applied

sciences
Article
Design and Evaluation of Uniform LED Illumination
Based on Double Linear Fresnel Lenses
Hoang Vu 1 , Ngoc Minh Kieu 2,3 , Do Thi Gam 4 , Seoyong Shin 1, * , Tran Quoc Tien 2,3, * and
Ngoc Hai Vu 5, *
1 Department of Information and Communication Engineering, Myongji University, 116 Myongji-ro,
Cheoin-gu, Yongin, Gyeonggi-do 17058, Korea; vuhoangims@gmail.com
2 Institute of Materials Science, Vietnam Academy of Science and Technology, 18 Hoang Quoc Viet, Cau Giay,
Hanoi 1000, Vietnam; minhkn@ims.vast.ac.vn
3 Vietnam Academy of Science and Technology, Graduate University of Science and Technology,
18 Hoang Quoc Viet, Cau Giay, Hanoi 11307, Vietnam
4 Vietnam Academy of Science and Technology, Center for High Technology Development,
Hanoi 11307, Vietnam; honggam@htd.vast.vn
5 Faculty of Electrical and Electronics Engineering and Phenikaa Institute for Advanced Study,
Phenikaa University, Yen Nghia, Ha-Dong District, Hanoi 12116, Vietnam
* Correspondence: sshin@mju.ac.kr (S.S.); tientq@ims.vast.ac.vn (T.Q.T.);
hai.vungoc@phenikaa-uni.edu.vn (N.H.V.)




Received: 31 March 2020; Accepted: 4 May 2020; Published: 7 May 2020


Abstract: Redistribution of LED radiation in lighting is necessary in many applications. In this

article, we propose a new optical component design for LED lighting to achieve a higher performance.
The design consists of a commercial collimator and two linear Fresnel lenses. The LED radiation is
collimated by a collimator and redistributed by double linear Fresnel lenses to create a square-shaped,
uniform distribution. The linear Fresnel lenses design is based on Snell’s law and the “edge-ray
principle”. The optical devices are made from poly methyl methacrylate (PMMA) using a high-speed
computer numerical control (CNC) machine. The LED prototypes with complementary optics were
measured, and the optical intensity distribution was evaluated. The numerical results showed
we obtained a free-form lens that produced an illumination uniformity of 78% with an efficiency
of 77%. We used the developed LED light sources for field experiments in agricultural lighting.
The figures of these tests showed positive effects with control flowering criteria and advantages of
harvested products in comparison with the conventional LED sources. This allows our approach in
this paper to be considered as an alternative candidate for highly efficient and energy-saving LED
lighting applications.

Keywords: geometric optical design; illumination design; light-emitting diodes; non-imaging optics

1. Introduction
In recent years, LEDs have been used in many different fields. With the highlight of energy
savings, long life, and safety for users [1,2], LEDs are replacing other traditional lamps. However,
LEDs have disadvantages of a wide beam angle (from 120◦ to 150◦ ) depending on the type of LEDs [3]
and have Lambertian distributions. Illumination should be redistributed using an integrated primary
optical system or secondary optical system to control distribution shape and uniformity [4].
Utilizing free-form optics is a new trend to design secondary optical components for LEDs.
The advantages of free-form optics are their unique design, compact size, and precise irradiation
control [5]. In order to design a free-form lens, various methods have been proposed. For example,
the Miñano group developed the simultaneous multiple-surface method (SMS method) [6] to design

Appl. Sci. 2020, 10, 3257; doi:10.3390/app10093257 www.mdpi.com/journal/applsci

Appl. Sci. 2020, 10, 3257 2 of 13

Appl. Sci. 2020, 10, x FOR PEER REVIEW 2 of 13

a collimator for LEDs [7]. This method redistributes light from the source to the reflective surface
produce
to produce a parallel
a parallelbeam
beam with a beam
with a beam angle
angleof of
± 1.5°.
±1.5This method
◦ . This method also uses
also a wave-front
uses a wave-front theory
theoryof
geometrical optics [8] to design a Fresnel lens with combined functions
of geometrical optics [8] to design a Fresnel lens with combined functions of LED collimation and of LED collimation and
redistribution for
redistribution for uniform
uniform illumination.
illumination. Wang Wang et al. used
et al. used aa light
light energy
energy mapping
mapping methodmethod to to make
make aa
compact freeform
compact freeformlens lenscreate
createanan eveneven distribution
distribution in manyin many different
different shapesshapes [5]. In addition,
[5]. In addition, the optimal the
optimal numerical method can also be used to design free surfaces for LED
numerical method can also be used to design free surfaces for LED arrays, such as in Yu et al. [9]. All of arrays, such as in Yu et
al. [9].
the aboveAll methods
of the above
workmethods
very well work very well the
in controlling in controlling
distributionthe distribution
shape shape and
and producing producing
a uniform light
a uniform light distribution, but they are not flexible in different types of
distribution, but they are not flexible in different types of LEDs. Equations for lens surface calculationsLEDs. Equations for lens
surface calculations have parameters dependent on the light source.
have parameters dependent on the light source. Therefore, the lens structure must be recalculated Therefore, the lens structure
must be
when recalculated
applied whentypes
to different applied to different types of LEDs.
of LEDs.
In this paper, we present a method
In this paper, we present a method to fabricate to fabricate an optical
an optical component
component for a single
for a single LED, which
LED, which can be
can be applied to many types of LEDs with different beam angles to
applied to many types of LEDs with different beam angles to make a uniform distribution. The single make a uniform distribution.
The single
LED can be LED can be aLED
a high-power high-power LED or LED
or chip-on-board chip-on-board
(COB LED). LEDThe(COB LED). collimator
commercial The commercial lens is
collimator lens is placed on top of the single LED to collect and collimate
placed on top of the single LED to collect and collimate the light output from the LED. The collimated the light output from the
LED. is
beam The collimated by
redistributed beam
two is redistributed
linear Fresnel lensesby two[10] linear
to create Fresnel lenses
a uniform [10] to create
distribution on thea uniform
receiver,
distribution on the receiver, as shown in Figure 1. The surface of
as shown in Figure 1. The surface of the linear Fresnel lens is designed by using the edge-raythe linear Fresnel lens is designed
by using the
principle [11]edge-ray
and Snell’sprinciple
law. The [11] and Snell’s
parameters oflaw. The parameters
the linear Fresnel lens ofare
thenot
linear Fresnel lens
dependent on the areLED
not
dependent
chips and theon collimators,
the LED chips and poly
where the collimators, where poly(PMMA)
methyl methacrylate methyl methacrylate
material is used (PMMA)
to producematerial
the
is used to produce the lens. The results include the simulation of uniformity
lens. The results include the simulation of uniformity and efficiency of the system, comparison of and efficiency of the
system, comparison
simulations, of simulations,
and experimental and experimental
results. Finally, an experimentresults. Finally, anwas
on chrysanthemum experiment
conductedon to
chrysanthemum was conducted to control flowering. Other experiments
control flowering. Other experiments on flowering control do not present how the uniformity of the on flowering control do not
present
light how has
source the uniformity
an effect onofplant the light source
growth, andhasmostan effect
studies onhave
plantfocused
growth,on andthemost studies have
wavelength and
focusedofon
power thesources
light wavelength and power
[12]. Therefore, of study
this light sources
will be a[12]. usefulTherefore,
referencethis study will
to achieve bequality
better a usefulof
reference to achieve better quality of agricultural
agricultural products by improving lighting uniformity. products by improving lighting uniformity.

Figure 1. The principle of our proposed uniform LED illumination system.

Figure 1. The principle of our proposed uniform LED illumination system.
2. Module Design and Simulation
2. Module Design and Simulation
2.1. Module Design
2.1. Module Design
In this section we present the theory of designing a uniform illumination system for single LEDs.
We propose a novelwe
In this section uniform
presentillumination
the theory ofsystem consisting
designing of three
a uniform parts: LED,
illumination collimator,
system and
for single two
LEDs.
perpendicular
We propose a novellinearuniform
Fresnel illumination
lenses, as shown as consisting
system in Figure 2.ofThe
threecollimator
parts: LED,was used to collimate
collimator, and two
all light from the
perpendicular LEDFresnel
linear source.lenses,
The collimated
as shown beam from the
as in Figure collimator
2. The lenswas
collimator wasused
redistributed using
to collimate all
double linear
light from theFresnel lenses to
LED source. The provide uniform
collimated beamlight distribution
from over the
the collimator lensgiven
was target.
redistributed using
double linear Fresnel lenses to provide uniform light distribution over the given target.

Appl. Sci. 2020, 10, x; doi: FOR PEER REVIEW www.mdpi.com/journal/applsci

Appl. Sci. 2020, 10, 3257 3 of 13
Appl. Sci.
Appl. Sci. 2020,
2020, 10,
10, xx FOR
FOR PEER
PEER REVIEW
REVIEW 33 of
of 13
13

Figure 2.
Figure 2. Structure
2. Structureofof
Structure ofLED
LED
LED illumination
illumination
illumination system
system composed
composed
system ofLED
of an
composed of an LED
an LED chip,
chip,chip, collimator,
collimator, and double
and double
collimator, and double
linear
linear Fresnel
Fresnel
linear Fresnel lenses.
lenses.lenses.

Commercial
Commercial collimators
Commercial collimators are
collimators are cone-like
are cone-like lenses.
cone-like We
We chose
lenses. We
lenses. chose aaa collimator
chose collimator from
collimator from Led-link
from Led-link Optics
Led-link Optics Inc.
Optics Inc. [8]
Inc. [8]
[8]
for
for our
for our experiment.
our The
experiment. The
experiment. image
The image
image ofof the
the commercial
of the collimator
commercial collimator
commercial lens
lens is
collimator lens is shown
is shown in
shown in Figure
in Figure 3.
Figure 3. Collimators
3. Collimators
Collimators
are often
are often
are designed
often designed for
designed for point
for point sources
point sources
sources andand
and aaa fixed
fixed
fixed wavelength
wavelength to
wavelength to create
to create ideal
create ideal collimated
collimated beams,
ideal collimated while
beams, while
beams, while
LED chips
LED chips
LED are
chips are usually
are usually 1 × 1 mm
usually 11 ××1 mm
mm oror larger.
or larger. Therefore,
larger. Therefore, the output beams from the collimator are
Therefore, the output beams from the collimator are slightly slightly
divergent. Details
divergent. Details
divergent. will
Details will be
will be presented
be presented
presented in in the
in the simulation
the simulation part.
simulation part.
part.

Figure 3. Illustration of collimator using in this research: (a) Principles of Total Internal Reflection (TIR)
Figure
Figure
lens; 3.AIllustration
(b)3. Illustration
commercialofofcollimator
collimator
collimator using in
using
lens. in this
this research:
research: (a)
(a) Principles
Principles of
of Total
Total Internal
Internal Reflection
Reflection
(TIR) lens;
(TIR) lens; (b)
(b) A
A commercial
commercial collimator
collimator lens.
lens.
The collimated beam needs to be distributed to create uniform illumination. We used the double
linearThe
The collimated
collimated
Fresnel lensesbeambeam needs to
needs to be
as a redistribution be distributed
distributed
module, to to create
the create
design uniform
uniform
method illumination.
illumination.
as shown inWe We used4.
used
Figure the
the double
Adouble
single
linear Fresnel
linear Fresnellens
Fresnel lenses
lenses canasas aa redistribution
redistribution
redistribute module,beam
module,
the collimated the design
the design method
in onemethoddirection asin
as shown
shown in Figure
in
the receiver Figure 4.as
area,4. Ashown
A single
single
linear
linear
in Fresnel
Fresnel
Figure lens can
lens
4a. Although can redistribute
redistribute
the incident the the collimated
beam collimated
does not beambeamain
have in one direction
one
uniform direction in the
in
distribution, the the
receiver
receiver
Fresnel area,
area,lens asconsists
as shown
shown
in many
in
of Figuresmall
Figure 4a. Although
4a. Although
grooves, the the incident
where incident beam does
beam
each groove does not have
not
is similar have aa uniform
to a linear uniform
convex distribution,
distribution,
lens. Whenthe the Fresnel
theFresnel
collimated lens
lens
consists
consists
beam of many
of
passes many
through small
small grooves,
thegrooves, where lens,
where
linear Fresnel each these
each groove
groove is similar
is
grooves similar
will toto
splitaa linear
linear convexbeam
convex
the incident lens. (a
lens. When
Whendivided the
the
collimated
collimated
beam beam
can bebeam passesto
passes
considered through
through
have a thethe linear
lineardistribution)
uniform Fresnel lens,
Fresnel lens,andthese
these grooves
grooves
spread it over will
willthe split
split the incident
the
receiver incident beam (a
beam
in one direction. (a
divided
divided
When twobeam
beam
linear can
can be considered
be considered
Fresnel lenses areto to have a uniform
have a uniform
perpendicular to eachdistribution)
distribution)
other, theyand and spread
willspread
direct the it over
it over the receiver
the receiver
collimated beam in in
to
one direction.
one
spread direction.
across the When
When two linear
two
receiver linear
in Fresnel
twoFresnel
directions. lenses are
lenses are perpendicular
perpendicular to to each
each other,
other, theythey willwill direct
direct thethe
collimated
collimated
We appliedbeamthe
beam to spread
to spread
“edge-ray across
across the receiver
the
principle”receiver in two
in two path
and optical directions.
directions.
length conservation to calculate the curves
We applied
We
of Fresnel applied the “edge-ray
the
lens grooves. “edge-ray principle”
principle”
The calculation and
isand
shownoptical
optical path length
path
in Figure length conservation
4b andconservation
consists of three to calculate
to calculate
steps. the the curves
curves
of Fresnel
of Fresnel
Step 1: lens
lensSome grooves.
grooves. The
initialThe calculation
calculation
conditions is shown
is
of light shown
source in Figure
in Figure
(collimated 4b and
4b and consists
consists
light beam) ofsuch
of threeas
three steps.
steps.
left edge AL BL ,
Step
rightStep
edge1: 1:ASome
Some
R initial
BR ofinitial conditions
conditions
the segment, of light
of
position light source (collimated
source
of Fresnel (collimated
lens, and receiver light beam)
light beam) such as
such
size (illumination as left edge 𝐴𝐴
leftspace:
edge 𝐿𝐿𝐵
RR 𝐵𝐿𝐿‘),,
right
right
are edge 𝐴𝐴𝑅𝑅𝐵𝐵𝑅𝑅 of
edge
selected. of the
the segment,
segment, position
position of of Fresnel
Fresnel lens,
lens, andand receiver
receiver size size (illumination
(illumination space: space: RR RR ‘)‘)
are selected.
are selected.
Step 2: The ray to the left edge of AL BL passes through a linear Fresnel lens groove and refracts to
Stepside
Step
the right 2: The
2: The rayreceiver
ray
of the to the
to the left
left edgeRof
edge
(point of 𝐴𝐿𝐿𝐵𝐵𝐿𝐿ray
‘). 𝐴The passes
passes through
to thethrough
right edge aa linear
linear Fresnel
of ARFresnel
BR passes lens
lens grooveaand
groove
through and
groove refracts
refracts
and
to the
to the right
refracts right side
side
to the ofside
of
left the receiver
the receiver (point R
(point
of the receiver R(point
‘). The
‘). TheR).rayThese
ray to the
to thetwo right
right edge
edge
rays of 𝐴𝐴𝑅𝑅𝐵𝐵𝑅at
of
intersect passes
𝑅 passes
the focal through
through aa groove
groove
line. Similarly,
and refracts
and refracts to to the
the left
left side
sideof of the
the receiver
receiver (point
(point R).R). These
These two two raysrays intersect
intersect at at the
the focal
focal line.
line. Similarly,
Similarly,

Appl. Sci.
Appl. Sci. 2020,
2020, 10,
10, x;
x; doi:
doi: FOR
FOR PEER
PEER REVIEW
REVIEW www.mdpi.com/journal/applsci
www.mdpi.com/journal/applsci
Appl. Sci.
Appl. Sci. 2020,
2020, 10,
10, 3257
x FOR PEER REVIEW 44of
of13
13
Appl. Sci. 2020, 10, x FOR PEER REVIEW 4 of 13

the 𝐴𝑛 𝐵𝑛 rays between the 𝐴𝐿 𝐵𝐿 and 𝐴𝑅 𝐵𝑅 rays must also intersect at the focal line. Based on the law
the A
the 𝐴𝑛B𝐵𝑛 rays theA𝐴𝐿B𝐵𝐿and
rays betweenthe and A𝐴𝑅B𝐵𝑅 rays
raysmust
mustalso
alsointersect
intersectat
atthe
thefocal
focal line.
line. Based
Based on
on the
the law
n n pathbetween
of optical length (OPL) Lconservation,
L R R we have Equation (1): law
of optical path length (OPL) conservation, we have Equation
of optical path length (OPL) conservation, we have Equation (1): (1):

𝑛1 |𝐴|A𝐿𝐿L𝐵𝐵 + 𝑛
B𝐿L||| ++ 0 |𝐵 F𝐹|
|BL𝐿𝐿𝐹||== 𝑛1 |𝐴 𝐵𝑅 || +
|A𝑅R𝑅𝐵 + n𝑛0 |B
|𝐵𝑅F𝐹|| == n𝑛11|𝐴
|𝐴𝑛B𝐵𝑛||++ n𝑛0|B|𝐵𝑛F|𝐹| (1)
𝑛n11|𝐴 𝐿 𝑛n00|𝐵 = 𝑛n11|𝐴 B
𝑅R| + 𝑛 00|𝐵R𝑅 𝐹| = 𝑛1
|A 𝑛 𝐵𝑛 | + 𝑛00 |𝐵𝑛 𝐹|
n n n (1)
(1)
where 𝑛0 , 𝑛1 are refractive indices of air and materials of Fresnel lens, respectively. Based on the
where n𝑛00, ,n1𝑛1are
where refractive
are refractiveindices of air
indices of and materials
air and of Fresnel
materials lens, lens,
of Fresnel respectively. BasedBased
respectively. on theon
above
the
above equation,
equation, all coordinates
all coordinates of the
of the segment segment surface
surfacesurface curves
curves curves are calculated.
are calculated.
above equation, all coordinates of the segment are calculated.
Step3:3:Repeat
Step Repeatthethe calculation
calculation process,
process, and and
all allother
the the other
groovesgrooves
of linear of linear lens
Fresnel lens are
Step 3: Repeat the calculation process, and all the other grooves ofFresnel are obtained
linear Fresnel lens are
obtained
by the same by procedure.
the same procedure.
obtained by the same procedure.

Figure 4. Design method of linear Fresnel lens (a) A single linear Fresnel lens redistributes light to a
Figure 4. Design method of linear Fresnel lens (a) A single linear Fresnel lens redistributes light to a
receiver. (b) Design for a segment of a linear Fresnel lens. (c) The lights were spread in two dimensions
receiver. (b) Design for a segment of a linear Fresnel lens. (c) The
receiver. The lights
lights were
were spread
spread in two dimensions
dimensions
by double Fresnel lenses.
by double Fresnel lenses.
Because a linear
Because linear Fresnel
Fresnel lenslens distributes
distributes lightlight in in only
only oneone direction,
direction, we we placed
placed two two linear
linear Fresnel
Fresnel
Because aa linear Fresnel lens distributes light in only one direction, we placed two linear Fresnel
lenses
lenses perpendicularly to propagate light in two directions over the lighting area, as shown in Figure
lenses perpendicularly
perpendicularlyto topropagate
propagatelight lightinintwotwodirections
directionsover over the
thelighting
lighting area, as as
area, shown
shown in Figure
in Figure 4c.
4c.
So, So, a two-dimensional
a two-dimensional redistribution
redistribution will
willwill
create create a
a square squareshapeshape
on the on the receiver.
receiver.
4c. So, a two-dimensional redistribution create a square shape on the receiver.
The focal
The focal line is important in this linear Fresnel lens. The LED light was focused on the focalfocal
line
The focal line is important in this linear Fresnel lens. The LED light waswas
line is important in this linear Fresnel lens. The LED light focused
focused on theon focal
the line
and
line was spread
and spread
was spread across the
across receiver RR’.
the receiver Figure 5
RR’. Figure shows the
5 shows design process
the design of the novel lens in Matlab.
and was across the receiver RR’. Figure 5 shows the design processprocess of thelens
of the novel novel lens in
in Matlab.
Some initial
Matlab. Some conditions
initial of the light
conditions of the source
light such as
source suchFresnel
as lens radius
Fresnel lens (r), number
radius (r), numberof grooves
of grooves (a),
Some initial conditions of the light source such as Fresnel lens radius (r), number of grooves (a),
distance
(a), from
distance the light source to the receiver (d), refractive index (n), receiver size (RR’), and so forth
distance fromfrom the light
the light source source
to the toreceiver
the receiver (d), refractive
(d), refractive indexindex (n), receiver
(n), receiver size (RR’),
size (RR’), and so and
forthso
were
forth selected. Based on these initial conditions, the coordinates of the Fresnel lens surface in one
were were selected.
selected. BasedBased on these
on these initial
initial conditions,
conditions, thethe coordinates
coordinates of of
thetheFresnel
Fresnellens
lenssurface
surface in in one
one
directionwas
direction was designed.
designed.
direction was designed.

Figure 5. The flowchart of the design process of the linear Fresnel lens.

We designed aFigure
linear5.Fresnel lensesofsystem
The flowchart for aprocess
the design 5W power
of the COB
linear LED, whose
Fresnel lens. wavelength was
Figure 5. The
630 nm and had a collimator flowchart
lens of the
diameter design
of 100 mm. process of the linear
The lighting Fresnel
distance waslens.
2 m, and the lighting
area was 2 m × 2 m. The number of grooves of a linear Fresnel lens affects uniformity and efficiency [10].
Appl. Sci. 2020, 10, x; doi: FOR PEER REVIEW www.mdpi.com/journal/applsci
Appl. Sci. 2020, 10, x; doi: FOR PEER REVIEW www.mdpi.com/journal/applsci
Appl. Sci. 2020, 10, x FOR PEER REVIEW 5 of 13

We designed a linear Fresnel lenses system for a 5W power COB LED, whose wavelength was
630
Appl.nm
Sci. and
2020,had a collimator lens diameter of 100 mm. The lighting distance was 2 m, and the lighting
10, 3257 5 of 13
area was 2 m × 2 m. The number of grooves of a linear Fresnel lens affects uniformity and efficiency
[10]. Based on the previous results in [10], we selected a linear Fresnel lens with a groove number of
Based
20. Tableon 1the previous
shows resultssystem
the optical in [10],specifications.
we selected a linear Fresnel lens with a groove number of 20.
Table 1 shows the optical system specifications.
Table 1. The double Fresnel lenses design parameters.
Table 1. The double Fresnel lenses design parameters.
Design Parameters
Design
Type of LEDParameters COB
Type of collimator
Type of LED TIR
COB
Refractive index of the
Type of collimator Fresnel lens 1.49
TIR
Thickness
Refractive of the
index Fresnel
of the lens
Fresnel lens 5 mm
1.49
Thickness
Number of theofFresnel
of grooves lens lens
the Fresnel 520
mm
Number of grooves
Diameter ofFresnel
of the the Fresnel
lens lens 100 20
mm
Diameter of the Fresnel lens 100 mm
2.2. Simulation: Illumination Performance and Tolerance Analyses
2.2. Simulation: TMIllumination Performance and Tolerance Analyses
LightTools optical simulation software was used to simulate the structure of the proposed
TM optical simulation software was used to simulate the structure of the proposed LED
LED LightTools
lighting system. The structure of the simulation system is shown in Figure 6. A square lighting
lighting system. The structure
shape was chosen as an example of the simulation the
to investigate system is shown of
performance in Figure 6. A square
the designed lighting
lighting shape
system. As
was chosenabove,
described as an example
the LEDtolighting
investigate the performance
system consisted of of the designed
a single LED, alighting system.
commercial As described
collimator lens,
above,
and theofLED
a set lighting
double linearsystem
Fresnelconsisted
lenses to of a single LED,
redistribute lightaon
commercial collimator
the receiver. A 5W, 630lens,
nmand
COBa set
LEDof
double linear Fresnel lenses to redistribute light on the receiver. A 5W, 630 nm COB
was used in simulations with a luminous flux of 200 lm. Fresnel lenses and the collimator lens were LED was used in
simulations
made of polywith a luminous
methyl flux of 200
methacrylate lm. Fresnel
(PMMA) lenses
material. and the
Double collimator
Fresnel lenseslens were made
produced of poly
a uniform
methyl methacrylate
distribution (PMMA)
of illumination on amaterial.
2 m × 2 mDouble
receiver. Fresnel lenses was
The receiver produced
locateda 2uniform
m awaydistribution of
from the light
illumination on a 2 m × 2 m receiver. The receiver was located 2 m away from the light source.
source.

Figure 6. The simulation parameters using LighttoolsTM software.

Figure 6. The simulation parameters using LighttoolsTM software.
The simulation results of ray tracing of the collimator lens are shown in Figure 7. Figure 7a shows
The simulation
the illumination results of in
distribution raythe
tracing of the
receiver. collimator
With lensdistance
a lighting are shownof 2inm,
Figure 7. Figure 7a shows
the distribution on the
the illumination
receiver area has distribution
a radius of 15incm.
the Figure
receiver.
7b With
shows a the
lighting distance
intensity of 2 m, the
distribution distribution
of the on the
far-field receiver.
receiver
It showsarea
thathas
theafull
radius of 15
width at cm. Figure 7b shows
half-maximum the intensity
intensity angle was 1.2◦ . The of
distribution theflux
total far-field receiver.
was 198.4 lm.
It shows
Figure 7cthat the fulldisplays
intuitively width atthehalf-maximum intensitythe
rays passing through angle was 1.2°.
collimator The total flux was 198.4 lm.
lens.
Figure 7c intuitively displays the rays passing through the collimator lens.

Appl. Sci. 2020, 10, x; doi: FOR PEER REVIEW www.mdpi.com/journal/applsci

Appl. Sci. 2020, 10, x FOR PEER REVIEW 6 of 13

Appl. Sci. 2020, 10, 3257 6 of 13

Appl. Sci. 2020, 10, x FOR PEER REVIEW 6 of 13

Figure 7. Simulation performance of collimator lens: (a) light distribution of output beam; (b) viewing
angle of7.LED
Figure with collimator
Simulation at 1.2°;
performance and (c) raylens:
of collimator tracing of LED
(a) light with collimator.
distribution of output beam; (b) viewing
Figure
angle of7.LED
Simulation performance
with collimator ◦ ; and
at 1.2of collimator
(c) raylens: (a) of
tracing light
LEDdistribution of output beam; (b) viewing
with collimator.
Simulation
angle of LED with collimator at 1.2°; distribution
results of the optical of theof COB
and (c) ray tracing LEDcollimator.
LED with with the collimator and double
Fresnel lenses are
Simulation shown
results ofin Figure
the optical8. It shows the of
distribution illumination
the COB LED withwith
a sizetheofcollimator
2 × 2 m square on the
and double
targetSimulation
Fresnel surface.
lenses The distribution
areresults
shown ofinthe formed
Figure
optical ashows
squaretheradiation
8. Itdistribution of the pattern
illumination
COB LED withwith
with aasize
size ofcollimator
of
the 21.8
×2 m×
m1.8 m at double
square
and height
on the
of 2 m,
target which
surface. was
The in agreement
distribution with
formed the a expected
square shape.
radiation
Fresnel lenses are shown in Figure 8. It shows the illumination The uniformity
pattern with a of
size illumination
of 1.8 m× on
1.8 m the
at target
height
size of 2 × 2 m square on the
surface
of
target was 78%
2 m, surface.
which was
Thewith a minimum
indistribution
agreement intensity
with
formed of 28radiation
theaexpected
square lux and The
shape. maximum intensity
uniformity
pattern with is 48
1.8lux.
of illumination
a size of The
onmefficiency
m× 1.8 the target
at height
was
of 2 77%.
surface Illumination
was
m, which 78%
was with auniformity
minimum
in agreement is calculated
intensity
with of 28bylux
the expected Equation The(2):
and maximum
shape. intensity
uniformity is 48 lux. The
of illumination on efficiency
the target
was 77%.was
surface Illumination
78% with auniformity
minimum is calculated
intensity bylux
of 28 Equation (2):
and maximum intensity is 48 lux. The efficiency
was 77%. Illumination uniformity is calculated by 𝐼Equation (2):
= Imin
𝑚𝑖𝑛
u𝑢= (2)
(2)
I𝐼avg
𝑎𝑣𝑔
𝐼𝑚𝑖𝑛
where I is the minimum intensity value and 𝑢
I = is the averaged intensity value of all the pixels.(2)
where 𝐼min avg 𝐼𝑎𝑣𝑔
𝑚𝑖𝑛 is the minimum intensity value and 𝐼𝑎𝑣𝑔 is the averaged intensity value of all the pixels.

where 𝐼𝑚𝑖𝑛 is the minimum intensity value and 𝐼𝑎𝑣𝑔 is the averaged intensity value of all the pixels.

Figure 8.
Figure Light distribution
8. Light distribution on
on an
an LED
LED receiver
receiver using
using aa collimator
collimator and
and double
double linear
linear Fresnel
Fresnel lenses.
lenses.

To
To evaluate
evaluate the losses
losses in
in the
thedistribution the system, wewe inserted two optical surface receivers in
in the simulation
Figure 8. Light on system, inserted
an LED receiver usingtwo optical and
a collimator surface
doublereceivers the lenses.
linear Fresnel simulation
process,
process, as shown in Figure 9. The system losses were from the loss of the collimator and the
as shown in Figure 9. The system losses were from the loss of the collimator and the loss
loss of
of
the double
the double Fresnel
Fresnelthe
To evaluate lenses.
lenses.
lossesThe
The loss
in loss of the collimator
of the collimator
the system, we insertedwas
wastwofrom
from the material
the material
optical absorption
surface absorption and
receivers inand the
thethe Fresnel
Fresnel
simulation
losses.
losses.
process, The
The loss
asloss
shownof the
of the collimator
in collimator
Figure 9. Thewassystem
was 1.56 Lux,
1.56 Lux, as show
as
lossesshow in
werein Figure
Figure
from 7b.
the7b. The
lossThe collimator
of collimator
the losses
losses
collimator arethe
are
and denoted
denoted
loss of
by
by L
𝐿 1 .
. Using
Using LightTools™
LightTools™ software,
software, L 𝐿
1 can
canbebecalculated
calculated by
byEquation
Equation (3):
(3):
the double Fresnel lenses. The loss of the collimator was from the material absorption and the Fresnel
1 1
losses. The loss of the collimator was
Flux1.56 Lux, source
o f LED as show in Figure
− Flux 7b. The1 collimator losses are denoted
on Receiver
L1 =𝐹𝑙𝑢𝑥 𝑜𝑓
by 𝐿1 . Using LightTools™ software, 𝐿1 𝐿𝐸𝐷
canFlux
be calculated
𝑠𝑜𝑢𝑟𝑐𝑒 − 𝐹𝑙𝑢𝑥by Equation
𝑜𝑛 𝑅𝑒𝑐𝑒𝑖𝑣𝑒𝑟(3):
1 (3)
𝐿1 = o f LED source (3)
𝐹𝑙𝑢𝑥 𝑜𝑓 𝐿𝐸𝐷 𝑠𝑜𝑢𝑟𝑐𝑒
𝐹𝑙𝑢𝑥 𝑜𝑓 𝐿𝐸𝐷 𝑠𝑜𝑢𝑟𝑐𝑒 − 𝐹𝑙𝑢𝑥 𝑜𝑛 𝑅𝑒𝑐𝑒𝑖𝑣𝑒𝑟 1
𝐿1 = (3)
𝐹𝑙𝑢𝑥 𝑜𝑓 𝐿𝐸𝐷 𝑠𝑜𝑢𝑟𝑐𝑒
Appl. Sci. 2020, 10, x; doi: FOR PEER REVIEW www.mdpi.com/journal/applsci

Appl. Sci. 2020, 10, x; doi: FOR PEER REVIEW www.mdpi.com/journal/applsci

LighTools™ software, 𝐿2 can be calculated by Equation (4):

𝐹𝑙𝑢𝑥 𝑜𝑛 𝑅𝑒𝑐𝑒𝑖𝑣𝑒𝑟 1 − 𝐹𝑙𝑢𝑥 𝑜𝑛 𝑅𝑒𝑐𝑒𝑖𝑣𝑒𝑟 2

𝐿2 = (4)
𝐹𝑙𝑢𝑥 𝑜𝑛 𝑅𝑒𝑐𝑒𝑖𝑣𝑒𝑟 1
Appl. Sci. 2020, 10, 3257 7 of 13

Figure 9. Illustration of the simulation structure for loss analysis and ray-tracing analysis.
Figure 9. Illustration of the simulation structure for loss analysis and ray-tracing analysis.
The loss of double Fresnel lenses was from the material absorption, the Fresnel losses, and the
By calculating
geometrical loss of Equations
the Fresnel(3) andGeometrical
lens. (4), respectively, wethe
loss of found L1 =lens
Fresnel 0.7%occurred
and L2 =when
22.3%. The total
a non-ideal
loss is the sum
collimated beam 𝐿1 and
of was 𝐿2 and
applied can be
to this calculated
design. by Equation
Although (5):
the theoretical design was optimized for an
ideal parallel beam, this design still worked with a highly collimated beam, with a slight reduction
in optical efficiency and uniformity. 𝐹𝑙𝑢𝑥 In𝑜𝑓
this research,
𝐿𝐸𝐷 𝑠𝑜𝑢𝑟𝑐𝑒 the selected
− 𝐹𝑙𝑢𝑥 collimator
𝑜𝑛 𝑅𝑒𝑐𝑒𝑖𝑣𝑒𝑟 2 provided a beam with a
divergence of 1.2◦ , so there 𝐿total
were=some staggered could not reach the receiver. The loss of(5)
𝐹𝑙𝑢𝑥rays that𝑠𝑜𝑢𝑟𝑐𝑒
𝑜𝑓 𝐿𝐸𝐷 the
double Fresnel lenses was 44.26 Lux. The double Fresnel lenses are denoted by L2 . Using LighTools™
With Lthe
software, 2 can be calculated
parameters given byabove,
Equation the(4):
efficiency was found to be 77%, including the material
absorption, the Fresnel losses, and the geometrical loss of the Fresnel lens.
Tolerance is an important Fluxfor
issue on Receiver
free-form Flux onInReceiver
1 −lenses. 2
this study, the effects of the angle
L2 = (4)
Flux on Receiver 1
between the collimator lens and double linear Fresnel lens were analyzed in terms of uniformity and
efficiency as shown Equations
By calculating in Figure 10.
(3) Figure
and (4),10a–c shows the
respectively, weeffects
found of = 0.7%
L1the line and L2 = 22.3%.
distributions along
Thethe X
total
axis
loss and
is theYsum
axis of across
L1 and theL2center
and canof be
thecalculated
illuminated by area for dT
Equation angles of 2°, 3°, and 4°. Figure 10d
(5):
shows the results of simulations of the efficiency and the uniformity as a function of angle tolerance.
When the deflection angle dT of the Flux o f LED source − Flux on Receiver 2
Ltotal = elements increased from 0° to 4°, the efficiency and the uniformity (5)
were reduced accordingly. Because the change Flux o f LED
in angle source optical devices changes the incident
between
angle of the collimated light beam to the linear Fresnel lens, the distribution will be skewed to the
With the parameters given above, the efficiency was found to be 77%, including the material
corresponding angle, reducing uniformity. The permissible deviation of dT is 1.8° as long as the
absorption, the Fresnel losses, and the geometrical loss of the Fresnel lens.
uniformity and the efficiency are kept at 90%.
Tolerance is an important issue for free-form lenses. In this study, the effects of the angle between
the collimator lens and double linear Fresnel lens were analyzed in terms of uniformity and efficiency
as shown in Figure 10. Figure 10a–c shows the effects of the line distributions along the X axis and Y
axis across the center of the illuminated area for dT angles of 2◦ , 3◦ , and 4◦ . Figure 10d shows the results
of simulations of the efficiency and the uniformity as a function of angle tolerance. When the deflection
angle dT of the elements increased from 0◦ to 4◦ , the efficiency and the uniformity were reduced
accordingly. Because the change in angle between optical devices changes the incident angle of the
collimated light beam to the linear Fresnel lens, the distribution will be skewed to the corresponding
Appl. Sci. 2020, 10, x; doi: FOR PEER REVIEW www.mdpi.com/journal/applsci
angle, reducing uniformity. The permissible deviation of dT is 1.8◦ as long as the uniformity and the
efficiency are kept at 90%.
Appl. Sci. 2020, 10, x FOR PEER REVIEW 8 of 13
Appl. Sci. 2020, 10, 3257 8 of 13
Appl. Sci. 2020, 10, x FOR PEER REVIEW 8 of 13

Figure 10. Light distribution on a receiver with (a) dT = 2°; (b) dT=3°; and (c) dT=4°. (d) Efficiency and
Figure 10. Light distribution on a receiver with (a) dT = 2◦ ; (b) dT=3◦ ; and (c) dT=4◦ . (d) Efficiency and
uniformity depend
Figure 10. depend
Light on deflection
distribution angle
on aangle dT.
receiver
uniformity on deflection dT. with (a) dT = 2°; (b) dT=3°; and (c) dT=4°. (d) Efficiency and
uniformity depend on deflection angle dT.
In addition,
In addition,another
another tolerance
tolerance waswas
alsoalso considered
considered in thisinstudy,
this which
study,iswhich is the
the angle of the angle of the
collimated
collimated
beam.InThe beam.
addition,
simulation The simulation
another
resultstoleranceresults of the
was alsoand
of the efficiency efficiency
considered and the uniformity
in thisasstudy,
the uniformity as a function
whichofisbeam
a function of
the angle of beam
of the
the
angle of the
collimatedbeam
collimated collimated
beam.dθ The beam dθ of 2°,
2◦ , 4◦ , and results
of simulation 4°, and
6◦ are shown 6° are shown
of the efficiency in Figure
and theWhen
in Figure 11a–c. 11a–c.
uniformity When
the beam the
as a angle
functionbeam angle
wasofwider,
beam
was
angle
the wider,
of the
square the square shape
collimated
illumination illumination
beamchanged shape
dθ of 2°, changed
4°, and
more, and6° more,
shownand
areuniformity
the the efficiency
inand
Figureuniformity
11a–c. alsoand
When efficiency
the
decreased. beamThis also
angle
is
decreased.
was wider,
explained This
by the is explained
squareangle
the beam by the
illumination beam
related toshape angle related
changed
incident angle of to
more, incident
and to
the light the angle of the
theuniformity light
Fresnel lens. andWhento the
efficiencyFresnel
the beamalso
lens.
angle When
decreased. the
This
is wider, beam
the anglelight
isoutput
explainedis wider,
by thethe
does output
beam
not light
angle
focus on does
focalnot
related
the to focus
so iton
incident
line, the
is anglefocal line,
of the
impossible so
tolight it to
controlis impossible
the
theFresnel
shape
to
of control
lens. the
distribution shape
When theinbeam of distribution
angle is wider,
two directions. in
The the two directions.
output beam
maximum light does The
anglenot maximum
was ◦
focus beam
on the
3 , and 90% angle
focal was
line, soand
efficiency 3°, and 90%
it isuniformity
impossible
efficiency
to control
was and
obtained.the uniformity was obtained.
shape of distribution in two directions. The maximum beam angle was 3°, and 90%
efficiency and uniformity was obtained.

Figure 11. Light distribution on a receiver with (a) dθ = 2◦ ; (b) dθ = 4◦ ; and (c) dθ = 6◦ . (d) Efficiency
and uniformity depend on the angle of the collimated beam.
Appl. Sci. 2020, 10, x; doi: FOR PEER REVIEW www.mdpi.com/journal/applsci

Appl. Sci. 2020, 10, x; doi: FOR PEER REVIEW www.mdpi.com/journal/applsci

Appl. Sci. 2020, 10, x FOR PEER REVIEW 9 of 13

Figure
Appl. Sci. 2020,11.
10, Light
3257 distribution on a receiver with (a) dθ = 2°; (b) dθ = 4°; and (c) dθ = 6°. (d) Efficiency9 of 13
and uniformity depend on the angle of the collimated beam.

3. Experiment

3.1.
3.1. Manufacturing
Manufacturing Process
Process of
of Double
Double Fresnel
Fresnel Lenses
Lenses
We
We used
used aa commercial
commercial collimator
collimator lenslens from
from Led-link
Led-link Optics
Optics Inc.
Inc. [13]
[13] to
to collimate
collimate allall light
light from
from
the
the LED. A 5W COB LED with 630 nm wavelength was selected. Redistribution module consisted of
LED. A 5W COB LED with 630 nm wavelength was selected. Redistribution module consisted of
two
two linear
linear Fresnel
Fresnel lenses
lenses perpendicular
perpendicular to to each
each other.
other. Optical
Optical elements
elements were
were made
made of of poly
poly methyl
methyl
methacrylate
methacrylate (PMMA).
(PMMA). The The module
module size size was
was determined
determined by by aa commercial
commercial collimator
collimator lenslens with
with aa
diameter of 100 mm. Based on this, the two linear Fresnel lenses had 100 mm
diameter of 100 mm. Based on this, the two linear Fresnel lenses had 100 mm diameters, respectively. diameters, respectively.
The
The lens
lens system
system waswas designed
designed forfor aa lighting
lighting area of 22 ×× 22 m,
area of m, and
and the
the distance
distance from
from the
the light
light source
source to
to
the target was
the target was 2 m. 2 m.
The
The prototype
prototypewas made
was madebased on the
based onsimulated FresnelFresnel
the simulated lens structure. We changed
lens structure. Wethe simulation
changed the
file into a CAD
simulation filefile,
intogenerated
a CAD file,G-code, and processed
generated G-code, and theprocessed
lens by high-speed
the lens byCNC milling. CNC
high-speed CNCProxxon
milling.
MF70 was usedMF70
CNC Proxxon to realize the prototype
was used to realize [14]. The basic[14].
the prototype information
The basicofinformation
this method ofisthis
presented
methodin is
reference [15]. We used an acrylic sheet (PMMA) with a thickness of 5 mm
presented in reference [15]. We used an acrylic sheet (PMMA) with a thickness of 5 mm to create to create the grooves in the
the
linear
grooves Fresnel
in the lenses. The CNC
linear Fresnel milling
lenses. The machine
CNC millinghas an accuracy
machine hasofan2accuracy
µm, so we of 2applied
µ m, so awe polishing
applied
process to make the lenses clearer. Figure 12 shows the two Fresnel lenses after
a polishing process to make the lenses clearer. Figure 12 shows the two Fresnel lenses after polishing. polishing. Efficiency
measurement results showed
Efficiency measurement that the
results linear Fresnel
showed that thelenses
linearafterFresnel
polishing had aafter
lenses transmission
polishing efficiency
had a
up to 95% with a 650 nm laser.
transmission efficiency up to 95% with a 650 nm laser.

lenses.
Figure 12. Photograph of the fabricated linear Fresnel lenses.

3.2.
3.2. Prototype
Prototype Test
Test
We
We made
made the the optical
optical distribution
distribution measurement
measurement system
system shown
shown inin Figure
Figure 13.
13. Osram
Osram bpw21
bpw21
photodiode
photodiode [16] was used to collect optical energy from the light source, and this photodiode was
[16] was used to collect optical energy from the light source, and this photodiode was
located
located alongalongthe theXYXYaxis,
axis,which
whichcould
couldbebemoved
moved and
andcontrolled byby
controlled computer.
computer.WeWemade a program
made a programfor
the
for photodiode
the photodiode to sample at consecutive
to sample locations
at consecutive 4 cm apart.
locations The measurement
4 cm apart. systemsystem
The measurement had an had
area anof
m ×of2 2m.mSo,
2area the square detector system had 50 × 50 pixels. The value of the pixel ranged
× 2 m. So, the square detector system had 50 × 50 pixels. The value of the pixel ranged from 0 to
Appl.(8Sci.
255 2020, 10, x FOR PEER REVIEW 10 of 13
from 0bits analog
to 255 to analog
(8 bits digital converter (ADC)), which
to digital converter quantitatively
(ADC)), representsrepresents
which quantitatively the intensity
theof light.
intensity
of light.

Appl. Sci. 2020, 10, x; doi: Figure

FOR PEER
13. REVIEW www.mdpi.com/journal/applsci
Experimental setup of light distribution measurements.
Figure 13. Experimental setup of light distribution measurements.

The simulated intensity and the measured intensity were normalized to compare simulation and
test results. Figure 14 shown the Experiment results of the distribution measuring system. Figure 14a
shows the measured optical distribution, and Figure 14b shows a comparison of the line distributions
along the X axis across the center of the illuminated area.
Appl. Sci. 2020, 10, 3257 10 of 13
Figure 13. Experimental setup of light distribution measurements.

Thesimulated
The simulatedintensity
intensityand
and the
the measured
measured intensity
intensity were
were normalized
normalizedto
to compare
compare simulation
simulationand
and
testresults.
test results. Figure
Figure 14
14 shown
shown the
the Experiment
Experiment results
resultsof
ofthe
the distribution
distributionmeasuring
measuringsystem.
system. Figure
Figure14a
14a
shows the measured optical distribution, and Figure 14b shows a comparison of the line distributions
shows the measured optical distribution, and Figure 14b shows a comparison of the line distributions
along the
along the XXaxis
axisacross
acrossthe
thecenter
centerof ofthe
theilluminated
illuminatedarea.
area.

Figure 14. Experiment

Figure 14. Experiment results
results of
of the
the distribution
distribution measuring
measuring system.
system. (a)
(a) Distribution
Distribution on
on the
the target
target
surface. (b) Comparison of the simulation and experiment distribution along the X axis across the
surface. (b) Comparison of the simulation and experiment distribution along the X axis across the
center of the target surface.
center of the target surface.

Most of the linear distributions along the X axis for simulation and experiment results were
Most of the linear distributions along the X axis for simulation and experiment results were
consistent with each other. The experiment achieved a uniformity of 75% with a minimum value of 110
consistent with each other. The experiment achieved a uniformity of 75% with a minimum value of
and maximum value of 243, where uniformity was calculated by Equation (2). The difference between
110 and maximum value of 243, where uniformity was calculated by Equation (2). The difference
simulation and test results can be explained by many factors. Firstly, the exact value of the PMMA
between simulation and test results can be explained by many factors. Firstly, the exact value of the
refractive index could be different from the simulation. In addition, the size of the engraving tools may
PMMA refractive index could be different from the simulation. In addition, the size of the engraving
cause quality degradation. Creating lens morphologies may not entirely be accurate with this design.
tools may cause quality degradation. Creating lens morphologies may not entirely be accurate with
Increasing the accuracy of engraving tools helps to achieve a more precise morphology. However,
this design. Increasing the accuracy of engraving tools helps to achieve a more precise morphology.
studies have also shown that if an engraving tool is too small, it can deteriorate the quality of the work
However, studies have also shown that if an engraving tool is too small, it can deteriorate the quality
surface [11].
of the work surface [11].
3.3. Outdoor Test
3.3. Outdoor Test
An LED with an optical component in this study was applied to control chrysanthemum flowering.
An LED with
Chrysanthemum is aan optical plant,
short-day component
which in this study
flowers was the
only when applied to control
provided chrysanthemum
photoperiod is shorter
flowering.
than Chrysanthemum
the “critical period” [17].is We
a short-day plant,
used night which
break flowers only
phenomenon [12]when the in
as show provided photoperiod
Figure 15 to control
is shorter than the
chrysanthemum “critical We
flowering. period”
made[17]. Weflash
a short usedof night
lightbreak
in thephenomenon
middle of the [12]
nightasso
show in Figure
the plant would15
to control chrysanthemum flowering. We made a short flash of light in the middle
behave as if it had been exposed to a long day. The differences of chrysanthemum growth between of the night so the
plant would
controlled behaveand
flowering as uncontrolled
if it had been exposedare
flowering to shown
a longinday. The 15b
Figures differences
and 15c, of chrysanthemum
respectively. When
growth between controlled flowering and uncontrolled flowering are
chrysanthemums were under uncontrolled flowering, it flowered at a time of no demand. shown in Figure 15b and Figure
Flower sizes
were small, and the quality was also low. Chrysanthemums under controlled flowering had a longer
Appl. Sci. 2020, 10, x; doi: FOR PEER REVIEW
vegetative process, resulting in taller chrysanthemums. With time-controlledwww.mdpi.com/journal/applsci
flowering, the quality of
the product improved.
Appl. Sci.Sci.
Appl. 2020, 10,10,
2020, x FOR PEER
x FOR REVIEW
PEER REVIEW 11 11
of 13
of 13

15c, respectively.
15c, respectively. When
When chrysanthemums
chrysanthemums were
wereunder
underuncontrolled
uncontrolled flowering, it flowered
flowering, at at
it flowered a time
a time
of of
nono demand.
demand. Flower
Flowersizes were
sizes were small, and
small, andthethe
quality was
quality was also
alsolow. Chrysanthemums
low. Chrysanthemums under
under
controlled
controlled flowering
flowering
Appl. Sci. 2020, 10, 3257
had a
had longer
a vegetative
longer process,
vegetative process,resulting in
resulting taller
in chrysanthemums.
taller chrysanthemums. With
With time-
11 oftime-
13
controlled flowering, the quality of the product improved.
controlled flowering, the quality of the product improved.

Figure 15. Night break phenomenon (a) Principle; (b) Chrysanthemum with controlled flowering by
Figure 15.15.
Figure Night break
Night phenomenon
break phenomenon(a)(a)
Principle; (b)(b)
Principle; Chrysanthemum with
Chrysanthemum controlled
with flowering
controlled byby
flowering
LED. (c) Chrysanthemum without controlled flowering.
LED. (c) Chrysanthemum without controlled flowering.
LED. (c) Chrysanthemum without controlled flowering.
The flowering process takes 20 to 30 days after stopping the night break phenomenon.
The
Theflowering
flowering process takes 2020 to to
3030days after stopping thethenight break phenomenon. The
The wavelength that isprocess
suitabletakes
for illumination daysis after stopping
630 nm [18]. In night
this break
experiment,phenomenon.
we used two The
wavelength
wavelength thatthatis issuitable forforillumination
suitable illumination is is630
630nmnm [18]. In Inthis
[18]. thisexperiment,
experiment, weweusedused twotwo
flowerbeds of chrysanthemums. The first flowerbed was under a 5W COB LED with 630 nm wavelength
flowerbeds
flowerbeds of ofchrysanthemums.
chrysanthemums. TheThefirst flowerbed
firstthe
flowerbed was
wasunder
under a 5W 5WCOB LED
LEDwith with630
630nm
with optical components we designed, and second flowerbed wasa under COB
the same LED with anm
wavelength
wavelength withwith optical
opticalcomponents
components wewe designed,
designed, and
andthethe
second
second flowerbed
flowerbed was under
was under thethe
same
same
traditional parabolic reflector. Chrysanthemums were illuminated at night with a projection time of
LED with witha and
atraditional parabolic reflector. Chrysanthemums were illuminated at atnight with a2 a
2 hLED
per day, traditional
the distance parabolic
betweenreflector.
stems was Chrysanthemums
10 × 16 cm. Thewere area of illuminated
the flowerbed night
was 4withm
projection time of 2 h per day, and the distance between stems was 10 × 16cm. The area of of
the
(2 × 2 m). Thus, 250 flowers were planted in each flowerbed. After two months, the results are shownthe
projection time of 2 h per day, and the distance between stems was 10 × 16cm. The area
flowerbed was 4 m (2 × 2 m). Thus, 250 flowers were planted in each flowerbed. After two months,
2
in flowerbed
Figure 16. was 4 m (2 × 2 m). Thus, 250 flowers were planted in each flowerbed. After two months,
2

thethe
results areare
results shown
shown in in
Figure
Figure16.16.

Figure 16. Chrysanthemums under controlled flowering (a) By LED with a parabolic reflector. (b) By
Figure 16.16.
Figure
LED with Chrysanthemums
our Chrysanthemums under controlled
under
optical component. flowering
controlled (a)(a)
flowering ByBy
LED with
LED a parabolic
with reflector.
a parabolic (b)(b)
reflector. ByBy
LED
LEDwith
withourour
optical component.
optical component.
As shown in Figure 16a, chrysanthemums under LED with parabolic reflectors had non-uniform
AsAs
heights; shown
theshown in in
plants Figure
on the16a,
Figure edgechrysanthemums
16a,of under
the area were shorter
chrysanthemums under LED
than
LEDwith
the parabolic
plants
with reflectors
at the
parabolic had
center, which
reflectors non-uniform
had results in a
non-uniform
heights; the
lower-quality plants on
heights; the commercial the
plants on the edge of
product. the
edge ofOn area
thethe were
areaother shorter
werehand, than the plants
chrysanthemums
shorter at the
than the plantsunder center,
at the LED which
with
center, results in in
our results
which optical
a lower-quality
components
a lower-quality commercial
showed commercial product.
quite uniform OnOn
heights,
product. thethe
other
which ishand,
other chrysanthemums
important
hand, in the market.
chrysanthemums under LED
Specific
under with our
numerical
LED with our
optical
results components
are shown in showed
Table 2. quite uniform heights, which is important
optical components showed quite uniform heights, which is important in the market. Specificin the market. Specific
numerical
numerical results
resultsareare
shown
shown in in
Table 2. 2.
Table
Table 2. Results of chrysanthemums after 2 months of testing.

Parameter Table 2. Results

Table of of
2. Results
LED chrysanthemums after
chrysanthemums
with 2 months
after
Double Fresnel ofLED
2 months
Lenses testing.
of testing.
with Parabolic Reflector
Number ofParameterParameter
stems surveyed LED with
LED double
with 216 Fresnel
double lenses
Fresnel lenses LED with
LED parabolic
with 207 reflector
parabolic reflector
Rate
Numberof early
Number of offlowering
stems
stemssurveyed
surveyed 4.2%
216216 5.9%
207207
Lowest Rateheight
of of
Rate of the
early
early stems
flowering
flowering 75 cm
4.2%
4.2% 58
5.9% cm
5.9%
Highest
Lowest height
Lowestheight of
height the
of of stems
thethestems
stems 84 cm
75 75
cmcm 80
cmcm
58 58 cm
Average height of the stems 82 cm 72 cm
Average
Appl. Sci.Sci.
Appl. 2020, diameter
10,10,
2020, x; doi: of
FOR
x; doi: the
FOR flower
PEERPEERREVIEW
REVIEW 7.8 cm 7.3 cm
www.mdpi.com/journal/applsci
www.mdpi.com/journal/applsci

Based on the data in Table 2, we can see that both samples had good results in controlled flowering;
more than 94% of plants had no flowers before stopping the night break illumination. The LED with
Appl. Sci. 2020, 10, 3257 12 of 13

our optical components was better than the LED with a parabolic reflector for controlling the flowering
effect in terms of uniformity in height. Also, the overall height of flowers under our light source was
higher than those under the conventional LED source by 12%. The flower size was also 6% larger in
our case.
A uniform light distribution on the cultivated area using our proposed design is a very promising
solution in the field of high-value-added agriculture. Our method can be applied to many types of
LEDs, reducing costs and achieving high efficiency.

4. Conclusions
In this paper, double linear Fresnel lenses made of PMMA with 20 grooves each were fabricated
using high-speed CNC milling. The method of making the lenses demonstrated its capability for
high-volume-based manufacturing. Design, simulation, fabrication, and measurement of double linear
Fresnel lenses were discussed. The achieved uniformity of the redistributed irradiation was about 75%,
which is consistent with the simulation result. An outdoor test of the performance of our design for
agricultural lighting and a comparison between our light source and a conventional light source were
conducted. It shows our proposed design exhibits great potential for commercialization.

Author Contributions: S.S.; T.Q.T., and N.H.V. proposed the ideas and designed the simulation and experiments.
H.V. and N.M.K. performed simulations and experiments and analyzed the data. D.T.G. collected agricultural
experiment results. All authors have read and agreed to the published version of the manuscript.
Funding: This research was funded by Vietnam Academy of Science & Technology under Grant Number
QTKR01.01/18-19, National Science & Technology Program for Development of Center Highlands (Vietnam) under
Grant Number TN18/C08 and Ministry of Science and Technology of Vietnam under Grant Number NĐT.46.
KR/18.
Conflicts of Interest: The authors declare no conflicts of interest.

References
1. Giang, D.T.; La, T.L.; Tien, T.Q.; Duong, P.H.; Tong, Q.C. A simple designed lens for human centric lighting
using LEDs. Appl. Sci. 2020, 10, 343.
2. Giang, D.T.; Pham, T.S.; Ngo, Q.M.; Nguyen, V.T.; Tien, T.Q.; Duong, P.H. An Alternative Approach for High
Uniformity Distribution of Indoor Lighting LED. IEEE Photonics J. 2020, 12. [CrossRef]
3. E. Fred Schubert Light-Emitting Diodes, 2nd editio; Cambridge University Press: Cambridge, UK, 2006.
4. Ding, Y.; Liu, X.; Zheng, Z.; Gu, P. Freeform LED lens for uniform illumination. Opt. Express 2008, 16, 12958.
[CrossRef] [PubMed]
5. Wang, K.; Chen, F.; Liu, Z.; Luo, X.; Liu, S. Design of compact freeform lens for application specific
light-emitting diode packaging. Opt. Express 2010, 18, 413. [CrossRef] [PubMed]
6. Benı´tez, P. Simultaneous multiple surface optical design method in three dimensions. Opt. Eng. 2004, 43,
1489. [CrossRef]
7. Miñano, J.C.; Benítez, P.; Liu, J.; Infante, J.; Chaves, J.; Wang, L. Applications of the SMS method to the design
of compact optics. Opt. Model. Des. 2010, 7717, 77170I.
8. Wang, G.; Wang, L.; Li, F.; Kong, D. Design of optical element combining Fresnel lens with microlens array
for uniform light-emitting diode lighting. J. Opt. Soc. Am. A 2012, 29, 1877. [CrossRef] [PubMed]
9. Yu, X.; Wei, X.; Zhang, O.; Zhang, X. Research on illumination uniformity of high-power LED array light
source. AIP Conf. Proc. 2018, 1971. [CrossRef]
10. Vu, N.H.; Pham, T.T.; Shin, S. LED uniform illumination using double linear fresnel lenses for energy saving.
Energies 2017, 10, 2091. [CrossRef]
11. Ries, H.; Rabl, A. Edge-ray principle of nonimaging optics. J. Opt. Soc. Am. A 1994, 11, 2627. [CrossRef]
12. Shimomachi, T. Effect of Night Break Treatment on Growth and Flowering in. Metab. Clin. Exp. 2011, 8521,
303–308.
13. Lens by LED. Available online: https://www.ledlink-optics.com/lens/ (accessed on 6 March 2020).
14. High Speed CNC Milling. Available online: https://www.proxxon.com/en/micromot/27110.php (accessed on
6 March 2020).
Appl. Sci. 2020, 10, 3257 13 of 13

15. Blalock, T.; Medicus, K.; DeGroote Nelson, J. Fabrication of freeform optics. Opt. Manuf. Test. XI 2015,
9575, 95750H.
16. BPW 21-Photodiode. Available online: https://www.osram.com/ecat/Metal Can® TO39 Ambient Light Sensor
BPW 21/com/en/class_pim_web_catalog_103489/global/prd_pim_device_2219533/ (accessed on 18 April 2020).
17. Higuchi, Y. Florigen and anti-florigen: Flowering regulation in horticultural crops. Breed. Sci. 2018, 68,
109–118. [CrossRef] [PubMed]
18. Tripathi, S.; Hoang, Q.T.N.; Han, Y.J.; Kim, J. Il Regulation of photomorphogenic development by plant
phytochromes. Int. J. Mol. Sci. 2019, 20, 6165. [CrossRef] [PubMed]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).