Deutscher Luft- und Raumfahrtkongress 2016

DocumentID: 420207

ASPECTS ON CONCEPTUAL AND PRELIMINARY HELICOPTER DESIGN

A. Krenik, P. Weiand,
DLR Institute of Flight Systems, 38108 Braunschweig, Germany
Abstract
The German Aerospace Center (DLR) is currently developing a multidisciplinary design framework for conventional and
unconventional rotorcraft. This program is the result of the project EDEN (Evaluation and DEsign of Novel rotorcraft
concepts), which is a cooperation of the Institute for Flight Systems, the Institute for Aerodynamics and Flow Technology
and the Institute of Structures and Design. The complexity of rotorcraft design requires the development of several
subject-specific tools. Here, the components are continuously refined and published to the project group. This framework
delivers a high level of modularity that makes the arrangement and testing of the process very flexible. The process
starts with an empty workspace and the minimum input of payload mass, range, flight speed, number of rotor blades and
a rotor configuration. The rotor configurations include also the unconventional types. The fundamental concept behind
the layout of the tools inside the process is demonstrated especially the use of scaling and optimization loops. The
implementation of more than one main rotor is in a special focus. Hence, the toolbox is dealing with different
configurations and different mission requirements, the process itself can be called generic design.

NOMENCLATURE R rotor radius, m

U blade tip speed, m s
a speed of sound, m s
V air velocity, m s
c chord, m
Vground ground speed, m s
c chord ratio
γ Lock number
cT rotor thrust coefficient
δb relative foil thickness
c lα lift curve slope
ε energy ratio
d foil thickness, m κ ov overlapping factor
2
g gravitational acceleration, m s μ advance ratio
l rotor shaft spacing, m ρ air density, kg m3
m rotor disc area overlapping factor ρb blade mass density, kg m3
mb blade linear mass distribution, kg m σ rotor solidity
m̂f fuel co n s umption, kg km Ω rotor frequency, rad s
m f fuel mass flow, kg s
s thrust share coefficient
sfc specific fuel co n sumption, μ g J
t flight endurance, min
vi induced velocity, m s
wd descent speed, m s
Erot rotational energy of the rotor(s), J
Etrans translational energy of the helicopter, J
Ib moment of inertia of the blade, kg m2
Mf fuel mass, kg
Mp payload mass, kg
MOE operating empty mass, kg
MP power source mass, kg
MTO take-off mass, kg
Nb number of blades
Nr number of rotors
PTO power at take-off mass, kW
PZF power at zero-fuel mass, kW
R range, km

©2016 1
 Deutscher Luft- und Raumfahrtkongress 2016

1. INTRODUCTION According to Roskam [4] 65% of the final life cycle costs
The design of rotorcraft is comparable to the design of are determined after the end of the conceptual design
fixed-wing aircraft, a highly multidisciplinary process. The phase. After the end of the preliminary design phase 85%
technical layout and scaling of an aircraft with the respect of the upcoming costs cannot be influenced any more.
to its Top Level Aircraft Requirements (TLARs) has a Therefore the objective is a process chain that combines
sustainable impact on the performance of the resulting the tools of the conceptual phase and as far as possible
configuration. Potential modifications in the design can, as the preliminary phase. Here, the methods used alternated
later they occur, only be implemented with an increase of between applying design tasks and analysis tasks. The
time and costs. essential performance is the functional arrangement of the
The basic performance of a design process is to tools according to their fidelity and computational time
concretize the information in order to increase the fidelity demand.
of the results with every step. In the standard references In resent years several aerospace research organizations
for aircraft design like Raymer [1] or Nicolai [2] the design presented approaches for multidisciplinary design
process is divided into three main phases, the conceptual processes in the field of rotorcraft engineering.
design, the preliminary design and the detailed design. Mentionable are the works of NASA [5], NLR [6], Georgia
Other references, like Layton [3] extend the three phases Tech [7] and ONERA [8]. In the last 10 years the German
with a concept study at the beginning and a design Aerospace Center developed design programs
proposal between the preliminary and the detailed phase. respectively process chains which dealt only with the
In the 1980s and still in the 1990s there was the doctrine design of fixed wing aircraft.
to freeze a design after every phase. Results of a previous In 2011 the project RIDE1 started to build up a process
design phases were not allowed to be changed in the chain for Helicopter design. DLR internal project partners
following phases. With the introduction of computational were the Institute of Flight Systems, the Institute of
multidisciplinary design processes the borders of the Aerodynamics and Flow Technology, both in
phases became blurred. If required tools can give a Braunschweig, and the Institute of Structure and Design in
feedback to preceding tools, then the process has to be Stuttgart. The objective was to build an integrated,
repeated to the originating tool. automated and multidisciplinary design process, with only
The most crucial determination of the external mission requirements, starting from scratch. The project
configuration occurs at the end of the conceptual design had a strong focus on the development of a data set that
phase. At this point the configuration and the major allows a further processing of a rotorcraft design to high
dimensions are defined. The most utilized rotorcraft fidelity analysis and simulation methods in order to get
configuration has one main rotor and one tail rotor to information about the flight dynamics and flight properties
compensate for the main rotor torque. Furthermore, of a virtual rotorcraft. In the process new calculation tools
helicopter can have more than one main rotor (usually just were developed as well as many tools from the fixed wing
two) being a representative of the so-called design were adapted to the boundary conditions of
unconventional configurations. This includes only 4 rotorcraft design. The conception behind these activities
different types: coaxial, intermeshing, side-by-side and was to gather knowledge about the design process of
tandem rotors arrangement, FIG 1. rotorcraft and create a toolbox for generic and virtual
design. In this project with a few exceptions the only
configuration that was analysed was the classic
configuration with one main and one tail rotor.
In 2014 the results and findings of project RIDE were
conventional
processed into the new project EDEN2. Basic work
packages of project EDEN are the further development of
the process chain and the extension of the tools to novel
respectively unconventional rotorcraft configurations. The
design of a rotorcraft with more than one main rotor is an
coaxial inherent capability of the toolbox. Here, the tandem rotors
are partly overlapping and the coaxial rotors fully
overlapping. By adding lifting surfaces and thrust
producing elements like propellers or jet engines all
synchropter helicopter types can be adapted to a compound
(intermeshing) configuration. The adequate calculation of the
interferences of more than one main rotor and other
compound elements is a continuing task in the EDEN
project.
side-by-side
(lateral)

tandem
(longitudinal)

1 Rotorcraft Integrated Design an Evaluation

FIG 1. Different types of helicopters 2 Evaluation and DEsign of Novel rotorcraft concepts

©2016 2
 Deutscher Luft- und Raumfahrtkongress 2016

2. DESIGN LOOP AND INITIALIZATION as structure, avionics, hydraulics, transmission, engine

3 and so on. The 2nd station – aerodynamics – can be run
The design process loop starts with the so-called TLARs through in parallel to the 1st, even though it is not the case
consisting of only three parameters: in FIG 2, since there is no dependency between them. The
- cruising speed (not maximum speed) 3rd station estimates engine power required, the engines’
masses and the fuel consumption characteristics.
- payload (which has to be transported at the
defined cruising speed) The 4th station is the 1st nested loop, in which the
necessary amount of fuel referring to the TLARs and the
- range (which has to be fulfilled at the defined chosen engine is calculated.
cruising speed, transporting the defined payload )
The 5th station is the 2nd nested loop dealing with the rotor
The initialization of the process creates the first dataset. parameter optimization, such as blade dimensions and
The basis for this initial step is a statistical dataset of 159 blade loading.
existing helicopters, accumulated during the project RIDE.
From here, all necessary parameters are evaluated to The 6th station is the 3rd nested loop and optimizes the
create an initial dataset, including the fuselage geometry blade twist at the cruising speed for the minimum power
choice and scaling. requirement.
Then, the created dataset is processed to a scaling loop. The design loop continues until significant changes in the
FIG 2 shows the basic layout. It iterates the maximum take take-off mass disappear.
of mass, which is the sum of the operating empty mass, At the end the converged design can be fed to the
payload mass and the fuel mass (1). envelope calculation with a wide array of flight parameters.
(1) MTO  MOE  Mp  Mf Depending on the TLARs first dimensioning, both the
three-dimensional expansion of the helicopter and the
mass fractions are determined. FIG 3 shows a plot of the
Basic Transportation Requirements:
operating empty mass over the payload mass. In addition
Cruising Speed, Payload, Range FIG 4 gives a plot of the fuel mass over the payload mass.
In both cases the fractions of the take-off mass are
First Parameter Estimation parameterized by a potential approach. Depending on the
from Statistics primary design mission the variance in fuel mass to the
best-fit curve can be moderate or high.
4
Geometry Choosing and Scaling x 10
fit
2.5

Complete Covergence of
Envelope Caclulation T/O-Mass
2
MOE [kg]

1.5
Mass Break-Down
of 1
Aircraft Components

0.5
Aerodynamics of
Drag producing 0
Elements 10
3 4
10
Mp [kg]

Power Source FIG 3. Statistical data of the operation empty mass over
Choosing
the payload mass with best-fit curve
7000
fit
Covergence of Fulfillment of 6000
Fuel-Mass Range Requirement
5000

4000
Mf [kg]

Fulfillment of
Covergence of Boundary Conditions
Rotor-Mass (Rotor Optimization) 3000

2000
Fulfillment of
Covergence of 1000
Boundary Conditions
Power (Blade Twist Optimization)
0
3 4
10 10
Mp [kg]
FIG 2. Flowchart of the design loop
FIG 4. Statistical data of the fuel mass over the payload
mass with a best-fit curve
Inside the loop the 1st station is the update of the operating
empty mass by breaking it into component masses, such Following the given TLARs the dimensions and positions
of the rotors and the stabilizers are predicted for a desired
configuration. In the data set, Rotor 1 is the main rotor for
3
Top Level Aircraft Requirement

©2016 3
 Deutscher Luft- und Raumfahrtkongress 2016

the standard configuration, the front rotor for the tandem 3. ITERATIVE PROCESSING AND OPTIMIZATION
or the upper rotor in case of the coaxial configuration, left
for the synchropter and side-by-side configuration. Rotor 2 The design loop and the optimization algorithms are
is consequently the tail rotor, rear rotor, lower rotor or right implemented within the distributed, workflow-driven
rotor. Every design approach considers right from the integration environment RCE4, developed by DLR.
beginning at least two rotors. Other compensations for 3.1. Mass Estimation of Components
main rotor torque, like NOTAR or Fenestron are not
included for now. The more sophisticated prediction of the operating empty
mass is conducted by calculating the masses of the
The size of the fuselage, respectively the body is derived several components. Therefore formulas by Beltramo [9],
from the required cabin volume. A generic fuselage is Layton [3] and Prouty [10] are available. In addition
scaled according to the required cabin. The fuselage is Palasis [11] has already analyzed a combination of the
arranged under the rotors with the centre of the payload formulas by Beltramo and Layton. The weight models of
mass in the centre of gravity of the configuration. The the AFDD (Aero Flight Dynamics Directorate) presented
fuselage body can be separated into “FuselageFront” with by Johnson [5] are most recent and have highest coupling
cockpit and the avionics, the “FuselageMid” with the cabin between geometric and performance parameters with the
and the “FuselageRear” with a possible ramp or a cargo respective component weight. Small changes in the
hold. FIG 5 shows the arrangement of the components geometrical dimensions result in recognizable changes of
and the component parts. The dimensions of the front and the empty mass. Consequently the size of the rotors and
rear fuselage body are at minimum derived by standard the required installed power is scaled with the take-off
aspect ratios but vary by adding requirements for mass. A best of combination of the available models is still
retractable nose gears or enhanced avionics at the front or in progress.
special loading characteristics at the stern ramp.
After calculating the updated maximum take-off mass a
convergence check is conducted. The data set is iterated
until mass, size and flight performance is consistent and
the loop is converged. Iterations to minimize deviation
below 1% of the take-off mass are not reasonable since
the accuracy of the utilized tools is recognizable lower.
The framework software needs an additional feedback
from all nested loops to insure they have finished their
computation before. In total three nested loops were used
inside the outer scaling loop. The previous initialization
and the subsequent calculation of the flight envelope are
arranged straight forward, see FIG 2.
3.2. Fuselage Aerodynamics
From the adapted external configuration a 3D-model is
created according to FIG 5. The approach based on the
work of Kunze [12] then calculates the aerodynamic
FIG 5. Component parts and assembly of the generic properties of the specific 3D-model by the use of a
fuselage approach modified version of VSAERO [13]. The program uses a
linearized 3D-panel-method with a coupled viscous solver.
While the scaling of the rotors is related to the maximum
FIG 7 shows the pressure distribution on the surface of a
take of mass, the fuselage is scaled from the payload
fuselage up to the separation line. In general the
volume. The correlation of the regression curve and the
helicopter fuselage has a region of separated flow on the
statics, presented in FIG 6 is very good. By including other
backside below the tail boom, due to its blunt shape.
geometric ratios the prediction of an average cabin shape
Vortexes of separated flow strongly increase the parasite
from the take-off mass gives reasonable initial results.
drag reducing the flight performance in terms of increasing
35
fit
the required power. Therefore flow separation has to be
30 taken into account. The automated modelling of the
streamlines by the use of wake panels showed to be
25 problematic and less robust. Another method is to assume
a constant pressure on the surface panels with separated
Lfus [m]

20
flow. The line of separation is predicted by boundary layer
15
code integrated into VSAERO. The value for the pressure
10 coefficient behind the separation line can be calculated or
manually set. The calculation of the clean fuselage surface
5
is extended by some handbook methods to take additional
0
3 4
parts like the rotor hub, landing gears and more into
10 10 account. The aerodynamic drag is the basis for the flight
MTO [kg]
performance calculation.
FIG 6. Statistical data of the fuselage length over the
maximum take-off mass with a regression curve

4
Remote Component Environment

©2016 4
 Deutscher Luft- und Raumfahrtkongress 2016

In the hover case of the coaxial arrangement the lower

rotor is highly affected by the contracted downwash of the
upper rotor as can be seen as a concentric disc of 90% of
the radius. In forward flight the downwash of the upper
rotor is swept back downstream, reducing the affected
area of the lower rotor disc. In the hover case of the
tandem arrangement (FIG 9) the fore rotor is highly
affected by the downwash of the aft rotor, as the aft rotor
is above the fore rotor. In forward flight the intermeshing
area shrinks due to downstream inclination of both rotor
wakes – in principle due to the same reasons as with the
coaxial type.
FIG 7. Results of the calculation of the pressure The impact on power consumption compared to the in
distribution on the helicopter fuselage reality non-existing but calculated case of non-interacting
treatment of the rotors is shown in FIG 10. The increased
3.3. Rotary Wing Aircraft Calulations power consumption is highest at hover decreasing almost
All required flight mechanics calculations within the design to negligible values at cruising speed. To take the rotor
loop are performed by the Helicopter Overall Simulation head penalty into account each rotor shaft is modeled as a
Tool (HOST). The word tool at the end of the acronym is a cylinder in a flow and the rotor blade area between the
clear understatement, because the code behind it can be rotor shaft and the aerodynamic beginning (beginning of
considered as a full-fledged virtual rotorcraft design and the airfoil) is modeled as a 2-dimensional drag producing
evaluation operating system. The addressed functions of cylinder section.
HOST in the process chain are only a fraction of its
possibilities. The main task here is to perform trim ↑µ
calculations referring to the TLARs with continuously

induced velocity
varying datasets. The main output is the required power
consumption.

m/s
The special nature of all multiple main rotor configurations
is the existence of mutual rotor-rotor interactions, which
are most pronounced at hover and low flight speeds. For
this reason models of super positioned induced velocity Fore Rotor Fore Rotor
fields were developed and applied with HOST. The basic
approach is to perform one calculation without mutual
rotor interactions and to use the results to shape the Aft Rotor Aft Rotor
additional velocity field as an overlay and to perform a ↑µ
second trim calculation with the synthesized overlay. The
factors for the induced velocity increase were derived from
Johnson [16]. FIG 8 and FIG 9 show the induced velocities
at hover and in forward flight for a coaxial and a tandem
configuration after (double) trim calculations.

↑µ
induced velocity

FIG 9. Resulting induced velocity fields in hover (left)

and forward flight (right) at Ω=424 rpm, µ=0.08,
m/s

2.2 tons take-off mass for the tandem rotor

arrangement
8
coaxial
Upper Rotor Upper Rotor 7
tandem

Lower Rotor Lower Rotor 5

↑µ
ΔP [%]

0
0 0.05 0.1 0.15 0.2 0.25 0.3
μ [-]

FIG 10. Increase of power consumption due to mutual

FIG 8. Resulting induced velocity fields in hover (left) rotor interactions, relative to the non-interacting
and forward flight (right) at Ω=424 rpm, µ=0.08, case
2.2 tons take-off mass for the coaxial rotor
arrangement

©2016 5
 Deutscher Luft- und Raumfahrtkongress 2016

3.4. Estimation of Engine Mass and Fuel range calculation. With this approach there is no need to
Consumption estimate the absolute required power for the concerning
flight regimes of the preliminarily designed helicopter.
The adequate choice of the power source within the
iterative preliminary design loop is not a matter of 3.5. Endurance and Range Calculation
browsing a catalogue of available engines. Instead the
evaluation of the engines database is paired with the At this point of the design loop the dataset of the
database of the helicopters. The ratio of the overall helicopter is defined. Now, the flight mechanics
engine’s mass related to the maximum take-off mass of characteristics at steady flight conditions – to be fulfilled as
each helicopter from the paired databases is shown in FIG set up from the basic requirements – like cruising speed
11. with the defined payload and range at cruising speed can
be calculated.
0.12
light light
light medium For that purpose the endurance and consequently the
0.1
light heavy
medium range calculation can be easily performed and furthermore
heavy
super heavy the required fuel mass can be accurately determined, in
[-]

mean fit
0.08 contrast to the roughly estimated amount before (shown in
TO(max)

FIG 4).
M /M

0.06
The endurance calculation is accomplished with the so-
P

0.04
called “mean fuel flow”-method, which was found to be
very accurate throughout the velocity envelope, requiring
0.02
only 2 trim-calculations (1. with a full fuel tank and 2. with
10
3
10
4
10
5
zero fuel mass) for each flight regime. The mean fuel flow
MTO(max) [kg]
is:
FIG 11. Fraction of the installed power mass related to
the maximum take-off mass f 
m
1
m 
2 f  TO
m f ZF
  
The helicopters were subdivided into 6 weight classes,
starting with the maximum value of 2 tons for the “light-
(2)
 

1 fengine PTO M TO  
  

light”, doubling the maximum value for each following
higher class up to the “super-heavy” (maximum of 64
2  f 
 engine ZF  ZF
P M  M TO  M f   

tons). For each weight class the data was averaged and a
The resulting flight duration then is simply:
fit-function was applied.
It can be seen that the mean fraction of the installed power Mf
(3) t flight 
mass related to the maximum take-off mass decreases f
m
throughout the first 5 weight classes, just slightly
increasing at the few realized “super-heavy” helicopters. The range is calculated as follows:
The explanation for this trend might be the circumstance
that in general the effectiveness of the shaft-turbine (4) R  Vground  t flight
engines increases with the overall size, also conditioned
with the technological issue of more experience with the With this output the required fuel mass can be found by
lighter weight classes, in contrast to the older few application of an iterating and converging loop with the
individuals of the “super-heavy” class. following formulation at which the following outputs from
HOST trim-calculations have to converge:
The same procedure as before was applied for the
estimation of the specific fuel consumption at take-off M f,old
power setting (FIG 12). The trend here is almost the same 
OutputsHOST : Ractual
as with the engine’s masses.
R
150  required
light light
140 light medium
light heavy The difference of the actual and required range is
medium
130 heavy
super heavy
calculated in the iteration loop:
mean fit
[μ g/J]

120
(5) ΔR  Ractual  Rrequiered
110
TO
sfc

100 The average fuel consumption per unit flight range is:
90
(6) m̂f  Mf,old Ractual
80

70
10
3
10
4 5
10
The improved fuel mass setting for the required range
MTO(max) [kg] from (5) and (6) then is:

FIG 12. Specific fuel consumption at take-off power (7) Mf,new  Mf,old  m
ˆ f  ΔR
setting
The new fuel mass is then handed over into the trim-
The specific fuel consumption for the continuous power calculation from the beginning of the loop. Mostly it takes 3
regimes is in most cases 8 to 9% higher than that of the iterations to fulfill the requirement (in terms of calculated
take-off regime. So the values of the fit-functions are range) with a deviation of less than 2 km.
multiplied with the factor of 1.1 for the endurance and

©2016 6
 Deutscher Luft- und Raumfahrtkongress 2016

At that point the fuel tank is exactly filled and the overall 16
mass is (in most cases) reduced, compared to the
14
estimation from the initiation, based on statistics.
12
The next step in the loop is the rotor parameter 10
determination to fulfill the set up boundary conditions and

R [m]
8
in general to optimize them in terms of one or more
selected criteria. 6

4
3.6. Rotor Parameter Optimization
2
The rotor parameter optimization follows the “call and 0
response”-approach, meaning that there are only three
3 4
10 10
MTO(max) [kg]
parameters of the rotor which are the calling inputs
(number of blades is set and the rotor blade twist is
FIG 14. Rotor radius from helicopter database
considered later in the design loop):
 Rotor Speed
- rotor blade chord
The rotational frequency or rotor speed (FIG 15) of the
- rotor blade length
rotor has reciprocal relation to the rotor radius, as the tip
- rotational frequency of the rotor speed is mostly set within a relative thin margin:
On the other side the dependent functions from inputs U
serve as responses. For that purpose the module ParRot5 (8) Ω
R
was developed and coupled with the optimization
algorithm COBYLA6 [14]. The optimization goal is to 60

achieve the lightest rotor blade. 50

3.6.1. Design Parameters (inputs)

40
Ω [rad/s]

The main three rotor design parameters are initially set

from the evaluation of the statistical database and their 30

lower and upper boundaries form the permitted design 20

space for iterative optimization process. In the following
the design input parameters extracted from the database 10
are presented.
0
 Chord
3 4
10 10
MTO(max) [kg]
The chord (FIG 13) has a slightly degressive trend (even
though the logarithmic scaling distorts it to the opposite) FIG 15. Rotational frequency of rotors from helicopter
from the database. database
1  Number of Blades

0.8 The only parameter, which is not varied throughout the

design and optimization loop is the number of rotor blades
0.6
per rotor, because of the natural number type, which is not
suitable for most of the optimization algorithms and
c [m]

0.4
because it is a matter of preferences, as can be seen in
FIG 16. The simplest rotor of two blades was used for
take-off masses of up to 8 tons (Bell 214-ST). The widest
0.2
range have the 3-bladed rotors with take-off masses of up
to 13 tons (with coaxial rotors of the Ka-27). The single
0
10
3 4
10 rotors end at 3.6 tons of the S-62. The 3-bladed rotor is
MTO(max) [kg] preferred by all tandem configurations and (except for the
X² and S-97 from Sikorsky) is still standard to the coaxial
FIG 13. Chord from helicopter database configurations among Kamov’s designs. It can be said that
 Radius any number of blades between 2 and 4 can be chosen in
the weight category between 1 and 10 tons. Above that 5
The rotor radius (FIG 14) follows a more precise trend to 8 should be preferred.
than the chord, but also degressive with increasing take-
off mass.

5
Parametrization of the Rotor
6
Constrained Optimization BY Linear Approximation

©2016 7
 Deutscher Luft- und Raumfahrtkongress 2016

7
upper boundary
6
0.1
5

σ [-]
Nb [-]

3 0.05

2
lower boundary
1
0
0 10
3 4
10
3 4
10 10 MTO(max) [kg]
MTO(max) [kg]

FIG 18. Rotor solidity from helicopter database

FIG 16. Number of blades per rotor from helicopter
database  Blade Loading
So, on the one hand there are three parameters of the Blade loading is derived from division of the thrust
rotor, which are to be found optimal for a specified flight coefficient by the rotor solidity, describing the aerodynamic
condition (or design point) and on the other hand there are workload of each blade:
(some) rival and conflicting parameters (among those the
inputs themselves), which are constricted by boundary s  MTO g
(11) CT σ 
conditions with the aim of receiving only one solution. This ρ c Nb R U2
is true for example for the rotor length and the rotor
frequency in terms of blade tip speed. To address the multirotor configurations properly the trust
sharing coefficient is introduced. At equal thrust sharing of
3.6.2. Design Characteristics (outputs) two rotors it would be 0.5. But both the tandem and
The design characteristics from the variation of the afore coaxial arrangement have one rotor, which produces more
mentioned parameters, or in that case - responses, are than 50% of the take-off mass thrust, so the coefficient is
described in the following. set to 0.6, as it corresponds to the results by Bourtsev et
al. from [15].
 Chord Ratio
In FIG 19 it can be seen that the majority of the helicopters
The ratio of the blade chord to the blade radius throughout the weight classes is set for a value of
somewhere between 0.07 and 0.10 (in hover). The
(9) c c R maximum value of the blade loading in the design loop
has no recognizable trend, hence can be treated as a should not exceed 0.12 as it is considered as boundary
constant bandwidth throughout the whole database, as value for rotor stall [16].
can be seen in FIG 17. So the boundary conditions can be
roughly set to 0.04 and 0.08.
upper boundary
0.1
C /σ [-]
T

0.1 lower boundary

upper boundary
0.05
c [-]
-

0.05
0
3 4
lower boundary 10 10
MTO(max) [kg]

0
10
3
10
4
FIG 19. Blade loading from helicopter database
MTO(max) [kg]
 Blade Tip Speed
FIG 17. Chord ratio from helicopter database
The blade tip speed has also an almost universal
 Rotor Solidity character, varying between values of 180 m/s and 220 m/s
(FIG 20). Even though there are many successful designs
The rotor solidity is the ratio of all blade planforms related
with speeds of 150 m/s but mostly only with light and ultra-
to the rotor disc area:
light designs, which are not designed for the speed of
Nb c R Nb c more than 200 km/h and therefore do not exceed high
(10) σ  advance ratios either.
π R2 πR

It has a degressive increase with the take-off mass as can

be seen in FIG 18. In general the boundary conditions can
be set to 0.03 and 0.12.

©2016 8
 Deutscher Luft- und Raumfahrtkongress 2016

250 0.5
upper boundary
upper boundary
200 0.4

lower boundary
150 0.3
U [m/s]

μc [-]
lower boundary
100 0.2

50 0.1

0 0
3 4 3 4
10 10 10 10
MTO(max) [kg] MTO(max) [kg]

FIG 20. Blade tip speed from helicopter database FIG 22. Advance ratios from helicopter database
 Mach Number at Cruising Speed  Lock Number
The Mach number at cruising speed is determined by the The Lock number is an important parameter characterizing
set up blade tip speed: the flapping behavior of the rotor blade:

(12) Mac 
 Vc  U (14) γ
ρ c l αc R4
a Ib
In most cases a value of 0.85 can be defined as good It represents the ratio of the aerodynamic and inertial
target value for the design, as values above cohere with forces acting on the blade. Lower values of the Lock
high profile drag at the advancing blade side and therefore number are associated with a high (and in most cases
high power consumption and vibrations. In FIG 21 it can preferred) inertial characteristic of the blade, as the
be seen that there is only a thin margin of values where flapping moment of inertia is in the denominator of the
the majority of the designs is spread. equation.
1
upper boundary In FIG 23 there is only a selection of the database shown,
as the data concerning the blade masses and the
0.8
moments of inertia are more difficult to acquire or simply
lower boundary
not published.
0.6
Mac [-]

15
R-22
0.4

Mi-4 Mi-17 Mi-26

0.2 10
BO-105 CH-47
γ[-]

0 UH-1 Mi-24 CH-53

3 4
10 10 AS-365
MTO(max) [kg] 5 EC-135

FIG 21. Mach number at blade tip from helicopter

database
0

3 4
Advance Ratio at Cruising Speed 10 10
MTO(max) [kg]
The advance ratio at cruising speed is the rival parameter
to the Mach number as the blade tip speed is in the FIG 23. Lock Number from reduced helicopter database
denominator of the equation: Nevertheless, the Lock numbers are found to have values
Vc between 5 and 12, larger with increasing take-off masses.
(13) μc 
U  Energy Ratio
In most cases a value of 0.3 can be defined as good target Another parameter concerning the responses of the rotor
value for the design at cruising speed to have some speed optimization focuses on the autorotational behavior of the
reserves. The advance ratio can be seen as a soft border, helicopter. The main idea behind this is to consider the
compared to the Mach tip speeds. Though high values kinetic energy of the rotor(s) at nominal rotor speed
cohere with high vibration levels as it describes the related to the translational kinetic energy of the helicopter
diameter of a circle of reversed flow on the retreating as a whole in vertical autorotation.
blade side of the rotor disc. In FIG 22 it can be seen that
the majority margin is between 0.25 and 0.4. Erot 1 2  Nr  Nb  Ib  Ω2
(15) ε 
Etrans 1 2  MTO w 2d

The vertical descent velocity at ideal autorotation can be

estimated as 1.8 times the induced velocity in hover [16]:

©2016 9
 Deutscher Luft- und Raumfahrtkongress 2016

MTO  g The area ratio of the profiled section is usually a function

(16) w d  1.8  v i  1.8  of the relative thickness, but can be assumed constant to:
2  ρ  κ ov  π  R2
Ab
The overlapping factor for two rotors configurations is (21) Ab   0.7
δb  c 2
calculated from Johnson [16] to:
The mass distribution of the blade is then:
2
(17) κ ov 
2m (22) mb  δb  c 2  Ab  ρb
For the coaxial rotor arrangement the value of m is 1 and With it, the blade mass is calculated to:
hence the overlapping factor is the square root of 2. For
the tandem arrangement the factor is a function of the (23) Mb  mb  R
rotor shaft spacing:
The moment of inertia of each blade with a correction
 2
factor of 1.1 (to fit the moments of inertia in the database)
2  l  l  l  
(18) m  arccos    1    due to non-uniformity of the blade mass distribution
π 2R 2R 2R 
  becomes:

With these formulas equation (15) becomes: 1

Mb  1.1 R 
2
(24) Ib 
3
Erot Nr  Nb  Ib  Ω2
(19) ε  The last equations were applied on the (reduced)
Etrans M2TO  g
1.62  helicopter database to get an insight on the blade mass
ρ  κ ov  π  R2
density trend. In FIG 25 it can be seen that the density is
In FIG 24 the energy ratio values are shown for the same very high for the shown light helicopters, continuously
reduced database. decreasing to almost one third of the heaviest Mi-26,
compared to the maximum value of the R-22. For the rotor
3
Mi-4
optimization the density between the upper and lower best
2.5 UH-1 fit curves is used.
1200
R-22 upper fit
2 lower fit
1000
AS-365
ε [-]

1.5
Mi-17 800 BO-105
EC-135
ρb [kg/m3]

Mi-24 EC-135
1 CH-53
R-22 BO-105 Mi-26 Mi-4
CH-47 600
Mi-24 CH-47
0.5 UH-1 CH-53
400
Mi-17 Mi-26
AS-365
0
3 4
10 10 200
MTO(max) [kg]
0
3 4
10 10
FIG 24. Energy ratio from reduced helicopter database MTO(max) [kg]

It can be seen that the energy ratios of the selected

FIG 25. Blade mass density from the reduced helicopter
database start at values of 1. Helicopters with known as
database
“good at autorotations”-behavior such as UH-1 and Mi-4
exhibit values above 2 and representatives of the “bad at
autorotations”-behavior such as the light R-22 hold values 3.7. Blade Twist
of 1. With a high energy ratio (above 2) it is possible that
the helicopter should be able to perform a horizontal The blade twist of the rotor is optimized for the lowest
autorotational descent to the ground. With a value of 1 this required power needed at the required speed. The only
is not feasible, because it is impossible to use the entire restriction is that the collective blade angle may not
rotational energy for the complete deceleration of the exceed 20° (to not further decrease the effectiveness of
descending helicopter. The rotor thrust breaks down due the rotor at hover, since the inner sections of the blade will
to stall at 60-70% of the nominal rotor speed [16]. get into the overpitched stall).

 Blade Mass Density After the first successful rotor optimization loop the rotor
masses and with it the take-off mass changed. So the
The last parameter concerning the rotor design and whole design is handled back to the 1st station of the
optimization is the blade mass density, which is also design loop again. In that case the re-calculation of the
analyzed from the (reduced) helicopter database. fuselage mass, as the geometry of the fuselage due to
rotor parametrization changed as well.
The parametrization of the blade geometry follows a
simple approach considering the relative blade thickness The 2nd station is the re-calculation of the fuselage
as constant: aerodynamics.
d The 3rd station is the power source estimation and
(20) δb   0.12 consequent refilling to fulfill the unchanged requirements
c
from the initialization.

©2016 10
 Deutscher Luft- und Raumfahrtkongress 2016

The 4th station is the 2nd entry into the rotor optimization the original value, so is the installed power, indicating less
loop, and so on. parasite drag than the original.
The results of the multirotor configurations had reached
4. RESULTS almost the maximum set blade loading of 0.1 with the
maximum tip speed of 220 m/s. The main difference and
In the project EDEN the design loop and its components hence influence on the rotor optimization of the multirotor
are constantly improved with the aim to achieve the ability configurations was found for the deviation of the empty
to perform preliminary design not only for the 3 main mass compared to the originals – the re-design became
helicopter types, but also for their derivates into compound 16-28% lighter. With it, the required power and the needed
helicopters. fuel mass also dropped below the original values.
To demonstrate the interim abilities of the design loop the Mi-8

following helicopters from three configurational types were

Ka-226
CH-46

subjected to a “re-design”: 1.2

1.1

- Mil’s Mi-8 1

0.9
- Kamov’s Ka-226

relative change [-]

- Boing Vertol’s BV-107 (known as CH-46)
From each helicopter the known data of payload capacity,
cruising speed and the corresponding range were used as
TLARs for the EDEN design loop. All helicopters were
subjected to the same boundary conditions within the rotor
optimization process. In FIG 26 the rotor optimization of
the Mi-8 is shown as overall time history. 0
c [m] R [m] Ω [s- 1] -
c [-] C /σ [-] U [m/s] Mac [-] μc [-]
T
0.6
0.4 FIG 27. Rotor optimization results of the re-design
c [m]

0.2
1. Optimization 2. Optimization 3. Optimization
0 Mi-8
Ka-226
CH-46
5 10 15 20 25 30 35 40 45
iteration
1.1
1
11
R [m]

0.9
relative change [-]

10
0.8
9

5 10 15 20 25 30 35 40 45
iteration

24
Ω [rad/s]

22
20

5 10 15 20 25 30 35 40 45 0
iteration Vc [km/h] R [km] Mp [kg] MTO [kg] MOE [kg] Mf [kg] Pc [kW]

FIG 26. Rotor optimization process of the Mi-8 re-design FIG 28. Mass and performance results of the re-design
Comparing the initial with the final values, it can be seen
that every parameter increased. The main changes were
th 5. CONCLUSION AND OUTLOOK
made in the first optimization call (until 17 iteration). The
radius increased up to 11 meters, the rotor frequency had The objective of the project EDEN is a generic design of a
the tendency to compensate it, regarding all the boundary rotorcraft according to a given set of TLARs. It has been
margins. The algorithm also tried negative values for the demonstrated that different configurations were designed
chord (which cannot be prevented, due to the search for a for a given minimal set of mission relevant requirements.
global optimization goal), but instantly returned back. In general it can be concluded that the rotor optimization
Comparing the following 2 calls with the first, it can be presented here finds very reasonable parameter sets,
seen that all parameters had a more relaxed behavior. even with wide margin boundary conditions.
The re-design results of all three helicopters compared to  The porting to the DLR’s software framework
the original values are shown in FIG 27 and the results of RCE including the networking of the three
the masses and installed power in FIG 28. One of the institutes was achieved and nearly all tools are
striking results is the blade tip speed, which hit the upper reintegrated.
boundary of 220 m/s. The blade radius of the Mi-8 was
reduced by 10%, the chord is increased by almost the  On the basis of the new framework an iterative
same factor. All other rotor characteristics are very close design loop was created, which includes the rotor
to the original values. The resulting take-off mass is optimization loop.
almost identical to the original value, the empty mass now  A concept for the generic helicopter design for
is 2% higher. The needed fuel mass is almost 15% below different configurations was developed and
implemented.

©2016 11
 Deutscher Luft- und Raumfahrtkongress 2016

The design loop itself showed good results, but the main [13] B. Maskew, “Program VSAERO theory document -
difficulty is the estimation of the operating empty mass and NASA CR-4023”, Redmond, WA, 1987.
the estimation of the aerodynamic drag, which in turn
results in underestimation of the installed power and less [14] M. J. D. Powell, “A view of algorithms for optimization
required fuel mass. Following steps have to be performed without derivatives”, Cambridge, UK, 2007.
in the near future: [15] Bourtsev, B.N., Ryabov, V.I., Selemenev, S.V.,
Butov,V.P. “Helicopter Wake Form Visualization
 The prediction of the operation empty mass (for
Results and their Application to Coaxial Rotor
unconventional configurations) has to be
Analysis in Hover”, 27th European Rotorcraft Forum,
improved.
Moscow, Russia, September 11-14, 2001.
 The full integration of the generic structural
[16] Johnson, W., “Rotorcraft Aeromechanics”,
design including the FEM analysis has to be
Cambridge University Press, 2013.
completed.
 Improvement of overall drag prediction methods.

6. REFERENCES
[1] D. P. Raymer, “Aircraft Design: A Conceptual
Approach”, Reston, VA, American Institute of
Aeronautics and Astronautics, 2006.
[2] L. Nicolai and G. Carichner, “Fundamentals of
Aircraft and Airship Design, Volume 1“, Reston, VA,
American Institute of Aeronautics and Astronautics,
2010.
[3] D. M. Layton, “Introduction to Helicopter Design”,
AIAA Professional Studies Series, Salinas, CA,
1992.
[4] J. Roskam, “Airplane Design”, Lawrence, KS,
Design, Analysis and Research Corporation
(DARcorporation), 1985.
[5] W. Johnson, “NDARC - NASA Design and Analysis
of RotorCraf,”, Moffett Field, CA, 2009.
[6] J.-F. Boer and J. Stevens, “Helicopter Life Cycle
Cost reduction through pre-design optimisation”, 32nd
European Rotorcraft Forum, 12-14 September 2006.
[7] A. Khalid and D. P. Schrage, “Helicopter Design Cost
Minimization using Multidisciplinary Design
Optimization”, 63rd Annual Forum and Technology
Display of the American Helicopter Society, May 1-3,
2007.
[8] P. Basset, A. Tremolet, F. Cuzieux, C. Schulte, D.
Tristrant, T. Lefebvre, G. Reboul, F. Richez, S.
Burguburu, D. Petot, and B. Paluch, “The
C.R.E.A.T.I.O.N. project for rotorcraft concepts
evaluation : The first steps”, 37th European Rotorcraft
Forum, September 13-15, 2011.
[9] M. N. Beltramo and M. A. Morris, “Parametric study
of helicopter aircraft systems costs and weights”,
Hampton, VA, 1980.
[10] R. W. Prouty, “Helicopter Performance, Stabililty, and
Contro”, Malabar, FL, Krieger Publishing Company,
1990.
[11] D. Palasis, “Erstellung eines Vorentwurfsverfahrens
für Hubschrauber mit einer Erweiterung für das
Kipprotorflugzeug”, Düsseldorf, 1992.
[12] P. Kunze, “Parametric Fuselage Geometry
Generation and Aerodynamic Performance
Prediction in Preliminary Rotorcraft Design”, 39th
European Rotorcraft Forum, 2013.

©2016 12