sensors
Article
Design and Optimization for Mounting Primary Mirror with
Reduced Sensitivity to Temperature Change in an Aerial
Optoelectronic Sensor
Meijun Zhang 1,2 , Qipeng Lu 1, *, Haonan Tian 1 , Dejiang Wang 1 , Cheng Chen 1 and Xin Wang 3






Citation: Zhang, M.; Lu, Q.; Tian, H.;
Wang, D.; Chen, C.; Wang, X. Design
and Optimization for Mounting
Primary Mirror with Reduced
Sensitivity to Temperature Change in
an Aerial Optoelectronic Sensor.
Sensors 2021, 21, 7993. https://
doi.org/10.3390/s21237993
Academic Editors: Mi Wang,
Hanwen Yu, Jianlai Chen and
Ying Zhu
Received: 23 October 2021
Accepted: 26 November 2021
Published: 30 November 2021
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2021 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 (https://
creativecommons.org/licenses/by/
4.0/).
Sensors 2021, 21, 7993. https://doi.org/10.3390/s21237993
1 Key Laboratory of Airborne Optical Imaging and Measurement, Changchun Institute of Optics,
Fine Mechanics and Physics, Chinese Academy of Sciences, Changchun 130033, China;
zhangmeijun@ciomp.ac.cn (M.Z.); tianhaonan@ciomp.ac.cn (H.T.); wangdj04@ciomp.ac.cn (D.W.);
chengengeng@126.com (C.C.)
2 University of the Chinese Academy of Sciences, Beijing 100049, China
3 School of Mechanical and Aerospace Engineering, Jilin University, Changchun 130025, China;
wxin@jlu.edu.cn
* Correspondence: luqipeng@126.com; Tel.: +86-0431-8617-6160
Abstract: In order to improve the image quality of the aerial optoelectronic sensor over a wide range
of temperature changes, high thermal adaptability of the primary mirror as the critical components is
considered. Integrated optomechanical analysis and optimization for mounting primary mirrors are
carried out. The mirror surface shape error caused by uniform temperature decrease was treated as
the objective function, and the fundamental frequency of the mirror assembly and the surface shape
error caused by gravity parallel or vertical to the optical axis are taken as the constraints. A detailed
size optimization is conducted to optimize its dimension parameters. Sensitivities of the optical
system performance with respect to the size parameters are further evaluated. The configuration of
the primary mirror and the flexure are obtained. The simulated optimization results show that the
size parameters differently affect the optical performance and which factors are the key. The mirror
surface shape error under 30 ◦ C uniform temperature decrease effectively decreased from 26.5 nm to
11.6 nm, despite the weight of the primary mirror assembly increases by 0.3 kg. Compared to the
initial design, the value of the system’s modulation transfer function (0◦ field angle) is improved from
0.15 to 0.21. Namely, the optical performance of the camera under thermal load has been enhanced
and thermal adaptability of the primary mirror has been obviously reinforced after optimization.
Based on the optimized results, a prototype of the primary mirror assembly is manufactured and
assembled. A ground thermal test was conducted to verify difference in imaging quality at room and
low temperature, respectively. The image quality of the camera meets the requirements of the index
despite degrading.
Keywords: primary mirror; sensitivity analysis; optimization design; modulation transfer function;
thermal deformation resistance
1. Introduction
Aerial optoelectronic sensors have the advantages of flexibility, high resolution and
strong timeliness, which are widely used in the fields of topographic survey, emergency
rescue, and national defense security, etc. [1–3]. Structural design of an aerial optoelectronic
sensor is not only subject to strong constraints of volume and weight. Furthermore, it
needs to have good environmental adaptability, be capable of withstanding a wide range
of temperature change (−40~+55 ◦ C), shock, vibration, overload and other mechanical
environments impact, so as to output clear images. The primary mirror, as the critical
component of a high-resolution optoelectronic sensor, is sensitive to environmental factors,
such as temperature variation and external acceleration. How to reduce the sensitivity of
the supporting structure of the primary mirror to ambient temperature change becomes
https://www.mdpi.com/journal/sensors
Sensors 2021, 21, 7993 2 of 15

an issue worthy of serious consideration. A stable and reliable support structure plays an
important role in ensuring image quality.
There has been a lot of research on design and optimization for mounting the mirrors.
The parameter optimization or topology optimization method is widely used to optimize
the structural parameter on a predefined configuration [4–7]. Shao Mengqi et al. [8] estab-
lished a size parameter optimization model to optimize the optical quality of the mirror.
Contribution of size parameters to optical performance was further evaluated with the help
of ray-tracing, multidisciplinary optimization and finite element analysis software. Kihm
Hagyong et al. [9,10] proposed a design method to optimize a primary mirror and its flexure
separately in a space telescope. A multi-objective genetic algorithm was implemented to
optimize the design parameters which met the design goals and reduced the optimization
time by an order of magnitude. Liu Shutian et al. [11] presented a design process of primary
mirror based on topology optimization, in which structural compliance was treated as
the objective function to minimize the optical surface deviation due to self-weight and
polishing pressure loading. According to the simulation result, the configuration of a
lightweight mirror for a large-aperture space telescope was obtained. Hu Rui et al. [12]
developed a method for designing the flexure for mounting the primary mirror. A topology
optimization was firstly adopted to the extraction of conceptual configuration. Then a
detailed shape optimization was conducted to optimize the dimension parameters. To
reduce the sensitivity of a lightweight mirror to the mount location, the eccentricity of the
hyperbola curve was introduced as the objective function in the topology and parameter
optimization by Jiang Ping et al. [13]. An optimal lightweight design for a zerodur primary
mirror with an outer diameter of 566 mm and supporting bipod flexure was studied by
Chen YIcheng et al. [14], in which the lightweight ratio and the surface shape error for the
deformed surface under polishing pressure and gravity was achieved.
At present, there are few studies on the optimization of mirrors and their support-
ing structures together in aerial photoelectric sensors, despite most of this research on
the optimization for lightweight mirrors and its supporting structures being focused on
space-borne or ground-based telescope applications. Due to the difference in the working
environment of the aviation photoelectric remote sensor and the aerospace remote sensor,
the performance requirements are different. The operating temperature of the internal opti-
cal system of the aerospace remote sensor can usually be maintained within ±5 ◦ C through
the internal thermal control facility. However, the working temperature of the optical
system inside the aerial remote sensor can usually be maintained within −40~+55 ◦ C with
poor thermal control measures or none. Meanwhile, the design or optimization process for
supporting the mirror is usually separated into two independent parts: mirror design and
flexure design. In fact, the design or optimization of the mirror is coupled with the flexure
design. It is difficult to obtain better optimization results by just studying the optimization
of the mirror itself or its supporting structure alone.
In this paper, the optimization design of the mirror and its supporting structure is
combined and studied together. High thermal deformation resistance of the primary
mirror of an aerial optoelectronic sensor is considered for better adaptation in the low
or high temperature working environment. A detailed size optimization of the primary
mirror and the flexure is conducted together to optimize their dimension parameters. The
process of optomechanical analysis and optimization for the primary mirror assembly are
developed in sequence. The mirror surface shape error caused by uniform temperature
decrease was treated as the objective function and the fundamental frequency (the first-
order natural frequency) of the mirror assembly and the surface shape error caused by
gravity parallel or vertical to the optical axis are taken as the constraints. The influence of
size parameters on the mirror surface shape error is analyzed. As a result, the configuration
of the primary mirror and the flexure is obtained. Compared with the initial design, the
optical performance of the system has been significantly improved.
 Sensors 2021, 21, x FOR PEER REVIEW 3 of 16

Sensors 2021, 21, 7993 3 of 15

the primary mirror and the flexure is obtained. Compared with the initial design, the op-
tical performance of the system has been significantly improved.

2.2.Performance
PerformanceRequirements
Requirements
The aerial
The aerial optoelectronic
optoelectronic sensor
sensorcontains
containsa avisible light
visible camera
light camerawhich adopts
which the Cas-
adopts the
segrain optical
Cassegrain system
optical in this
system in paper. The visible
this paper. light camera
The visible is integrated
light camera in the spherical
is integrated in the
envelopeenvelope
spherical space, andspace,
the focal
andlength is 1000
the focal mm,
length is which
1000 mm, is easily
which influenced by the thermal
is easily influenced by
disturbances
the and externaland
thermal disturbances acceleration. As the core As
external acceleration. component of the optical
the core component system,
of the the
optical
primarythe
system, mirror needs
primary to adapt
mirror needsto the temperature
to adapt change of the
to the temperature environment
change to the great-
of the environment
estthe
to extent andextent
greatest minimize the degradation
and minimize of imageofquality.
the degradation Figure 1Figure
image quality. shows1 shows
the optical
the
optical
model ofmodel of the camera
the camera which iswhich is established
established in the ray-tracing
in the ray-tracing software
software Zemax forZemax for
this study.
this study.

Figure1.1.Optical
Figure Opticalmodel
modelof
ofthe
thevisible
visiblelight
lightcamera.
camera.

Usually,the
Usually, theoptical
opticalaxis
axisof ofthe
themirror
mirrorisisplaced
placedhorizontally
horizontallyon onthe
theair-floating
air-floating plat-
plat-
form, which
form, which is conducive
conducive to toassembly
assemblyand andoptical
opticaltesting.
testing. In In
thisthis
situation, the the
situation, optical per-
optical
formance can be adjusted to the best state, almost close to the diffraction
performance can be adjusted to the best state, almost close to the diffraction limit. In fact, limit. In fact, the
the optical system is installed on the airplane, in actual use scenarios, and most of theit
optical system is installed on the airplane, in actual use scenarios, and most of the time
is in itthe
time working
is in the working state of squint
state imaging.
of squint This inconsistency
imaging. This inconsistency between alignment
between status
alignment
and usage
status statusstatus
and usage will inevitably, more more
will inevitably, or less,orlead
less,to a decline
lead in theinperformance
to a decline the performanceof the
optical
of system.
the optical Namely,
system. the supporting
Namely, structure
the supporting of the primary
structure mirror needs
of the primary mirrorto needs
have suf-
to
ficient
have rigidity rigidity
sufficient to resist to gravity
resist deformation or deformation
gravity deformation caused by
or deformation external
caused accelera-
by external
tion [15]. Furthermore,
acceleration [15]. Furthermore,the working temperature
the working of the of
temperature optical system
the optical insideinside
system the aerial
the
aerial optoelectronic sensor in this paper can usually be maintained
optoelectronic sensor in this paper can usually be maintained within a wide range of tem- within a wide range
of temperature
perature changechange(−10~+50 °C) with◦an
(−10~+50 C) active-passive
with an active-passive
combinedcombined thermal
thermal control control
measures.
measures.
In the airborneIn theworking
airborneenvironment,
working environment, the temperature
the temperature environment environment
has a more has a more
direct im-
direct
pact on impact on thequality
the image image of quality of the system
the optical optical system
compared compared
to othertoenvironmental
other environmental
factors
factors
such assuch as gravity.
gravity. Therefore,
Therefore, the supporting
the supporting structure
structure of theofprimary
the primary
mirrormirror
needsneeds to
to have
have adequate compliance to resist
adequate compliance to resist thermal deformation. thermal deformation.
AAconvenient
convenientmeasure measureto toquantify
quantifythe theoptical
opticalperformance
performanceof ofthe
themirror
mirrorisisRMS
RMS (root-
(root-
mean-square) of the mirror surface shape error. The RMS
mean-square) of the mirror surface shape error. The RMS can be obtained by Zernike can be obtained by Zernike
pol-
polynomial
ynomial fitting, fitting,
whichwhich is expressed
is expressed byby Equation
Equation (1):(1):
v
u NN
=
RMS = t w
RMS 
wiδδ2i , ，
∑
2
δ i u= −
ui z− z−1 −z z−
2 −
z z3 (1)
u
δ =
i i i i 1 2 3 (1)
i =1
i =1

where N is the number of the nodes of the mirror face, wi is weight of the i-th node, ui is
where N is the number of the nodes of the mirror face, wi is weight of the i-th node, ui is
the sag displacement of the i-th node, z1, z2, z3 represent the piston, tilt and power term of
the sag displacement of the i-th node, z1 , z2 , z3 represent the piston, tilt and power term of
the deformed surface, respectively [12].
the deformed surface, respectively [12].
Sensors 2021, 21, 7993 4 of 15

Due to the orthogonality of the Zernike polynomials over a unit circle, they are con-
venient to use to describe the aberration in optical engineering [16,17]. The mathematical
description of Zernike polynomials for a deformed surface are defined by Equation (2):
∞ ∞ n
∆Z (ρ, θ ) = A00 + ∑ An0 R0n (ρ) + ∑ ∑ Rm
n ( ρ )[ Anm cos( mθ ) + Bnm sin( mθ )] (2)
n =2 n =1 m =1

where ∆Z(ρ,θ) denotes the deformed surface, ρ is the normalized radius on the reflection
surface, θ is the azimuth angle, Anm and Bnm are Zernike coefficients in x direction and y
direction, the variables n and m are radial and circumferential wave number, respectively.
The radial dependence Rmn ( ρ ) of the Zernike polynomials can be expressed by Equation (3):

(n−m)/2
(−1)k (n − k)!
Rm
n (ρ) = ∑ k![(n + m)/2 − k ]![(n − m)/2 − k ]!
r (n−2k) (3)
k =0

The goal of this paper is to further optimize the thermal adaptability of the supporting
structure of the primary mirror. Thus, the mirror surface shape error caused by uniform
temperature decrease was treated as the objective function, and the fundamental frequency
of the mirror assembly and the surface shape error caused by gravity parallel or vertical
to the optical axis are taken as the constraint. The RMS of the mirror under gravity and
thermal load separately should be less than λ/40 (where λ = 632.8 nm). The mass of the
primary mirror component should be less than 1.6 kg including the primary mirror, invar
sleeves and the flexures. The fundamental frequency of the primary mirror component
should be more than 200 HZ, which indicates the stiffness to resist the gravity deformation
or external acceleration.
Some performance requirements are requested in our camera. Generally, there should
be a certain safety margin in the structural design [18]. According to the decomposition of
the camera specifications and applications, the performance requirements of the primary
mirror assemblies are listed in Table 1.

Table 1. Performance requirements of the primary mirror assembly.

Total Mass ≤1.6 kg

RMS error under 1G gravity parallel or vertical to the optical axis ≤8 nm
RMS error under uniform temperature decrease 30 ◦ C ≤15.6 nm
Fundamental frequency ≥200 HZ
Operating temperature 20 ± 30 ◦ C
The primary mirror’s tilt under 1G gravity vertical to the optical axis ≤20”
The primary mirror’s decenter under 1G gravity vertical to the optical axis ≤20 um

3. Optimal Design of the Primary Mirror Assembly

3.1. Initial Design of the Primary Mirror Assembly
Classical configurations of lightweight mirror include closed back, semi-closed back
and open back. The open back configuration has the advantages of easy processing and
high lightweight ratio, which is very suitable for an aerial camera [19]. Consequently, an
open back configuration is employed in this article. The mirror usually consists of faceplate
and ribs from behind. A common lightweight pocket shape could be triangular, round,
hexagon or sector holes [20]. Due to the use of SiC material with high specific stiffness,
sector holes are adopted for the purpose of further improving the lightweight ratio in this
article. The ribs with certain thickness support the thin faceplate form behind, which not
only improves the stiffness and cut down weight of the base, but also provides channels
for heat conduction. Thermal influence form the surrounding environment can be partially
weakened by this means [8].
A typical three-point back support is applied for the primary mirror in this study on
account of its simplicity and reliability, as shown in Figure 2. The diameter of the primary
Sensors 2021, 21, x FOR PEER REVIEW 5 of 16

channels for heat conduction. Thermal influence form the surrounding environment can
Sensors 2021, 21, 7993 5 of 15
be partially weakened by this means [8].
A typical three-point back support is applied for the primary mirror in this study on
account of its simplicity and reliability, as shown in Figure 2. The diameter of the primary
mirror is
mirror is 190
190 mm,
mm, and and the
thediameter
diameter of of the
theinner
innerhole
holeisis4444mm.
mm.The Themirror
mirrorsurface
surfaceof ofthe
the
primary mirror is a parabolic surface with high steepness, and the radius of curvatureis
primary mirror is a parabolic surface with high steepness, and the radius of curvature
397.1
is 397.1 mm.
mm. TheTheradial
radial position
position ofofthe supporting
the supporting hole
holeofof
thetheprimary
primary mirror
mirror is is
located
located at
thethe
at equivalent
equivalent radius of gyration.
radius of gyration.To minimize
To minimizethe mirror surface
the mirror distortion
surface when the
distortion op-
when
ticaloptical
the axis of axis
the mirror
of the is no longer
mirror is novertical,
longer the axial position
vertical, the axialofposition
the supporting connecting
of the supporting
surface of the primary mirror is approximately located near the center
connecting surface of the primary mirror is approximately located near the center of gravity of gravity plane
[20,21]. The primary mirror assembly includes the primary mirror,
plane [20,21]. The primary mirror assembly includes the primary mirror, the invar sleeves the invar sleeves and
the flexures.
and The The
the flexures. massmassof the
ofprimary
the primarymirror assembly
mirror assemblyis 1.3iskg
1.3and
kg the
andlightweight
the lightweight ratio
is 45.8%
ratio for the
is 45.8% forinitial lightweight
the initial design.design.
lightweight The threeTheinvar
threesleeves are separately
invar sleeves bonded
are separately
into the into
bonded support holes of holes
the support the mirror,
of the and the and
mirror, threetheflexures are also respectively
three flexures fastened
are also respectively
fastened on thewith
on the sleeves sleeves with The
screws. screws. The primary
primary mirror ismirror is indirectly
indirectly supported supported
by threeby three
parallel
parallel
flexures through support holes at the back of the mirror. The invar sleeves and three and
flexures through support holes at the back of the mirror. The invar sleeves flex-
three flexures
ures are are circularly
circularly symmetrical symmetrical
around the around
opticalthe optical
axis. axis. The
The chosen chosenofmaterial
material the flexureof
the flexure
is TC4, is TC4,
which has which
a high has a high
specific specificThe
stiffness. stiffness.
selectedThe selectedofmaterial
material the sleeveof the sleeve is
is artificially
artificially
prepared, prepared,
which haswhich has acoefficient
a similar similar coefficient
of thermal of thermal
expansion expansion
(CTE) to(CTE) to SiC.
SiC. The The
chosen
chosen
material material properties
properties of the primary
of the primary mirror mirror
assembly assembly
are shownare shown
in Table in2.Table 2.

Figure 2.
Figure 2. Schematic
Schematic of
of the
the primary
primary mirror
mirror assembly.
assembly.

Flexure
Table 2.elements are usually
Material properties adopted
of the mirrorto produce elastic deformation, so as to absorb
assembly.
the strain energy caused by mounting and thermal affect which are a convenient means
CoefficientTypical
of Thermal
Components Materialto mount
Young’a Modulus
mirror with high-performance
(Gpa) Density (g/cm3requirements.
) flexuresPoisson’
used consist
Ratio of
Expansion
a notch hinge, cross hinge or leaf type. A rotationally symmetric (10−6 /K)
leaf hinge provides com-
Primary mirror SiC pliance in defined 350 axes for the purpose 3.0 of precise motion2.5 control, which is easily
0.25 formed
Invar sleeve 4J36 from cylindrical 143 tubes or circular disks
8.12 by electrical discharge
2.43 0.25 type of
wire-cutting. This
Flexure TC4 flexure provides 107 three freedom, one 4.44 8.9
translation and two rotations [22–24]. 0.34

TableFlexure
2. Material properties
elements of the mirror
are usually assembly.
adopted to produce elastic deformation, so as to absorb
the strain energy caused by mounting and thermal affect which are a convenient means
Young’ Modulus Density Coefficient of Ther- Poisson’
toComponents
mount a mirror Material
with high-performance requirements. Typical flexures used consist
(Gpa) (g/cm3) mal Expansion(10−6/K) Ratio
of a notch hinge, cross hinge or leaf type. A rotationally symmetric leaf hinge provides
Primary mirror SiC 350 3.0 2.5 0.25
compliance in defined axes for the purpose of precise motion control, which is easily
Invar from
formed sleeve 4J36 tubes or circular
cylindrical 143 disks by8.12 2.43 wire-cutting.0.25
electrical discharge This
Flexure TC4 107 4.44 8.9 [22–24]. 0.34
type of flexure provides three freedom, one translation and two rotations

3.2. Establishment of Finite Element Model (FEM)

FEA is employed to determine the deformation of the primary mirror in the following
scenarios: (1) deformation in the case of 30 ◦ C uniform temperature decrease; (2) self-
 Sensors 2021, 21, x FOR PEER REVIEW 6 of 16

Sensors 2021, 21, 7993 3.2. Establishment of Finite Element Model (FEM) 6 of 15
FEA is employed to determine the deformation of the primary mirror in the following
scenarios: (1) deformation in the case of 30 °C uniform temperature decrease; (2) self-
weight deformation when the optical axis of the mirror is horizontal or vertical. The es-
weight deformation when the optical axis of the mirror is horizontal or vertical. The
tablishment of the finite element model is put forward through the finite element pre-
establishment of the finite element model is put forward through the finite element pre-
processing software Hypermesh. Shell elements are used for mesh generation, with a total
processing software Hypermesh. Shell elements are used for mesh generation, with a total
number of 45,784 elements and 46,585 nodes. Rigid elements are established to simulate
number of 45,784 elements and 46,585 nodes. Rigid elements are established to simulate
glue and screw connections. The mirror surface shape error RMS (removal of piston, tilt
glue and screw connections. The mirror surface shape error RMS (removal of piston, tilt
and power) is converted into a response equation in DRESP2 format to perform Zernike
and power) is converted into a response equation in DRESP2 format to perform Zernike
polynomial fitting of the deformed mirror surface, and then imported into Optistruct
polynomial fitting of the deformed mirror surface, and then imported into Optistruct solver
solver for size optimization analysis. The design variables involved in this article are di-
for size optimization analysis. The design variables involved in this article are divided
vided into two types: gauge variables and shape variables. Gauge variables are particular
into two types: gauge variables and shape variables. Gauge variables are particular
casesof
cases ofsize
sizevariables,
variables,where
wherethethedesign
designvariables
variablesare
aretwo-dimensional
two-dimensionalproperties.
properties.The The
shape optimization technique is an automated way to modify structural
shape optimization technique is an automated way to modify structural shape based on shape based on
predefined shape variables to find the optimal shape with the aid of a morph
predefined shape variables to find the optimal shape with the aid of a morph toolbox. The toolbox. The
sensitivitiesof
sensitivities ofthe
theresponse
responsevalue
valueto tothe
thedesign
designvariables
variablesarearesolved
solvedwith
withthethemethod
methodofof
feasible direction (MFD). MFD is a gradient-based algorithm, which is
feasible direction (MFD). MFD is a gradient-based algorithm, which is effective when effective whenthe the
sensitivities of the system response, with respect to the design variables, can
sensitivities of the system response, with respect to the design variables, can be computed be computed
easilyand
easily andinexpensively;
inexpensively; therefore,
therefore, ititisisconvenient
convenientforforunderstanding
understandingthe theeffect
effectofofthe
the
design variables and will most likely find the optima. Figure 3 presents the
design variables and will most likely find the optima. Figure 3 presents the finite element finite element
modelof
model ofthe
theprimary
primarymirror
mirrorassembly.
assembly.

(a) (b)
Figure3.3.Finite
Figure Finiteelement
elementmodel
modelofofthe
theprimary
primarymirror
mirrorassembly.
assembly.(a)
(a)Front
Frontview.
view.(b)
(b)Back
Backview.
view.

3.3.
3.3.Optimal
OptimalDesign
DesignFormulation
Formulation
The goal of the optimization
The goal of the optimization ofof
thethe
mirror
mirrorcomponent
component is toisobtain the the
to obtain design parameter
design param-
set that will minimize the mirror surface shape error (RMS ) caused by 30 ◦ C uniform
eter set that will minimize the mirror surface shape error (RMS T T) caused by 30 °C uniform
temperature
temperaturedecrease,
decrease,while
whilemeeting
meetingthe theconstraint
constraintrequirements
requirementsin inTable
Table1.1.The
Themirror
mirror
surface ◦ C uniform temperature decrease is set to be
surfaceshape
shape error
error (RMS
(RMSTT)) caused
causedby by3030°C uniform temperature decrease is set to be the
the objective
objective function
function with
with thethe purpose
purpose of characterizing
of characterizing thethe thermal
thermal adaptability
adaptability of the
of the pri-
primary mirror assembly. The initial working temperature of the primary
mary mirror assembly. The initial working temperature of the primary mirror assembly mirror assembly
isisset ◦ C. The reason is that it is convenient for optical
settotonormal
normalroomroomtemperature
temperature20 20 °C. The reason is that it is convenient for optical
parts
partsandandstructural
structuralparts
partsprocessing
processingand andoptical
opticalassembly
assemblyand andalignment
alignmentunder
underroom
room
temperature conditions. The optimization formulation of the primary
temperature conditions. The optimization formulation of the primary mirror componentmirror component
can
canbebedefined
definedininthe
thefollowing
followingexpressions.
expressions.
The objective function and the constraints are shown in Equation (4):
The objective function and the constraints are shown in Equation (4):
Minimize f ( X ), f ( X ) = RMST ,
f ind X = (di , v j ) T f or i = 1, 2, 3, . . . , 9; j = 1, 2, . . . , 6.
S.T. M ≤ M0
RMSGR ≤ δ0 , RMSGZ ≤ δ0 (4)
f1 ≥ f0
dilow ≤ di ≤ diup ,
v jlow ≤ v j ≤ v jup .
Sensors 2021, 21, 7993 7 of 15

where di is the thickness of the i-th mirror rib or faceplate and vj denotes dimensions related
to compliance of the flexure. M is the current mass of the primary mirror component and
M0 (M0 ≤ 1.6 kg) is the upper limit of the mass allocated by total performance index.
RMSGR , RMSGZ are the values of the mirror surface shape error caused by gravity parallel
and vertical to the optical axis, respectively. δ0 (δ0 ≤ 8 nm) is the upper limit of the mirror
shape error. f 1 is the current fundamental frequency, and f 0 (f 0 ≥ 200 HZ) is the lower limit
of the fundamental frequency. dilow and diup are the lower limit and upper limit of di . vjlow
and vjup are the lower limit and upper limit of vj .
In finite element based structural optimization, the structural fundamental formulas
for linear static analysis is defined by Equation (5):

Ku = F (5)

where K is the stiffness matrix and u is the displacement vector to be determined, and F is
the applied force vector including nodal force, gravity, thermal load, etc. Differentiating
this with respect to the design variable vector X (X = (di , vj )T ),

∂K ∂u ∂F
u+K = (6)
∂X ∂X ∂X
The sensitivity of the displacement vector u with respect to the design variable vector
X can be calculated as follows:
∂u ∂F ∂K
= K −1 ( − u) (7)
∂X ∂X ∂X

The response vector G (G = (RMST , RMSGX , RMSGY , RMSGZ , M)T ) calculated from
the displacements vector is written as:

G = PT u (8)

where PT is the corresponding coefficient matrix with respect to the i-th response function.
The sensitivities of the response vector G with respect to the design variable vector X
can be expressed by Equation (9):

∂G ∂P T ∂u
= u + PT (9)
∂X ∂X ∂X

3.4. Results and Discussion

In order to achieve optimal optical performance, the structure of the primary mirror
and the flexures are parameterized together. The parameterized optimization model
includes nine design variables for the primary mirror and six variables for the flexure. d1 –d6
respect the thickness of the mirror ribs, and d7 denotes the thickness of the mirror faceplate.
v1 –v6 signifies dimensions related to compliance of the flexure. These parameters can be
adjusted during the design process. The thickness of the mirror faceplate, the reinforcing
ribs and the flexible leaf hinge involves mechanical processing and manufacturing limit,
and a certain safety margin should be left. Consequently, the minimum thickness of
the mirror ribs is 3 mm, and the faceplate is 4 mm separately for the initial lightweight
design. Figure 4 depicts the parameterized optimization model of the primary mirror and
the flexure.
 manufacturing limit, and a certain safety margin should be left. Consequently, the mini-
Sensors 2021, 21, 7993
mum thickness of the mirror ribs is 3 mm, and the faceplate is 4 mm separately for the
8 of 15
initial lightweight design. Figure 4 depicts the parameterized optimization model of the
primary mirror and the flexure.

(a) (b)
Figure 4.
Figure 4. Design
Design variables
variables for
for size
size optimization.
optimization. (a)
(a) Primary
Primary mirror.
mirror. (b)
(b) Flexure.
Flexure.

The
The objective function
function and andthetheconstraints
constraintssettings
settingscan
canrefer
referto to Equation
Equation (4).(4).
TheThe
op-
optimization problem
timization problem is issolved
solvedbybymeans
meansof of Optistruct
Optistruct solver, in which,RMS
in which, RMSTT,,RMSRMSGRGR,,

RMS
RMSGZ Mand
GZ,, M andf1 fare
1 are treated
treated as response
as the the response function.
function. TableTable 3 summarizes
3 summarizes thebound,
the lower lower
bound, upper initial
upper bound, bound,and initial and optimization
optimization results ofresults of the
the design design variables.
variables. The final
The final optimized
optimized
primary mirrorprimary mirrorcan
assembly assembly can be
be achieved byachieved
updatingby theupdating the corresponding
corresponding geometry di-
geometry
mensions.dimensions.
After multiple iterations, the optimized value of d3, d4, d5, d6 and d8 reached their
Table 3. Initial
respective and optimized
lower values of d1,
bounds, whereas the d9
design
andvariables.
v5 reached their respective upper bounds.
d1 represents the thickness of the supporting hole rib of the primary mirror obtained the
Lower Bound Upper Bound Initial Value Optimized
Design
upper Variable
bound so that it has sufficient stiffness
(mm) (mm)to resist gravity deformation.Value
(mm) The thickness
(mm)
of the mirror faceplate d7 increased from the initial value of 4 mm to the optimized value
d1 4.0 8.0 4.0 8.0
of 6.6 mm. In addition, the size d2, v3 and v4 increased, on the contrary, v1 and v6 de-
d2 3.0 6.0 3.0 5.4
creaseddto 3
varying degrees. 3.0 6.0 3.0 3.0
d4 3.0 6.0 3.0 3.0
Table 3. dInitial
5 and optimized
3.0values of the design
6.0 variables. 3.0 3.0
d6 3.0 6.0 3.0 3.0
DesigndVar-7
Lower Bound
4.0 Upper Bound 8.0 Initial 4.0 Optimized
6.6
iabled8 (mm) 2.0 (mm) 6.0 Value (mm) 2.0 Value (mm)
2.0
d
d1 9 4.0 3.0 8.0 6.0 4.0 3.0 8.06.0
d2v1 3.0 1.1 6.0 7.6 3.0 6.0 5.45.8
v2 1.1 7.6 6.0 6.0
d3v 3.0 9.5 6.0 15.5 3.0 12.5 3.015.1
3
d4v4 3.0 13.5 6.0 19.5 3.0 16.5 3.016.7
d5v5 3.0 16.5 6.0 22.5 3.0 19.5 3.022.5
d6 v 6 3.0 6.5 6.0 12.5 3.0 9.5 3.07.7
d7 4.0 8.0 4.0 6.6
After
d8 multiple iterations,
2.0 the optimized
6.0 value of d ,
2.0 d ,
3 4 5 6 d , d and d reached
8 2.0 their
respective
d9 lower bounds,
3.0 whereas d , d
1 6.09 and v 5 reached 3.0 their respective upper
6.0 bounds.
d1 represents
v1 the thickness
1.1 of the supporting
7.6 hole rib of 6.0
the primary mirror obtained
5.8 the
upperv2 bound so that 1.1
it has sufficient stiffness
7.6 to resist gravity
6.0 deformation. The
6.0 thickness
of the v3 d7 increased from
mirror faceplate9.5 15.5the initial value 12.5
of 4 mm to the optimized15.1 value of
6.6 mm. In addition, the size d2 , v3 and v4 increased, on the contrary, v1 and v6 decreased
to varying degrees.

3.4.1. Analysis of Wavefront Aberration Characteristics

The Fringe Zernike polynomials is a reordered subset of the Standard Zernike terms,
which are applied to fit the deformed mirror surface. The 37 Zernike coefficients (remov-
 v4 13.5 19.5 16.5 16.7
v5 16.5 22.5 19.5 22.5
v6 6.5 12.5 9.5 7.7

Sensors 2021, 21, 7993 9 of 15

3.4.1. Analysis of Wavefront Aberration Characteristics
The Fringe Zernike polynomials is a reordered subset of the Standard Zernike terms,
which are applied to fit the deformed mirror surface. The 37 Zernike coefficients (remov-
piston and
ing the piston and tilt)
tilt) under 30 ◦°C
under 30 uniform temperature
C uniform temperature decrease
decrease before
before and
and after
after the
the
optimization were obtained as as shown
shown in
in Figure
Figure 5.5. The result indicated that
that the
the low-order
low-order
aberration
aberration power
power (term
(term 4)4) was
was the
the dominating
dominating partpart of
of thermal
thermal deformation,
deformation,which
whichsignifi-
signif-
cantly
icantlydiminished
diminishedafter
afteroptimization
optimizationand
andcancan be
be corrected
corrected by realignment and refocusing.
Meanwhile, the second trefoil aberration B (term (term 20)
20) is
is the
the secondary
secondary largest
largest aberration
aberration
which
which is
is of
of high-order
high-order aberration,
aberration, and it
it is
is difficult
difficult to
to be
be eliminated
eliminated by
by realignment
realignment and
and
refocusing.
refocusing. In addition, other aberrations are reduced correspondingly after optimization.

Figure 5.
Figure 5. Zernike coefficients
Zernike under
coefficients 30 °C
under ◦ C uniform
30uniform temperature decrease
temperature before and
decrease afterand
before the after
opti-
mization.
the optimization.

3.4.2. Sensitivity Analysis of Size Parameters

The method
method of feasible
feasible direction
direction is is adopted
adopted to to solve
solve the sensitivity
sensitivity of responses
responses value
value
to the design variables
variables as mentioned
mentioned in in Section
Section 3.2.3.2. The sensitivity
sensitivity of the response
response function
function
with
with respect
respect toto the
the design
design variable
variable can
can be be expressed
expressed by by Equation
Equation (9). (9). Figure
Figure 66 shows
shows the
the
normalized
normalized percentage contributions of the size parameters to the response functions in
percentage contributions of the size parameters to the response functions in
four
four cases.
cases. From
FromFigure
Figure5a,5a,the
themirror
mirrorsize sized1d1,
, d5d5,
, d7 d7,
, andand thethe
flexure
flexure d8 have
sizesize significant
d8 have signifi-
influence on the ◦ C uniform temperature
cant influence onmirror surface
the mirror shapeshape
surface error error
(RMS(RMS T ) caused by 30by
T) caused 30 °C uniform temper-
decrease. From From
ature decrease. FigureFigure
6b, the 6b,mirror size dsize
the mirror 2 , d5d2,andd5the andflexure size dsize
the flexure 9 , v4 d9,
andv4v6and
exert
v6
greater impactimpact
exert greater on theon mirror surfacesurface
the mirror shape error
shape(RMS errorGR ) caused
(RMS by gravity vertical to the
GR) caused by gravity vertical
optical axis. From
to the optical Figure
axis. From 6c, the
Figure 6c,flexure
the flexure d8 , vd8,
size size 3 andv3 andv4 have obvious
v4 have obvious contribution
contributionto
the fundamental frequency. From Figure 6d, the mass of the
to the fundamental frequency. From Figure 6d, the mass of the primary mirror componentprimary mirror component
mainly
mainly depends
depends on on the
the mirror
mirror size d1 , dd5
size d1, and dd7.
5 and 7 . The
The flexure
flexure sizesize vv1 and vv2
1 and 2 have
have relatively
relatively
weak
weak effects on the four response functions above. In summary, each size exerts aa certain
effects on the four response functions above. In summary, each size exerts certain
impact
impact onon the
the performance
performance in in different
different cases.
cases. The The sensitivity
sensitivity analysis
analysis is is very
very useful
useful for
for
understanding which factors are the key, which provides a reference for the structural
understanding which factors are the key, which provides a reference for the structural
optimization design of aerial optoelectronic sensors.
optimization design of aerial optoelectronic sensors.
3.4.3. Performance Evaluations and Comparisons
After multiple iterations, the optimization response function under gravity parallel
and vertical to the optical axis and thermal load converged step by step as a whole. The
iteration history of the mirror surface shape error in three cases is plotted in Figure 7. The
initial and final values of the system performance responses are summarized in Table 4.
The mirror surface shape error RMST (removing the piston, tilt and power terms) caused
by 30 ◦ C uniform temperature decrease drastically diminishes from 26.5 nm to a minimum
value 11.6 nm (less than λ/40), which is better than the performance requirements in
Table 1. Namely, thermal deformation resistance of the primary mirror assembly has been
obviously improved. Figure 8 depicts the primary mirror surface error map under 30 ◦ C
uniform temperature decrease before and after optimization. The mirror surface shape
error RMSGR under gravity vertical to the optical axis rises from 6.3 nm to 8.0 nm, which
Sensors 2021, 21, 7993 10 of 15

still meets performance requirements as mentioned in Table 1. Furthermore, the mirror

surface shape error RMSGZ under gravity parallel to the optical axis increases from 1.3 nm
to 1.4 nm, which is far less than the design requirements. Moreover, the fundamental
frequency of the primary mirror assembly decreases from 594.1 HZ to 203.1 HZ so that
dynamic stiffness is sacrificed to a certain extent. However, it also satisfies the dynamic
stiffness requirements. Furthermore, the mass of the primary mirror assembly increased
Sensors 2021, 21, x FOR PEER
from REVIEW
1.3 kg to 1.6 kg mainly depends on the increase in thickness of the primary mirror 10

faceplate which is still acceptable.

(a) (b)

(c) (d)
Sensors 2021, 21, x FOR PEER REVIEW Figure of
Figure 6. Sensitivity 6. size
Sensitivity of size(a)parameters.
parameters. The mirror(a) The mirror
surface surface
shape error shape
RMS error
T . (b) 11
TheRMS ofT.16(b) The m
mirror
surface
surface shape error shape
RMS error RMSGR. (c) Fundamental frequency. (d) Mass of the primary component.
GR . (c) Fundamental frequency. (d) Mass of the primary component.

3.4.3. Performance Evaluations and Comparisons

After multiple iterations, the optimization response function under gravity par
and vertical to the optical axis and thermal load converged step by step as a whole.
iteration history of the mirror surface shape error in three cases is plotted in Figure 7
initial and final values of the system performance responses are summarized in Tab
The mirror surface shape error RMST (removing the piston, tilt and power terms) ca
by 30 °C uniform temperature decrease drastically diminishes from 26.5 nm to a minim
value 11.6 nm (less than λ/40), which is better than the performance requirements in T
1. Namely, thermal deformation resistance of the primary mirror assembly has been
viously improved. Figure 8 depicts the primary mirror surface error map under 3
uniform temperature decrease before and after optimization. The mirror surface s
error RMSGR under gravity vertical to the optical axis rises from 6.3 nm to 8.0 nm, w
still meets performance requirements as mentioned in Table 1. Furthermore, the m
surface shape error RMSGZ under gravity parallel to the optical axis increases from 1.3
to 1.4 nm, which is far less than the design requirements. Moreover, the fundamenta
Figure
Figure 7. Iterative
7. Iterative
quency history
history
of the of
of the the mirror
mirror
primary surface
surface
mirror shape
shape error.
error.
assembly decreases from 594.1 HZ to 203.1 HZ so tha
namic stiffness is sacrificed to a certain extent. However, it also satisfies the dynamic
Table 4. Comparison of initial and optimal primary mirror component design.
ness requirements. Furthermore, the mass of the primary mirror assembly increased
1.3 kgTerms
to 1.6 kg mainly depends on the increase
Initial Value in thickness of theValue
Optimized primary mirror
plate which
RMST (nm) is still acceptable. 26.5 11.6
RMSGR (nm) 6.3 8.0
RMSGZ (nm) 1.3 1.4
f1 (HZ) 594.1 203.1
Sensors 2021, 21, 7993 11 of 15
Figure 7. Iterative history of the mirror surface shape error.

Table 4. Comparison of initial and optimal primary mirror component design.

Table 4. ComparisonTerms
of initial and optimal primary mirror
Initialcomponent
Value design.
Optimized Value
RMST (nm)
Terms Initial Value 26.5 11.6
Optimized Value
RMSGR (nm) 6.3 8.0
RMST (nm) 26.5 11.6
RMSGR RMS (nm) GZ (nm) 6.3 1.3 8.0 1.4
RMSGZ (nm) f1 (HZ) 1.3 594.1 203.1
1.4
f 1 (HZ) Mass (kg) 594.1 1.28 203.11.59
Mass (kg) 1.28 1.59
Lightweight ratio (%) 45.8 33.3
Lightweight ratio (%) 45.8 33.3
MTF
MTFT (0 T (0° field angle)
◦ field angle) 0.15 0.15 0.210.21
MTFT (+0.24 ◦
MTFT (+0.24° field angle)
field angle) 0.13 0.13 0.170.17
−0.24T◦(−0.24°
MTFT (MTF field angle)
field angle) 0.14 0.14 0.160.16
MTFT (+0.34 ◦ field angle) 0.09 0.100.10
MTFT (+0.34° field angle) 0.09
MTFT (−0.34◦ field angle) 0.10 0.11
MTF (−0.34°
MTFGR (0◦Tfield field angle)
angle) 0.46 0.10 0.430.11
MTFGZ (0 field angle) angle)
MTF ◦ GR (0° field 0.47 0.46 0.460.43
MTFGZ (0° field angle) 0.47 0.46

(a) (b)
Figure 8.
Figure 8. Surface
Surface error
error map
map under
under 30
30 ◦°C uniform temperature
C uniform temperature decrease.
decrease. (a)
(a) Initial
Initial design.
design. (b)
(b) Opti-
Opti-
mized design.
mized design.

There are various

There variousways
waystotoevaluate
evaluatethethe
optical performance
optical of the
performance ofsystem. MTF (mod-
the system. MTF
ulus of theofOTF)
(modulus is employed
the OTF) in thisin
is employed paper to assess
this paper the optical
to assess performance
the optical of the camera
performance of the
after theafter
camera Zernike coefficients
the Zernike of the of
coefficients deformed surface
the deformed are directly
surface imported
are directly into into
imported the ray-
the
tracing software
ray-tracing Zemax
software Zemax [25].[25].
TheTheMTFMTF values described
values below
described areare
below thethe
average
averagevalue
valuein
in the tangential and sagittal planes [26–28]. The image size of the visible light detector
is 1920 × 1080 and the pixel size is 5.5 um corresponding to the cutoff frequency of MTF
is about 90 cycles per mm. The field of view (FOV) the camera is 0.6◦ (azimuth) × 0.34◦
(pitch). As shown in Figure 9, in wavelength bandwidth of 450~700 nm, the value of
MTF (0◦ field angle) under 30 ◦ C uniform temperature decrease has declined significantly
from 0.47 to 0.15 (before optimization) compared with optical design result (regardless of
structural distortion) even though Zernike terms of the piston, tilt and power are subtracted.
Compared to the initial design, the value of MTF (0◦ field angle) is improved from 0.15 to
0.21. The values of MTF (±0.24◦ , ±0.34◦ field angle, normalized ±0.7, ±1 field points) are
inordinately improved after optimization as listed in Table 4. In other words, the optical
performance of the camera under thermal load has been enhanced after optimization and
thermal adaptability of the primary mirror has been obviously reinforced. The values of
MTF (0◦ field angle) in the case of gravity parallel and vertical to the optical axis have
slightly declined.

Sensors 2021, 21, 7993
Compared to the initial design, the value of MTF (0° field angle) is improved from 0.15 to
0.21. The values of MTF (±0.24°, ±0.34° field angle, normalized ±0.7, ±1 field points) are
inordinately improved after optimization as listed in Table 4. In other words, the optical
performance of the camera under thermal load has been enhanced after optimization and
thermal adaptability of the primary mirror has been obviously reinforced. The values of
12 of 15
MTF (0° field angle) in the case of gravity parallel and vertical to the optical axis have
slightly declined.

9. Comparison
Figure 9. Comparisonofofthe
themodulus
modulusofof
thethe OTF
OTF (MTF, 0◦ field
(MTF, 0° field angle)
angle) curves
curves of initial
of initial and opti-
and optimized
mized design
design under
under 30 30 °C uniform
◦ C uniform temperature
temperature decrease.
decrease.

4. Experimental Verification
4. Experimental Verification
Based
Based on on the
the above
above optimized
optimized design
design results,
results, the
the prototype
prototype of of the
the primary
primary mirror
mirror
assembly was manufactured and assembled as shown in Figure
assembly was manufactured and assembled as shown in Figure 10. The precision 10. The precision of
of the
the structure
structure should
should be guaranteed,
be guaranteed, especially
especially the planarity
the planarity ofmounting
of the the mounting surface
surface is sug-is
suggested to be less than 3 nm. The invar sleeves were separately bonded
gested to be less than 3 nm. The invar sleeves were separately bonded into the support into the support
holes
holes of
of the SiC primary
the SiC primary mirror
mirror with
with the
the help
help of
of positioning
positioning tooling
tooling using
using epoxy
epoxy adhesive
adhesive
(GHJ-01Z),
(GHJ-01Z), and the flexures were also respectively fastened on the sleeves with bolts.
and the flexures were also respectively fastened on the sleeves with bolts. High
High
and
and low
low temperature
temperature cycle
cycle tests
tests and
and aa vibration
vibration test
test was
was used
used to
to release
release the
the stress
stress of
of the
the
primary
primary mirror
mirror component
component during
during machining,
machining, assembly
assembly andand gluing.
gluing.
Interferometry was performed on the primary mirror to acquire high measurement
precision with a ZYGO interferometer working at 632.8 nm wavelength. Zernike terms of
piston, tilt, power and coma which could be adjusted by alignment between mirror and
interferometer were removed. An optical compensator containing two spherical lenses was
adopted to correct the wavefront for measuring the aspheric surface of the primary mirror.
The photograph of the interferometry scenario of the SiC primary mirror with the optical
axis horizontal is displayed in Figure 11. The primary mirror surface shape error RMS after
Sensors 2021, 21, x FOR PEER REVIEW 13 of 16
processing and coating is less than 0.025 λ at room temperature of 23 ◦ C, which meets the
performance requirement. The interferometry test result is shown in Figure 12.

Figure10.
Figure 10.Prototype
Prototypeof
ofthe
theprimary
primarymirror
mirrorassembly.
assembly.

Interferometry was performed on the primary mirror to acquire high measurement

precision with a ZYGO interferometer working at 632.8 nm wavelength. Zernike terms of
piston, tilt, power and coma which could be adjusted by alignment between mirror and
interferometer were removed. An optical compensator containing two spherical lenses
was adopted to correct the wavefront for measuring the aspheric surface of the primary
mirror. The photograph of the interferometry scenario of the SiC primary mirror with the
optical axis horizontal is displayed in Figure 11. The primary mirror surface shape error
 piston, tilt, power and coma which could be adjusted by alignment between mirror and
interferometer were removed. An optical compensator containing two spherical lenses
was adopted to correct the wavefront for measuring the aspheric surface of the primary
mirror. The photograph of the interferometry scenario of the SiC primary mirror with the
optical axis horizontal is displayed in Figure 11. The primary mirror surface shape error
Sensors 2021, 21, 7993 13 of 15
RMS after processing and coating is less than 0.025 λ at room temperature of 23 °C, which
meets the performance requirement. The interferometry test result is shown in Figure 12.

Sensors 2021, 21, x FOR PEER REVIEW 14 of 16

Figure 11.
Figure Interferometryscenario
11. Interferometry scenario of
of the
the SiC
SiC primary
primary mirror
mirror with
with the
the optical
optical axis
axis horizontal.
horizontal.

Figure12.
Figure 12.Interferometry
Interferometrytest
testresult
resultof
ofthe
theprimary
primarymirror.
mirror.

Whensubjected
When subjectedto toaawide
widerange
rangeof ofambient
ambienttemperature
temperaturechanges,
changes,thetheimaging
imagingquality
quality
of the
of the high-resolution optical
optical system
systemwillwilldegrade
degradedifferently.
differently.A A ground
ground temperature
temperature ex-
experiment
periment was was conducted
conducted to to verify
verify difference
difference in imaging
in imaging quality.
quality. The The environmental
environmental tem-
temperature
perature change change simulation
simulation waswas carried
carried outout
in ainhigh-
a high-
andand low-temperature
low-temperature testtest cham-
chamber.
ber. A domestically
A domestically produced
produced targettarget
whichwhich
was was engraved
engraved withwith
blackblack
and andwhitewhite strips
strips was
was used to identify the resolution of the camera. Figure 13a shows
used to identify the resolution of the camera. Figure 13a shows the imaging photos taken the imaging photos
taken ◦ C.
by theby the visible
visible cameracamera as mentioned
as mentioned in Section
in Section 3.1 attemperature
3.1 at room room temperature
of 20 °C.ofIt20is not
Itdifficult
is not difficult to identify
to identify that the that the minimum
minimum distinguishable
distinguishable stripe
stripe is the 23rdis the 23rdThen
group. group.
the
Then
camera thewas
camera was aput
put into into
high anda high
low and low temperature
temperature test chambertest chamber and the tempera-
and the internal internal
temperature
ture is set tois−10 to −
set °C 10 ◦ C without
without thermalthermal
control control facilities.
facilities. After placed
After being being placed
for twoforhours,
two
hours, the internal temperature environment reached a state of thermal
the internal temperature environment reached a state of thermal equilibrium. The target equilibrium. The
target image obtained by the visible system is shown in Figure 13b. It
image obtained by the visible system is shown in Figure 13b. It can be seen that the small- can be seen that
the
est smallest discernible
discernible stripe isstripe is the
the 16th 16th group
group under under low temperature
low temperature condition.
condition. The
The result
result demonstrates that although the image quality of the camera
demonstrates that although the image quality of the camera degrades, it still meet thedegrades, it still meet
the requirements
requirements of the
of the index.
index.

(a) (b)
 camera was put into a high and low temperature test chamber and the internal tempera-
ture is set to −10 °C without thermal control facilities. After being placed for two hours,
the internal temperature environment reached a state of thermal equilibrium. The target
image obtained by the visible system is shown in Figure 13b. It can be seen that the small-
Sensors 2021, 21, 7993 est discernible stripe is the 16th group under low temperature condition. The14result of 15
demonstrates that although the image quality of the camera degrades, it still meet the
requirements of the index.

(a) (b)
Figure13.
Figure 13.Target
Targetimages
imagesacquired
acquiredatatdifferent
differenttemperatures.
temperatures.(a)
(a)AtAtthe temperatureofof2020◦ C.
thetemperature °C.(b)
(b)At
At
thetemperature
the temperatureofof−−10
10 ◦°C.
C.

5.5.Conclusions
Conclusions
The
Thepurpose
purpose of of this article
article isistotofurther
furtheroptimize
optimize thethe thermal
thermal adaptability
adaptability of sup-
of the the
supporting structure
porting structure of the
of the primary
primary mirror.
mirror. Integrated
Integrated optomechanical
optomechanical analysis
analysis andand op-
optimi-
timization
zation for for
the the primary
primary mirror
mirror assembly
assembly are are
put put forward.
forward. The The mirror
mirror surface
surface shape shape
error
error
caused caused by uniform
by uniform temperature
temperature decrease decrease is treated
is treated as the as the objective
objective function,
function, and theandfun-
the fundamental frequency of the mirror assembly and the surface shape
damental frequency of the mirror assembly and the surface shape error caused by gravity error caused by
gravity parallel or vertical to the optical axis are taken as the constraints.
parallel or vertical to the optical axis are taken as the constraints. A detailed size optimi-A detailed size
optimization is conducted
zation is conducted to optimize
to optimize its size
its size parameters.
parameters. TheThe contribution
contribution of size
of size parame-
parameters
ters
to the surface shape error, fundamental frequency and mass are further evaluated.The
to the surface shape error, fundamental frequency and mass are further evaluated. The
configuration
configurationofofthe theprimary
primarymirrormirrorand andthetheflexure
flexureareareacquired.
acquired.The
Theresults
resultsdemonstrate
demonstrate
that
thatthe
thesize
sizeparameters
parameterssignificantly
significantlyhave havean aneffect
effectononthe
theoptical
opticalperformance
performanceand andwhich
which
dimension
dimension parameters
parameters are
are of
of the
the critical
critical factors.
factors. The
The mirror
mirror surface
surface shape
shape error
error caused
caused by
by 30 ◦ C uniform temperature decrease effectively decreased from 26.5 nm to 11.6 nm,
although the weight of the primary mirror assembly increases by 0.3 kg. The mirror surface
shape error under gravity parallel and vertical to the optical axis grows slightly. The funda-
mental frequency of the primary mirror assembly decreases from 594.1 HZ to 203.1 HZ, but
still satisfies the dynamic stiffness requirements. The value of the system’s MTF (0◦ field
angle) is improved from 0.15 to 0.21 after optimization. The values of MTF (±0.24◦ , ±0.34◦
field angle, normalized ±0.7, ±1 field points) are inordinately improved after optimiza-
tion. In other words, the optical performance of the camera under thermal load has been
enhanced and thermal adaptability of the primary mirror has obviously been reinforced.
Finally, a prototype of the primary mirror assembly is manufactured and assembled. A
ground thermal test was developed to verify variation of imaging quality at room and low
temperature, respectively. The image quality of the camera meets the requirements of the
index despite degradation. This research provides a meaningful reference for the structural
optimization design of aerial optoelectronic sensors.

Author Contributions: Conceptualization, Q.L.; Data curation, M.Z.; Formal analysis, M.Z.; Funding
acquisition, H.T.; Investigation, M.Z.; Methodology, Q.L. and X.W.; Project administration, D.W.;
Resources, D.W.; Software, M.Z.; Supervision, Q.L. and X.W.; Validation, C.C.; Visualization, M.Z.;
Writing—original draft, M.Z.; Writing—review & editing, M.Z. All authors have read and agreed to
the published version of the manuscript.
Funding: This research was funded by the National Natural Science Foundation of China, grant
number 61905240.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Acknowledgments: The authors are thankful for the support from the Key Laboratory of Airborne
Optical Imaging and Measurement, Changchun Institute of Optics, Fine Mechanics and Physics,
Chinese Academy of Sciences.
Conflicts of Interest: The authors declare no conflict of interest.
Sensors 2021, 21, 7993 15 of 15

References
1. Gao, Y.; Zhang, B.; Chen, L.; Xu, B.; Gu, G. Thermal design and analysis of the high resolution MWIR/LWIR aerial camera. Optik
2019, 179, 37–46. [CrossRef]
2. Liu, F.; Cheng, Z.; Jia, P.; Zhang, B.; Liu, X.; Hu, R. Impact of Thermal Control Measures on the Imaging Quality of an Aerial
Optoelectronic Sensor. Sensors 2019, 19, 2753. [CrossRef] [PubMed]
3. Cheng, Z.; Sun, L.; Liu, F.; Liu, X.; Li, L.; Li, Q.; Hu, R.; Liu, L. Engineering Design of an Active–Passive Combined Thermal
Control Technology for an Aerial Optoelectronic Platform. Sensors 2019, 19, 5241. [CrossRef] [PubMed]
4. Xue, Z.-P.; Wang, C.-X.; Yu, Y.; Wang, P.-P.; Zhang, H.-Y.; Sui, Y.-Y.; Li, M.; Luo, Z.-Y. Integrated optomechanical analyses and
experimental verification for a thermal system of an aerial camera. Appl. Opt. 2019, 58, 6996–7005. [CrossRef]
5. Xie, Y.; Mao, X.; Li, J.; Wang, F.; Wang, P.; Gao, R.; Li, X.; Ren, S.; Xu, Z.; Dong, R.; et al. Optical design and fabrication of an
all-aluminum unobscured two-mirror freeform imaging telescope. Appl. Opt. 2020, 59, 833–840. [CrossRef] [PubMed]
6. Li, C.; Ding, Y.; Lin, C.; Wei, Y.; Zheng, Y.; Zhang, L. Optomechanical design and simulation of a cryogenic infrared spectrometer.
Appl. Opt. 2020, 59, 4642–4649. [CrossRef]
7. Liu, G.; Guo, L.; Wang, X.; Wu, Q. Topology and parametric optimization based lightweight design of a space reflective mirror.
Opt. Eng. 2018, 57, 075101. [CrossRef]
8. Shao, M.; Zhang, L.; Jia, X. Optomechanical integrated optimization of a lightweight mirror for space cameras. Appl. Opt. 2021,
60, 539–546. [CrossRef]
9. Kihm, H.; Yang, H.-S. Design optimization of a 1-m lightweight mirror for a space telescope. Opt. Eng. 2013, 52, 091806. [CrossRef]
10. Kihm, H.; Yang, H.S.; Moon, K.; Yeon, J.; Lee, H.S.; Lee, Y. Adjustable bipod flexure for mounting mirrors in a space telescope.
Appl. Opt. 2012, 51, 7776–7783. [CrossRef]
11. Liu, S.T.; Hu, R.; Li, Q.H.; Zhou, P.; Dong, Z.G.; Kang, R. Topology optimization-based lightweight primary mirror design method
of a large-aperture space telescope. Appl. Opt. 2014, 53, 8318–8325. [CrossRef]
12. Hu, R.; Liu, S.; Li, Q. Topology-optimization-based design method of flexures for mounting the primary mirror of a large-aperture
space telescope. Appl. Opt. 2017, 56, 4551. [CrossRef] [PubMed]
13. Jiang, P.; Zhou, P. Optimization of a lightweight mirror with reduced sensitivity to the mount location. Appl. Opt. 2020, 59,
3799–3805. [CrossRef]
14. Chen, Y.-C.; Huang, B.-K.; You, Z.-T.; Chan, C.-Y.; Huang, T.-M. Optimization of lightweight structure and supporting bipod
flexure for a space mirror. Appl. Opt. 2016, 55, 10382. [CrossRef]
15. Zhang, L.; Wang, T.; Zhang, F.; Zhao, H.; Zhao, Y.; Zheng, X. Design and optimization of integrated flexure mounts for unloading
lateral gravity of a lightweight mirror for space application. Appl. Opt. 2021, 60, 417–426. [CrossRef]
16. Mahajan, V.N. Zernike Circle Polynomials and Optical Aberrations of Systems with Circular Pupils. Appl. Opt. 1994, 33, 8121–8124.
[CrossRef]
17. Mahajan, V.N. Zernike Polynomial and Wavefront Fitting. In Optical Shop Testing; John Wiley & Sons: Hoboken, NJ, USA, 2006.
[CrossRef]
18. Liu, X.; Tian, X.; Zhang, W.; Zhang, B.; Cheng, Z.; Fu, L.; Wang, Z. Lightweight design of high volume SiC/Al composite mirror
for remote camera. Optik 2019, 188, 64–70. [CrossRef]
19. Yoder, P.R. Opto-Mechanical Systems Design; Taylor & Francis: Boca Raton, FL, USA, 2006; pp. 420–423.
20. Zhou, P.; Xu, S.; Yan, C.; Zhang, X. Research on neutral surface of lightweight, horizontally supported mirror. Opt. Eng. 2018,
57, 025107. [CrossRef]
21. Zhou, P.; Zhang, D.; Liu, G.; Yan, C. Development of space active optics for a whiffletree supported mirror. Appl. Opt. 2019, 58,
5740–5747. [CrossRef] [PubMed]
22. Liu, B.; Wang, W.; Qu, Y.J.; Li, X.P.; Wang, X.; Zhao, H. Design of an adjustable bipod flexure for a large-aperture mirror of a space
telescope. Appl. Opt. 2018, 57, 4048–4055. [CrossRef]
23. Smith, S. Flexures Elements of Elastic Mechanisms; Gorden and Breach Science: Amsterdam, The Netherlands, 2000; pp. 308–310.
24. Kihm, H.; Yang, H.S.; Lee, Y. Opto-mechanical analysis of a 1 m lightweight mirror system. J. Korean Phys. Soc. 2013, 62, 1239–1256.
[CrossRef]
25. Bavinder, R.; Ravindra, B.; Chatterjee, S. Opto-thermal analysis of a lightweight mirror for solar telescope. Opt. Express 2013, 21,
7065–7081. [CrossRef]
26. Yuan, L.; Xie, J.; He, Z.; Wang, Y.; Wang, J. Optical design and evaluation of airborne prism-grating imaging spectrometer. Opt.
Express 2019, 27, 17686–17700. [CrossRef] [PubMed]
27. Bai, Y.; Li, J.L.; Zha, R.W.; Wang, Y.; Lei, G.Z. Catadioptric optical system design of 15-magnitude star sensor with large en-trance
pupil diameter. Sensors 2020, 20, 5501. [CrossRef] [PubMed]
28. Fan, X.; Zou, G.; Qiu, Y.; Pang, Z.; Zhao, H.; Chen, Q.; Pan, Y.; Yuan, H.; Gangyi, Z. Optical design of the visible telescope for the
SVOM mission. Appl. Opt. 2020, 59, 3049–3057. [CrossRef] [PubMed]