The relationship between angular

velocity and power efficiency of a

helicopter in hover

Research Question:​ What is the relationship between angular velocity and

power efficiency of a twin bladed, single rotor helicopter system in hover?

Word Count:​ 3998

Subject:​ Physics
Contents

1 Introduction 3

2 Background Information 4
2.1 Induced velocity and thrust . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Power and power efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

3 Extending Power Efficiency 8

4 Slipstream Correction for Pitch Angle 9

5 Determining Thrust Coefficient 11

5.1 Determining inflow angle of attack . . . . . . . . . . . . . . . . . . . . . . . . 12
5.2 Determining lift-curve gradient . . . . . . . . . . . . . . . . . . . . . . . . . . 14
5.3 Determining thrust coefficient . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

6 Determining Power Coefficient 19

6.1 Determining drag coefficient . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
6.2 Determining power coefficient . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

7 Determining Experimental Power Coefficient 24

7.1 Experimental Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
7.2 Experiment procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
7.3 Theoretical vs experimental power coefficient . . . . . . . . . . . . . . . . . . . 25

8 Determining Power Efficiency 27

9 Evaluation and Conclusion 28

10 Works Cited 30

11 Appendix 31
11.1 Appendix 1: Simplifying Reynolds Number . . . . . . . . . . . . . . . . . . . . 31
11.2 Appendix 2: Conversion between Angular Velocity and Reynolds Number . . . 32

1
11.3 Appendix 3: Determination of Lift-curve Gradient . . . . . . . . . . . . . . . . 33
11.4 Appendix 4: Determination of Thrust Coefficient . . . . . . . . . . . . . . . . 34
11.5 Appendix 5: Determination of Drag Coefficient . . . . . . . . . . . . . . . . . 35
11.6 Appendix 6: Determination of Power Coefficient . . . . . . . . . . . . . . . . . 36
11.7 Appendix 7: Raw Experimental Data . . . . . . . . . . . . . . . . . . . . . . . 37
11.8 Appendix 8: Determination of Power Efficiency . . . . . . . . . . . . . . . . . 44
11.9 Appendix 9: Arduino Code for Modulating Motor RPM . . . . . . . . . . . . . 45

2
1 Introduction
Helicopters are used for their ability to take off and land vertically, qualifying them for a host
of tasks, namely: tourism, aerial observation and medical transportation (“Helicopter Career
Info”, 2017). An intriguing physics phenomenon of a helicopter, is its potential to hover. A
twin bladed single rotor helicopter, consumes 60 - 85% more power to hover than with forward
flight, making the manoeuvre inefficient (Lombardi, 2017). A vast amount of literature has
been published for optimising blade shape and fuselage weight, in order to enhance efficiency.
However, the role of a rotors angular velocity has never been addressed.
Hence, this essay attempts to answer the question “What is the relationship between angu-
lar velocity and power efficiency of a twin bladed single rotor helicopter system, in hover?” The
essay seeks to correct Froude’s momentum and Drzweicki’s blade element theory, to obtain a
theoretical model for power efficiency in terms of angular velocity. In order to test validity of
the theoretical model, an experiment is devised to evaluate the correlation between theoretical
and empirical power data.
Firstly, the essay delves into Froude’s model and realises the necessity of thrust and power
coefficient in expressing power efficiency. To accommodate NACA 0015 aerofoil geometry used
within the experiment, coefficients are corrected by integrating small blade elements along the
blade using Drzewiecki’s model. Thereafter, a dependence between air resistance and thrust
coefficient is established and incorporated using XFOIL simulations. The simulation allows to
compare thrust and power coefficient against angular velocity with industrial specifications,
providing insights into hypothetically inefficient, ideal and efficient ranges for power interms
of angular velocity. Correspondingly, the absolute uncertainty for coefficients are found to be
substantially large; losing confidence with the theoretical model. Hence, to affirm if the inves-
tigation is concurrent with empirical data, an experiment is devised to simulate a helicopter
rotor; obtaining data for power coefficient. The empirical and theoretical power coefficient
establish a strong correlation, which implied the uncertainties accumulated as a consequence
of extensive mathematical calculations. Ultimately, the calculated coefficients were substituted
in the model, yielding a clear relationship between power efficiency and angular velocity.
This research question is worthy of investigation, as it advances the understanding and
provides impetus to the research in power performance of aerial craft’s. Commercially, we

3
observe the demand of various clients including: airlines and hobbyists who desire to minimise
battery drain time and maximise usage, leading to the central question of power efficiency.

2 Background Information
Power dissipated is product of force imparted on air and mean air velocity. Therefore, in order
to establish a theoretical model for power efficiency, we must inspect airflow velocity and the
different forces in a rotor system.
A hovering helicopter is considered at rest, since the body has zero acceleration and so
zero net-force. The airscrew of twin bladed single rotor helicopter rotates around central z-
axis, propelling air in a downwards vortex known as the slipstream, figure 1. The slipstream
is governed by an inverse relationship termed the venturi effect, where the decreasing cross-
sectional area (A), increases air velocity (v̄) (Halliday et al., 2014). In the early 20th century,
William Froude combined venturi effect with Bernoulli’s principle; formulating an equation for
the axial force (z-plane) lifting a helicopter - thrust (T ).

Figure 1: A slipstream for twin bladed single rotor helicopter.

4
 According to Bernoulli’s principle, the rotor system experiences dynamic pressure (Pd )
inside and static pressure (Ps ) outside the slipstream boundary, which always adds to a constant
(1). The dynamic pressure, characteristically exhibits proportionality with airflow velocity:
Pd ∝ v̄ (Zhao et al., 2019). We assume air density is ρ = 1.23kgm−3 for simplicity, although
the value varies with elevation.

ΣP = Ps + Pd = constant

1
ΣP = Ps + ρv̄ 2 = constant (1)
2

The airflow velocity increases along slipstream from initial velocity (u = 0ms−1 ) above to final
velocity (v) below the airscrew, proportionally increasing dynamic pressure. Therefore, the
cross-sectional area decreases from Au above to Av below the airscrew by the venturi effect,
proportionally decreasing static pressure from P1 to P2 respectively; preserving the constant
relationship in (1).
1
Above airscrew: ΣP = P1 + ρu2
2
1
Bellow airscrew: ΣP = P2 + ρv 2
2
1 1
P1 + ρu2 = P2 + ρv 2
2 2

Rearranging the equation to find pressure differential experienced by airscrew.

1
P1 − P2 = ρ(u2 − v 2 )
2

1
Dynamic pressure differential: ∆Pd = P2 − P1 = ρv 2 (Since, u = 0ms−1 )
2

2.1 Induced velocity and thrust

Froude argued, since pressure is quotient of net-force and cross-sectional area, the greater
dynamic pressure beneath the blade causes an upwards directed net-force, figure 2. This net-
force is termed thrust, acting orthogonal to airflow on the airscrew (2) (Venkatesan, 2012).

5
The airscrew area is assumed to be circular where blade’s radius is R, denoted by Ai = πR2 .

ΣForcenet = Pd (Ai )

1
T = Ai ρv 2
2
1
T = ρπR2 v 2 (2)
2

Figure 2: Froude uses Bernoulli’s principle and properties of venturi effect to model thrust (T ) in a rotor system.

Conversely, Froude proposed an alternative argument where air particles are the frame
of reference. Air particles have negligible mass, requiring mass flow rate (ṁ) defined as the
fluid mass passing airscrew per unit time. Mass can also be considered as volumetric density
(m = V ρ), where velocity of air positioned at rotor is known as induced velocity (vi ).

∆m
ṁ =
∆t

Vρ
ṁ = =⇒ ρAi vi
t

Elastic collision occurs between air particles and airscrew for conserving momentum in sys-

6
tem, producing an equal and opposite axial thrust force (Gessow & Myers, 1985). Substituting
ṁ in Newton’s second law, we obtain thrust (3). We assume ideal gas properties are preserved
in such high pressure situations for simplicity.

F = ma =⇒ ṁ(∆v̄)

T = ρAi vi (v − u)

T = ρπR2 vi v (Since, u=0ms−1 ) (3)

Realising (2) and (3) model thrust and also incorporate similar variables except for velocity,
equating the two expressions yields an important velocity identity (4).

1
T = ρπR2 vi v = ρπR2 v 2
2

2vi = v (4)

Thereby, substituting velocity identity within (2) and (3) we are able to derive unified
thrust equation (5). Moreover, by isolating induced velocity we determine an equation for
velocity at the airscrew (6). This model for force and velocity is Froude’s momentum theory
(Gessow & Myers, 1985).
1
T = ρπR2 vi (2vi ) = ρπR2 (2vi )2
2

T = 2ρπR2 vi2 (5)

s
T
vi = (6)
2ρπR2

2.2 Power and power efficiency

Recalling from earlier, power (P̄ ) can now be defined as the product of axial thrust force
orthogonal to airflow and induced air velocity shown in (7).

P̄ = F v =⇒ T vi (7)

Power efficiency is the ratio between useful and total power, expressed in percentage. In

7
fluid mechanics, useful power is represented by (7), as here the fluid behaves in an ideal system
without viscosity. Viscosity measures a fluids resistance to deformation (Adminstration, 2020).
Total power refers to power measured experimentally (P̄exp ), where aerodynamic losses due to
resistive forces are considered. Hence, expanding for power efficiency yields (8):

Useful Power T vi
η= × 100 =⇒ ¯ × 100 (8)
Total Power Pexp

The exploration in deriving power efficiency has laid the foundation for this essay. This foun-
dation allows incorporating the important concept of angular velocity.

3 Extending Power Efficiency

Investigations in fluid mechanical systems incorporate coefficients, as they allow simplifying
power performance equations (Gessow & Myers, 1985). The thrust coefficient (cT ) recognises
ratio between total thrust produced and cross-sectional area (9). Similarly, power coefficient
recognises ratio between power required and cross-sectional area (9). These coefficients recog-
nise blade velocity in circular motion, vblade = ωR, where ω represents the blade’s angular
velocity. Also notice, these equations are non-dimensional thus, have no unit.

Thrust coefficient: T = cT ρπR2 (ωR)2 Power coefficient: P̄ = cP ρπR2 (ωR)3

T P̄
cT = cP = (9)
ρπR2 (ωR)2 ρπR2 (ωR)3

In order to substitute the coefficients for power efficiency, we first require to express induced
velocity in terms of the coefficients. Therefore, induced velocity from (6) simplifies to (10).

s s
T cT ρπR2 (ωR)2
vi = =
2ρπR2 2ρπR2

r
cT
vi = ωR (10)
2

Substituting induced velocity (10), thrust and power coefficient in (8), we obtain a simplified

8
form of power efficiency (11).

s 3
c3T cT2
η= × 100 = 70.7 × (11)
2c2P cP

The sections hereafter are based on calculating thrust and power coefficients to determine
a relationship between power efficiency and angular velocity (ω).

4 Slipstream Correction for Pitch Angle

Pitch angle/ pitch (θ) refers to the
angle a blade makes with the hor-
izontal (x-axis). Empirical analy-
sis confirms, thrust cannot be pro-
duced without pitch (Venkatesan,
2012). Unified thrust in (5) as-
sumes a zero pitch system hence, it
fails to satisfy properties changing
with pitch. These properties are
illustrated in figure 3 and table 1.

Figure 3: A modified approach to the slipstream model.

Properties changed Description

Angular velocity (ω) A blade rotating through air, is similar to air
flowing across a stationary blade with angu-
lar velocity (ω) and tangential velocity (ωr),
depending on the radial position of blade (r).
Inflow velocity (vi and ωr) Along with induced velocity (z-plane), air
also has tangential velocity: ωr (x-y plane).
The vector sum of velocities is resultant ve-
locity, vR (12).
Table 1: Properties that change in a system when implementing a blade pitch angle

q
vR = vi2 + (ωr)2 (12)

Due to these properties, pitch comprises of inflow angle of attack (Inflow AoA (φ)) and
angle of attack (AoA (α)), demonstrated in figure 4 and (13). The inequality θ > φ and θ > α

9
holds true, hence by small angle approximation the quotient of induced and tangential velocity;
tan φ approximately equal φ (14). As inflow AoA is small, so is induced velocity and therefore
resultant velocity in (12) is best approximated by tangential velocity (15).

θ =φ+α (13)

vi
tan φ = ≈φ (14)
ωr

vR ≈ ωr (15)

Figure 4: A diagram in which (a) deconstructs the angles in relation with the blades, (b) interprets
the blades free body diagram.

The force perpendicular to airflow is lift (L) and a new force parallel to airflow also exists,
termed drag (D), figure 4. The vertical component of lift and drag produces thrust. As the
inflow AoA is insignificant (φ ≈ 0), trigonometric ratios reduce to cos φ → 1 and sin φ → 0.
Hence, thrust approximately equals lift (16).

T = L cos φ + D sin φ

10
 T ≈L (16)

5 Determining Thrust Coefficient

Realise, the force orthogonal to airflow is lift, corresponding with Bernoulli’s principle (2).

1
L = Aρv̄ 2
2

Substituting resultant velocity (15) in lift, we recognise as radial position (r) increases from hub
(r = 0) to tip (r = R), lift also increases with each blade element (dr). This implies lift is non-
uniformly produced along the blade. The blade element theory suggests, the cumulative sum of
all blade elements is total lift and hence, we integrate (Gessow & Myers, 1985). Furthermore,
the blade element area is rectangular and therefore product of infinitesimal length (dr) and
chord width (C), figure 3.
Z R
1
L= Cρ(ωr)2 dr
0 2

Incorporating an empirical factor known as lift coefficient (cL ) allows accounting for NACA
0015 aerofoil geometry used in our experiment (see section 7). The lift coefficient is a first order
function when mapped with AoA (α) and passes through origin (α, cL ) : (0,0) (Venkatesan,
2012). The function is y = mx ⇐⇒ cL = aα, where ’a’ is the lift-curve gradient. We make
AoA subject of (13) and further substitute inflow AoA (14) for lift.

Z R Z R
1 1
L= Cρ(ωr)2 cL dr = Cρ(ωr)2 aαdr
0 2 0 2

vi
α = θ − φ =⇒ θ −
ωr
Z R
1 vi
L= Cρ(ωr)2 a(θ − )dr
0 2 ωr

We simplify lift and remove constants outside the integral. Since, lift only considers force from
one blade, multiplying by 2 yields total lift. As lift-thrust identity (15) holds true, total lift
approximates thrust (17).

Z R
1
L = Cρa θ(ωr)2 − vi (ωr)dr
2 0

11
 Z R
T = Cρa θ(ωr)2 − vi (ωr)dr (17)
0

Lastly, we can now determine the corrected thrust coefficient using (9), by substituting for
r
corrected thrust (17). Realise by dividing the fraction, we obtain a new variable, r̄ = R

(Venkatesan, 2012). Hence, boundary conditions change; lower limit limr→0 r̄ = 0 and upper
limit limr→R r̄ = 1.

RR
T Cρa 0
θ(ωr)2 − vi (ωr)dr
cT = 2 2
=⇒
ρπR (ωR) ρπR2 (ωR)2

Z 1
Ca vi
cT = (θ(r̄)2 − (r̄))dr̄
πR 0 ωR
vi vi
To further simplify, we assume inflow AoA φ = ωr
≈ ωR
. This assumption is valid as mag-
nitude for tangential velocity along the blade (ωr) approximates to at the tip (ωR), since
dynamic pressure is significantly larger at tip due to their proportionality. Substituting this
approximation and solving the integral we obtain thrust coefficient (18).

1 1
Ca θr̄3 φr̄2
Z 
Ca 2
cT = (θr̄ − φr̄)dr̄ = −
πR 0 πR 3 2 0

Ca θ φ
cT = ( − ) (18)
πR 3 2

Since, chord width (C = 0.0180 ± 0.00500m), radius of blade (R = 0.0900 ± 0.00500m) and
pitch (θ = 0.157 ± 0.00900rad) are measured variables in the experiment (see section 7), we
require to know inflow angle of attack and lift-curve gradient to determine thrust coefficient.

5.1 Determining inflow angle of attack

vi vi
We still assume φ = ωr
≈ ωR
to keep inflow AoA constant with radial position, otherwise the
investigation becomes convoluted. Notice, (10) from section 3 can be rearranged to the form
vi
ωR
, which by our assumption approximates inflow AoA.

r r
cT vi cT
vi = ωR =⇒ φ ≈ =
2 ωR 2

12
Substituting thrust coefficient (18) and lift-curve gradient, a = 2π (explained in section 5.2),
we express inflow AoA as a quadratic.

cT Ca θ φ
φ2 = = ( − )
2 2πR 3 2

2πR 2 1 θ
( )φ + ( )φ − = 0
Ca 2 3
R 1 θ
( )φ2 + ( )φ − = 0
C 2 3

Replacing values measured in experiment (see section 7) forms a quadratic (19).

5.00φ2 + 0.500φ − 0.0520 = 0 (19)

The steps for error propagation are demonstrated bellow.

∆R
Fractional uncertainty for numerator. R
= 0.00500
0.0900
= 0.0560m
∆C 0.00500
Fractional uncertainty for denominator. C
= 0.0180 = 0.278m
Sum of fractional uncertainties for numerator ∆R ∆C
R
+ C
= 0.334
and denominator is that for coefficient of φ2 .
Absolute uncertainty for coefficient, rounded
∆ CR = 1.70m
to 3.sig.fig.

Fractional uncertainty for y-intercept equals ∆θ 0.00900

θ
= 0.157
= 0.0573m
fractional uncertainty for pitch.
Absolute uncertainty for y-intercept is
∆ 3θ = 0.00300m
rounded to 3.sig.fig similar to y-intercept.

(5.00 ± 0.170)φ2 + 0.500φ − (0.0520 ± 0.00300) = 0

We obtain an original, maximum, minimum value for inflow AoA after implementing the
uncertainties and graphically solving them with domain (0,∞). The original quadratic (19)
yields φ = 0.0640rad (figure 5a).

5.00φ2 + 0.500φ − 0.0520 = 0

13
The maximum quadratic yields φ = 0.0740rad (figure 5b).

3.30φ2 + 0.500φ − 0.0490 = 0

The minimum quadratic yields φ = 0.0560rad (figure 5c).

6.70φ2 + 0.500φ − 0.0550 = 0

(a) y = (5.00)φ2 + (0.500)φ − (0.0520) = 0 (b) y = (3.30)φ2 + (0.500)φ − (0.0490) = 0 (c) y = (6.70)φ2 + (0.500)φ − (0.0550) = 0

Figure 5: Graphs to determine inflow AoA (φ).

Hence, a good approximation for inflow AoA is the mean (φ̄) of range and for uncertainty
is unbiased standard deviation (σφ ) demonstrated below.

φ1 + φ2 + φ3 0.056 + 0.064 + 0.074

φ̄ = = = 0.0650
3 3
s
(φ1 − φ̄)2 + (φ2 − φ̄)2 + (φ3 − φ̄)2
σφ =
N −1
r
(0.056 − 0.065)2 + (0.064 − 0.065)2 + (0.074 − 0.065)2
σφ = = 0.00903
2

Therefore, inflow AoA, φ = 0.0650 ± 0.00903 to 3.sig.fig.

5.2 Determining lift-curve gradient

The lift-curve gradient accounts for resistive forces caused by the blade’s (aerofoil) geome-
try. Air resistance results from skin friction and form drag creating turbulence, figure 6.
Turbulence decreases tangential velocity, proportionally decreasing dynamic pressure (Admin-
istration, 2020). By Froude’s argument this decreases thrust.
The industrially averaged magnitude for lift-curve gradient is a = 2π, for all angular veloc-

14
ities and aerofoil (Scott, 2018). The lift-curve gradient of lift coefficient against AoA (cL vs α)
graph can be found using XFOIL; a fluid simulation programme. XFOIL develops unique
environment for varying angular velocities of a rotor based on Reynolds number (Re), which
predicts airflow patterns (Halliday et al., 2014). After undergoing a mathematical process (see
Appendix 1), Reynolds number is found using simplified form (20).

Re = 1266.98(ωR) (20)

Figure 6: Air resistance forces which occur decreasing resultant velocity.

To determine Reynolds number, we require to define angular velocities to simulate in

XFOIL. The experiment in section 7, uses revolutions per minute (RPM min−1 ) as a rela-
tive measure for angular velocity. The experiment uses a motor with an operational range
from 0 − 7000 RP M , where tests are conducted in increments of 250 RP M . We will therefore
simulate within the operational range and the same increments. RPM determines frequency
2π
of revolution, hence conversion to angular velocity requires the factor 60
. A sample set of
conversions is shown in table 2, where values are rounded to RP M of lowest significant figure.
Substituting angular velocities, we obtain Reynolds numbers displayed in sample table 2 (see

15
Appendix 2). Error propagation is unnecessary as XFOIL does not recognise error, decreasing
confidence in lift-curve gradient.

Revolutions per Angular Tangential Reynolds

minute (RP M ) velocity (ω) velocity (ωR) Number (Re)
/min−1 /rad s−1 /m s−1 -
0 0 0 0
250 26 2.4 30̄00
500 50 5 6000
750 79 7.1 90̄00
1000 100 10 10000
... ... ... ...
Table 2: A table determining Reynolds number

Reporting Reynolds number to XFOIL, a simulation is produced in one tab (see figure 7a)
and a cL vs α table in the other. The table is then converted to a graph (see figure 7b), which
uses 26 rads−1 as an example. A linear regression line is graphed to verify linearity, which as
indicated by Pearson correlation coefficient r = 0.998 is very strong. However, a systematic
error is observed as the regression line intersects y-axis at (α, cL ) : (0, −0.007), but not the
origin. The difference between y-intercepts is negligible, hence the graph can be considered
reliable. The lift-curve gradient for 26 rads−1 is a = 1.47 to 3.sig.fig. The remaining angular
velocities are found with a similar process, shown in sample table 3 (see Appendix 3).

(a) An illustration of XFOIL simulation for 26rads−1 . (b) A lift cL vs αc graph.

Figure 7: Simulation and analysis of data from XFOIL.

16
 Angular Tangential Reynolds Lift-curve
velocity (ω) velocity (ωR) Number (Re) gradient (a)
/rad s−1 /m s−1 - -
0 0 0 0
26 2.4 3000 1.47
50 5 6000 1.28
79 7.1 9000 1.23
100 10 10000 1.22
... ... ... ...
Table 3: A table determining lift-curve gradient

Data from table 3 is quantitatively demonstrated as a graph in figure 8. The mean for all lift-
curve gradient points is represented as a mean line (atheo = 4.81). An industrially agreed mean
line is also illustrated (aemp = 2π = 6.28). The data points < atheo (0−222rads−1 ) shows a the-
oretically inefficient range, as a great proportion of work is lost to turbulence, which decreases
dynamic pressure. Similarly, data points between atheo and aemp (222 − 240rads−1 and 431 −
733rads−1 ) indicates an ideal range, as predominant amount of work is used to generate thrust.
Lastly, data points > aemp (240 − 431rads−1 ) represents an efficient range, as resistive forces
insignificantly impact dynamic pressure and therefore, most work generates thrust.

Figure 8: The relationship between angular velocity and lift-curve gradient.

17
5.3 Determining thrust coefficient

Recalling (18), we can now find the thrust coefficient with respect to angular velocity. However,
the thrust coefficient is theoretically deduced and fails to recognise mechanical losses for e.g.
bearing friction. This requires to implement a constant k = 1.75 (Venkatesan, 2012).

Ca θ φ
cT = k ( − )
πR 3 2

Substituting values for the variable’s, we obtain thrust coefficient demonstrated in table 4
(see Appendix 4). The thrust coefficient is calculated to 3.sig.fig, as the variable with smallest
significant figure. Uncertainty for the thrust coefficient are found with steps bellow.

Ca
Fractional uncertainty for multiplicand ( πR ) is Ca
∆ πR
Ca = 0.334 derived from section 5.1.
that for CR , as other variables have zero uncertainty. πR

Fractional uncertainty for 3θ is that for pitch. ∆θ

θ
= 0.00900
0.157
= 0.0573
Absolute uncertainty for 3θ . ∆ 3θ = 0.00300
Same steps for absolute uncertainty of φ2 . ∆ φ2 = 0.00400
Absolute uncertainty for 3θ − φ2 . ∆ 3θ − φ
2
= 0.00300 + 0.00400 = 0.00700
∆ θ3 − φ 0.00700
Fractional uncertainty for multiplier. θ
−φ
2
= 0.0200
= 0.350
3 2

Fractional uncertainty for thrust coefficient is sum

∆cT
of the fractional uncertainty for multiplicand and cT
= 0.334 + 0.350 = 0.684
multiplier.
Percentage uncertainty for thrust coefficient ∆cT % = 68.4%

Air screw Air screw angles Angular Lift-curve Thrust

geometry (R, C) (θ, φ) velocity (ω) gradient (a) coefficient (cT )
/m /rad /rad s−1 - -
±0.00500, ±0.00500 ±0.00900, ±0.00900 - - 68.4%
0 0 0
26 1.47 0.00359
50 1.28 0.00313
0.0900, 0.0180 0.157, 0.0650
79 1.23 0.00301
100 1.22 0.00299
... ... ...
Table 4: A table demonstrating the thrust coefficient cT based on it’s constituent angular velocity (ω).

18
 The data from table 4 is quantitatively represented as a graph in figure 9. The trend closely
corresponds with lift-curve gradient against angular velocity graph in figure 8. The mean of
all thrust coefficient data points is illustrated through a mean line. Although error bars are
substantially large, the inefficient range is positioned below mean line and opposite is true for
ideal and efficient ranges providing confidence with our speculations.

Figure 9: The relationship between angular velocity and thrust coefficient.

6 Determining Power Coefficient

The power determined in (7) is unsatisfactory, as it omits variables such as, resultant velocity
and air resistance. Hence, we will consider a new approach, which focuses on the revolution of
airscrew. Power is tangential force (F ) experienced by an aerofoil along a distance (s) divided
by time elapsed (∆t) to travel the distance (see figure 10a).

Fs
P̄ = F v =
∆t

The magnitude of arc length travelled can be expressed as s = βR, where the aerofoil
β
revolves an angle β in ∆t time. Notice, ∆t
is definition of angular velocity. Power is therefore

19
 product of tangential force and velocity.

F βR
P̄ = = F (ωR)
∆t

Additionally, considering air as frame of reference,

tangential force is the horizontal component of lift and
drag, figure 10b. The small angle approximation for
inflow AoA simplifies sin φ → φ and cos φ → 1. Lift
and drag can be expanded using blade element theory,
integrating along the blade’s length. The expanded
expression for drag is equivalent to lift, except for
drag coefficient (cD ) (Gessow & Myers, 1985). Figure 10: A modified approach for power coefficient.

Z R  Z R
P̄ = ωR (L sin φ + D cos φ)dr = ωR (Lφ + D)dr
0 0

Z R Z R
1 2 1
P̄ = ωR ( Cρ(ωr) cL )φdr + Cρ(ωr)2 cD dr
0 2 0 2
 Z R Z R 
1 2 2
P̄ = CρωR aφ θ(ωr) − vi (ωr)dr + (ωr) cD dr
2 0 0

We multiply power equation by two, in order to consider both blades (21). Substituting
power within power coefficient from (9) and dividing by denominator, we observe the variable
r
r̄ = R
. The boundary conditions change (see Section 5).

 Z R Z R 
2 2
P̄ = CρωR aφ θ(ωr) − vi (ωr)dr + (ωr) cD dr (21)
0 0

h R i
R RR
P̄ CρωR aφ 0 θ(ωr)2 − vi (ωr)dr + 0 (ωr)2 cD dr
cP = =
ρπR2 (ωR)3 ρπR2 (ωR)3
Z 1 Z 1
Ca 2 C 3
cP = φ (θr̄ − φr̄)dr + (r̄ cD )dr
0 πR 0 πR

Notice, the first integral simplifies to thrust coefficient (18). Solving the integral, we obtain
power coefficient (22).
Z 1
C 3
cP = φcT + (r̄ cD )dr
0 πR

20
 C
cP = φcT + cD (22)
4πR

The power coefficient can only be found, after determining drag coefficient.

6.1 Determining drag coefficient

The drag coefficient incorporates resistive forces acting on a NACA 0015 aerofoil. Similar to
lift coefficient, drag coefficient is found using XFOIL. After inserting the calculated Reynolds
numbers (table 3), a cD vs α table is generated. Since, AoA is α = θ − φ = 0.157 − 0.0650 =
0.0920rad, we select drag coefficients at α = 0.0920rad for all angular velocities displayed in
sample table 5 to 3 sig.fig (see Appendix 5). Uncertainty is not calculated as XFOIL doesn’t
recognise error, decreasing confidence with drag coefficient.

Angular Tangential Reynolds Lift curve Drag

velocity (ω) velocity (ωR) Number (Re) gradient (a) coefficient
(cD )
/rad s−1 /m s−1 - - -
0 0 0 0 0.0933
26 2.4 30̄00 1.47 0.0756
50 5 6000 1.28 0.0683
79 7.1 90̄00 1.23 0.0642
100 10 10000 1.22 0.0616
... ... ... ... ...
Table 5: A table demonstrating drag coefficient (cD ) based on its constituent angular velocity (ω).

6.2 Determining power coefficient

We can now calculate power coefficient (22). Recall from section 5.3, in order to consider
mechanical losses we incorporate constant k = 1.75.

C
cP = k(φcT + cD )
4πR

Substituting for the variables, we obtain power coefficient demonstrated in sample table 7
(see Appendix 6). Uncertainty calculations are displayed bellow to 3 sig.fig, similar to variable
with lowest significant figure. However, we make an exception and increase significant figures
of angular velocity to 3 sig.fig, for maximising accuracy and precision when analysing data.

21
 ∆φ 0.00903
Fractional uncertainty for inflow AoA. φ
= 0.0650
= 0.138rad
Sum for the fractional uncertainties of φ and cT ∆φcT
φcT
= 0.138 + 0.684 = 0.822
is that for φcT .
Absolute uncertainty for φcT varies as cT
changes with angular velocity. Example for φcT = 0.000233 ∆φcT =
ω = 26.2 and rest shown in sample table 6. 0.000233(0.822) = 0.000195

C
Fractional uncertainty for 4πR cD is equivalent ∆C
R
C = 0.334 from section 5.1.
to that for CR . R

C
Absolute uncertainty for 4πR cD varies, as cD
C C
changes with angular velocity. Example for c
4πR D
= 0.00149 ∆ 4πR cD =
ω = 26.2 and rest shown in sample table 6. 0.00149(0.344) = 0.000496

C
Absolute uncertainty for φcT + 4πR cD (22) is
sum of the absolute uncertainties for φcT and C
C
∆φcT + 4πR cD =
c . Example for ω = 26.2 and rest shown
4πR D
0.000195 + 0.000496 = 0.00691
in sample table 6.
C
Fractional uncertainty for φcT + 4πR cD is then
C
∆φcT + 4πR cD 0.00691
allowed. Example for ω = 26.2 and rest shown C
φcT + 4πR cD
= 0.00172
= 0.402
in sample table 6.

C
Fractional uncertainty for φcT + 4πR cD is equiv- ∆cP
C
∆φcT + 4πR cD
cP
= C
φcT + 4πR cD
alent to that for cP .
Absolute uncertainty for cP varies with ω. Ex-
ample for ω = 26.2 and rest shown in sample cP = 0.00301 ∆cP =
table 7. 0.00301(0.402) = 0.00121

C C C ∆cP
ω φcT ∆φcT c
4πR D
∆ 4πR cD φcT + 4πR cD ∆φcT + cP
C
c
4πR D

26.2 0.000233 0.000195 0.00149 0.000496 0.00172 0.000691 0.402

52.4 0.000204 0.000170 0.00120 0.000402 0.00141 0.000572 0.407
78.5 0.000195 0.000164 0.00109 0.000363 0.00128 0.000526 0.411
104 0.000194 0.000162 0.00102 0.000341 0.00122 0.000504 0.414
... ... ... ... ... ... ... ...

Table 6: Determining the absolute uncertainties of variables

22
 Angular Power Absolute
velocity (ω) coefficient (cP ) uncertainty ∆cP
rads−1 - -
0 0 0
26.2 0.00301 0.00121
52.4 0.00246 0.00100
78.5 0.00224 0.000921
104 0.00213 0.000881
... ... ...
Table 7: Determining the theoretical power coefficient and the coefficients absolute uncertainty

The data in table 7 is represented as a graph in figure 11. For a system to be power efficient,
the thrust and power to area ratio must strictly maximise and minimise respectively. The
mean of all power coefficient data points are represented as a mean line. The angular velocities
between 0 − 183rads−1 are power inefficient, as they demonstrate a low thrust (see figure 9),
but a large power coefficient. This suggests most power is used to overcome resistive forces.
Moreover, angular velocities between 183 − 733rads−1 in retrospect are not power efficient, as

Figure 11: The relationship between angular velocity and power coefficient

23
they have a large thrust and power coefficient. Here, most power is used to generate thrust
than lost to air resistance and therefore, has ideal efficiency contrary to speculations.

7 Determining Experimental Power Coefficient

I devised an experiment to observe power dissipated, later converted to power coefficient in
revolving a twin bladed airscrew by a brushless DC motor (BLDC motor). Thrust measure-
ments for determining thrust coefficient are not accounted, as the measuring scale provided
misleading results (see figure 12). The experiment is necessary to validate whether theoretical
data (see section 5.3 and 6.2) corresponds with experimental data. This is important, to find a
reason for large error bars seen with the coefficients, sources of error or confounding variables.

7.1 Experimental Setup

In 1920-33, the National Advisory Committee for Aeronautics (NACA) designed and tested
various standardised aerofoils (Allen, 2017). I selected their NACA 0015 aerofoil, for its sym-
metric geometry and ease to 3D print. The aerofoil was printed with dimensions stated in
table 8 and went on top of BLDC motor.

Aerofoil Geometry Properties

Blade length: Radius R = 0.0900 ± 0.00500m
NACA 0015 Blade length: Chord width C = 0.0180 ± 0.00500m
Total angle: Pitch angle θ = 0.157 ± 0.00900rad
Table 8: A description for the properties of NACA 0015 aerofoil

The experiment setup is represented in figure 12. An Arduino Mega 2560 microcontroller
used a preloaded code (see Appendix 9) to change RPM of motor (Nedelkovski, 2019). A
tachometer measures the frequency of reflections from a reflective adhesive on the motor, to
determine the motors RPM. A power analyser measured the total power used by load. A
camera positioned above the setup, recorded the displayed readings.

7.2 Experiment procedure

• Independent variable: Angular velocity measured in RPM.

• Dependent variable: Power dissipated in the system.

24
 Figure 12: The experimental setup for determining the power coefficient.

A physical limitation of the motor is, it harshly vibrated between 0 < ω ≤ 314rads−1
and 681 ≤ ω < 733rads−1 deviating the photo tachometer readings. I realised to minimise
systematic error, the testing range should be between 314 ≤ ω ≤ 681rads−1 . Additionally, the
angular velocity readings never remained constant while performing preliminary tests. Hence,
I decided the experiment would be conducted in intervals of 250RPM (26.2rads−1 ) beginning
from 301 < ω ≤ 327rads−1 and ending with 668 < ω ≤ 694rads−1 . Each interval would be
tested 5 times, for 15 seconds with a stopwatch. The data was collected from video recordings
and is displayed in Appendix 7.

7.3 Theoretical vs experimental power coefficient

2π
The RPM were converted to angular velocity using the factor 60
(see Section 5.2). The angular
velocities of an interval is best demonstrated by the mean. The mean for power measured within
the interval, is suitable for the same reasoning. Such means are demonstrated in sample table
9 for first 5 intervals. Recall, the power coefficient (9). Substituting the values from table 9
yields the experimental power coefficient, also shown in sample table 9.

25
 P̄
cP =
ρπR2 (ω̄R)3

Angular Mean angular Mean Experimental power

velocity velocity (ω̄) ¯ )
power (Pavg coefficient (cP )
rads−1 rads−1 kgm2 s−3 -
301 < ω ≤ 327 318 2.8 0.00389
327 < ω ≤ 353 343 3.4 0.00363
353 < ω ≤ 380 367 3.7 0.00326
380 < ω ≤ 406 397 4.3 0.00302
406 < ω ≤ 432 420 4.9 0.00290
... ... ... ...
Table 9: Determining the experimental power coefficient.

The experimental and theoretical data points from table 7 and 9 are illustrated as a graph
in figure 13, with ordinates drawn to indicate the experimented range. Both curves are close
to perfect fit, as the experimental outcomes are more exaggerated deviating around 314 and
680rads−1 . Nevertheless, both power coefficients have similar properties such as, a curve fit

Figure 13: An analysis between angular velocity and power coefficient

26
 which monotonically decreases with angular velocity. Their gradient ( dcdωP ) converges to 0
from 550 to 600rads−1 as both curves either seem to or form a plateau. Error bars are also
useful, as experimental data follows the trend well within the error bars of theoretical data.
Hence, the theoretical model closely resembles outcomes from experiment.

8 Determining Power Efficiency

Realise, we can finally determine power efficiency (11). Substituting the trust and power
coefficient, yields power efficiency displayed in sample table 10 (see Appendix 8).

3
c2
η = 70.7 × T
cP

Uncertainty for power efficiency is propagated with steps displayed below.

3
3
∆cT2
Fractional uncertainty for numerator cT . 2
3 = 32 ( ∆c
cT
T
) = 1.50(0.684) = 1.09
cT2
Fractional uncertainty for denominator cP is
demonstrated in table 6

Fractional uncertainty for power efficiency (η)

is sum of the fractional uncertainties for nu- 3
3
∆η ∆cT2
merator cT2 and denominator cP . An example η
= 3 + ∆c
cP
P
= 1.09+0.00301 = 1.49
cT2
is ω = 26.2rads−1 and rest are shown in table
10.

The absolute uncertainty for power efficiency

can then be found. An example is ω = η = 5.05 ∆η = 5.05(1.49) = 7.54
26.2rads−1 and rest are shown in table 10.

Angular Power Fractional Absolute

velocity (ω) efficiency (η) uncertainty ( ∆η
η
) uncertainty (∆η)
rads−1 kgm2 s−3 - kgm2 s−3
26.2 5.05 1.49 7.54
52.4 5.04 1.50 7.54
78.5 5.20 1.50 7.80
105 5.42 1.50 8.15
... ... ... ...
Table 10: Calculations for determining the power efficiency of our twin bladed single rotor helicopter system.

27
 Table 10 is then converted to a power efficiency against angular velocity graph in figure 14.
The general trend line increases and plateaus in a regular succession. Power efficiency at

Figure 14: The relationship between power efficiency and angular velocity.

smaller angular velocities (0−157rads−1 ) are an order of magnitude inefficient than at larger
ones (314 − 733rads−1 ) confirming the inefficient and ideal ranges, respectively. The maximum
efficiency peaks at 54.6%, just under the maximum possible efficiency by Betz limit at 59.3%
(Burton, 2009). Negative uncertainties cannot exist. As aforementioned, large uncertainties
are caused by exhaustively using multiple equations to reach this relationship. Hence, the data
quality is not compromised.

9 Evaluation and Conclusion

In investigating the research question, we decided to determine power efficiency with the thrust
and power coefficient, resulting in many interesting observations.
Firstly, we found a limitation with Froude’s model, in that it assumed a zero pitch system,
omitting internal forces and velocities. Hence, modifications such as blade element theory were
made to implement pitch, enhancing the correlation between empirical and theoretical data.

28
 In determining the thrust coefficient, XFOIL allowed to find lift-curve gradient and an anal-
ysis with industrial specification. This categorised angular velocity into an inefficient, ideal
and efficient range, where smaller angular velocities in comparison to larger angular veloci-
ties required a substantial amount of power to overcome resistive forces. The power coefficient
showed similar efficiencies, except for the efficient range. The experiment validated the theoret-
ical ranges and also realised the large uncertainties were of a purely mathematical consequence.
Hence, the inefficient and ideal ranges were confirmed as maximum power efficiency peaked
only at 54.6%. However, a limitation with experiment was the small experimented range, not
allowing a full comparison with theoretical model.
In conclusion a clear proportionality is visible between angular velocity and power efficiency
for a twin bladed single rotor helicopter in hover. As we increase angular velocity, power
efficiency increases then plateaus and repeats the same trend once again.
This investigation is not completely accurate, due to uncertainties and limitations. A part
of these uncertainties arise from assumptions made with the theoretical model. For instance,
assuming thrust linearly increases along the blade. Empirically, the tip produces negligible
thrust, as the substantial dynamic pressure differential occurs from 20 to 80% of the blade
(Adminstration, 2020). Secondly, a vortex known as induced drag is formed at the blade’s
tip, which moderately increases air resistance. This was not implemented in XFOIL adding
unaccounted systematic error (Adminstration, 2020). Thirdly, the motor harshly vibrated at
the first (301 < ω ≤ 327rads−1 ) and last interval (668 < ω ≤ 694rads−1 ), influencing the
tachometer and power analyser, causing random error in data. This may explain deviations at
start and end of the experimental power coefficient. However, I realised, seeking for a perfect
match with experimental data is ambitious, as the investigation then becomes an extraneous
mathematical process, reducing its ability to explain physics.
This investigation raises many questions, including one which initially inspired me: what
angular velocity against power efficiency relationship can geometrically varied aerial vehicles
such as, bicopter and tricopter demonstrate? Irregularly positioning rotor systems, makes the
slipstream geometry unique and transcends the scope of this investigation. Many interesting
problems arise, as the rotors must produce variable thrust to balance while hovering. Such an
investigation would be incredibly interesting.

29
10 Works Cited

Administration, F. (2020). ​Rotorcraft Flying Handbook​ (pp. 2-28). BN Publishing.

Allen, B. (2017). ​NACA Airfoils​. National Aeronautics and Space Administration.

Retrieved 24 December 2020, from
https://www.nasa.gov/image-feature/langley/100/naca-airfoils.

Burton, T. (2009). ​Wind energy​ (1st ed., p. 6). Wiley.

Gessow, A., & Myers, G. (1985). ​Aerodynamics of the helicopter​ (8th ed., pp. 46-65).
Ungar.

Halliday, D., Resnick, R., Walker, J., & Halliday, D. (2014). ​Fundamentals of physics
(11th ed., pp. 386-412). Wiley.

Helicopter Career Info​. Heliventures.co. (2017). Retrieved 4 September 2020, from

https://www.heliventures.co/new-helicopter-pilot-resources/helicopter-career-info/.

Lombardi, F. (2017). ​Understanding Helicopter Power Requirements: The Power

Struggle - Rotor & Wing International​. Rotor & Wing International. Retrieved 25
December 2020, from
https://www.rotorandwing.com/2017/10/10/understanding-helicopter-power-require
ments-power-struggle/.

Nedelkovski, D. (2019). ​Arduino Brushless Motor Control Tutorial​.

HowToMechatronics. Retrieved 26 June 2020, from
https://howtomechatronics.com/tutorials/arduino/arduino-brushless-motor-control-tu
torial-esc-bldc/.

Scott, J. (2018). ​Lift Coefficient & Thin Airfoil Theory.​ Aerospaceweb. Retrieved 28
November 2020, from
http://www.aerospaceweb.org/question/aerodynamics/q0136.shtml#:~:text=Accordi
ng%20to%20the%20ideal%20aerodynamics,%CE%B1)%20is%20equal%20to%20
2%CF%80.&text=According%20to%20Thin%20Airfoil%20Theory,(CL)%20goes
%20up.

Venkatesan. (2012). ​Hovering Theory:Blade Element Theory.​ Lecture, Indian Institute

of Technology Kanpur.

Zhao, D., Han, N., Goh, E., Cater, J., & Reinecke, A. (2019). ​Wind turbines and
aerodynamics energy harvesters​ (pp. 1-5). Elsevier.
11 Appendix
11.1 Appendix 1: Simplifying Reynolds Number

Reynolds number measures airflow patterns, based on the properties of air. In fluid

mechanics one form Reynolds number can take is displayed bellow. Here, 𝜈 is kinematic

viscosity, 𝑣𝑅 is resultant velocity of airfoil and 𝐶 is chord length. Kinematic viscosity of air at

sea level and normal room temperature is: 𝜈 = 1.42 × 10−5 .

𝐶𝑣
𝑅𝑒 =
𝜈

We substitute the value for variables from section 7 and use resultant velocity (15).

However, realise 𝑣𝑅 = 𝜔𝑟 ≈ 𝜔𝑅, as dynamic pressure is substantially large at the tip and

approximately equal to that across the blade. Hence, by venturi effect tangential velocity at

tip is also approximately equal to that across the blade.

(0.0180)𝜔𝑟
𝑅𝑒 =
0.0000142

(0.0180)𝜔𝑅
𝑅𝑒 = = 1266.98𝜔𝑅
0.0000142

31
11.2 Appendix 2: Conversion between Angular Velocity and Reynolds Number

Revolution per Angular Tangential Reynolds

minute (RPM) velocity (ω) velocity (ωR) Number (Re)

/𝒎𝒊𝒏−𝟏 /𝒓𝒂𝒅𝒔−𝟏 /𝒎𝒔−𝟏 -

0 0 0.0 0
250 26 2.4 2991
500 52 4.7 5982
750 79 7.1 8972
1000 105 9.4 11963
1250 131 11.8 14954
1500 157 14.1 17945
1750 183 16.5 20935
2000 209 18.8 23926
2250 236 21.2 26917
2500 262 23.6 29908
2750 288 25.9 32899
3000 314 28.3 35889
3250 340 30.6 38880
3500 367 33.0 41871
3750 393 35.3 44862
4000 419 37.7 47853
4250 445 40.1 50843
4500 471 42.4 53834
4750 497 44.8 56825
5000 524 47.1 59816
5250 550 49.5 62806
5500 576 51.8 65797
5750 602 54.2 68788
6000 628 56.5 71779
6250 654 58.9 74770
6500 681 61.3 77760
6750 707 63.6 80751
7000 733 66.0 83742

32
11.3 Appendix 3: Determination of Lift-curve Gradient

Angular Tangential Reynolds Lift-curve

velocity (ω) velocity (ωR) Number (Re) gradient (a)

/𝒓𝒂𝒅𝒔−𝟏 /𝒎𝒔−𝟏 - -
0 0.0 0 0.00
26 2.4 2991 1.47
52 4.7 5982 1.28
79 7.1 8972 1.23
105 9.4 11963 1.22
131 11.8 14954 1.23
157 14.1 17945 1.25
183 16.5 20935 3.06
209 18.8 23926 4.28
236 21.2 26917 5.14
262 23.6 29908 6.14
288 25.9 32899 7.13
314 28.3 35889 7.66
340 30.6 38880 7.60
367 33.0 41871 7.26
393 35.3 44862 6.84
419 37.7 47853 6.51
445 40.1 50843 6.02
471 42.4 53834 5.69
497 44.8 56825 5.42
524 47.1 59816 5.27
550 49.5 62806 5.25
576 51.8 65797 5.26
602 54.2 68788 5.36
628 56.5 71779 5.44
654 58.9 74770 5.52
681 61.3 77760 5.62
707 63.6 80751 5.70

33
 11.4 Appendix 4: Determination of Thrust Coefficient

Air screw geometry Aircrew Angles Angular Gradient of Thrust

(𝑹, 𝑪) (𝜽, 𝝓) velocity (ω) lift-curve (a) Coefficient (𝒄𝑻 )
/𝒎 /𝒓𝒂𝒅 /𝒓𝒂𝒅𝒔−𝟏 - -

± 0.00500, ±0.00500 ±0.00900, ±0.00900 - - 68.40%

0 0.00 0.00000
26 1.47 0.00359
52 1.28 0.00313
79 1.23 0.00301
105 1.22 0.00299
131 1.23 0.00301
157 1.25 0.00305
183 3.06 0.00747
209 4.28 0.01044
236 5.14 0.01253
262 6.14 0.01497
288 7.13 0.01738
314 7.66 0.01870
340 7.60 0.01854
0.157, 0.0650
0.0900, 0.0180 367 7.26 0.01772
393 6.84 0.01670
419 6.51 0.01588
445 6.02 0.01470
471 5.69 0.01387
497 5.42 0.01321
524 5.27 0.01287
550 5.25 0.01281
576 5.26 0.01284
602 5.36 0.01308
628 5.44 0.01327
654 5.52 0.01347
681 5.62 0.01371
707 5.70 0.01391
733 5.82 0.01420

34
11.5 Appendix 5: Determination of Drag Coefficient

Angular Tangential Reynolds Lift-curve Drag

velocity (ω) velocity (ωR) Number (Re) gradient (a) coefficient (𝒄𝑫 )

/𝒓𝒂𝒅𝒔−𝟏 /𝒎𝒔−𝟏 - - -
0 0.0 0 0.00 0.0000
26 2.4 2991 1.47 0.0933
52 4.7 5982 1.28 0.0756
79 7.1 8972 1.23 0.0683
105 9.4 11963 1.22 0.0642
131 11.8 14954 1.23 0.0616
157 14.1 17945 1.25 0.0598
183 16.5 20935 3.06 0.0664
209 18.8 23926 4.28 0.0658
236 21.2 26917 5.14 0.0623
262 23.6 29908 6.14 0.0568
288 25.9 32899 7.13 0.0490
314 28.3 35889 7.66 0.0421
340 30.6 38880 7.60 0.0376
367 33.0 41871 7.26 0.0345
393 35.3 44862 6.84 0.0321
419 37.7 47853 6.51 0.0302
445 40.1 50843 6.02 0.0289
471 42.4 53834 5.69 0.0275
497 44.8 56825 5.42 0.0266
524 47.1 59816 5.27 0.0256
550 49.5 62806 5.25 0.0249
576 51.8 65797 5.26 0.0241
602 54.2 68788 5.36 0.0235
628 56.5 71779 5.44 0.0229
654 58.9 74770 5.52 0.0224
681 61.3 77760 5.62 0.0219
707 63.6 80751 5.70 0.0215

35
11.6 Appendix 6: Determination of Power Coefficient

Angular Power Absolute

velocity (ω) coefficient (𝒄𝑷 ) uncertainty 𝚫𝒄𝑷
/𝒓𝒂𝒅𝒔−𝟏 - -
0 0 0
26 0.00172 0.001209
52 0.00141 0.001001
79 0.00128 0.000921
105 0.00122 0.000881
131 0.00118 0.000860
157 0.00115 0.000847
183 0.00154 0.001328
209 0.00173 0.001605
236 0.00181 0.001772
262 0.00188 0.001954
288 0.00191 0.002111
314 0.00188 0.002172
340 0.00180 0.002116
367 0.00170 0.002008
393 0.00160 0.001889
419 0.00151 0.001793
445 0.00142 0.001668
471 0.00134 0.001576
497 0.00128 0.001505
524 0.00124 0.001463
550 0.00123 0.001451
576 0.00122 0.001447
602 0.00122 0.001464
628 0.00123 0.001476
654 0.00123 0.001491
681 0.00124 0.001510
707 0.00125 0.001524

36
11.7 Appendix 7: Raw Experimental Data

# 2875 < rpm Power 3125 < rpm Power 3375 < rpm Power 3625 < rpm Power
≤ 3125 ≤ 3375 ≤ 3625 ≤ 3875
1 3017 2.9 3267 3.2 3417 3.3 3736 4.2
2 3017 3.1 3269 3.3 3434 3.5 3737 4.3
3 3018 3.0 3269 3.3 3436 3.1 3738 4.4
4 3019 3.3 3270 3.3 3445 3.5 3739 4.2
5 3020 3.1 3271 3.4 3447 3.8 3740 4.3
6 3022 2.7 3271 3.3 3450 3.8 3740 4.3
7 3022 2.8 3272 3.4 3453 3.5 3742 4.4
8 3023 2.9 3272 3.4 3453 3.7 3742 4.3
9 3029 2.8 3272 3.3 3459 3.7 3743 4.3
10 3030 2.8 3273 3.1 3460 3.7 3753 4.2
11 3039 3.2 3275 3.4 3461 3.6 3758 4.3
12 3040 3.2 3276 3.4 3462 3.8 3759 4.1
13 3042 2.4 3277 3.4 3464 3.6 3760 4.3
14 3042 3.0 3278 3.3 3472 3.3 3761 4.2
15 3043 2.7 3282 3.3 3474 3.8 3761 4.2
16 3043 2.7 3292 3.6 3478 4.0 3761 4.2
17 3044 2.4 3296 3.4 3496 3.8 3762 4.2
18 3044 2.8 3298 3.3 3500 3.3 3764 4.4
19 3044 2.6 3298 3.6 3516 3.7 3765 4.2
20 3045 2.4 3299 3.4 3517 3.4 3765 4.2
21 3045 2.8 3300 3.4 3518 3.8 3766 4.3
22 - - - 3536 3.6 3767 4.2
23 - - - - 3538 3.6 3768 4.3
24 - - - - 3540 3.8 3773 4.4
25 - - - - 3541 3.8 3851 4.5
26 - - - - 3546 3.8 3854 4.3
27 - - - - 3547 3.7 3855 4.3
28 - - - - 3548 3.7 3855 4.4
29 - - - - 3548 3.7 3856 4.4
30 - - - - 3548 3.8 3857 4.4
31 - - - - 3549 3.8 3863 4.4
32 - - - - 3549 3.9 3864 4.4
33 - - - - 3550 3.8 3866 4.4
34 - - - - 3552 3.8 3867 4.4
35 - - - - 3553 3.6 3867 4.4
36 - - - - 3556 3.8 3875 4.4
37 - - - - 3559 3.8 - -
38 - - - - 3567 3.8 - -
39 - - - - 3575 3.9 - -

37
# 3875 < rpm Power 4125 < rpm Power 4375 < rpm Power 4625 < rpm Power
≤ 4125 ≤ 4375 ≤ 4625 ≤ 4875
1 3964 4.9 4141 5.5 4478 5.6 4699 6.4
2 3965 4.9 4157 5.2 4479 5.7 4705 6.5
3 3966 4.8 4161 5.3 4482 5.6 4708 6.2
4 3971 4.9 4165 5.4 4484 5.7 4708 6.2
5 3973 5.2 4165 5.4 4485 5.8 4709 6.4
6 3973 4.9 4167 5.2 4485 5.6 4710 6.5
7 3974 4.9 4167 5.5 4486 5.6 4711 6.5
8 3974 4.9 4169 5.2 4486 5.5 4712 6.4
9 3975 4.9 4169 5.3 4488 5.6 4712 6.5
10 3975 5.1 4259 5.3 4488 5.5 4713 6.0
11 3975 5.1 4261 5.1 4489 5.5 4713 6.5
12 3977 4.7 4262 4.9 4489 5.7 4714 6.5
13 3977 4.8 4263 5.5 4490 5.6 4714 6.0
14 3978 4.9 4263 5.5 4494 5.6 4715 6.2
15 3978 4.8 4263 5.4 4535 6.3 4716 6.5
16 3978 4.8 4264 5.4 4537 6.0 4716 6.4
17 3979 4.9 4264 5.5 4539 6.3 4717 6.5
18 3980 4.9 4264 5.5 4541 6.0 4719 6.6
19 3980 4.9 4264 5.3 4541 5.7 4719 6.5
20 3981 5.1 4264 5.3 4546 6.0 4720 6.0
21 3981 4.7 4265 5.4 4547 5.9 4720 6.4
22 3981 5.0 4265 5.2 4547 5.8 4721 6.5
23 3983 4.8 4265 5.2 4548 6.3 4723 6.0
24 3984 4.8 4266 5.5 4549 6.3 4724 5.9
25 3984 4.8 4266 5.6 4550 6.1 4725 6.2
26 4024 4.8 4266 5.4 4551 6.3 4726 6.0
27 4031 4.9 4266 5.5 4551 5.9 4728 6.0
28 4032 4.9 4267 5.6 4552 6.0 4729 6.0
29 4037 4.8 4267 5.8 4552 5.8 4731 6.1
30 4038 4.9 4267 5.4 4554 6.0 4732 6.4
31 4038 4.9 4269 5.2 4555 6.3 4732 6.5
32 4043 4.8 4270 5.7 4555 5.9 4734 6.0
33 4044 4.9 4271 5.6 4556 6.0 4735 6.2
34 4045 4.8 4271 5.5 4556 5.9 4862 6.6
35 4045 5.1 4271 5.2 4557 5.9 4864 6.8
36 4047 4.8 4274 5.3 4557 5.9 4866 6.6
37 4047 4.9 - - 4558 6.0 4872 6.8
38 4048 5.0 - - 4558 5.9 4873 6.6
39 4048 4.8 - - 4559 6.0 4874 6.5

38
40 4048 4.7 - - 4559 5.8 4875 6.7
41 4049 4.8 - - 4559 5.6 - -
42 4050 5.0 - - 4561 5.9 - -
43 4051 4.9 - - 4561 5.7 - -
44 4053 5.1 - - 4561 5.8 - -
45 4054 5.0 - - 4561 5.9 - -
46 4055 4.9 - - 4562 6.2 - -
47 4056 4.8 - - 4562 5.9 - -
48 4063 4.8 - - 4562 5.9 - -
49 - - - - 4563 5.8 - -
50 - - - - 4563 6.0 - -
51 - - - - 4563 5.7 - -
52 - - - - 4564 6.0 - -
53 - - - - 4565 6.2 - -
54 - - - - 4565 6.0 - -
55 - - - - 4565 5.8 - -
56 - - - - 4565 5.9 - -
57 - - - - 4565 5.8 - -
58 - - - - 4565 6.0 - -
59 - - - - 4566 6.0 - -
60 - - - - 4566 6.0 - -
61 - - - - 4566 5.9 - -
62 - - - - 4566 5.7 - -
63 - - - - 4566 5.9 - -
64 - - - - 4567 5.8 - -
65 - - - - 4571 6.0 - -
66 - - - - 4572 5.8 - -
67 - - - - 4574 5.7 - -
68 - - - - 4580 5.9 - -

# 4875 < rpm Power 5125 < rpm Power 5375 < rpm Power 5625 < rpm Power
≤ 5125 ≤ 5375 ≤ 5625 ≤ 5875
1 4864 6.3 5126 7.3 5417 8.5 5680 9.4
2 4869 6.5 5127 7.2 5418 8.3 5681 9.3
3 4873 6.1 5127 7.2 5419 8.9 5682 9.7
4 4876 6.5 5176 7.5 5419 7.8 5682 9.3
5 4878 6.6 5177 7.4 5422 8.2 5685 9.4
6 4880 6.7 5179 7.5 5428 7.0 5689 9.5
7 4881 6.8 5180 7.5 5520 8.6 5687 9.2
8 4882 6.7 5180 7.0 5529 8.4 5691 9.5
9 4882 6.5 5181 7.3 5531 8.8 5698 9.3
10 4884 6.6 5182 7.4 5569 9.0 5699 9.5
11 4886 6.7 5184 7.4 5571 8.5 5676 9.1

39
12 4886 6.5 5185 7.1 5572 9.1 5679 9.5
13 4886 6.6 5185 7.4 5575 8.8 5682 8.5
14 4887 6.3 5187 7.4 5576 9.1 5681 11.0
15 4888 6.7 5232 7.8 5576 8.6 5680 7.9
16 4901 6.8 5233 7.6 5577 9.1 5679 8.5
17 4926 6.6 5236 7.6 5578 8.3 5683 8.0
18 4929 6.4 5236 7.9 5578 9.0 5684 8.9
19 4960 6.4 5237 7.7 5580 8.8 5685 9.3
20 4975 6.5 5237 7.8 5591 8.8 5686 9.5
21 4983 6.6 5238 7.6 5591 9.0 5692 10.1
22 4983 6.8 5238 7.9 5595 9.0 5690 9.7
23 4983 6.5 5239 7.6 5595 8.8 5691 9.1
24 4984 7.0 5240 7.7 5595 9.0 5695 8.5
25 4984 7.0 5242 7.6 5596 8.5 5696 8.5
26 4986 6.7 5247 7.8 5609 8.8 5699 9.6
27 4986 6.9 5248 7.7 5614 8.9 5701 9.1
28 4987 6.6 5248 7.6 5616 8.8 5703 9.3
29 4993 7.0 5249 7.7 - - 5703 9.7
30 4993 6.8 5250 7.7 - - 5768 10.4
31 4994 6.9 5251 7.6 - - 5774 9.6
32 4998 6.8 5254 7.9 - - 5789 10.0
33 4998 6.8 5254 8.2 - - 5791 9.3
34 4999 6.7 5254 7.7 - - 5795 9.1
35 5061 7.1 5254 7.6 - - 5793 9.6
36 5062 6.8 5255 7.3 - - 5796 10.0
37 5063 6.9 5256 7.8 - - 5797 10.2
38 5066 6.8 5257 7.8 - - 5801 9.3
39 5068 7.2 5257 7.7 - - 5806 9.0
40 5070 7.1 5258 7.7 - - 5810 8.5
41 5074 6.9 5258 7.3 - - 5812 8.5
42 5079 7.0 5261 7.4 - - 5811 10.5
43 5080 6.8 5262 7.8 - - 5817 9.2
44 5085 7.1 5267 7.8 - - 5815 8.7
45 5086 7.1 5268 7.6 - - 5814 9.5
46 5088 6.9 5269 7.7 - - 5821 8.5
47 5090 6.7 5269 7.6 - - 5751 8.9
48 5091 7.1 5270 7.6 - - 5753 9.7
49 5092 6.9 5273 7.5 - - 5750 9.4
50 5096 7.1 5274 7.6 - - 5749 9.1
51 5097 7.1 5275 8.1 - - 5741 9.0
52 5099 7.0 5276 7.6 - - 5746 9.0
53 5100 7.2 5276 7.9 - - 5745 8.9

40
54 5101 7.0 5277 7.6 - - 5754 9.5
55 5103 7.0 5279 7.7 - - 5753 9.3
56 5107 7.3 5342 8.2 - - 5747 9.1
57 5108 7.4 5343 8.0 - - 5741 8.7
58 5108 7.1 5349 7.9 - - 5740 9.4
59 5121 7.2 5351 7.8 - - 5742 9.1
60 5122 7.2 5353 8.0 - - 5741 9.1
61 5122 7.4 5354 8.2 - - 5740 9.3
62 5123 7.2 5359 7.9 - - 5737 9.1
63 5124 7.2 5361 7.9 - - 5738 9.1
64 5124 7.3 5362 7.9 - - 5736 8.9
65 - - - - - - 5732 9.1

# 5875 < rpm ≤ 6125 Power 6125 < rpm ≤ 6375 Power 6375 < rpm ≤ 6625 Power
1 5890 9.3 6126 10.5 6419 11.1
2 5893 9.5 6128 10.3 6466 12.1
3 5923 9.6 6129 10.5 6472 11.4
4 5928 9.5 6130 10.3 6473 11.1
5 5929 9.6 6136 10.5 6479 11.5
6 5930 9.6 6140 10.1 6483 11.7
7 5931 9.9 6144 10.1 6500 11.5
8 5932 9.6 6151 10.0 6511 11.9
9 5933 10.1 6153 10.5 6521 11.4
10 5933 9.9 6156 10.4 6529 12.3
11 5935 9.7 6210 10.5 6532 11.7
12 5937 9.6 6210 10.2 6542 12.1
13 5937 9.5 6211 10.5 6543 11.6
14 5937 9.7 6212 10.5 - -
15 5938 9.4 6215 11.5 - -
16 5939 9.4 6216 10.8 - -
17 5939 9.8 6218 11.3 - -
18 5939 9.2 6218 11.5 - -
19 5940 9.9 6218 10.4 - -
20 5940 9.6 6219 10.9 - -
21 5941 9.7 6219 11.5 - -
22 5942 10.7 6219 10.5 - -
23 5942 10.1 6222 10.3 - -
24 6008 10.4 6223 11.0 - -
25 6009 10.9 6225 11.1 - -
26 6009 10.3 6225 11.0 - -
27 6010 10.8 6226 9.2 - -
28 6010 9.6 6226 10.7 - -
29 6011 9.6 6227 10.9 - -

41
30 6011 9.9 6229 10.9 - -
31 6012 9.7 6229 10.7 - -
32 6013 10.2 6230 11.3 - -
33 6014 10.5 6230 11.7 - -
34 6014 10.2 6230 11.8 - -
35 6014 9.8 6230 10.9 - -
36 6015 10.2 6230 10.5 - -
37 6016 9.9 6230 10.3 - -
38 6017 9.8 6231 11.3 - -
39 6017 10.5 6232 12.7 - -
40 6032 10.0 6233 11.2 - -
41 6036 9.6 6233 11.3 - -
42 6039 10.7 6233 10.6 - -
43 6042 9.6 6234 12.4 - -
44 6042 11.0 6237 10.1 - -
45 6043 10.5 6240 10.8 - -
46 6046 10.2 6242 11.3 - -
47 6046 9.7 6243 10.2 - -
48 6048 9.9 6244 11.2 - -
49 6049 10.5 6246 10.2 - -
50 6049 10.4 6263 10.7 - -
51 6050 9.9 6266 11.2 - -
52 6051 10.8 6274 10.8 - -
53 6051 9.7 6277 10.2 - -
54 6052 10.2 6278 11.0 - -
55 6052 10.4 6280 11.8 - -
56 6052 10.1 6285 10.7 - -
57 6052 10.5 6290 11.2 - -
58 6052 10.3 6291 11.2 - -
59 6053 10.9 6293 10.7 - -
60 6053 10.0 6294 11.0 - -
61 6054 10.6 6295 10.7 - -
62 6054 10.5 6297 11.1 - -
63 6055 10.4 6297 11.9 - -
64 6055 11.0 6298 11.2 - -
65 6056 10.5 6362 11.2 - -
66 6057 10.3 6367 10.7 - -
67 6058 10.8 6367 10.7 - -
68 6059 10.2 6368 11.5 - -
69 6059 10.6 - - - -
70 6059 10.4 - - - -
71 6060 10.4 - - - -

42
72 6062 10.5 - - - -
73 6062 10.5 - - - -
74 6065 10.6 - - - -
75 6066 9.5 - - - -
76 6070 10.8 - - - -
77 6071 10.5 - - - -
78 6073 9.9 - - - -
79 6075 9.6 - - - -
80 6080 9.8 - - - -
81 6081 9.9 - - - -
82 6083 10.8 - - - -
83 6083 10.0 - - - -
84 6084 9.8 - - - -
85 6085 10.2 - - - -
86 6087 10.2 - - - -
87 6093 10.6 - - - -
88 6096 10.2 - - - -
89 6097 10.2 - - - -
90 6097 10.2 - - - -

43
11.8 Appendix 8: Determination of Power Efficiency

Angular Power Fractional Absolute

velocity (ω) Efficiency (η) 𝚫𝛈
uncertainty ( 𝛈 ) uncertainty (∆𝜼)

/𝒓𝒂𝒅𝒔−𝟏 𝒌𝒈𝒎𝟐 𝒔−𝟑 - 𝒌𝒈𝒎𝟐 𝒔−𝟑

0 0 1.090 0
26 5.05041 1.492 7.53647
52 5.03806 1.497 7.54115
79 5.19595 1.501 7.79749
105 5.41957 1.504 8.15253
131 5.66842 1.508 8.54589
157 5.90947 1.511 8.92700
183 16.9106 1.582 26.7590
209 24.9627 1.622 40.4835
236 31.3804 1.651 51.8054
262 39.4344 1.685 66.4387
288 48.5075 1.722 83.5170
314 54.8055 1.748 95.8195
340 56.5343 1.760 99.5000
367 56.0104 1.765 98.8327
393 54.6171 1.766 96.4554
419 53.4560 1.767 94.4715
445 50.8639 1.764 89.7014
471 49.2979 1.763 86.8983
497 47.8633 1.761 84.2870
524 47.4140 1.762 83.5577
550 47.6857 1.765 84.1640
576 48.2416 1.769 85.3168
602 49.3993 1.774 87.6109
628 50.3016 1.777 89.4075
655 51.2764 1.782 91.3497
681 52.3214 1.786 93.4234
707 53.1726 1.789 95.1214

44
11.9 Appendix 9: Arduino Code for Modulating Motor RPM

int VoltPin = A0;

int SignalPin = 9;
int readVal;
float Volt;
#include <Servo.h>
Servo ESC;

void setup() {
ESC.attach(9,1000,2000);
Serial.begin(9600);
}
void loop() {
realVal = analogRead(VoltPin);
realVal = map(realVal, 0, 1023, 0 180);
ESC.write(realVal);
}

45