Harrier aircraft, which indicated that considerable time was spent operating in the speed range at which this mode was prevalent.
Analysis and practical testing was able to show that control of this flutter mode was highly dependent on the tribology of the snubber faces. During operation, the snubbers can become locked due to centrifugal and aerodynamic loads on the blade. Excitation of the stall flutter mode can reverse the loading on the snubbers, leading to a situation whereby higher-order modes become excited due to stick-slip motion as the snubber faces move relative to each other.
The snubber faces are coated with tungsten carbide and a dry film lubricant, and will lock when the critical coefficient of friction, determined by the angle of the snubber face is exceeded. For the existing design, having a snubber angle of 20°, this value is 0.36. It was indicated that loss or lack of lubrication between the two faces can impact the behavior of the mode.
In order to understand this phenomenon, a friction wear test was initiated at the Rolls-Royce sponsored University Technology Centre (UTC) at Oxford University. This test was set up to simulate the relative motion between two snubbers in service and was performed at a frequency in line with what would be seen in service. The rig used is shown in the image on the left.
The results from the rig enabled an understanding of the bedding-in process and indicated that the level of friction between the two faces increases as the dry film lubricant wears away and debris builds up. The company proposed a number of design changes to the blade to mitigate this issue: the angle of the snubber face was increased to 25° to make the adjacent faces slide more easily and an improved under-platform damper was introduced to control the dynamic response of the blade.
Having identified the design solution, the company’s efforts turned to proving the change. It was established that the design changes would be validated by both rig testing and engine running. The validation process commenced with a frequency survey of the revised blade standard and the results compared with the data previously available from the existing standard. With the results of the frequency survey in line with pre-test FE predictions, the next step was a fatigue test of the new blade. Six sample blades were tested to establish each of their failing fatigue strengths.
Each blade was mounted on a vibration table and restrained at the inner dovetail fixing and the snubber to accurately represent engine conditions. The rig used is shown in the figure below right.
From the testing, it was demonstrated that the basic strength of the new blade was within the scatter of the legacy blades and that no new potential failure modes had been introduced.
Rolls-Royce followed a structured process to establish the verification and validation requirements for the revised blade. It was determined by the engineers that a number of engine tests would be necessary to gather all the evidence required to clear the new blade for operational duty.
The last production engine was shipped from the Rolls-Royce Filton factory in 2005 and the testbed was mothballed in 2010. Therefore establishing an engine test capability was the first major challenge. The first of the engine tests, to calibrate the testbed, was completed in May 2013.
With the testbed deemed serviceable, the program to validate the design change could begin in earnest. The first test of the structured validation program was a build of the engine, incorporating its existing baseline configuration, against which the intended design improvements could be assessed. The build of the engine incorporated two systems to measure the dynamic response of the blade.
The first system is the blade tip timing, which uses optical probes to detect the arrival time of the tip of the blade at a number of points around the casing. The system measures the speed of the engine spool and differences in the time at which the tip of the blade passes the probe compared with the expected timing, and can be used to determine the amplitude and frequency of any blade resonance.
The second system is the FM grid, and is similarly used to measure blade dynamic response with a magnet attached to the blade tip and a conductive wire grid around the engine casing. The motion of the blade tip causes a signal to be induced in the grid, which is proportional to the frequency and displacement of the blade tip and from which the blade dynamic characteristics can be deduced. Subsequent engine builds and testing were aimed at understanding the impact of the individual improvements in the proposed design changes.
As discussed earlier, the propensity for flutter is dependent on the lubrication of the snubber and other contact faces on the blade. The gap between adjacent snubber faces is also a major contributor to blade lock up. In order to promote flutter during the test, the fan assembly was built with dry film lubricants removed and snubber gaps at the bottom limit of tolerance. Fan blade flutter can be influenced by a number of factors, the two principal being engine inlet distortion and blockage downstream of the fan. During the test program the blockage in the front (cold) nozzles of the engine was varied to further promote flutter.
The first of these tests was completed in September 2013 and featured the existing blade fitted with the revised damper. The purpose of the test was to establish the individual benefit of the damper, with the testing completed in the flutter promoting configuration.
The next test, again in the flutter promoting configuration, featured both the revised blade and the damper. The final test, completed in August 2014, featured the revised blade and damper in the production standard configuration with all surface lubrication reapplied. The results of the test program clearly showed that the redesign of the blade and the incorporation of the new damper led to a great reduction in flutter amplitude of the blade. In the final production standard configuration, flutter has been all but eliminated.
Despite the twin technical and logistical challenges, the Rolls-Royce team compiled a comprehensive understanding of the issue and devised a substantial design improvement, which is currently being evaluated by the customer.
Tim Williams is chief engineer – Pegasus, transport and helicopter engines, with Rolls-Royce plc, based in Filton, UK