Seeking Design Solutions through Decoupling Problems – A Global Engineering Example Mahmut F Aksit Engineering Across the Borders In the race for higher power and efficiency, the challenges faced by design engineers get more complicated every day. It can be said that everything is being pushed to the limit. On the other hand, technological advances present opportunities and recourses that were not available before. Communication and transportation technologies reached to a level that multinational engineering teams can be formed tapping into the talents available across the globe. Different education systems bias engineering students –one way or the other- towards their established point of view. This may limit the ability of engineers to see a wider picture that may include solutions which do not fit in their established vision of the design space. In today’s global marketplace, common and obvious solutions are everywhere. The secret to success lies with unconventional/novel design approaches. Many international companies tap into the global talent pool, hiring engineers across the globe. While forming multinational teams at home was the early phase of this approach, the new trend is to form research centers at various locations in the developing countries. The second approach is getting popular as it also helps minimize costs in cases where there are lower wages abroad. This paper presents a real life example how such a multinational team developed a simple solution for a challenging application through decoupling an engineering problem. Problems with Contradicting Performance Parameters Challenging design problems typically involve multiple aspects, coupled performance parameters and contradicting objectives. For example, decreasing pressure while increasing temperature will not be possible when dealing with a gas medium in an enclosure. Interacting performance parameters prevent reaching a solution through a one-at-a-time approach. Common optimization schemes will provide an optimum design point where contradicting objectives reconcile. The problem starts when this compromised optimum does not meet the design goals. For such problems, the solution cannot be improved without decoupling the contradicting performance parameters. However, as in the pressuretemperature example, decoupling may not always be possible or easy. On the other hand, there may be totally different, not-so-obvious ways to decouple the problem. Having different view points will increase the chances of finding such nontrivial decoupling options. As discussed in the previous section, having engineers from different backgrounds will be one way to widen design options. A similar design solution case will be summarized below.
Nozzle Segments
Cooling Air Side
Radial
FSN
TD Axial
Hot Gas Flow Figure 1. Typical first stage nozzle – transition duct (can) junction [1]. The Problem Designing seals for gas turbine combustor exit locations is very challenging. Multiple objectives, like long life and improved sealing have to be met under challenging loads at elevated temperatures. Special seal designs are needed to bridge the gap between transition ducts (cans) and first stage nozzles (FSN). Exited by combustor dynamics, transition ducts (TD) move in both axial and radial directions relative to first stage nozzles. As illustrated in Figure 1, the original design called for axial and radial seal slots. The design intent was to allow the seal to slide in these slots while mating parts are free to sustain relative motion. However, the thermal misalignments cause neighboring nozzle pieces to move relative to each other, jamming the seal in the nozzle slot. With seals stuck between the nozzle segments, steady beating by the transition ducts causes heavy wear on both the seals and the duct frames. The problem is accentuated by the fact that the seal warps under an uneven thermal profile causing localized contact points. This increases wear rates resulting in total loss of seal thickness at some localized sections. When this problem was brought to the GE Global Research Center Advanced Seals Team, the design goal was set to increase the existing 8000 service hours to 12000. With existing severe wear problems prospects of further life expansion looked dim. The Solution The common approach to such wear life extension problems would be increasing seal thickness that will provide more sacrificial material, thereby increasing wear life. However, this is a coupled problem where increasing seal thickness undesirably increases seal stiffness which –in turn- results in tip-toe rigid motion of the seal causing even greater concentrated contact loads. The thicker seal option does not provide any solution as earlier designs resulted in heavy damage to transition duct frame and more localized seal wear. Damage to duct frame is even more detrimental as it increases operational risk, and requires more expensive and time consuming repair procedures.
The design challenge was assigned to a global team with engineers from the USA, Turkey and India. The vicious-circle of added seal thickness for added life resulting in more stiffness and more wear could be broken with a solution that adds seal thickness for wear life without adding stiffness. At operating temperatures well beyond 1000 F, only high temperature metal alloys could endure such exposures while maintaining their elasticity. However, increased metal thickness meant more stiffness. The vision of the multinational team reached design areas where previously discarded as offlimits by seal designers due to inherent leakage issues. The idea was to use woven metal structures for added sacrificial metal thickness without increasing seal stiffness. Woven structures can sustain compressive contact loads across their weave thickness, while they do not have much bending stiffness. However, metal cloth structures are porous medium which are not good candidates if you are seeking leakage performance. A solution was found by decoupling the wear issue from the sealing issue. Flexible metal shims are used to block the leakage flow while outer layers of densely woven metal
Pinch Angle Cloth-Shim Assembly
Seal Frame
cloth provide a sacrificial wear layer without adding much stiffness. Since the problem was decoupled, separate design optimization work could be performed
Aft frame radial clearance
Cantilever Length
for both the cloth and shim selections. While shim material and thickness was selected for oxidation life and pressure load capacity, a separate optimization work conducted for metal cloth layer for weave type, fiber
Slot Engagement
Preload/ Interference
Figure 2. Metal cloth seal
size,
and
mesh
density
to
meet
different
objectives for leakage, wear and oxidation life [2]. Results well
were beyond
expectations. Using the best of both worlds, a composite seal structure with flexible sealing lip has been designed (Figure 2). The composite structure allowed the seal to flex and distribute the contact loads across the seal-slot interface. Lower contact loads reduced both seal and duct frame wear dramatically. Field tests proved that seals were almost intact after 12,600 hrs of operation (Figure 3). While initially aiming for 12,000 service hours, the team certified composite metal cloth seal designs for 24,000 hours in most gas turbines. In addition to
Figure 3. Cloth seal after 12600 hrs of service
tripled wear life success, there was one more benefit which was not the original design team intent; improved sealing. The new flexible seal structure was conforming to thermally distorted mating surfaces so well that drastic leakage performance
improvements were achieved. The new seals quickly migrated to many other sealing locations between shroud and nozzle segments [3]. The tests indicated 4). Although large leakage reductions are most welcome for sealing, less leakage means less cooling for the seal, and a larger risk of overheating. However, seeping cooling air flow through the porous metal cloth layer proved useful to keep embedded structural shim even cooler than before [4]. With superior life and sealing performance many different cloth seal versions were developed for various sealing applications across many gas turbines resulting in well over 50 patents worldwide. Today, metal cloth seals are standard in most GE gas turbines sold in the market.
1.6
Normalized Eq. Gap (mil/mil)
up to three fold leakage reductions in some locations (Figure
1.4 1.2
Baseline
1
Offset & Skew
0.8 0.6 0.4 0.2 0 Rigid Seal
Cloth Seal
Figure 4. Leakage performance comparison under baseline and misalignment conditions
Conclusion Most real life problems involve nonlinear and implicit performance parameter interactions. Although established optimization schemes will provide a design optimum where contradicting parameters meet, most of the time such compromised optimums will not put your design ahead of the global competition. To break bounds and go beyond the common limitations on your design, you need fresh view angles and unbiased ideas. Truly, global teams with talents from different educational backgrounds have a better chance to find the out-of-the box, game changing designs. With talents across the globe, a diverse engineering team is the key to success.
AUTHOR Mahmut F Aksit, Faculty of Engineering and Natural Sciences, Sabanci University, Istanbul, Turkey,
[email protected] REFERENCES [1] Akşit, M.F., B.S. Bagepalli, and S. Aslam, “Life and Performance Optimization of High Performance Combustor Cloth Seals,” Proceedings of 36th AIAA/ASME/SAE/ASEE Joint Propulsion Conference, Huntsville, Alabama, AIAA Paper 00-3510, 2000. [2] Dinç, O.S., B.S. Bagepalli, C. Wolfe, M.F. Akşit, and S. Calabrese,, “A New Metal Cloth Stationary Seal For Gas Turbine Applications”, Proceedings of 33rd AIAA/ASME/SAE/ASEE Joint Propulsion Conference, Seattle, Washington, AIAA Paper 97-2732, 1997. [3] Aksit, M.F., R.E. Chupp, O.S. Dinc, and M. Demiroglu, “Advanced Seals for Industrial Turbine Applications: Design Approach and Static Seal Development,” AIAA Journal of Propulsion and Power, 18, 1254-1259 (2002). [4] Doğu, Y., M.F. Aksit, B.S. Bagepalli, J. Burns, B. Sexton, and I. Kellock, “Thermal and Flow Analysis of Cloth-Seal in Slot for Gas Turbine Shroud Applications,” Proceedings of 34th AIAA/ASME/SAE/ASEE Joint Propulsion Conference, Cleveland, Ohio, AIAA Paper 98-3174, 1998.
