1、04FTM10The Failure Investigation and Replacementof a Large Marine Gearby: P. Hopkins, UK MOD Defence Logistics Organisation, B.A. Shaw,University of Newcastle, J. Varo, David Brown Textron, A. Kennedy, UKMinistry of DefenceTECHNICAL PAPERAmerican Gear ManufacturersAssociationThe Failure Investigatio
2、n and Replacement of a LargeMarine GearP.Hopkins,UKMODDefenceLogisticsOrganisation,B.A.Shaw,UniversityofNewcastle,J.Varo,DavidBrownTextron,A.Kennedy,UKMinistryofDefenceThe statements and opinions contained herein are those of the author and should not be construed as anofficial action or opinion of
3、the American Gear Manufacturers Association.AbstractFollowing a routine inspection of HMS INVINCIBLEs main propulsion gearboxes, cracking was identified onthe starboard main wheel teeth. This article presents a summary of the subsequent work, including: thepalliativerepair,inspectionregime,riskreduc
4、tionmeasures,failureinvestigation,thedesignandmanufactureof replacement gears and the permanent repair.Copyright 2004American Gear Manufacturers Association500 Montgomery Street, Suite 350Alexandria, Virginia, 22314October, 2004ISBN: 1-55589-833-51The Failure Investigation and Replacement of a Large
5、 Marine Gear P. Hopkins Marine Propulsion Systems IPT, Warship Support Agency, UK MOD Defence Logistics Organisation B. A. Shaw Design Unit, University of Newcastle, UK J. Varo David Brown Textron,UK A. Kennedy UK Ministry of Defence, Graduate Engineer Introduction The three Carrier Vertical Strike
6、(CVS) platforms in service with the UK Royal Navy (RN) provide the platform and facilities for the command and control of maritime and joint forces. They are 209m in length, displace 22,000 tonnes and have a complement of 685 ships company and 386 Air Group personnel. Powered by four 18MW Olympus ga
7、s turbines, they are capable of 28kts (see Figure 1). The CVS class was originally designed to carry large numbers of anti-submarine helicopters and to be protected from air attack by the Sea Dart weapons system. Sea dart has now been removed and the flight deck extended to increase the space availa
8、ble for parking aircraft. These versatile ships now deploy regularly with RAF Harrier GR7 strike aircraft aboard, as part of a mixed air group that can be tailored to operational needs. Air groups also include Sea Harrier FA2s, providing air defence, Sea King, Merlin and Chinook helicopters. With th
9、ese capabilities, the high value of the CVS platform has been demonstrated in many events since the 1980s. Figure 1 - HMS Invincible. Figure 2 - CVS Gearbox during build. Each CVS is fitted with two main propulsion gearboxes, the largest and most complex in RN service, designed and manufactured by D
10、avid Brown, now part of the Textron group. The triple reduction, reversing, double helical tandem design transmits power from either one or two Olympus gas turbines to a fixed pitch propeller. Manoeuvring ahead and astern is achieved by means of fluid couplings and for high power ahead operation dri
11、ve 2Figure 3 Flank pitting on a cracked tooth. is transmitted through two SSS clutches. The gearbox weighs 170 tonnes and contains 19 gear elements, the largest being the main wheel, which is 3m in diameter and weighs 22 tonnes (see Figure 2). In December 1999, during the Chief Engineers supercessio
12、n inspection of HMS INVINCIBLEs stbd gearbox, ships staff identified damage to a number of teeth on the main wheel. A full dye penetrant non-destructive examination (NDE) of the main wheel by David Brown revealed cracks originating from surface pits in 11 teeth. These varied in length up to a maximu
13、m of 30cm - see Figure 3. Palliative Repair Following an initial assessment it was decided to remove the cracks by in-situ dressing, to establish the full extent of the problem and to prevent any further crack growth. An iterative process of grinding and NDE was undertaken, to ensure that no crack i
14、ndications remained (see Figure 4). This resulted in the removal of 9 complete teeth and parts of 2 teeth (see Figure 5). Figure 4 NDE whilst removing a crack. Figure 5 Dressing to remove tooth sections and areas of pitting. A full and detailed inspection of the port main wheel revealed no defects c
15、onfirming that the problem was isolated to the starboard main wheel. Reducing the Risk of Further Damage It was clear that revised propulsion limitations were required if the risk of further gear tooth failures or deterioration was to be minimised. Therefore, to reduce the loads being experienced by
16、 the 3damaged main wheel teeth, revised single and twin engine torque limits were calculated, based on the loss of tooth contact area and an assessment of the allowable stress to theoretically assure infinite life. For the undamaged gearbox the torque at which additional engines were brought on-line
17、 (change-up criteria) was reduced, to avoid maximum tooth stressing at intermediate ship speeds. Whilst this did increase engine running hours and fuel consumption, it provided a margin of error against failure. The original twin engine maximum limitation was retained, but for urgent operational use
18、 only and a lower routine operational maximum was introduced to reduce the risk further. A revised procedure was also introduced to ensure that a good balance of torque was achieved between the Olympus gas turbines, particularly at high torques. With the torsion meter measuring output torque at the
19、main shaft, it had been possible for one engine to be working harder than the other, effectively overloading the main wheel teeth, but for this to go unnoticed because the total shaft torque was within limits. With these revised limitations and the close control of transients, the risk of further to
20、oth failure was assessed as small, but the ship was still able to meet her operational requirements. Subsequently a monitoring system was installed, which captured torque readings on a continuous basis, to enable the cause of any transients to be investigated and avoided. Monitoring the Defects To p
21、rovide early warning of any further deterioration, an enhanced inspection regime was devised. The defects were initially monitored following post repair trials and then by a rigorous three monthly inspection, undertaken using dye penetrant NDE by a specialist team consisting of MoD, David Brown and
22、Design Unit. Other monitoring techniques, including vibration analysis, were considered, but in tests this had not been able to reliably identify and trend deterioration. It was considered that new cracks would not propagate at a rate that would not be identifiable by the imposed regime of dye penet
23、rant NDE. The task of full scale dye penetrant inspection on a large wheel is time consuming and unpleasant to carry out, but it was felt that this technique gave the best opportunity of providing full coverage of all teeth with good accuracy for identification and monitoring of the crack types know
24、n to be forming. Although other methods are available for crack detection (e.g. ultrasonics and eddy current), these were not as easy to apply with confidence in-situ. Being a relatively simple technique to apply, dye penetrant was also made available as a backup during ships staff visual inspection
25、s. With the altered running procedures in place, further deterioration did not occur which suggested that the new restrictions had been pitched at about the right level. Failure Investigation Having managed the immediate situation through inspection, remedial work and changed operating protocol in o
26、rder to control and prevent any further degradation of the main wheel, the next stage was to gain a full understanding of the cause. This was tackled through investigating the basic material properties and failure mechanism, gaining a full understanding of operational characteristics and relating th
27、is to the gear design intent with a reassessment of the gear design using up-to-date techniques. Figure 6 Flank initiated pitting leading to bending fatigue crack growth. 4Figure 7 SEM images showing flank pitting leading with associated crack growth (top image is secondary electron imaging and the
28、lower is backscattered electron imaging). Metallurgical Examination Using sections from the failed portions of teeth, a full material investigation was carried out by Design Unit to determine and deficiencies in basic material properties that could account for the occurrence of the failure and to id
29、entify the mode of failure from fracture surfaces. The specified material condition is through hardened to a U condition (approximately 290Hv minimum). Hardness measurements taken from polished cross sections through the failed portions of tooth revealed a hardness level in the region of 280Hv. This
30、 compared with manufacturing records showing a hardness of approximately 300Hv. Examination of the microstructure did not show anything unexpected for the condition of material and no large inclusions were observed showing the steel to be of the quality level intended. Examination of the fracture su
31、rfaces showed clear indications of fatigue crack growth indicating that the crack had taken sometime to initiate and propagate to give the final failures. The main bending crack that led to failure was found to initiate from below the surface and this was tracked to the presence of flank surface ini
32、tiated pitting that was seen fairly widespread in the regions of tooth cracking (see Figures 6). Examination of the fracture surfaces in an SEM with EDX (for chemical analysis) confirmed the visual and optical microscopy observations of fatigue initiated from below surface flank pitting. Close exami
33、nation of the pitting and the main fatigue crack initiation sites did not reveal the presence of any unusual material defects such as large oxide inclusions that could account for the failure (see Figures 7 and 8). In Figures 7 and 8, the lower image compares backscattered imaging with the normal se
34、condary imaging in the upper image. In this imaging mode oxides would appear black. It was therefore concluded that the material condition was not the primary cause of failure as it was within the specified condition used in the original design conditions and that no unexpected quality issues had re
35、sulted in a lower that expected material strength. The mechanism of failure was traced to a primary contact fatigue failure resulting in flank pitting that then led on to a secondary initiation of a bending (flank initiated) fatigue crack that ultimately resulted in the major tooth bending fatigue f
36、ailure observed. From these investigations, it was then necessary to identify what had resulted in gear flank contact stressing sufficient to result in contact fatigue pitting that had been identified as the primary cause of failure. CVS Operating Philosophy To minimise fuel usage and engine operati
37、ng hours CVS tended to operate on the minimum number of engines at all times. Starting with one engine driving through one gearbox with the other shaft trailing, then two engines, then three, then four. The tendency was to drive up to near maximum torque on each engine before adding the next one. Be
38、cause of the design of the gearing, with each engine driving through separate drive trains onto the main wheel, each drive train and most crucially the teeth on the main wheel would therefore tend to see something near their maximum design loading most of the time. 5Figure 8 SEM images showing fatig
39、ue cracks growing from flank pitting (bottom image BEI). The situation was exacerbated by the fact that the maximum power design point was for all four engines in use. In any other mode, the engines that are driving are producing their power at lower speeds and therefore commensurately higher torque
40、s. In this situation, particularly with gas turbines, it is much easier to introduce transients to the gearing. When accelerating or manoeuvring the vessel the risk of accidentally overloading the power train increases. The situation is further exacerbated by the fact that torque is only measured at
41、 the output shaft. When driving on two engines, it is generally assumed that the inputs are balanced, but this was not necessarily the case. Gas turbine optimisation could and did drift and if so, when at maximum two engine torque, one drive will inevitably be overloaded. HMS Invincible Operating Hi
42、story The operating history of INVINCIBLE was reviewed and compared with other ships of the class and the following were identified as potentially significant differences: Being the first of class, INVINCIBLE had accumulated considerably more operating hours than the other ships of the class. Discus
43、sion with experienced operators identified significant differences in operating conditions between the 1980s and the 1990s. Changing deployment patterns, from predominantly anti-submarine warfare operations in the windy North Atlantic, to mainly fixed wing operations in the high temperature and low
44、winds of the Gulf and Adriatic resulted in more high power use to achieve optimum aircraft launch. It was also found that undesirable transient over-torques had been experienced during operation. Material History MPS211 and David Browns undertook a review of the material history of the gearing, whic
45、h highlighted the following: In 1983 pitting was discovered on the main wheel, at the forward end of the forward helix. An investigation identified that this had been caused by the main wheel aft side plate bolts becoming loose, which caused a lack of radial support at the aft end of the rim resulti
46、ng in a transfer of load to the forward end of the helices. Combined with pinion torsional wind-up and bending, the maximum load intensities were now all at the forward end of the forward helix. At this time, replacement of the side plate fasteners arrested the pitting, but the pitting was left alon
47、e as it was felt that pitting of this kind was reasonably safe on through hardened wheels of this type. In the early 1990s, slight progression of the pitting was noted which was probably caused by increased usage at high power. At some time during the mid/late 1990s cracks initiated from some of the
48、 pits, probably due to a small number of overloads. Once initiated, these progressed during normal operation to the extent discovered in late 1999. Peer Review A gearing peer review group, comprising MOD and gearing experts from David Brown, Alstom Gears, VSEL and Design Unit met to review the desig
49、n, operating history and nature of failure. They agreed that during design it would not have been realised that the configuration and way to achieve fuel 6economy would lead to high tooth loading continuously and that this situation would not have been recognised by the design codes of the time. They concluded that the main wheel was the weak link in a CVS gearbox and that it appeared to have insufficient capacity for its current duty. In this case, the design was now considered to have been operating above its fatigue limits, confirmed by the fact
