FORCES AND MOMENTS IN CFD ANALYSIS

Authors
Ing. Zdeněk Říha, PhD., Institute of Geonics ASCR v.v.i., Email: [email protected]
Ing. Josef Foldyna, CSc., Institute of Geonics ASCR v.v.i.

Anotace
V příspěvku je pojednáno o nalezení působiště celkové síly v případě, že známe
celkovou sílu a celkový moment. S daným typem úlohy se setkáváme, když řešíme proudění
pomocí CFD programu. Správná interpretace působiště síly je za daných podmínek velmi
důležitá v případě dalších navazujících analýz.

Annotation
The paper shows how to find the force point of action, when we know the total force
and the total moment. It is usual case if we solve fluid flow with help of a CFD code. Correct
interpretation of force´s act is very important for next possible analysis.

1. Introduction
Various CFD codes offer possibility of calculating of the total force and of the total
moment to the given wall surface, see [3]. Sometimes, the knowledge of these values could be
important for subsequent design of the given mechanical part, for example dimensioning of a
rotating shaft with vanes in a mixer, etc. The total force and the total moment knowledge are
important especially in cases when we cannot use some FEA code (e.g. ANSYS) in next
design step. The paper describes relationships between calculated forces and moments. We
will also try to find correct interpretation of forces and moments calculated in a CFD code for
subsequent strength analysis.

2. Calculation of forces and moments on given surfaces

In CFD analysis, every surface is divided into elementary areas. This is consequence
of geometry meshing, when we divide given domain to cells where base equations of flow are
solved.

Figure 2.1 – Elementary forces description

In fluid flow calculations, we can decompose total elementary force FET to normal and
tangential direction with regards to small elementary area, where force acts, see figure 2.1.
Normal force created by pressure is called pressure force FEP. Tangential force created by
viscosity is called viscous force FEV. Then, the total elementary force has to be vector sum of
the pressure force and of the viscous force.
→ → →
F ET = F EP + F EV (2.1)
Described total force can be also fragmented (FEx, FEy, FEz) according to directions of
existing coordinate system x, y, z. We obtain components of total elementary force and we
can write following relation:
→ → → →
F ET = i FEx + j FEy + k FEz . (2.2)
Small area depicted on the picture 2.1 can be part of a wall surface, where we are going to
determine total force FT. If given wall surface contains “m” small areas where forces act, we
can write following equations:
m
FTx = ∑ F Ex , (2.3)
E =1
m
FTy = ∑F Ey , (2.4)
E =1
m
FTz = ∑F Ez , (2.5)
E =1
→ → → →
F T = i FTx + j FTy + k FTz . (2.6)
Forces FTx, FTy, FTz are components of total force FT. Vectors i, j, k in equations (2.2) and
(2.6) are unit vectors, which define given directions x, y, z of the given coordinate system.
Now, we have calculated “non-existing” total force acting on given wall surface. Values from
equations (2.3) (2.4) and (2.5) represent standard output from above mentioned CFD codes.
We are able to calculate inclination of line, where “non-existing” total force lies with respect
to axes of given coordinate system. Unfortunately, we do not know coordinates of any point
of action, which lies on the above mentioned line. In other words, we know size of total force
and its direction, but we do not know where the total force FT acts. It is consequence of
simplification, when we substitute “m” real elementary forces by one “non-existing” total
force. Nevertheless, sometimes it is necessary to know total forces for next analysis, e.g.
dynamic analysis of whole system, where our solved piece represents only a part of the
assembly. Therefore, correct interpretation of the “non-existing” total force is very important.
Let’s find the total moment MT given by above mentioned elementary forces. Moment can be
defined as consequence of the force acting. Generally speaking, the moment equals to vector
product of the position vector and the force vector, see [2].
→ → →
M = r× F (2.7)
The equation (2.7) can be written in matrix form:
→ → →
i j k
rx ry rz . (2.8)
Fx Fy Fz

Then, we can find components of moment using Sarrus´s rule for the determinant calculation.
→ → →
M x = i r y Fz − i rz F y (2.9)
 → → →
M y = j rz F x − j rx Fz (2.10)
→ → →
M z = k rx F y − k r y F x (2.11)
Moment is vector sum of its components.
→ → → →
M = Mx+ M y+ Mz (2.12)
Now, we come back to the wall surface with elementary areas and elementary forces. We will
calculate moment of the elementary forces to given point MET. We can directly use equations
(2.9), (2.10) and (2.11) because components of position vector are known, see figure 2.2.
Let´s suppose that we can define global coordinate system, where the moment is calculated
and local coordinate system with the same orientation as global coordinate system on the
elementary area, where elementary forces act, see figure 2.2. Position vector rET defines
distance between booth coordinate systems. According to above mentioned equations, we can
write equations for components of the elementary moment MET:

Figure 2.2 – Elementary moments description

M Ex = rEy FEz − rEz FEy (2.13)

M Ey = rEz FEx − rEx FEz (2.14)

M Ez = rEx FEy − rEy FEx (2.15)

Total components of moments can be identically calculated as components of the total forces
which act on given elementary areas, see equations (2.3), (2.4) and (2.5).
m
M Tx = ∑ M Ex (2.16)
E =1
 m
M Ty = ∑M Ey (2.17)
E =1
m
M Tz = ∑M Ez (2.18)
E =1
Then, the total moment equals:
→ → → → → → →
M T = i M Tx + j M Ty + k M Tz = M Tx + M Ty + M Tz (2.19)
Values given by equations (2.16), (2.17) and (2.18) represent standard output from CFD
codes.
Calculated total moment MT is real existing value, because there is no simplification
introduced into the calculation. The total moment MT is direct consequence of elementary
forces acting on given elementary areas of the observed wall surface. Now, we have
calculated the total force FT on given surfaces wall and the total moment MT in the given
point. In other words, we have standard output of forces and moments from a CFD code.

3. Location of the total force and its moment

We can find point of action of the total force FT if we use equations (2.9)-(2.11)
describing relations between forces and moments. When we know the moment and the force,
then, we have three equations and three variables. If we are finding componets of the posotion
vector, it is possible to show that the system of equations (2.9)-(2.11) is singular. The system
of equations should describe line where total force is lying. However, there is very important
condition, which has to be fulfilled. Equations (2.9)-(2.11) describe points, which can create
line of the total force if vector of the total force FT and vector of the total moment MT are
perpendicular to each other. If this condition is not valid, then we can calculate more than one
line where the total force can act. Thus, such interpretation of the total force and total moment
is incorrect.

Figure 3.1 – 2D case, the total force and the total moment
Let´s find cases, where the total force FT and the total moment MT always will be
perpendicular to each other. In 2D space, vectors always have to be perpendicular, because
force FT and position vector rT always lie in plane where we solve fluid flow. Thus, the
moment MT always is perpendicular to plane (and force FT) where we solve fluid flow. Let´s
suppose that 2D geometry is defined in XY plane, see figure 3.1. In this case we can find
point of action using equation (2.11). One coordinate has to be selected as parameter ”p”
because only one equation (2.11) exists and second coordinate will be calculated. Then, we
can write:
y= p (3.1)
M + FTx • p
x = Tz (3.2)
FTy
We can see that the point of action is easy to find for 2D case. Equations (3.1) and (3.2) also
describe line, where total calculated force can act. Feature of the force line is that moment is
zero here, because position vector is always zero.
In 3D space, it can be that vectors MT and FT are not perpendicular to each other.
Nevertheless, let´s suppose that vectors of the total moment and of the total force calculated
by some CFD code are perpendicular to each other in 3D space. It means that cosϕ of angle
ϕ between force and moment vectors is zero (ϕ =90ο). This can be checked by equation (3.3),
see [1]. We can see that cosϕ also is zero if scalar product of above listed vectors is zero.
→ →
FT • M T FTx • M Tx + FTy • M Ty + FTz • M Tz
cos ϕ = → →
= (3.3)
FT • M T FTx2 + FTy2 + FTz2 • M Tx
2
+ M Ty2 + M Tz2

Figure 3.2 – 3D case, the total force and the total moment

Then, we can find point of action using equations (2.9)-(2.11). Equation system is singular. It
means that we have to choose one coordinate as a parameter “p” and calculate the rest of
coordinates.
x= p (3.4)
 FTy • p − M Tz
y= (3.5)
FTx
M Ty + FTz • p
z= (3.6)
FTx
If we choose coordinate “x” as a parameter “p”, see (3.4), we can derive rest of coordinates
from equations of the moment (2.10) and (2.11), see equations (3.5) and (3.6). Equation 2.9 is
not used in this case. Equations (3.4), (3.5) and (3.6) describe points of the spatial line, where
the total force FT lies and can act. Again, feature of the force line is that moment of the FT is
zero here, because position vector rT is always zero, see figure 3.2. There is also another
feature of the depicted line, if we move the total force FT along this line to any point lying on
this line, we always produce the same total moment MT to given point in the space (for
example point [0,0,0] in the global coordinate system, see figure 3.1 or 3.2).
We have solved determination of point of action for simple 2D and 3D cases. Now, let´s
suppose that the total moment MT and the total force FT vectors are not perpendicular to each
other.

4. General relation between the total force and the total moment.
Generally speaking, calculated total force and calculated total moment from CFD code
does not have to be perpendicular to each other. The total force FT is created by “m”
elementary forces FET. Every elementary force has line where it can act. Influence of “m”
elementary force´s lines is removed and substituted by one line of the total force FT. The
substitution of “m” lines by one line interrupts mathematical coupling between total force FT
and total moment MT. Therefore, we always should check perpendicularity between FT and
MT using the relation (3.3) for 3D case. If we will ignore above mentioned conditions of
perpendicularity, then we can easily obtain three parallel lines, where the total force FT could
act. Three parallel lines can be obtained, because we have three pairs of “independent”
equations (one coordinate is constant elective parameter), which describe line in space, where
the total force FT can act. This can be derived from relations (2.9)-(2.11). Of course, this
result is not acceptable, because given total force FT can have one line of action only.
Let´s solve case, when cosϕ is not equal to zero. It means that the vector of the total force FT
and vector of the total moment MT are not perpendicular to each other. Remember, that we
would like find point of action for above mentioned conditions. In other words, we would like
to find parametric equations for only one line, where the total force FT lies and acts.
We can define plane ρ using the total force vector FT and the total moment vector MT if we
move down the total force vector along position vector rT, see figure 3.2. We know that the
force is cause and the moment crates consequence of the force acting. We can perform
decomposition of the total moment MT to the total force FT direction and direction
perpendicular to the direction of the total force FT. We obtain moment MTF and moment MT0,
see figure 4.1. Following equations are valid:
→ → →
M T = M TF + M T 0 (4.1)
→
MT = 2
M Tx + M Ty2 + M Tz2 (4.2)

→
MT = 2
M TF + M T2 0 (4.3)
→
M TF = M T • sin ϕ (4.4)
 →
M T 0 = M T • cos ϕ (4.5)

Figure 4.1 – 3D case, the total force FT and the total moment MT in plane ρ

The moment MTF is directly produced by calculated total force FT and the moment MT0 is rest
given by non perpendicularity between the total force FT and the total moment MT. Now, we
know perpendicular moment MTF and we are able to find parametric equations of only one
line, where the total force FT lies. We can calculate size of the total moment MT using the
equation (4.2) and we can also calculate size of components MTF and MTO using equations
(4.4) and (4.5). If we want to find point of action, we have to find components of the moment
MTF in global coordinate system. Then, it will be possible to use equations (2.9)-(2.11), resp.
(3.4)-(3.6) and calculate point of action. One way how to find components of MTF is to
calculate size and components of position vector rs (see figure 3.1) in global coordinate
system. Then, we can use equations (2.9)-(2.11) for components of MTF calculation in global
coordinate system. The position vector rs is special vector, which allows to calculate the
moment using relation (2.7) resp. (4.6). This vector is perpendicular to the plane ρ.
Size of position vector rs can be calculated by following equation which is valid for vector
product, see [1]:
→ → →
M TF = r s • F T • sin ϕ ´ . (4.6)

As can be seen in Figure 4-1, ϕ´ equals 90o and sinϕ´ equals 1. Therefore, we can write:
→

→
M TF
rs = →
. (4.7)
FT

Now, we need calculate components of the position vector rs in global coordinate system. It is
possible to use direction cosines, see [1]. We can find direction cosines from vector product of
the total force FT and the total moment MT, because position vector rs is perpendicular to
plane ρ created by moments MT, MTF, MT0 and force FT. Let´s define vector FMT and its
components.
 → → →
FM T = F T × M T (4.8)
→ → 2 → 2

FM T = F T × MT (4.9)

FM Tx = FTy • M Tz − FTz • M Ty (4.10)

FM Ty = FTz • M Tx − FTx • M Tz (4.11)
FM Ty = FTx • M Ty − FTy • M Tx (4.12)
We can calculate searched direction cosines.
FM
cos α = → Tx , (4.13)
FM T
FM Ty
cos β = →
, (4.14)
FM T
FM Tz
cos γ = →
. (4.15)
FM T
Then components of position vector rs can be calculated using following equations:
→
rsx = rs • cos α (4.16)
→
rsy = rs • cos β (4.17)
→
rsz = rs • cos γ (4.18)

We have fully defined position vector rs. The calculated position vector rs is shortest distance
between lines, where the total force FT and the total moments MT act. Coordinates of the
position vector rs also define one point of action on line where the total force acts. Now, we
can calculate searched components of the moment MTF, if we install calculated components of
rS to equations (2.9) – (2.11), see equations (4.19)-(4.21).
M TFx = rsy FTz − rsz FTy (4.19)

M TFy = rsz FTx − rsx FTz (4.20)

M TFz = rsx FTy − rsy FTx (4.21)

If we substitute components of the position vector rs by general coordinates x, y, z we obtain
three parametric equations, which describe line where the total force FT lies. System of
equations is singular. In other words, we can find a lot of (infinitely) points on line, where the
total force FT can act. It means that we have to choose one coordinate as constant parameter
“p” and to calculate rest of coordinates:
y= p (4.22)
p • FTz − M TFx
z= (4.23)
FTy
M TFz + p • FTx
x= (4.24)
FTy
Equations (4.22)-(4.24) describe line of the total force FT in case that the calculated total force
and the calculated total moment MT are not perpendicular to each other. Now, we are able to
interpreted acting of total force for next possible analysis. We have found place where the
total force FT acts and moment MTF has to be zero. Nevertheless, we must not forget that
moment MT0 acts in point, where the total moment MT is calculated. For further possible
analysis, it is necessary to present the total force FT in given point of action “C” and the
moment MT0 in point “D”, where the total moment MT was calculated, see figure 4.2.

Figure 4.2 –possible interpretation of calculated results, example of beam

5. Example
Let´s go to take an example. We have solved some fluid flow using ANSYS FLUENT
CFD software. Geometry consist of rotating shaft with vanes, box, input pipe and output pipe.
Fluid flows from input into the output and it is also mixed by rotating vanes, see figure 5.1.
We would like to find point of action for correct interpretation of loading for next analysis.

Figure 5.1 – Mixer

It is easy to print out the total force FT and the total moment MT (in other words loading of
the rotating shaft with vanes) calculated by ANSYS FLUENT. Then, components are
following:
FTx = - 598,6 N; FTy = -469,4 N; FTz = -1120,5 N;
MTx = 1,17 Nm; MTy = 1,85 Nm; MTz = -75,2 Nm;
Total moment is calculated to point [0,0,0] (middle of the solved geometry on axis of
rotation).
Firstly, we should check if the total force FT and the total moment MT are perpendicular to
each other. When we use equation (3.3), then, cos(ϕ) equals 0,83. This implies that ϕ = 35,7
and tested vectors are not perpendicular to each other. We have to use layout listed in chapter
4.
Now it is necessary decompose total moment MT to direction of total force FT and direction
perpendicular to direction of total force. We can size of desired components using equations
(4.4) and (4.5). Then,
MTF = 43,93 Nm; MT0 = 61,03 Nm.
MTF is important moment which is perpendicular to total force line and is direct result of
action of the force FT. Components of this moment in absolute coordinate system fill
equations (4.19) – (4.21) which allow to find any point of action.
Let´s calculate size of position vector rs which is perpendicular to both moment MTF and
force FT. The result from equation (4.7) is following:
rs = 0,032 m.
Now, we will determine direction cosines of rs vector using FT x MT vector product and then
we will calculate rs vector components. Using of equations (4.8) – (4.15) we obtain following
values:
cos α = 0,63; cos β = -0,78; cos γ = -0,0039.
Then, equations (4.16) - (4.18) define components of rs vector.
rsx = 0,020 m ; rsy = -0,025 m; rsz = -0,0003m.
We can calculate components of MTF moment using equation (4.19)-(4.21)
MFTx = 28,15 Nm; MFTy = 23 Nm; MFTz = -24,67 Nm;
Now, we have collected all necessary variables and we are able to build equations (4.22)-
(4.24). In other words we are able to find point of action. For example, if we take “y = 0,03m”
then we obtain rest of coordinates according to equations (4.23) and (4.24)
x = 0,091m; z = 0,132m;
It is also necessary to calculate component of MTO moments. Firstly, we will calculate
direction cosines of FT according to equations (4.13)-(4.15), where we substitute size of FMT
vector and its component by size of the total force vector and its component. Then, we can
calculate components of MT0 with help of equations (4.16)-(4.18), where we substitute size of
rs by MT0 and we use calculated direction cosines of FT. Results are following:
MT0x = -26,97 Nm; MT0y = -21,15 Nm; MT0z = -50,49 Nm;
Components of vectors FT, MTF, MT0 and the point of action coordinates fully define loading
for next possible analysis.

6. Conclusion
The paper shows how to calculate point of action, if we know the total force FT and
the total moment MT to the given wall surface from a CFD code. We can use equations
presented in section 3, which come from equations in section 2, see (2.9)-(2.11). If we solve
2D case of fluid flow, we can always use components of the calculated total force FT and
components of the total moment MT in equations (2.9)-(2.11) and derive equations like (3.1)
and (3.2), because the total force vector FT and the total moment vector MT always are
perpendicular to each other. When we solve 3D fluid flow, firstly, we have to check if above
mentioned vectors are perpendicular to each other using the relationship (3.3). When
condition of perpendicularity is fulfilled for 3D case (cosϕ =0 resp. ϕ =90ο) we can do the
same as for 2D case. If condition of vector´s perpendicularity between the total force FT and
the total moment MT is not fulfilled, we should proceed according to the procedure presented
in the section 4. It means to find the real moment MTF produced by the total force FT. Finally,
the total force FT and its position have to be completed about the moment MT0 and its position
in presentation of loading.
5. References
[1] Rektorys, K., a kol.: Přehled užité matematiky I, publishng house Prometheus, Prague
2000, ISBN 80-7196-180-9,
[2] Florian, Z., Pellant, K., Schánek M.: Technická mechanika I – Statika,
http://www.umt.fme.vutbr.cz/_studium_/opory/Tech_mech_I.pdf ,
[3] Ansys-Fluent User´s guide, http://www.fluentusers.com/fluent/doc.