By Bjorn Fehrm
19 Jan 2015: There is a lot written about the fundamentals of how aircraft fly. It is something that fascinates people and it generates a high level of understanding of these fundamentals. The same is not true for the airline turbofan engines in use today; their detailed function remains a bit of black art.
To some extent this might be because what is exiting in the engines (the thrust) is generated behind closed doors. The only visible part of the process is a rotating fan face and sometimes a slight miss-colored exhaust out the other way. There is also at takeoff a funny buzzing sound interspersed with the general engine noise. Apart from that, the most that one sees is a round nacelle and that is it.
If one opens the lid of a modern turbofan, one finds a virtual fireworks of advanced technology. Aerodynamically engines cover a larger envelop than the aircraft it sits on. They go to full supersonic flow in the fan. The aerodynamics in the compressors and turbines are 3D shaped to a level which is not found on the aircraft. For reasons of manufacture, the tube section of the fuselage does not have Coca-Cola bottle-type waist, etc.. The technology around emission control when burning fuel efficiently is reaction chemistry at the highest level and finally the fact the turbines don’t melt from working in gasses almost double the temperature of their melting point is mind boggling.
Modern airliner turbofans are at least as exciting as the aircraft they serve, if not more so. As we’ve seen with the Boeing 787 and 747-8, the 737 MAX, the Airbus A350, A330neo, A320neo and potentially the A380neo, engines drive the airplane development; new airframe design might add 5-10% of efficiency—new engines add 10%-20%.
Here is a closer description of these technological wonders. We describe engines with the help of CFM’s new LEAP engine.
LEAPing from the CFM56 to new technology
In Part 2 or our series we described the LEAP-1B engine of the Boeing 737 MAX 8 on a general level. We will now dive deeper into the LEAP-1B and compare it to the engine it replaces, the CFM56-7B, used on the 737NG.
The company behind the engines, CFM International, a joint venture of GE Aviation and Snecma of Safran, is the world’s dominant manufacturer of engines for airliners with more than 27,400 engines sold and in service. In preparation for this part of our series, CFM described its philosophy when designing the heir to the world’s most successful aero engine, the CFM56. We were especially interested in how CFM would combine a 15% efficiency improvement with the reliability and maintainability levels that a sequel to a CFM56 would imply. Details are below.
To make it possible to follow the evolution from the CFM56-7 to the LEAP-1B and understand the design challenges involved, we have built a model of the LEAP in our engine modeling tool, GasTurb12 of GasTurb GmbH of Aachen Germany (http://gasturb.de/index.html). It is very powerful engine simulator and has excellent visualization graphics which we will use to try and understand the technology choices that CFM has used to produce the LEAP.
It shall be pointed out that these simulations are carried out with our own estimation of values for LEAP-1B in GasTurb. CFM has not provided any more information than what is to be found in their official communication and Power Points (as should be expected at this point of the program). We do these simulations to further the overall understanding of a state of the art engine like the LEAP and how the different parameters change during normal use, and not in order to reveal any of CFM’s proprietary design data (which we can’t as we don’t have it).
Turbofans and how they work
We will now go through the evolution of the LEAP-1B and learn how turbofans are specified, how they work and how CFM managed to improve the efficiency with 15% while retaining the reliability of the CFM56. The engine was briefly described in Part 2 of our Series. We will now dive deeper to see what parameters are critical and how CFM used its technology leadership in many areas to realize the engine with principally the same base architecture as CFM56. But before we go into the engine, let’s look at the requirements Boeing would have for a new engine for the 737 MAX.
To understand how an engine like the LEAP gets designed, we will follow the process from Boeing’s decision to do the 737 MAX series until design freeze of aircraft and engine. When Boeing decided to do a 737 MAX instead of a new single aisle aircraft, CFM had designed the LEAP-1A for the A320neo and the LEAP-1C for the COMAC C919. The requirements on the engine from A320, C919 and 737 are not the same. For instance, the needed maximum TakeOff thrust is 33klbf (lb force which we write as lbf) for A320 and 28klbf for the 737 MAX. Adapting the 737 engine from the A320 would have produced a suboptimal engine as the thrust requirements for the 737 are around 10% less than for the A320 series. Keeping the core would have meant the core was dimensioned for driving a fan to generate 33klbf of thrust and not 28klbf. The result is a loss in efficiency. The core should be the minimum required to do the job to make for an efficient engine.
The LEAP-1B is therefore a design 100% adapted to the requirements of the 737 MAX.
Here the important requirement points on a civil airliner turbofan engine:
Top of climb thrust
For many aircraft, the thrust needed to get to the initial cruise altitude is the most demanding requirement flow-wise for the engine and therefore it determines the aerodynamic paths (the efficient throughput of the air mass as described in Part 2) and thereby the size of the engine. Our airplane model shows the LEAP should produce around 6,000lbf at the last part of the climb to reach initial FL320-330 with an acceptable minimum climb speed for the heaviest member of the MAX family, the 88 tonnes MAX 9 (an engine is sized by the most demanding member of the family, then derated to serve lighter and smaller members). Ideally that thrust level shall be flat-rated to 15°C (27°) over normal or ISA temperatures. At FL320 this would mean 15°C over -48°C (-33°C or -28°F).
For a two- engine aircraft, which loses 50% of its power if an engine goes inoperative at takeoff in contrast to a four engined aircraft which loses only 25%, the takeoff requirements can be more demanding then the top-of-climb requirements. It is normally not the static thrust, i.e. when the aircraft is lining up at end of the runway, which is the problem. It then has both engines intact and has plenty of power. The most critical demand on the engine is when one is going inoperative directly after lift-off. The requirement is then that the aircraft shall be able to climb with a minimum 2.4% at the so-called safety speed, V2. For a MAX 9 at Max Takeoff Weight it would be around 145kts and the required thrust would be approximately 22.000lbf on the remaining engine. For our six tonnes lighter MAX 8, the requirement would be around 20,500 lbf at a speed of 140kts.
Flat rating thrust
Most engines are specified to maintain their takeoff and climb performance to ISA +15°C. Often the most stressing point for the engine temperature wise is therefore start from hot airports at the max rated thrust and max rated temperature. When we simulate the known and assumed values of the LEAP-1B, the V2 demand of 22.000 lbf at ISA +15°C comes out as the most challenging condition, it sizes the engine core processes that we need to drive the fan to produce 90% of the 22,000 lbf thrust. The remaining 10% thrust is a rest product from the core while producing the shaft horsepower for our fan.
The cruise thrust requirements are much lower. Still the engine has to work hard to satisfy these. At average cruise weight the MAX 9 needs 4,900 lbf and the MAX 8 4,500 lbf to overcome the aircraft drag at 35,000 feet and M 0.8. The fan requires the low pressure turbine to produce 9,800 hp to produce this thrust for the MAX9 and 9,200 hp for the MAX 8. This means we need two shaft hp to produce one lbf of thrust at cruise and we only needed 1.36 hp (total 30,000 hp) to produce the 22,000 lbf of thrust at the denser air at sea level. The engine is only producing less than 5,000 lbf but is still working hard aerodynamically and spinning at around 90% of full rpm to produce the thrust we need at cruise.
Specific fuel consumption
The cost of producing these 4,900 lbf (MAX 9) or 4,500 lbf (MAX 8) is measured in how much fuel is consumed per hour to produce one lb force (lbf) of thrust. At ground level and takeoff, the cost is around 0.3 lb/lbf/hr and at cruise around 0.6-0.65 for the present CFM56 engines. For the LEAP, CFM managed to reduce this a full 15% to around 0.53-0.56 lb/lbf/hr. The range at cruise is deliberate, it depends on many factors.
When engine manufacturers give the Thrust Specific Fuel Consumption, TSFC, one has to be careful with at what condition it is given. Many times the values of TSFC, e.g. as given in the ICAO Emission Databank (only takeoff, initial climb and landing TSFC, not cruise), are test stand TSFC with no nacelle losses and no power offtake for driving hydraulics for flight controls, electrical generators or bleed air for the cabin air conditioning . This increases the TSFC with 7-8% and therefore values that seem extremely good (low 0.5 for existing engines) are most of the time from test stands with no service functions to the aircraft included. Our simulations show that LEAP-1 achieves state-of-the-art values for this size of aero engines (it is harder to make a smaller engine efficient as gap tolerances don’t scale with smaller dimensions of compressor or turbine blades).
LEAP design choices
How did CFM achieve the 15% reduction while maintaining the CFM56 reliability level? To begin this analysis, let’s start with the question we put to CFM through their director of Strategic Communication, Jamie Jewell.
A: Narrowbody aircraft have different design objectives than widebody aircraft because of high cyclic operation, which can be up to 6-8 flights per day. The expectation for LEAP is 20,000 cycles on wing, first run, in order to achieve low maintenance cost. The message of the chart is that the cycle design and temperatures for LEAP have been selected for the right balance of performance and operating cost. High cycle temperatures resulting from high operating pressure ratio could drive down time on wing and increase maintenance cost.
A: Reliability in this sense means the operational metrics in terms of departure reliability, delays and cancellation rate, in flight shutdown rate.
A: The goal is from launch. Significant advance maturation work is being conducted, including 40,000 cycles of testing prior to EIS.
A: The maintenance cost per hour target for LEAP is expected to be comparable to the CFM56 family over its lifetime.
A: CFM is offering three levels of aftermarket support for LEAP: full risk transfer through a long-term rate-per-flight hour (RPFH) agreement, under the terms of which CFM guarantees the maintenance costs; Material Service Agreements (MSA); and Time & Materials. Our goal is to provide our customers with more flexibility through customized agreements and an open network (the details of which we have not yet finalized).
In essence this mean the LEAP is a true heir to the CFM56 with the same targets for reliability and maintainability. We will now look closer at how this is achieved. We reuse the picture from Part 2 but dive deeper on each point this time (as before, this is a picture of the LEAP-1A or C but the principal technologies are the same for LEAP-1B):
No 1 marks the fan and fan case. GE is the world leader in CFRP-based fans (first with GE90) and fan cases (first with GEnx). LEAP is now taking this technology further. The GE90 and GEnx blades are hand-laid up from CFRP prepreg cloth. Snecma, which has responsibility for the fan and low pressure parts in CFM, has formed a joint venture, Albany Engineered Composites Inc. (AEC, Rochester, N.H.) with one of US high tech weaving companies, Albany International. AEC has the technology to weave the Carbon fibers into a 3D mesh which when placed in a form get soaked in injected resin.
With this technique, fibers can be woven in an interlocking manner eliminating any risk of delamination associated with the discrete layers of prepreg in conventional blade layup. The technique can also be more automated, necessary for the high volumes that the LEAP will be produced in.
This process also gives very exact control of the blades properties like directional stiffness of the different parts and we understand that Snecma is using this to have the blades form themselves according to their working condition. A modern fan is a very advanced piece aerodynamically. Its tips are supersonic, the middle part has transonic aerodynamics and the inner part has subsonic aero. This requires different aerodynamic forms, from thin and pointy to blunt and round. In all 3D woven resin infused blade ensures a very exact blade with high efficiency and low weight. AEC is also producing the fan case in the same way, first forming the weave and then infusing the resin.
Right after the fan come the booster compressor. As it sits on the same diameter as the fan’s inner part it has subsonic aerodynamics and a low speed through the air. The pressure gain is therefore low, on the level of the fans average pressure ratio of around 1.4-1.5. Conversion efficiency, shaft hp to pressure gain, can still be as high as the fans efficiency using today’s advanced 3D aerodynamics, above 90%.
No 2 denotes the variable bleed ports on the outer part of the swan neck duct between booster and compressor. The fan will ingest debris like sand from the runway. It is important to keep that away from the fine dimension of gaps and seals in the core. CFM uses several techniques to ensure this:
– the spinner form throws most debris past the core entry, into the bypass.
– The small particles that do enter the core swan neck will then slide on the outer wall of the duct due to its curvature where the engine control computer, the FADEC, can angle the handling bleed doors slightly open to catch the particles that got ingested and route these to the bypass duct. This technique has been proven on the GE90 operating in sandy environments.
No3. Snecma is also responsible for the gearbox with accessories. It is mounted on the fan case to ease service access and slung to the side to maximize ground clearance.
No 4. The passing point of the responsibility in the core from Snecma to GE Aviation, where the booster swan neck leads into the High Pressure Compressor, HPC.
No 5. The high pressure compressor, HPC. It takes the air from a pressure raise of around 2 and raises that to above 40 before the air passes into the combustor. GE Aviation is known in the industry to be the masters of efficient high pressure rate compressors. Our simulations for cruise suggests the HPC for the MAX8 consumes around 9,500 hp to do that with a conversion efficiency (shaft energy to compressed air energy) of well above 90%. At its peak operating point it can raise the pressure 22 times, most of the time it is operating around 20 times however.
One of the effects of this high compression is that the air gets hot, at ground level and hot take-offs the HPC exit temperature can be a limiting factor. Our simulations with GasTurb suggest the HPC exit would be at around 700°C or 1,200°F. The last stages of the HPC is therefore made with heat resistant nickel based alloys whereas the early stages are made of Titanium. They can be machined in on piece (nickel based alloys are not easy to machine), so-called BLISKs, giving both aerodynamic (leakage) as well as manufacturing advantages.
At this point it can be instructive to understand that most values in a turbofan engine vary with the demanded thrust. Figure 3 is showing a diagram from a GasTurb simulation of the LEAP-1B at average cruise height. It shows what happens with specific fuel consumption (SFC), overall pressure ratio (OPR) and by-pass ratio (BPR) when the engine goes from en-route climb (5,300 lbf) to cruise (4,500 lbf) to flight idle (probably around 1,000 lbf).
The diagram shows several interesting aspects of turbofan design:
TSFC: this varies with every thrust setting of the engine, important is that the TSFC “bucket” has its minimum at cruise thrust (between 4,000-5,000 lbf).
OPR: The overall pressure ratio of an engine is a function of its rpm, high thrust = high rpm = high OPR. The important area is once again cruise where OPR would be around 40, a good value and one of the corner-stones in the high efficiency of the LEAP.
BPR: Finally the By Pass Ratio, BPR, is varying from 8.7 to 13-15 at flight idle. BPR is strongly coupled to how hard the engine is working, the higher thrust, the more air through the core (to generate shaft hp) and lower BPR. This is why higher rated members of an engine family always have lower typical BPR. The kink in the BPR is from the on-set of a handling bleed through the variable bleed ports after the booster. At low load the HPC gets to aggressively feed by the booster and therefore excess air is bleed into the bypass channel from (in our simulation) 2,200 lbf thrust. Should this not be done the HPC early stages would stall.
No6. On our journey through the engine the next station is the combustor. One of the problems with high pressure ratios in turbofans is that this creates high levels of NOx emissions. GE has developed a lean burn two-zone combustor series called TAPS to master. LEAP is using the latest incarnation of this technology, TAPS II. The TAPS combustor also has a very uniform distribution of the heat in the exit gasses to avoid hot spots in the combustor nozzle and the first high pressure turbine stage.
No 7. LEAP divides the driving of the high pressure compressor over two turbine stages instead of one for the CFM56. This is necessary as the work done in the HPC has doubled, from maximum pressure ratio 11 for the CFM56 to 22 for LEAP. To create shaft hp efficiently with the smallest core, the gasses entering the turbines shall be hot and at high pressure. State-of-the-art right now (GEnx-1B75 ) is 1,700°C or 3,150°F turbine entry temperature and about 55 in pressure raise from inlet to first turbine. As stated by CFM, LEAP has another optimization philosophy than GEnx. We have assumed around 1,550°C and a pressure rise of around 50 at the most extreme working point.
To withstand such temperatures, the turbine section requires cooling and quite a lot of it for the first stages. At the toughest operating point, which we found to be one engine out at V2 on a hot day, the turbines needs over 20% of the air produced by the compressors for cooling. This cooling air is consistently taken from the lowest compression level possible in the engine, from fan, booster, compressors and even combustor; the pressure need to be high enough to secure positive circulation in the cooling object at all conditions but not more, any more pressure and the air is hotter than needed (worse cooling result) and more engine work has been invested in creating the cooling air than necessary.
A turbofan is therefore riddled with cooling air off-takes. The CFM56 has about 10, from after the fan for nacelle and aircraft bleed air cooling to the HPT being cooled with air from the combustors peripheral airflow. The highest cooling demands are for critical flight cases like takeoff or one engine flight. For cruise, the demands are lower. It is therefore attractive to regulate the cooling air for the different load cases. Easiest is to regulate the air to the stator side but the latest airliner engines have also regulated the air to the turbine rotors. To check what regulated turbine cooling would give, we halved the turbine cooling air in our simulator, which gains the engine about 1% in TSFC.
Better still is to use materials that require very little cooling such as Ceramic Matrix Composites, CMC. GE is a leader in the use of CMC for turbine engines. GE has deployed it in their stationary gas turbines for years and the LEAP now forms the premier for deployment of CMC in airliner engines. The application is on the outer shroud of the first high pressure turbine, Figure 4.
This is a clever choice. Should CFM hit any trouble with such an application it should not be difficult to replace it with a conventional Nickel alloy shroud. The drawback would be higher weight and perhaps more important, an increase of the cooling air flow to the shroud, but the change of the section to a conventional layout would be pretty straight forward. We don’t expect this to happen, GE has major knowledge and investments in CMC technology, but it shows how one approach new technology in a prudent way.
No 8. The Low Pressure Turbine, LPT, is the real work area of the engine. This is where the thousands of horsepower are produced to drive the fan (and the booster compressor but it only consumes a fraction of the power). At hot day sea level takeoff, the MAX 8 would ask the LPT to furnish the fan+booster with 29,000 hp. It does this work by dividing the extraction of the power in the gasses over five turbine stages with flow turning stator guide vanes in-between. It is therefore natural the LPT is a major building block of a modern turbofan and it constitutes an important part of the engine weight. The first stages operate in high temperature (around 1,000°C or 1,800°F) and are implemented in heavier nickel based rotor technologies but when the temperature sinks the last stages are made with lighter Titanium-Aluminide alloy technology.
Monitoring for reliability
The low pressure turbine also houses the Exhaust Gas Temperature (EGT) probe part way into the LPT. EGT temperature is an important health monitoring tool for Turbofans. EGT gives a good indication of the wear state (erosion, deposits) of the whole turbine section. There is an EGT temperature level for a new engine and this temperature then gradually rises as wear and tear makes the engine less efficient. The FADEC then commands more fuel to be injected to achieve the commanded thrust, something that raises the EGT. At maximum allowed EGT level, the engine must be taken off wing for partial overhaul, often called performance restoration, where some of the core components are changed or if the cycle limits of the engine is reached for complete overhaul where most rotating parts of the engine is changed.
Typical EGT margins for new our overhauled engines to maximum allowed are 75-100°C for lower rated engines in a series and 50-70°C for the highest rated variant (EGT is typically given and shown in °C). For highly rated engines in a series (like the CFM56-7B27 for 737-900ER with 27klbf thrust), the 50°C margin can be consumed in less than 10k cycles whereas a lower rated engine like the CFM56-7B26 with 26klbf and a 80°C margin could stay on wing until the best part of 20,000 cycles, all depending on aircraft usage and working conditions.
We can expect the same would apply for the LEAP. A 26klbf rated LEAP-1B could stay on wing, e.g., double the time of a maximum rated LEAP-1B 28klbf for the MAX9 and so on. This is all speculation on our part and it only serves to show that within an engine series there are trade-offs between required performance and time between overhauls. The higher the thrust rating, the lower time on wing before a maintenance action is needed. This maintenance action is not necessarily a large workscope visit. Maintenance planning is all about getting just the right amount of work done on an engine pulled for performance reasons so that it can last until the next scheduled bigger workscope visit (which is governed by performed cycles for most engine parts).
The EGT measurement is far from the only tool for health monitoring of an in-service turbofan. When the engine is developed, a typical pattern of the FADEC values of more than 100 parameters in the engine is built. The engine is then monitored against this pattern when operating. Any change will be noticed by the manufacturer or service provider dependant on the operator’s chosen maintenance program. Should engine parameters start to fall outside the pattern, a warning will be raised and more data will be sought from the engine, such as detailed checks on oil samples taken and on wing boroscope inspections of suspected parts in the engine through the many boroscope port plugs that can be opened. If something is found that is outside of acceptable limits, a preventive shop visit will be planned to correct the problem.
Direct Drive or Geared architecture
There has been a lot written about the fact that the main competitor to CFM, Pratt & Whitney, has chosen a geared architecture for the engine that competes with the LEAP. A geared architecture has advantages when by-pass ratios grow well above 10. The LEAP-1B has a BPR around 9 and the brother engine for the A320/C919, LEAP-1A /-1C around 11. CFM could achieve all goals in this generation of engines when keeping the well known direct drive architecture, where all technologies needed were either proven in service with GEnx or well advanced in different technology programs.
There simply was no need to employ any new technologies like a geared design for CFM. The LEAP program could be a continuation of all that CFM56, GE90 and GEnx had taught GE and Snecma. This might change for the next generation of engines for short-to-medium haul aircraft. Right now CFM feels confident it has the right architecture for the sequel to the world’s most used airline engine, the CFM56.
We have taken a deeper dive into an airliner turbofan using our 737 MAX’s LEAP engine as an example of a state of the art implementation. The description has given examples of real and simulated data to illustrate the typical values for different areas of an engine of this size and technology. This has been for the benefit of the reader’s understanding rather than to try and accurately describe the LEAP-1B. We imagine this has value as it ads color and a bit of hard facts to the rather high level technical discussion the advanced technologies of a modern turbofan leads to.