Collaborating to Reduce Gas Turbine Fuel Burn

The pressure is on to reduce fuel burn for gas turbines of all types. The need is particularly acute for aircraft engines, in that fuel is a large component of operating costs of an airline, so much so that even the volatility in its price can mean the difference between profit and loss. So when airlines demand more fuel efficient aircraft, much of that requirement is passed along to the engine manufacturers. While reducing gas turbine fuel burn is a primary driver, carbon emissions are related, so reducing the fuel burn “kills two birds with one stone”.

The biggest factor affecting efficiency is the turbine entry temperature: increasing it increases the Carnot efficiency (or cycle efficiency), which is the theoretical efficiency one could achieve from the engine if everything could be done perfectly without losses. This is of course not achievable, but real efficiency tracks the Carnot efficiency, hence the desire to increase the temperature. Gas turbine engineers have been increasing operating temperatures ever since the earliest engines (by Whittle and von Ohain for aircraft) were developed.

The problem is the negative effect that high temperatures have on the “hot section” components, the combustor, transition ducting and the turbine which sits downstream of the combustor. High temperatures reduce component strength and life, consequently engineers have worked on many strategies for dealing with these high temperatures. These include advanced materials, thermal barrier coatings and use of cooling air to provide a barrier between the hot combustion gases and hot section component. All three of these are used in modern engines.

Gas Turbine Blade

A schematic of a high-pressure turbine blade as used in aircraft jet engines. An internal cooling system enables a temperature gradient across the walls of the blade. Cooling holes allow the cooling air to exit and cover the outer blade surface with a cooling film. Courtesy: Wikepedia

Use of cooling air is attractive since it can be adjusted according to engine operating conditions. But in using cooling air, the designer tries to strike a balance, using only the minimum amount of air necessary to get the job done. Cooling air is bled off the compressor, and using any at all serves to reduce efficiency (energy is required to compress the air, after all). In the turbine section of the engine, engineers work very hard to determine the best locations to introduce cooling. This has been a “hot” research topic for many years. Ron Bunker, a well-known heat transfer expert at General Electric, gave a very nice overview and summary of efforts in that field, at the 2014 ASME Turbo Expo. Ron showed a range of cooling designs, some of which have become very exotic, so much so that manufacturability has become a constraint. So it is clear that optimizing cooling is a formidable challenge indeed.

Real engines operate at high speeds and temperatures, and hence are inherently difficult to instrument. Making and testing a variety of designs is expensive and time consuming, and just not practical given development constraints. So that is where simulation comes in. But of course simulation must agree with reality. Recently, Honeywell, Stanford University and ANSYS have begun a collaboration that aims at addressing the cooling problem from a number of angles. Stanford contributes the latest experimental techniques (3D magnetic resonance velocimetry) for accurately determining turbine blade flow fields. Honeywell provides all the aspects of the engine expertise, including the critical geometry and operating conditions and ANSYS provides the simulation software.

A key factor limiting the ability to simulate turbine heat transfer is turbulence modeling, and its application. The structure of the flow in cooled turbines is complex. This is partly due to the range of scales in the flow field. At the large end of the spectrum is the bulk flow in the blade passage. But the cooling flows are small and localized, in the general situation emanating from small holes on the blade surface, at the trailing edge and in the tip region. The structures created by the cooling flows are a complexity on top of the already complex blade boundary layers, which in some cases may involve laminar to turbulent transition. While ANSYS offers the widest range of high-fidelity RANS (Reynolds Averaged Navier-Stokes), LES (Large Eddy Simulation) and Scale Adaptive models available, there is always room for improvement. The unique nature of this collaboration is that it will directly use cooled turbine experimental data to enhance both the models and the application of the models to realistic gas turbine operating conditions.

This is an exciting new collaboration, and we look forward to reporting our progress in the coming months.