DOI: 10.1002/mawe.200800293
4-5
Design, production and testing of a titanium
exhaust duct for noise and weight reduction
Helicopter exhausts are usually basic components not yet devel-
oped to achieve the best noise reduction performance. This needs to
change to continue complying with ever more restrictive noise reg-
ulations. A more efficient acoustic design could be reached using a
complex configuration with assembly of multiple walls and 3D op-
timized geometry. As the exhausts nozzle are usually built out of Ni
based alloys, the weight of such a component would increase. Such
an increase could have a negative effect on the payload of the air-
craft. Due to high temperatures and pressures of gas stream, Cp Ti-
tanium is limited for this application but heat resistant titanium al-
loys such as Beta21S and Ti 6 – 2-4 – 2 would be a good substitution
material as it provides significant weight reduction and capacity to
sustain further power increase of the Helicopter engines. An Euro-
pean Project named “HORTIA” was funded to demonstrate the fea-
1 Introduction
Titanium alloys show an excellent combination of high
strength at ambient and elevated temperatures, good corrosion
resistance and low specific weight. In aerospace products like
aircraft, satellites, helicopters, etc. Titanium alloys play their
full advantage in areas with high stresses and/or thermal loads
and/or aggressive environment. Titanium alloy parts can save
half of the mass compared with Steel- and Ni-based alloys.
In course of increased performance level of the engines and
under the hazard of future more demanding noise regulation
due to environmental concerns, the current exhaust design has
to be modified. In previous studies, e.g. SILENCER, the gen-
eral design of a noise reduction item included in an IN625
exhaust duct structure was tested. The design relies on the
Helmholtz principle, which affords a perforated inner skin
and a tight outer shell.
The approach of the actual joint development program that
is partly presented hereafter is targeting on weight reduction
with the introduction of Titanium Alloys Ti6 – 2-4 – 2 and be-
ta21S. The noise reduction performance should be in the same
range, 3,8 dB, as with the previous IN625 test article.
The path we followed to achieve the production of a “scale
1 mock-up” part started with some material research. Micro-
structures and mechanical properties of the selected alloys
were quantified and the suitable heat treatment requested
for the Hortia application was defined taking the temperature
1
FormTech GmbH, Weyhe, Germany
2 load carrying elements were defined by means of load and
Eurocopter, Aeroport International Marseille-Provence, Mar-
ignane Cedex, France
3
Turbomeca Direction Technique, Tarnos, France
4
Prodem, Cornebarrieu, France
5 mize the design of the component.
Exameca, Aéropole Pyrénées Cedex, France
298
Mat.-wiss. u. Werkstofftech. 2008, 39, No.
W. Beck1, D. Chaniot2 , T. Larrouy3, Y. Marcel4,
P. Yannovitch5
sibility of using Heat and Oxidation Resistant Titanium Alloys for
exhaust ducts in a noise and weight reduction objective. Studies
were conducted on these materials to evaluate their mechanical per-
formances and resistance against corrosion at high temperature for a
service application. A thermo-mechanical simulation of stress and
temperature fields validated the definition design of a new compo-
nent in an acoustically efficient configuration. An analysis of form-
ing and welding parameters led to the industrial manufacturing of a
complete exhaust with techniques like Super-Plastic Forming and
Laser Beam Welding. Finally the nozzle was successfully tested on
en engine test bed for endurance and acoustic performance.
Keywords: Helicopter exhaust, Noise reduction, Titanium alloys,
Superplastic Forming, SPF, Hot Forming, Heat resistant Titanium,
Welding, Forming simulation
range of the exhaust nozzle. After a design phase, simulation
and real testing for process parameter identification (hot form-
ing and assembling techniques) gave us key processing data.
Finally all elementary parts that are needed for the completion
of a real exhaust nozzle assembly were produced. The final
part was then tested for temperature distribution, endurance
and acoustic properties
2 Exhaust duct design
The envisaged exhaust duct model is for a Turbomeca gas-
engine installed in a bi-engine helicopter. The shape of the
existing single-wall duct is relatively simple.
The addition of an acoustic treatment onto a gas exhaust
nozzle implies a more complex design due to the double
skin design and a direct increase of the weight. The part count
is more than doubled. With help of Titanium alloys the weight
can be kept down to a level of about the same weight as the
conventional Ni-based duct. The design work was heavily in-
fluenced by the specific requirements of the new forming and
welding techniques like Superplastic Forming (SPF), Gas
Pressure Hot Forming (GPHF) and Laser Beam Welding
(LBW). With help of these processes, it is possible to keep
production cost of Titanium alloys parts at an acceptable level.
Especially the forming techniques made it simple to foresee
“integrated” parts with a complex shape. This avoids much
assembly work.
The service loads and the service temperatures were ana-
lyzed by simulation methods. The wall thickness and the
temperature predictions and the materialś data from sample
tests at high temperature. In order to guarantee the necessary
lifetime for this component the minimum thickness was set to
s = 0.8 mm. A modeling tool was developed in order to opti-
F 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 1. Design of the silent exhaust

Figure 2. Cone test description,

After identification and localization of potential critical
areas, some iteration loops lead to an optimization. Final de-
sign see Fig. 1. By means of suitable design modifications, the 3.2 SPF Parameters
heat flow could be reduced. The partś temperature and/or the
stress level was reduced considerably. The design tool was The HGPF and SPF properties of Ti6 – 2-4 – 2 were inves-
validated during the nozzle tests. tigated in detail with the cone cup test. The test parameters
ranged from 820  C up to 980  C and the strain rate from
1x10-4 up to 1x10-2 s-1. The results were plotted in a 3-D-Dia-
3 Forming process gram, Fig. 3. From the 3-D-chart, the optimum range of para-
meters can easily be identified.
From the selected final design every single part has to be
3.1 SPF versus HGPF process manufactured with sheet metal forming techniques. The pro-
cess studies resulted in various “hot forming”-configurations
Forming of Titanium alloy sheets is particularly difficult
because of the high strength of the material and the large
spring back at ambient temperature. At temperatures between
800 and 950  C however, the flow stress is much lower and the
resulting strain can be some hundred percent if the right set of
parameters, e.g. temperature and strain rate is applied. With
controlled gas pressure, the blank is forced into a mould.
The process has its name SPF if the material has SPF proper-
ties and is formed under application of SPF parameters. If the
process is put outside of the SPF range or if it is a non-SPF-
material, the name GPHF applies.
The suitable set of process parameters is defined by means
of the cone cup test, Fig. 2. The material is tested on the basis
of a test matrix which includes a set of temperatures and strain
rates. For each combination of temperature and strain rate the
cones are formed until failure.
The Optimum combination of temperature and strain rate
allows the highest cone height, This is an indication for the
best plasticity and maximum overall strain without necking.
From the results, it is possible to determine the optimum set of Figure 3. 3-D-plot of “Cone Cup Test”-results for SPF-parameter
applicable parameters for a given part geometry. selection

Figure 4. Typical itera-

tion loop for SPF parts

Mat.-wiss. u. Werkstofftech. 2008, 39, No. 4-5 Design, production and testing of a titanium exhaust duct 299

adapted to the different plasticity of the two different Ti-alloys
and in reference to the complexity of the desired shape.
An “industrialisation loop”, that included the initial tooling
design, forming simulation analysis, iterative modification of
tool geometry was done until a good compromise was
achieved, Fig. 4. The main target was industrial feasibility,
examples of elements considered are. size and performance
of press, extracting angle, etc. But also final part properties
as the thickness profile and cost analysis for future production
is by influenced the creation of the set of dies and jigs.
This work led to the actual definition of all parameters and
forming configurations, temperature, starting thickness of the
Titanium sheet and forming tools required to produce the re-
quested elementary parts.
3.3 Forming of parts
The selected final exhaust duct design resulted in 6 single
parts with different and somehow complex geometry to be
formed and assembled afterwards, Fig. 5. Concerning the
forming operations, the process studies ended up in various
“hot forming” configurations respecting the ductility limita-
tions of the different Titanium alloys in conjunction with
the geometry of the parts. Process variants include SPF,
HGPF and Hot forming with driver sheet.
During “industrialisation phase” the priority was the tech-
nical feasibility combined with the consideration of the cost of
300 W. Beck, D. Chaniot, T. Larrouy, Y. Marcel, P. Yannovitch
Figure 5. Single parts definition
future production. Starting from initial tool design based on
experience from other parts, it was possible to achieve itera-
tive optimisation supported by simulation analysis results for
the whole set of dies and tools. A good compromise has been
found to end up with a suitable thickness distribution on the
parts. The resulting tools have been built in Stainless steel and
Ni based alloy. The tools have a total mass ranging from
300 kg up to 1.2 Tons. Fig. 6 shows the machining of
one of the tools. With assistance of the simulation tool, the
cycle time and the pressure cycle could also be pre-deter-
mined.
The production sequence followed a standard path for this
type of production, Fig. 7. Forming of all the skins and walls
was conducted on Prodem SPF Press during several weeks. As
Figure 6. SPF die on
milling
Figure 7. Production sequence
Figure 8. SPF tool of 1/2 shell in action with
extraction and air cooling of the part
Mat.-wiss. u. Werkstofftech. 2008, 39, No. 4-5
an example, one can see the die for the half-skin production in
hot action with one part beginning to be extracted at more than
900  C and left for air cooling, Fig. 8.
4 Chemical machining of titanium
surfaces
Process and materials analyses previously conducted re-
sulted in a special decontamination sequence for alpha-case
removal. The alpha-case layer has to be removed because
it is the origin of brittle cracking. The sequence was optimized
for refractory Titanium alloys. Hydrogen adsorption and inter-
granular attack is a phenomenon resulting from the chemical
composition of the HF / HNO3 bath. For beta21S, a special
chemical treatment in optimized conditions is necessary to
achieve the proper milling rate, thickness homogeneity, sur-
face roughness, Fig. 9. This etching was successful in avoid-
ing all Inter Granular Attack and exceeding hydrogen uptake.
5 Welding process
The development of the welding processes to assemble the
whole exhaust duct was split into phases:
l Definition of welding parameters of each characteristic ma-
terial, material condition (base material and heat treated and
strained material), material combination (Ti6 – 2-4 – 2 and
beta21S) and welding configuration with samples.
l Study on samples similar to the real exhaust geometry
l Transfer of experience on the real part
For the given materials and welding conditions, three dif-
ferent welding processes were identified and investigated :
Resistance Spot Welding, GTAW and LASER, Fig. 10.
Figure 9. Different surface smoothness as a function of the chemi-
cal-mill process
Mat.-wiss. u. Werkstofftech. 2008, 39, No. 4-5
Figure 11. LBW of SPF’ed beta21S to beta21S
Figure 12. LBW of
SPF’ed Ti6 – 2-4 – 2
welded to Hot formed
beta21S
The welding operations and the QA requirements for the
exhaust duct are classified according Class 2, Further charac-
terization has been done by NDT. The samples and the parts
were X-rayed and underwent a penetrant inspection.
As an illustration of the total scope of welding investiga-
tions (including of all the three different welding processes
associated with similar or dissimilar materials), the results
for LBW are presented. For the Laser Beam Welded samples
in butt joint configuration, following micrographs (Fig. 11 and
12), show the integrity of the weld seam and the heat affected
zone (HAZ), this for the both welding of the same material and
for the combination of Ti6 – 2-4 – 2 with Ti beta 21S.
The welding experiments show the good performance of
the different applied welding techniques. Both the welding
of Ti-material of the same alloy and the combination of
Ti6 – 2-4 – 2 with beta21S proved to be successful. The addi-
tional tensile tests for each of the different combinations of
alloys and heat-treatment conditions clearly show an influ-
ence of the welding operation on tensile strength. As usual,
there is a small reduction of ultimate tensile strength
(UTS). Some hydrogen pick-up during GTAW and LBW of
beta21s cannot be avoided. This contamination leads to a re-
duction of mechanical performance characteristics. Tensile
Figure 10. Matrix for
welding investigations:
materials vs. process
Design, production and testing of a titanium exhaust duct 301
 Figure 13. Exhaust nozzle
after first campaign with
Max Take off Power (MTP)

strength is reduced by 15 %. Special attention must be paid

to unacceptable hydrogen content which results in brittle fail- Figure 14. First campaign. Exhaust nozzle after coloration test at
ure at low strength level. MTP
The specified work package gave satisfactory results and
met the QA requirements. The welding Processes were
then conducted on all elementary parts (fig 6) in order to build l The second campaign targeted on endurance tests. 1000 cy-
the final HORTIA nozzle. The assembly of the complete test cles at Max take-off power have been performed. A full
item in accordance with the construction plan and the given check of the nozzle did not reveal any defects.
tolerances was successfully completed. l The third campaign was realized to measure the acoustic
performance, Fig. 16. Four different engine running condi-
tions, Maximum continuous power, stabilised power, Twin-
engine take-off power and Idle power were applied. A com-
6 Testing and results parison between the HORTIA gas exhaust nozzle and an
existing helicopter ejector resulted in a significant noise re-
The investigation of material’s and process resulted in the duction in the spectrum of frequencies targeted by the de-
establishment of a huge database. The dependency of the ma- finition of the acoustic design.
terial properties on heat-treatment, forming, welding and ac-
cumulated load cycles, representing services requirements,
were stored in the database.
The completed demonstration part was equipped with
probes and tested in full scale on an engine test bed. Both
the temperature distribution under full engine load and the
emitted noise level were investigated.
Three different campaigns of engine bench tests were per-
formed with a Turbomeca engine.
l First campaign suited for thermo mechanical tests, Fig. 13
and 14. Temperatures and mechanical stresses were mea-
sured for three different engine running conditions ranging
from stabilized power to emergency power. The maximum
temperature that occurred was lower than 700  C. Fig. 15
shows the comparison between the predicted and measured
values. The predicted data were validated. A predictive de-
sign tool for such a component was not available before. Figure 16. Test set-up for acoustic measurements

Figure 15. Second campaign result. Validation of

temperature simulation with measured data

302 W. Beck, D. Chaniot, T. Larrouy, Y. Marcel, P. Yannovitch Mat.-wiss. u. Werkstofftech. 2008, 39, No. 4-5
7 Summary
The development work was finished successfully and pro-
vided the information for the potential replacement of heavy
Ni-base exhaust ducts by Titanium alloys. The Titanium al-
loys Ti6 – 2-4 – 2 and beta21S were investigated in detail con-
cerning high temperature strength, aging behaviour due to
long-time exposure, oxidation characteristic, materials prop-
erties after forming and welding etc.This was done in relation
to the mission spectrum of the helicopter and the defined en-
gine. Concerning the production techniques for the complex
shape of the noise damping exhaust duct, the applicable pro-
cesses and parameters for forming, chemical machining and
welding, were elaborated. Superplastic Forming and Gas Pres-
sure Hot Forming techniques were applied for the complex
shapes of the integral parts. The assembly was done by diffe-
rent welding processes such as resistance spot welding TIG
and laser beam welding.
The described work was conducted in the framework of EC
FP.5 under the project title HORTIA=“Heat and Oxidation Re-
sistant Titanium alloys Application.”
Mat.-wiss. u. Werkstofftech. 2008, 39, No. 4-5
8 Acknowledgements
The authors would like to thank the other partners of the
HORTIA project: EADS CASA, Cadiz, Spain; INASMET,
San-Sebastian, Spain; IST, Lisboa, Portugal; NLR, Amster-
dam, Netherlands; TNO, Rijswijk, Netherlands and TUHH,
Hamburg, Germany. The project HORTIA was embedded
in the research activities “New Perspectives in Aeronautics”
supported by the Growth Programme of the Commission of
the European Community in the framework of the Fifth Fra-
mework Programme for Research and Technological Devel-
opment.
Corresponding author: Y. Marcel, Prodem, 84, Route de Seilh,
31700 Cornebarrieu, France, email: y.marcel@bonnans.fr
Received in final form: January 31, 2008 [T 293]
Design, production and testing of a titanium exhaust duct 303