Blade Element Momentum

Theory
Thomas LAMBERT
November 22, 2021
Introduction
Blade Element Momentum Theory (BEMT)
▶ Blade Element Theory (BET) + Momentum Theory

Principle
▶ Momentum theory
▶ Calculate inflow

▶ Blade Element Theory Ω

▶ Calculate aerodynamic loads on a small r0
y dy
element of blade, dy R
▶ Integrate along the blade for the total
thrust and torque (and power)

1
General assumptions

Assumptions
▶ Flow is
▶ Incompressible
▶ Inviscid
▶ Irrotational
▶ Uniform
▶ Continuous flow velocity and pressure, except at the disk
▶ Isolated airfoils
▶ No interaction between the blades

2
Blade Element
Angles
R ▶ Blade pitch: θ
 
dL dD V∞ + vi
▶ Inflow angle: ϕ = tan−1
Ωy − ui
▶ True AOA: α = θ − ϕ

U Ωy − ui Forces
α ▶ dL = 12 ρ U2 c cl dy
ϕ
V∞ + vi ▶ dD = 21 ρ U2 c cd dy
θ
where cl and cd come from the sectional
characteristics of the blade element
3
Blade Element
dFx ▶ Force perpendicular to rotor disk
R
dL dD dFz = dL cos ϕ − dD sin ϕ

dFz ▶ Force in-plane with rotor disk

dFx = dL sin ϕ + dD cos ϕ

U Ωy − ui
α ▶ Thrust, Torque, Power
ϕ
V∞ + vi dT = Nb dFz = Nb (dL cos ϕ − dD sin ϕ)
θ dQ = Nb dFx y = Nb (dL sin ϕ + dD cos ϕ) y
dP = Nb dFx Ωy = Nb (dL sin ϕ + dD cos ϕ) Ωy

4
Non-dimensionalizations

▶ Element rel. position ▶ Thrust coefficient1

y U dT
r= ≈ dCT =
R ΩR ρA(ΩR)2

▶ Solidity ▶ Torque coefficient1

blade area N c
σ= = b dQ
disk area πR dCQ =
ρA(ΩR)2 R

1
US notation! Some EU textbooks define these coefficients with a factor 1/2 in the denominator.

5
Inflow factors
The axial inflow, vi , and azimuthal (swirl) inflow, ui , are unknown. We can replace them by
using the axial inflow factor and swirl factor:

Uax = V∞ + vi = (1 + a)V∞
Uθ = Ωy − ui = (1 − b)Ωy

The local flow velocity and inflow angle

q
U = ((1 + a)V∞ )2 + ((1 − b)Ωy)2
 
−1 (1 + a)V∞
ϕ = tan
(1 − b)Ωy
This formulation is only valid for cases where V∞ ̸= 0.
6
Momentum equations
The incremental thrust on the rotor annulus can be derived
from the mass flow rate through the annulus and twice the
velocity at that section:

dṁ = ρdA(V∞ + vi ) = 2πρ(V∞ + vi )y dy

y
so the incremental thrust on the annulus is
dy
2
dT = 2ρ(V∞ + vi )vi dA = 4πρV∞ a(1 + a)y dy

Likewise, the torque can be expressed as

2
dQ = 4πρV∞ b(1 + a)Ωy 3 dy

7
Nonlinear System
Nonlinear system
▶ Blade Element Theory

dT = Nb dFz = Nb (dL cos ϕ −dD sinϕ )

dQ = Nb dFx y = Nb (dL sinϕ +dD cosϕ) y
▶ Momentum Conservation
2 a (1+a )y dy
dT = 4πρV∞
2 b (1+a )Ωy 3 dy
dQ = 4πρV∞
where  
(1 + a)V∞
ϕ= tan−1
(1 − b)Ωy
8
Iterative procedure
For each blade element:

1. Make an initial guess for the inflow factors a and b

2. Calculate the induced angle (ϕ) and the effective angle of attack (α)
3. Calculate lift and drag using cl and cd polar2
4. Calculate dT and dQ with the Blade Element Theory
5. Update the inflow factors using dT and dQ in the Momentum Equations
6. (repeat 2-5 until convergence)
7. Determine total thrust, torque and power
2
in theory, cl (α, M, Re) and cl (α, M, Re) !

9
Numerical issues

Convergence of nonlinear system never guaranteed

Workaround
▶ Linearize the system
+ Simple analytical solution
- Lots of additional hypotheses
▶ Rewrite the system to have a single equation
+ Higher chances of convergence
- Mathematically difficult (see. appendix)

10
Assumptions
Assume
1. Out-of-plane velocity smaller than in plane velocity: (Vc + vi ) ≪ Ωy
2. Swirl velocity (ui ) negligible
3. Small induced angle ϕ
4. Drag coefficient much smaller than lift coefficient
The Blade Element Theory equations become
▶ dT ≈ dL
▶ dQ ≈ (ϕdL + dD)y
▶ dP ≈ (ϕdL + dD)Ωy

These assumption may not be optimal for propellers.

11
Non-dimensionalizations and inflow expression

▶ Swirl factor b = 0, as the swirl velocity is negligible

▶ The (axial) inflow ratio

Vc + vi
λ= = ϕr
ΩR

▶ Solidity
blade area N c
σ= = b
disk area πR

12
Integrated Rotor Thrust and Torque

After non-dimensionalization, the incremental forces must be integrated along the blade
1
σ 1 2
Z Z
CT = dCT dr = cl r dr
0 2 0
Z 1
σ 1
Z
CQ = dCQ dr = (ϕcl + cd )r 3 dr
0 2 0

13
Thrust approximation
Following steady linearized aerodynamics, the local lift becomes

cl = clα α = clα (θ − ϕ)

Therefore, the thrust coefficient becomes

Z 1
σ
CT = clα (θ − ϕ)r 2 dr
2 0
Z 1
σ
= clα (θr 2 − λr)dr
2 0

Spanwise variation of inflow λ(r) ?

14
Momentum equations

Incremental thrust on the annulus is still

y
dT = 2ρ(V∞ + vi )vi dA = 4πρ(V∞ + vi )vi y dy
dy
or in non-dimensional form

dCT = 4λλi rdr = 4λ(λ − λ∞ )rdr

15
Radial Inflow
The Blade Element Theory gave us previously
1 σc
dCT = σcl r 2 dr = lα (θr 2 − λr)dr
2 2
Combining the two results gives a quadratic equation
 σc  σc
lα
λ2 + − λ∞ λ − lα θr = 0
8 8
and solving for λ leads to
s 2  
σclα λ∞ σc σclα λ∞
λ(r, λ∞ ) = − + lα θr − −
16 2 8 16 2

16
Twist
At the center of the rotor, the linear speed is low
⇒ Inner portions of the blade produce little lift
⇒ Twist used to increase α near the root

No twist
20 −10◦ linear twist
−20◦ linear twist
Example

dL [N]
▶ No twist: θ = θtip 10
▶ Linear twist: θ = θtip + θtw (r − 1)

0
0 0.2 0.4 0.6 0.8 1
17
r [-]
Ideal Twist
There is a special solution of the inflow equation that gives a uniform inflow over the
blade span: θr = cst = θtip , or in term of twist distribution
θ(r) = θtip /r

▶ This special distribution is called ideal twist

▶ It corresponds to the minimum induced power for the rotor in hover or axial climb
▶ However, it can not be implemented at the root (→ ∞)
▶ The solution is to trim the twist near the root (i.e. limit it) or simply use a linear twist
that is close enough on the outboard of the blade
Stall!
18
Ideal Twist
Example
Same blades (σ = 0.10), different twist at constant rotor thrust CT = 0.008.
2
0.08 Ideal blade twist
No twist
0.06 1.5 −10◦ linear twist
−20◦ linear twist
λi [-]

cl [-]
0.04 1

0.02 0.5

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
r [-] r [-]

19
Tip Loss
High local inflow at the tip of the blades due to trailing vortices
⇒ Reduction of lift (tip loss)
⇒ Need to account for these 3D effects in the BEMT (2D method)
1
Prandtl’s tip-loss function
0.8
▶ Correction factor F
 
2 0.6
cos−1 exp−f

F [-]
F=
π 0.4 Nb = 2; ϕ = 0.2
Nb = 4; ϕ = 0.2
where 0.2 Nb = 2; ϕ = 0.05
  Nb = 4; ϕ = 0.05
Nb 1−r
f= · 0
2 rϕ 0.7 0.8 0.9 1
r [-] 20
Tip Loss

Prandtl’s tip-loss function

▶ The correction factor can be incorporated in the BEMT

dCT = 4Fλ(λ − λ∞ )rdr

▶ In this case, the inflow velocity becomes

s
λ∞ 2 σclα
  
σclα σclα λ∞
λ(r, λ∞ ) = − + θr − −
16F 2 8F 16F 2

▶ Iterative process as F = f (ϕ) = f (λ)!

21
Procedure for linearized system
For each blade element, calculate:

1. the inflow ratio, λ

p
2. the flow velocity, (V∞ + vi )2 + (Ωy)2
3. the induced angle (ϕ) and the effective angle of attack (α)
4. (iterate 1-3 to account for tip-loss)
5. lift and drag using cl and cd polar3
6. thrust, torque and power

Then sum for all the elements and multiply by the number of blades in order to get the
solution
3
for the rotor.
in theory, cl (α, M, Re) and cl (α, M, Re) !

22
Conclusion

Simple numerical tool that gives decent results for a rotor in isolation

▶ Multiple formulations and ways to solve the system

▶ Convergence not always guaranteed
▶ Results need to be interpreted and checked against the reality

Perfectly suited for pre-design phase of a rotor or propeller

23
Verify!

Do not forget to check and verify your results and their physical meaning!

▶ Sonic conditions at tip?

▶ Stalled parts of the blade near root?
▶ Polar used for cl and cd valid with calculated Re?
▶ ...

24
Validation i
Caradonna and Tung, 1981. [NASA TM-81232]
Exp. 5deg
0.6 Exp. 8deg Exp.
0.01 BEMT
BEMT 5deg
BEMT 8deg
0.4
cl [-]

CT [-]
0.01
0.2

0
0 0.2 0.4 0.6 0.8 1 0
2 4 6 8 10 12 14
r [-]
AOA [deg]

25
Validation ii
Knight and Hefner, 1937. [NACA TN-626]
·10−2
1 0.8
0.8
0.6
0.6
CT [-]

FM [-]
Exp. σ = 0.042
0.4
0.4 BEMT σ = 0.042
Exp. σ = 0.064
BEMT σ = 0.064
0.2 Exp. σ = 0.085 0.2 BEMT σ = 0.085
Exp. σ = 0.106
BEMT σ = 0.106
0 0
0 0.2 0.4 0.6 0.8 1 0 2 4 6 8 10 12
CQ [-] ·10−3 CT /σ [-] ·10−2
26
Download the MATLAB code and this presentation on

https://gitlab.uliege.be/thlamb/bemt

27
BEMT matlab code
This code is a calculation code. It mainly outputs the Thrust and Power of a given
configuration.
THIS IS NOT A FULL DESIGN METHODOLOGY that will give you a propeller that fits your
need in one click!

1. Pick an airfoil (see UIUC Airfoil Data Suite)

2. Generate polars using XFoil
3. Create a config file, indicating the propeller geometry and polars to use
4. Run the code and check the thrust and power of your configuration
5. Iterate on the propellers parameters until you have the desired performance

28
References I

Leishman, J. Gordon (2006). Principles of Helicopter Aerodynamics. 2nd ed. Cambridge

Aerospace Series 18. Cambridge ; New York: Cambridge University Press.
Seddon, John et al. (2011). Basic Helicopter Aerodynamics. en. Third Edition. Aerospace
Series. Chichester, UK: John Wiley & Sons, Ltd. DOI:
10.1002/9781119994114.fmatter.
Stahlut, Conor W. (2012). “Aerodynamic Design Optimization of Proprotors for
Convertible-Rotor Concepts”. Master Thesis. College Park (MD), USA: University of
Maryland.
References II

Stahlut, Conor W. and J Gordon Leishman (2012). “Aerodynamic Design Optimization

of Proprotors for Convertible-Rotor Concepts”. In: American Helicopter Society 68th
Annual Forum. Fort Worth, Texas, USA, pp. 1–24.
To go further...

Stahlut and Leishman (2012) proposed different ways to solve efficiently the BEMT system
without needing to make strong hypotheses

▶ No small angles
▶ Take into account swirl velocity
▶ Drag not negligible with respect to the lift
▶ True lift and drag (no linearized simplifications)

The methodology remains the same (equating BET and Momentum equations). But the
system is much more difficult to solve.
To go further...
The complete thrust equation is
1 p
dCT = σ ξ 2 + λ2 (cl ξ − cdλ)dr = 4KT |λ|λi r dr
2
and the torque (or power)
1 p
dCP = σ ξ 2 + λ2 (cl λ + cdξ)r dr = 4KP |λ|ξr 2 dr
2
a third equation can be used to link the inflow ratios
λ
tan(ϕ) =
ξ
Where KT and KP are the loss factors (resp. in-plane and out-of-plane) and ξ is the in-plane
velocity ratio
Ωy − ui
KT = 1 − (1 − F) cos(ϕ); KP = 1 − (1 − F) sin(ϕ); ξ =
ΩR
To go further...

This system is impossible to solve analytically and needs to be solved numerically.

BUT, the convergence is not guaranteed due to the presence of nonlinear, transcendental
equations.

Solutions:

1. Fixed-point iteration scheme with a very low relaxation factor in order to calculate the
values of λ and ξ iteratively.
2. Or modify the formulation to express everything in terms of ϕ and solve the system
using a bracketed method (e.g. bisection).