ME4550 – ME

Heliostat Final Design REPORT DESIGN

DANIEL BRAKHA, ETHAN CLAXTON, ASH NYGREN, JULIAN POSTAK,

ISAIAH SOICHER
Introduction
Heliostats are one of the many design solutions proposed to help bring
renewable energy into the modern age. Concentrated Solar Power (CSP)
plants operate by centering the sun’s rays into a central point from an entire
field, allowing for thermal and radiation collection at that central position.
The strongest benefit of this method is that it designs its energy storage
mechanism into the design, as the collector will continue to generate power
even when the sun is down due to the residual built-up heat. This design
requires a device that can accurately track the sun through the sky, and
constantly adjust to the perfect angle to redirect the sun into the collector,
which is where the heliostat comes in.

When designing a heliostat, there are many different variables that need to
be considered such as the cost, longevity, and efficiency of the device. Costs
will determine the upfront cost of the farm while the efficiency will determine
how much power it can generate and how quickly it will recoup its costs.
Finally, the longevity will act as an insurance, guaranteeing that the unit will
last longer before it needs maintenance or if there is a catastrophic weather
event. While costs can be controlled by simplifying a design, and longevity
can be managed by reducing the number of moving parts, the hardest part
of the design to optimize is its efficiency. Even small deviations from the
proper angle of projection to the controller will drastically reduce the amount
of energy being received, meaning that both the tracking system and control
systems need to be perfectly aligned.

One common issue found in modern heliostat designs is their use of gear
trains to control their projection angle. While gears are a fairly standard
solution, they come with a couple of pitfalls which can be very detrimental
for the heliostat’s efficiency. When selecting a gear, the designer is forced to
choose between high and low precision gearing, both of which will affect the
main principles for optimizing a heliostat listed above. Low-precision gearing
will allow for cheaper costs for the design, however low-precision gearing
often will result in backlash in the system. This backlash will result in the
system being imprecise in directing the sunlight and cause a huge drop in
efficiency. High-precision gearing will eliminate these backlash problems;
however, it makes it far more expensive, greatly increasing the costs, and
still has longevity risks as the gears wear over time.

1
Solution Approach
Our novel solution to the problem laid out above is to remove gears entirely
from our system. As you can see in Figure 1, our design hinges on its use of
three lead screws that will allow the system the flexibility to redirect in any
direction necessary, while distributing any loading evenly, and all with a high
degree of precision.

2
 Fig. 1. Full CAD Mock-up of Heliostat Design

The first component in our design is the mirror at the top. It will be 1 m^2
and weigh about 550N.

Fig. 2. Underside of Mirror CAD

Connected to the back of the mirror is a metal frame / rail which is welded to
the mirror to prevent bending and allows the lead screws to slide across the
back while supporting the weight. A ball joint is attached to the end of each
lead screw which allows the lead screw to rotate in proportion to the rails.

3
 Fig. 3. Motor in the Motor Mount in CAD Mock-up

The lead screw is powered by a female threaded motor that is mounted into
the motor mount. This motor will only need one horsepower, as the sun
moves slowly through the sky, meaning only a small total speed output is
needed.

Fig. 4. Tripod Base Bolted to the Ground in CAD Mock-up

This motor mount is then bolted into the tripod base, which supports the
total weight of the design. The tripod base is then bolted into the ground,
which is assumed to be a cement platform, using multiple cap screws. This

4
will hold the design in place, to ensure that even in high wind loading, the
design will remain sturdy.

Workload Distribution
Team Member Responsibilities

Daniel Brakha General Report Writing / Editing

Design Brainstorming +
Improvement

Calculations for Component,

Connections, and Power Analyses

Ethan Claxton General Report Writing / Editing

Design Brainstorming +
Improvement

Calculations for Component,

Connections, and Power Analyses

Ash Nygren General Report Writing / Editing

Design Brainstorming +
Improvement

Calculations for Component,

Connections, and Power Analyses

Julian Postak General Report Writing / Editing

Design Brainstorming +
Improvement

Calculations for Component,

Connections, and Power Analyses

Isaiah Soicher General Report Writing / Editing

Design Brainstorming +
Improvement

Full CAD + FEA Responsibilities

5
Write-up for FEA Report Sections

6
Analysis
Component Design
The main components of the heliostat system were analyzed for static and
fatigue failure. First, the 304 Stainless steel tripod base was isolated for
calculations.

Fig. 5. Tripod Base in CAD Mock-up

Due to the symmetry of the setup, one “leg” was analyzed with a fixed
geometry at the base and a load equal to 1/3 of the total maximum load
applied at the upper end. Using the reaction forces calculated at the base,
the stresses in the square cross-section beam were calculated:
− Ax Ay
σ axial= cos θ− sin θ=−217.85 kPa
A A
M a∗b / 2
σ bending= =20.17 MPa
I
From these values, the principal stresses were found, and the Von Mises and
Tresca methods were used to determine the safety factor against static
failure:
σ s=20.41 MPa

σ H =20.30 MPa

n y =10.04

7
Next, the tripod was analyzed for fatigue failure. The cyclic load was
determined between maximum wind load and zero wind load. First, the
endurance strength was calculated:
'
Se =k a k b k c k d k e S e =87.15 MPa

Further, the component was tested for theoretical infinite life:

σ m=10.10 MPa ; σ a =10.10 MPa

( )
−1
σa σm
nf = + =7.365
Se S ut

With a determined safety factor greater than one, infinite life of this
component against cyclic loading is theoretically possible.

Finally, the tripod legs were analyzed for buckling risk. Euler’s formula was
used, and the critical stress was compared to the maximum axial stress as
follows:
2
C∗π ∗E
σ cr = 2
=57.75 MPa
SR

σ cr
n buckling= =265.10
σ ax

With such a high safety factor, buckling is not expected to be a risk to the
tripod’s integrity.

The above calculations were then verified using SolidWorks Finite Element
Analysis. The results were as follows:

8
 Fig. 6. FEA Von Mises Analysis of Tripod Leg

SolidWorks FEA simulation was performed assuming a fixed geometry (green

arrows) at the base of the tripod leg to estimate the internal stresses
affecting each leg. The maximum design load, comprising of the weight of
the mirror assembly, lead screws, motors, motor mount frame, and wind
force was modeled as a distributed force across the face that the motor
mount is benched on, acting vertically. As demonstrated in Fig. 6, the internal
stresses experienced by this component are relatively minimal, and do not
approach the yield strength of the material. Fig. 7 shows the estimated
deformation of the motor mount arm under the maximum design load –
again, this deformation is minimal and should not affect the performance or
reliability of the component.

9
 Fig. 7. FEA Deflection Analysis of Tripod Leg

Next, the motor mount component was analyzed in a similar fashion. Again,
due to symmetry, one “arm” of the mount was inspected. The load on each
arm included the weight of a single motor and lead screw assembly, as well
as 1/3 of the mirror weight and 1/3 of the maximum wind load. The motor
mount is an AISI 304 steel component that attaches to the top of the tripod
base and secures the three lead screw linear actuators in 3 radially
positioned brackets that position the lead screw axes at 12 degrees from
vertical.

10
 Fig. 8. Motor Mount in CAD Mock-up

The principal stresses were found by computing the bending and shear
stresses based on the loads mentioned above:
σ 1=0.9318 MPa , σ 2=0 MPa , σ 1=−0.1230 MPa

These values were used to find the safety factor using MSST and DET
analysis, which were 194 and 205, respectively. Based on the geometry of
the component and the relatively small stresses it experienced, buckling risk
was not considered.

As with the tripod base, fatigue failure analysis was conducted based on a
load cycling between maximum wind load and zero wind load. The
endurance strength was again calculated using the same equation as above,
which yielded a value of 166.529 MPa.

The safety factor against fatigue failure was then calculated using the
following equation:

( )
−1
σa σm
nf = + =264
se sut

The safety factor of 264 suggests that infinite life of this component is
possible under the estimated load.

11
SolidWorks FEA simulation was performed assuming a fixed geometry (green
arrows) at the inner face of the motor mount frame to estimate the internal
stresses affecting the motor mount arm. The maximum design load,
comprising of the weight of the mirror assembly, lead screw, motor, and wind
force was modeled as a distributed force across the face that the motor is
benched on, acting in the direction of the lead screw’s axis of motion. As
demonstrated in Fig. 9 below, the internal stresses experienced by this
component are relatively minimal, and do not approach the yield strength of
the material. Fig. 10 shows the estimated deformation of the motor mount
arm under the maximum design load – again, this deformation is minimal
and should not affect the performance or reliability of the component.

Fig. 9. FEA Von Mises Analysis of Motor Mount

12
Fig. 10. FEA Deflection Analysis of Motor Mount

13
The final inspected component was the lead screw/ball joint assembly that
supports the mirror. A set of three lead screws with ball-joint ends are
attached to carriages that slide along fixed axes on the underside frame of
the mirror. Again, due to the symmetrical nature of the setup, a single lead
screw was analyzed. The component was modeled as a column fixed at its
base (where the lead screw enters the motor). The load, which consists of
1/3 the maximum wind load and 1/3 the mirror weight, in addition to the
mass of the lead screw (made of 1018 carbon steel) itself.

Fig. 11. Lead Screw/Ball Socket Carriage Assembly

First, the principal stresses were determined from the calculated reaction and
load forces:
σ 1=31.185 MPa

σ 2=0.9102 MPa

σ 3=0 MPa

14
Using MSST, the Tresca effective stress was found, and used to calculate the
safety factor against static failure:
σ S=σ 1−σ 3=31.185 MPa

Sy
η S= =11.062
σS

Next, buckling analysis was performed. The axial load was used to calculate
the critical load affecting each lead screw:
2
C π EI
Pcr = 2
=276.824 kN
L
The lead screws were verified to be classified as slender rods, and therefore
Euler’s Law was used to calculate the safety factor against buckling:
P cr
η= =274
P
Based on this extremely high safety factor, the lead screws were determined
to be at little risk of buckling failure.

Finally, the fatigue analysis was conducted as above, using load variation
between maximum and zero wind load. Se was once again calculated and
used to determine the safety factor against fatigue failure:

( )
−1
σa σm
ηf= + =6.265
Se S ut

This safety factor suggests that infinite life is possible, and fatigue failure is
not a significant concern for these components.

SolidWorks FEA simulation was performed assuming a fixed geometry (green

arrows) at the base of the lead screw (where it would meet the motor) to
estimate the internal stresses affecting the lead screw itself. The maximum
design load, comprising of the weight of the mirror assembly and wind force,
was modeled as a distributed force across the top face of the ball socket,
acting vertically. As demonstrated in Fig. 12, the internal stresses
experienced by this component are relatively minimal, and do not closely
approach the yield strength of the material. As expected, the higher stresses
are concentrated toward the end of the lead screw that is furthest from the
interface with the mirror – where the lead screw enters the linear actuator.
Fig. 13 shows the estimated deformation of the motor mount arm under the

15
maximum design load – this deformation is relatively small, reaching a
maximum of about 3mm at the top end of the screw, but should be
counteracted by the free travel of the ball socket on the mirror frame’s rail
system. Because of this, deformation of the lead screw should not affect the
performance or reliability of the component.

Fig. 12. FEA Von Mises Analysis of Lead Screw

16
 Fig. 13. FEA Deflection Analysis of Lead Screw

Connection Design
Motor Mount to Tripod Bolt Connection

For the connection between the motor mounting plate and the tripod, a
series of three bolts were used. Each bolt will have a nut at the end, with a
tripod in-between. This will create a bolt connection as depicted in Figure 14.
For this selection, an 80mm, zinc-plated, M14 grade 4 bolt will be used with a
7mm thick nut. The recommended torque for this bolt is 215303 Nmm. These
bolts will be used to keep the top of the assembly firmly secured to the
bottom, allowing for all the weight to be supported, and hold everything in
place during the potential strong winds that could cause extreme forces on
the joints.

Fig. 14. Free Body Diagram of Resulting Reaction Forces from Lead
Screw on Motor Mount Plate Bolt

The reaction forces on the motor mount caused by the lead screw were
calculated to be:

R x =204.793 [ N ]

R y =963.475 [ N ]

17
Using a washer with a diameter of 21mm and thickness of 1.75mm and a nut
that is 7mm thick and securing two plates with equal thicknesses of
25.375mm the following equations can be solved:
Lgrip =( 2 ⋅h plate ) + ( 2 ⋅hwasher )=54.25 mm

LBolt =L grip +hnut =61.25 mm

Sizing up to the next nearest size, the bolt's length is 80mm long. Since the
bolt is less than 125mm long the following equation was used to calculate
the length of the threads:
Lthread= ( 2⋅ d )+6=34 mm

This equation was used to find the stiffness of the bolt along with the
following equations using the table A8.1 for the A t value and table A-20 for
Ebolt:
l unthreaded =Lbolt −Lthread=46 mm

l threaded=L grip−l unthreaded =8.25 mm

2
π ⋅d 2
Aunthreaded = =153.94 m m
4
2
At =115 mm

Ebolt =210000 MPa

( A unthreaded ⋅ Athreaded ⋅ E bolt ) N

k bolt= =566708.8
( A threaded ⋅lunthreaded ) + ( A unthreaded ⋅lthreaded ) mm

18
 Fig. 15. Frustum Geometry of Motor Mount to Tripod
t 1=t 4=h washer=1.75 mm ; t 2=t 3=h plate =25.375 mm

t 1=t 4=h washer=1.75 mm ; t 3=t 4=h plate =25.375 mm D1=D4 =d washer =1.5⋅ d=21 mm ; D2=D3=2⋅t 1 ⋅ tan ( 30 )=23.0

E1=E 2=E3 =E4 =193000 MPa

Using these geometric values for the frustrum the following equation can be
used to calculate the stiffness of each layer where i corresponds to each
layer number. The moduli of elasticity are all from 304 stainless steel:

( 0.5774 ⋅ Ei ⋅ π ⋅ d ) 7 N 7 N
k i= ; k 1=k 4=2.48 ⋅10 ; k 2=k 3=5.68 ⋅10

( )
( ( 1.155 ⋅t i ) + Di−d ) ⋅ ( D i+ d ) mm mm
ln
( ( 1.155 ⋅t i ) + Di +d ) ⋅ ( Di−d )

( ) =2.31 ⋅10 mmN

−1
1 1 1 1 6
k m= + + +
k1 k2 k 3 k4

Using a load value from the reactions Rx=963.475N and a proof strength of
225MPa the following values were calculated:
k bolt
C= =0.197
k bolt +k m

A threaded ⋅ S p−F i
n L= =268
C⋅P

19
 Fi
n 0= =99
( 1−C ) ⋅ P

Seeing as the factors of safety are 268 and 99 for bolt static failure and bolt
loosening respectively, it is safe to assume that these bolts will be sufficient.
For the fatigue calculations the Pmin was assumed to be just the load of the
weight of the panel and the Pmax as the weight of the panel in addition to the
wind loading. This calculates out to get a fatigue loading safety factor of
62.75.
Pmax −Pmin Pmax + P min
P a= =451.6 N ; Pm= =511.9 N
2 2
k d=k e =1; k f =2.2

Sut =400 MPa ; S e =90.9 MPa

( S ut ⋅ A threaded )−F i
nf = =62.75

(( S
C ⋅ P a ⋅ ut + P m
Se ) )
Accounting for the shear on the bolt using the following equation it is very
small in comparison to the other loading:
4V
τ= =1.77 MPa
3A
Due to the low shear, it can be assumed that the fatigue loading and the
stiffness for this aspect of the bolt is inconsequential.

20
 Fig. 16. FEA Connection Analysis of Tripod Bolted to Motor Mount

FEA was completed using SolidWorks Simulation, where the motor mount
was bolted vertically to the tripod base using 3 alloy steel M14 hex bolts with
nuts. Under maximum load (2800N, purple arrows above), the factor of
safety of each bolt connection was calculated at 10.00, exceeding the target
FOS of 2.

21
Motor Mount to Motor Bolt Connection

Each motor will be bolted through the bottom face of the motor mount arm
using 3 equally spaced alloy steel M5 bolts, as shown in the figure below:

Fig. 17. Geometry of Motor to Motor Mount Connection

Assuming the load caused by the lead screw on the motor will be equally
shared by each bolt, the load will be a third of the weight and wind load in
only the y-direction:

R y =328.333 [ N ]

Using a washer with a diameter of 7.5mm and thickness of 1.75mm and a

nut that is 4mm thick and securing a cross-section of 51.53mm the following
equations can be solved:

Lgrip =( h plate ) + ( 2 ⋅h washer )=55.03 mm

LBolt =L grip +hnut =59.03 mm

22
Sizing up to the next nearest size, the bolt's length is 60mm long. Since the
bolt is less than 125mm long the following equation was used to calculate
the length of the threads:

Lthread= ( 2⋅ d )+6=16 mm

This equation was used to find the stiffness of the bolt along with the
following equations using the table A8.1 for the A t value and table A-20 for
Ebolt:
l unthreaded =Lbolt −Lthread=44 mm

l threaded=L grip−l unthreaded =11.03 mm

2
π ⋅d 2
Aunthreaded = =19.63 mm
4
2
At =14.2 mm

Ebolt =210000 MPa

( A unthreaded ⋅ Athreaded ⋅ E bolt ) N

k bolt= =66276.47
( A threaded ⋅lunthreaded ) + ( A unthreaded ⋅lthreaded ) mm

Fig. 18. Frustum Geometry of Motor-to-Motor Mount

t 1=t 4=h washer=1.75 mm ; t 2=20.42 mm ; t 3 =31.11mm

D1=D 4=d washer=1.5 ⋅d=7.5 mm ; D2 =D3=2 ⋅t 1 ⋅tan ( 30 )=9.52 mm

23
E1=E 2=E3 =E4 =193000 MPa

24
Using these geometric values for the frustrum the following equation can be
used to calculate the stiffness of each layer where i corresponds to each
layer number. The moduli of elasticity are all from 304 stainless steel. These
stiffnesses can be calculated for the full frustrum stiffness:

( 0.5774 ⋅ Ei ⋅ π ⋅ d ) 6 N 6 6 N
k i= ; k 1=k 4=3.95 ⋅10 ; k 2=2.03 ⋅10 ; k 3=1.85 ⋅10

( )
( ( 1.155 ⋅t i ) + Di−d ) ⋅ ( D i+ d ) mm mm
ln
( ( 1.155 ⋅t i ) + Di +d ) ⋅ ( Di−d )

( ) =649827.16 mmN
−1
1 1 1 1
k m= + + +
k1 k2 k 3 k4

Using a load value from the reactions Ry=328.333N and a proof strength of
225MPa the following values were calculated:
k bolt
C= =0.0926
k bolt +k m

A threaded ⋅ S p−F i
n L= =24.699
C⋅P
Fi
n 0= =8.2
( 1−C ) ⋅ P

Seeing as the factors of safety are 24.7 and 8.2 for bolt static failure and bolt
loosening respectively, it is safe to assume that these bolts will be sufficient.
For the fatigue calculations the Pmin was assumed to be just the load of the
weight of the panel and the Pmax is the weight of the panel in addition to the
wind loading:
Pmax −Pmin Pmax + Pmin
P a= =133.333 N ; Pm = =194.999 N
2 2
k d=k e =1; k f =2.2

Sut =400 MPa ; S e =90.9 MPa

( S ut ⋅ A threaded )−F i
nf = =44.72

(( S
) )
C ⋅ P a ⋅ ut + P m
Se

Since the nf is 44.72 it is safe to assume that these bolts will have infinite life
under cyclic loading.

25
 Fig. 19. FEA Connection Analysis of Motor Bolted to Motor Mount

FEA was completed using SolidWorks Simulation, where the motor was bolted
through the bottom portion of the motor mount using alloy steel M5 hex
bolts with nuts. Under maximum load (985N, purple arrows above), the
factor of safety of each bolt connection was calculated at 6.99, exceeding
the target FOS of 2.

26
Tripod into Concrete Bolt Connection

As visualized in figure 20, the base of each tripod leg is attached to a large
concrete base, utilizing a hexagonal cap screw. For this connection analysis,
we will be conducting static load analysis and fatigue analysis on one leg, as
the maximum and minimum forces on each leg will be similar.

Fig. 20. Cap Screw Geometry for Tripod to Ground Connection

This cap screw will support force in the X direction, creating shear on the
screw from the drag caused by the large mirror assembly. This assumption
would create the forces:
F max= A x =692.82 N

F min=−692.82 N

h=steel height=80 mm

On each of the three legs, found in the previous Component Design and
Analysis document, we will be using a 120 mm, zinc-plated, M10 grade 5.8
cap screw, which does not require a washer, bolted through our tripod leg
and into a large concrete base. From table A-30, the height, H, of an M10
hexagonal cap screw is 6.63mm.

The overall length of cap screw was found using the equation:

L>h+1.5 d ⇒ L>80+1.5 ( 10 ) ⇒ L> 95 mm .

27
Rounding up, this gave a cap screw total length, L, of 100 mm. Further
analysis was done on the cap screw to get the dimensions:

LT =2 d+ 6=2 ( 10 ) +6=26 mm .

l d =L−LT =74 mm .

d
l=h+ =85 mm .
2
l t =l−l d =11mm .
2
πd 2
Ad = =78.54 mm .
4
¿ table 8.1 : At =58 mm .

E=200 GPa .
F max
σ max= =8.82 MPa
Ad

F min
σ min= =−8.82 MPa
Ad

Since the cap screw we are using is zinc coated,

K=0.2

Fig. 21. Frustrum for Tripod Bolt

28
The conical frusta for this fastener are shown in figure 21, with the
dimensions:
l
t 1= =50 mm . ; t 2=h−t 1=30 mm . ; t 3=L−t 2−t 3=20 mm .
2
D1=D3 =1.5 d=15 mm .

D2=D3 +2 t 3 tan ( 30 )=38 mm .

To analyze the static load on this cap screw, we first had to find the stiffness
factors for the screw, each frustum, and the stiffness of the whole member.
Those were found with the equations where i corresponds to each layer:
A d At E 5
k b= =1.76697∗10 N / m.
A d lt + A t l d

( 0.5774 ⋅ Ei ⋅ π ⋅ d ) 5 N 5 N 5 N
k i= ; k 1=1.23305 ⋅10 ; k 2=1.00939 ⋅10 ; k 3=1.35300 ⋅10

( )
( ( 1.155 ⋅t i ) + Di−d ) ⋅ ( D i+ d ) m m m
ln
( ( 1.155 ⋅t i ) + Di +d ) ⋅ ( Di−d )

k m= ( 1 1 1 −1
+ +
k1 k2 k 3 )
❑ =3.93570∗104 N / m.

The fraction of the external load that the cap screw withstands is calculated
with the equation:
kb
C= =0.8178
k b +k m

Using the pre-determined factor of safety of 1.5 and the proof strength of 5.8
grade steel shown below, the preload was calculated with the equation:
S p=380 MPa

S p=380 MPa

A t S pL−F i
n L= =305.49
C Pmax

Fi
nO = =12344.62
( 1−C ) Pmax

Because both of our safety factors for static failure are well within our
margins of safety, we can be sure that our material selection will work

29
against static loads. The proof strength will be 380 MPa, and the ultimate
tensile will be 520 MPa for the rest of this section of the report to accurately
reflect the strength of the cap screw.

It is assumed that the threads are rolled, giving a stress concentration,

Kf=2.2, with the grade is 5.8, as shown in table 8.2.1. It is also assumed that
the system will be operating at standard operating conditions, leaving all of
our fudge factors negligible for strength calculations.
1
σ a= ( Pmax −Pmin )=8.82 MPa .
2
1
σ m= ( Pmax −Pmin )=0 MPa .
2

After finding these values Se can be found with the equation:

'
k d k e Se
Se =
Kf

Since the max load, Pmax, is < 1400 MPa, the assumption Se′=0.5Sut′can be
made.. Solving this equation, we get Se=118.18 MPa. With these values, we
can find the factor of safety against fatigue with the equation:
Sut At −F i
nf = =325.227

( S
( C ) Pa ut + P m
Se )
The factor of safety against fatigue is relatively high, meaning that there is
an extremely rare chance that this bolt fails due to fatigue. This factor of
safety suggests that a cheaper bolt could be used to save, whether it is a
lower grade or smaller diameter. We have decided to keep using this
stronger bolt, as the desert environment is often unpredictable, so having a
very high factor of safety will be a beneficial expense in case of any
catastrophic environmental events (such as a dust storm). Additionally, the
hardware presents a very small cost in comparison to the rest of the
components.

30
 Fig. 22. FEA for Tripod Bolt to Ground Full

Fig. 23. FEA for Tripod Bolt to Ground Single

FEA was completed using SolidWorks Simulation, where the tripod base was
bolted to the concrete base using 3 alloy steel M14 hex foundation bolts.
Under maximum load (3000N, purple arrows above), the factor of safety of
each bolt connection was calculated at 5.61, exceeding the target FOS of 2.

31
Mirror to Slides Weld Connection

As calculated in the previous Component Design Report, the forces on the

top of the lead screw, (transferred to the slides that are welded along both
sides to the mirror) are Px=782.5N in line with the weld and Py=863.8N
perpendicular to Px but still acting along the face of the mirror. The length of
the slides is d=350mm and they are b=50mm wide. If we have a weld depth of
h=50mm, we can find the throat length (t) of the weld to be 35.35mm and
use that to find the minimum area of the weld.
t=0.707 h=0.03535 m
2
A=2td=0.024745 m
With our loads, we can get τ for each load. Due to the load being in-line with
the weld, the indirect shear τ’’ for Px is assumed to be zero, as no net
moment will be generated on the welds. Additionally, the load Px is within the
welded area, so no real moment will be generated on the weld as the force
will be absorbed as direct shear.

P
τ '=
A
' '
τ x =31622.55 Pa; τ y =34908 Pa

'' Mr
τ =
J
'' ''
τ x =0 ; τ y =0

The resultant shear of τx and τy can then be found (note τi=τi'as there is no
indirect shear).

τ =√ τ 2y +τ 2x =47101.5 Pa

With a minimum safety factor of 2, the weld electrode was chosen.

0.577 S y
η=
τ
0.577 S y
2= ; S =163263 Pa
47101.5 ymin

Knowing the minimum yield strength of 163 MPa, any AWS electrode of
E60xx or above will work. Our base material is ASTM 1070 CD Steel, with

32
Sy=495[MPa]and Sut=640[MPa]. So that our weld material is at least as strong
as the base material, E100xx was chosen, with Sy=600[MPa] and
Sut=689[MPa]. This gives an actual safety factor for static weld failure of 7.35.

0.577 ( 600000 )
η= =7.35
47101.5
For this analysis, fully-reversed loading will be assumed, though the actual
cyclic loading on the component will likely be within a much smaller range
(within the analyzed range).
τ max=47101.5 Pa
τ min=−47101.5 Pa
τ max −τ min
τ a=
2
τ max + τ min
τ m=
2
τ a=47101.6 Pa ; τ m =0
We are assuming that the temperature does not get excessively high, as was
done in the last report, i.e., kd=1
b
k a=a Sut

Normally, welds are treated as as-forged surface finish, but with finishing
steps, they can be better surfaces. For this analysis, it is assumed that
finishing steps to smooth the surface have been taken and the surface can
be treated as a ground surface.
a=1.38

b=−0.067
With 90% reliability, ke=0.897. As the ultimate tensile strength is less than
1400[MPa], we can use the following.
'
Se =0.5 ⋅ Sut =413500 Pa
S se=k a k d k e S e =162387.307 Pa

With the endurance strength and with Kfs=2.7 for the end of a fillet weld, we
can use the Goodman equation.
S su=0.67 S ut =461630 Pa

( )
−1
K fs τ a K fs τ m
nf = + =1.276888205
S se Ssu

33
 Fig. 24. FEA analysis of welded sliders to mirror

FEA was completed using SolidWorks Simulation, where the mirror base was
fillet welded to the mirror backing using 1mm radius American Standard
edge welds. Under maximum load (2000N, purple arrows above), the
estimated weld size was calculated at 0.57mm, leading us to believe that the
1mm radius welds will be sufficient under maximum load.

34
Power Transmission
The power screw system is the main driving component, responsible for both
supporting the weight of the mirror and positioning it to any necessary angle
to match the sun’s position in the sky. To handle this task, the team selected
1018 carbon steel, 40mm diameter, 1m long lead screws. The ball joint
carriage to which the end of the lead screw is attached was treated as a
collar for the following analysis. Using values found in the previous
component design and analysis, the estimated maximum load on one lead
screw was found as follows:
Wind Load
Wwind = =800 N
3
Wmirror =185 N

Wtotal=985 N

The next important step in validating this power screw design is to verify self-
locking capability. It should be noted that this formula assumes static
conditions and may not be entirely valid under vibration. Because the system
is expected to experience regular and high-intensity load fluctuation due to
wind and other environmental factors, there will likely be non-insignificant
vibration affecting the system. However, under ideal conditions, the
relationship is as follows:

L∗cos ( α )
f≥ ⇒ 0.17 ≥ 0.0590∴ True
π∗d m

The calculated value is much lower than the required 𝑓 value, which
suggests that the power screw will remain self-locking, even under
considerable vibration.

Using the calculated torque to raise the motor, and targeting a load speed of
3 in/sec under maximum load, the power necessary to drive the screws is
calculated:
LoadSpd
AngSpd=
Lead
2 π ⋅ AngSpd ⋅T R
Praise = =0.824 hp
60 ⋅550 ⋅12
2 π ⋅ AngSpd ⋅ T L
Plower = =0.617 hp
60 ⋅550 ⋅12

35
With these values in mind, it seems clear that relatively low-power linear
actuators will be able to reliably provide the necessary torque to drive the
system to the level of the previously determined requirements. As stated
above, each lead screw operates independently of the others, and therefore
each will require its own motor. The added cost of using multiple motors can
be justified by the mechanical simplicity that this setup allows, which is
extremely important in the high-risk environments where heliostats are often
installed.

Finally, the power screw efficiency was calculated:

WL
e raise = =0.122
2 πT R

WL
e lower= =0.163
2π T L

The efficiency values were found to be just 12.2% while raising the load and
16.3% while lowering. These values are slightly lower than normal for lead
screws (the usual range is between 20-80%), and a different thread
geometry may increase efficiency. However, due to the low power
requirement, these numbers suggest that the current lead screw setup is
feasible with respect to the project.

The only power transmission setup for our design is the lead screw system,
in which 3 screw/linear actuator assemblies operate in conjunction to
effectively distribute the applied forces and allow for quick and accurate
positioning of our heliostat’s mirror. From the above calculations, despite
experiencing significant load, these motors will draw relatively little power
and should allow for sustainable installation and, further, efficient power
generation. The expected load forces used in these calculations are
comparable to winds in Category IV hurricanes, meaning that during
standard operation, all calculations were performed with loads far above the
expected normal operating range.

Overall Design Safety

After component and connection analysis, we have found that the heliostat
design discussed in this report has no static, cyclic, or connection failure
points. The critical static factor of safety for components is 10.04 for the
tripod base and for connections, the weld along the back of the mirror with a

36
safety factor of 7.35. In fatigue failure, all of the components and
connections showed infinite life potential, with the lowest fatigue factor of
1.277 at the weld, which still corresponds to theoretically infinite life, but
under the condition that the device does fail under a large number of cycles,
it will likely be at the weld. The bolts have the special factor of safety against
loosening, with the lowest being 8.2 for the bolts between the motor and the
motor mount component. Our FEA analyses found the lowest factor of safety
to be 5.61 for the tripod to concrete bolt. A table summarizing our design’s
safety is shown in Table 1.

Table 1. Factors of Safety for Design

Component Static Failure Fatigue Failure Buckling

FoS Fos Failure
(>1 = infinite FoS
life)
Tripod Base 10.04 7.365 265.1
Motor Mount 194 264 N/A
Lead Screw 11.062 6.265 274
Connection Static Failure Fatigue Failure Loosening FEA
FoS Fos Failure Analysis
(>1 = infinite FoS
life)
Motor Mount- 268 62.75 99 10
Tripod (Bolt)
Motor Mount- 24.7 44.72 8.2 6.99
Motor (Bolt)
Tripod- 305.5 325.2 12344.6 5.61
Concrete
(Bolt)
Mirror-Sliders 7.35 1.277 N/A Min.
(Weld) 1mm
weld
depth

The design initially targeted a minimum factor of safety of 1.5, which all
components and connections satisfy.

Cost Analysis
Creating a profitable design is important for a company to grow and remain
in business. The majority of the cost can be split into two categories:

37
material cost and manufacturing cost. While this does not account for scaled
production, assembly, and transportation costs, the bulk cost falls within
these two categories.

Table 2. Material and Component Costs

Component Qty. Unit Price Cost

Hexagonal Mirror 1 $360 $360
Hexagonal Backplate 1 $150 $150
Steel Track Assembly 1 $125.40 $125.40
M40 Lead Screw 3 $190 $570
Tripod Base 1 $453.20 $453.20
M40 Lead Screw Stepper Motors 3 $800 $2400
M5 Bolts 9 $0.50 $4.50
M14 Bolts 3 $3 $9
M10 Hex Cap Screws 3 $2.50 $7.50

The total material costs came to be $3939.60. This was calculated with the
current price of 304 stainless steel, large hexagonal mirrors, and each
additional component through the McMaster-Carr catalogue. The largest
price per unit as well as overall was for the motors, which was expected.

Manufacturing Costs

Labor Hour Price/Hour Cost

s
Welding 1 $60 $60
Milling 5 $70 $350
Bending 2 $100 $200

The total cost of all labor came to $610 per heliostat produced. These prices
were calculated by determining the difficulty and necessary quality of each
component and assigning the national average cost to each service.

Overall Costs

Combing the material, component, and manufacturing costs, the total cost of
the heliostat comes to $4,549.60 per unit. This price does not include
transportation and assembly as those vary depending on where the unit is
produced and where it will be used. Though this price is high for the industry
standard heliostat, our heliostat is different than the industry standard. When
a customer buys our heliostat, they are getting a robust system that can
withstand any environmental factors, especially those that are present within

38
desert environments. Our robust system will not blow away in a hurricane or
dust storm and will continue to work in strong winds. These reasons are why
we believe this heliostat is worth the price. Additionally, there is still
additional room within the design to decrease the cost, there are multiple
components with a factor of safety far above what is necessary, especially
considering the difficulty of conditions that we tested in. Because of this, it is
likely that the price can be significantly reduced in further iterations, by
removing excess material.

Conclusions
Our project is a large, tileable heliostat that is used to reflect the sun’s rays
into a tower to produce heat for electricity generation or heating homes. It
works by utilizing three lead screws to maneuver the mirror to reflect the sun
into the tower at all hours of the day. The base is a tripod made of 304 SS,
which is bolted into a cement base utilizing strong cap screws.

The largest challenge was finding a way to align the mirror, material
selection, and how to accurately determine the stress due to wind load on
the mirror assembly. As shown in our report, we decided to use lead screws,
as it was important that the mirror was able to move precisely, and since it
did not need to move fast (just tracking the sun), lead screws were able to
meet our criteria. Our heliostat will be used in arid, harsh environments,
most likely the desert, so material choice is extremely important. After
weighing the pros and cons of each main type of material (aluminum, cast
iron, steel, composites, etc.), we determined that 304 stainless steel would
meet the criteria of withstanding the tough environment, while keeping costs
relatively low. The toughest challenge in our analysis was understanding how
load was applied to the system. Since the center of gravity of our heliostat is
always close to center, we knew the largest stress would be wind drag. By
calculating the maximum angle of our heliostat, we were able to determine
this x-directional force.

The weakest component (component with the smallest factor of safety) in

our analysis is the tripod base with a static safety factor of 10.04, well above
the design requirement of 1.5. The weakest overall point is the weld
connecting the mirror backing to the slider frame, with a factor of safety of
7.35, still well within the 1.5 design requirement.

If we had more time on this project, we would have liked to have done
simulations of the heliostat movement and created a functioning prototype
of the system. This would have required electrical components as well as

39
control theory for each lead screw. A GUI to control the heliostat itself would
allow us to manually position the heliostat, as well as introduce real sun-
tracking capability with a lux sensor. Additionally, more work could have
been done to reduce the material use as we had such large factors of safety
in order to reduce production costs.

40
Appendices
Appendix A: Overall Assumptions and Limitations
Assumptions

The first major assumption made in our calculations was the determination of
a maximum wind load on the mirror. This was done in accordance with the
ASCE-7-14 standards for wind load calculation with respect to solar panels, a
comparable structure. The additional assumptions are explained in the report
above, but mostly consist of the acknowledgement of symmetry – analyzing
a single leg of a 3-leg structure, etc., as well as the use of fixed points in our
component analysis (fixing the end of the lead screw where it enters the
motor, etc.). The mock-up was and important part in understanding which
components we can simplify and which require more specific analysis. These
assumptions were critical in the successful completion of this project and its
relevant calculations.

41
 Fig. 25. Heliostat Final Mock-Up

Limitations

While the design explored in this report is very robust, there are some
limitations to its scope. First, the mirror itself is relatively small at 1 square
meter. Increasing the size of the mirror would likely improve system
efficiency. Additionally, several components are exposed to the harsh
environment that this system will likely exist in – the lead screws being a
major candidate for environmentally-driven failure (dust ingress, etc.). To
mitigate this risk, a flexible fabric cover would likely need to be installed
along the length of each lead screw, covering the motor assembly and the

42
ball joint/guide rail assembly. Aside from these main risks, the design is a
strong candidate for installation in CSP plants around the world.

Appendix B: Referenced Code/Calculations

Maple Code – Raising/Lowering Mirror with Power Screw

43
Maple Code - Tripod Static Failure Analysis

44
45
Maple Code – Lead Screw Static Failure Analysis

46
47
Maple Code – Connections

48
49
Questions
a) Our project is a large, tileable heliostat that is used to reflect the sun’s
rays into a tower to produce heat for electricity generation or heating
homes. It works by utilizing three lead screws to maneuver the mirror
to reflect the sun into the tower at all hours of the day. We designed
our heliostat to be modular, which allows for easy transportation and
setup without requiring specialized machinery. We also made the
design considerations to add robustness to all aspects of the heliostat,
which ensures that during high wind the mirror or other components do
not become projectiles.

b) While the heliostat will largely function in remote areas away from
public interaction, there are still some safety, environmental, and
economic risks to consider in our design. First, during transport and
installation, the technicians must use caution when moving and
assembling the large, relatively heavy steel components. Additionally,
the area around the installation will likely be disrupted geologically,
causing environmental impact that must be acknowledged and
planned for. The directed sunlight leaving the mirror must also be
accounted for – this concentration of energy, if misguided, could
possibly start wildfires or damage other components in the CSP array.
To mitigate the environmental impact, we chose to base our support
system on a tripod design, meaning the actual footprint of the heliostat
is relatively small, allowing for minimal ground disruption and allowing
small animals, which are prevalent in desert environments where many
heliostats are used, to travel freely through the array.

c) It was important for us to split up the project to our individual

strengths, as some of the group members are more comfortable with
certain aspects of the analysis than others. We began each analysis by
laying out constraints and brainstorming ideas. Afterwards we split up
parts, Isaiah and Daniel tackled most of the CAD and FEA, while Ethan,
Ash, and Julian approached the project with Maple and hand
calculations. All group members worked on each report writeup, Ethan
created the templates and overview for each project, and all other
group members filled in their individual components. Working together,
we created high-quality and accurate analyses of the component
design, connections, and power transmission.

50
 Our largest consideration during the project was during the first
analysis, component design analysis. We determined that if we were to
power one lead screw connected to the base of the mirror, it would
create a bending moment at the ball joint connection. This was unsafe
and could cause the lead screw to snap, a major safety hazard. To
eliminate this, we decided to add rails to the underside of the mirror,
eliminating this hazard and allowing the mirror and lead screw
assemblies to move freely.

d) If we had more time to work on the design, we would need to find ways
to eliminate excess weight, while maintaining a high factor of safety.
Based on our FEA and calculations above, we could likely reduce the
thickness of several steel components without significantly
compromising safety or functionality. Further analysis would be
necessary to make these cuts.

e) Most of the project was completed with material we learned in class,

especially all component design and stress analysis. One area where
we had to look elsewhere was for the maximum wind load on the
system. The American Society of Civil Engineers published guidelines
for designing structures to withstand this conglomeration of forces,
which we followed to determine the wind load used throughout our
calculations.

51