Received September 30, 2018, accepted October 15, 2018, date of publication November 26, 2018,

date of current version December 27, 2018.

Digital Object Identifier 10.1109/ACCESS.2018.2883363

Modal and Dynamic Analysis of a Tether

for a Nonequatorial Space Elevator
ZHENKUN WANG 1, NAIGANG CUI2 , YOUHUA FAN1 , AND DUN LIU2
1 School of Science, Harbin Institute of Technology, Shenzhen 518055, China
2 Institute of Astronautics, Harbin Institute of Technology, Harbin 150001, China

Corresponding author: Youhua Fan (yhfan@hit.edu.cn)

ABSTRACT To study the fundamental dynamics of the nonequatorial space elevator, a universal modeling
strategy is presented, which applies to any anchored latitude. The equations of motion are established on the
basis of the lumped mass model of the tether coupled with an interactional elevator. The relationship between
the maximum anchored latitude and the design stress is studied by solving the equilibrium equations. The
linearized equations of motion are formulated on the basis of the static equilibrium positions to analyze
the modal of the tether. Numerical simulation is performed to analyze the properties of the system with
a moving elevator. The results show that a positive correlation arises between the maximum anchored
latitudes and the design stress. The fundamental modal and the fifth-order modal are associated with the
latitudes considering the frequency and modal shape, respectively. Furthermore, the asymptotically stable
characteristics are verified. The amplitude of the residual oscillation is magnified with an increase in the
anchored latitude or the elevator speed.

INDEX TERMS Space elevator, nonequatorial, modal analysis, dynamic response.

I. INTRODUCTION completed in approximately 2037, and operations may begin

The concept of a traditional tethered space elevator involves a
long tether/ribbon with one end attached to the Earth’s surface
somewhere along the equator and the other end attached to a
counterweight that is located beyond geosynchronous (GEO)
orbit. Payloads can be transported to an arbitrary orbit without
using conventional transportation methods [1]–[3]. To be
viable, the tether must have a very high strength-to-density
ratio to withstand the large tension force applied by the
counteracting gravitational and centripetal loads. The current
tether materials considered are carbon nanotube composite.
Fundamentally, the taper function of a tether with a uniform
stress was first derived by Pearson [1], although the nominal
strain was not considered. Subsequently, Cohen and Misra [4]
derived another taper function that considered the tether’s
longitudinal deformation. Since that time, some dynamic
analyses on the traditional space elevator have been carried
out. Such dynamic analyses include system stability on the
basis of the modal analysis [4], [5], dynamic response with
a moving elevator on the tether [5]–[8] and the suppres-
sion of oscillation by control or optimization of the elevator
motion [7], [9], among others. Thus far, the research on tradi-
tional tethered space elevators has achieved rewarding results;
the construction of the space elevator system is expected to be
thereafter [10].
To date, the vast majority of studies have focused on the tra-
ditional space elevator being anchored on the equator because
the equilibrium of the tether is easiest to understand in this
case. However, the heavily occupied geosynchronous orbit
increases the risk of collision [5], [11]. In addition, the tether
should avoid areas involving the most intense radiation of
the Van Allen belts [12], which is particularly important for
the transportation of humans. In fact, nonequatorial space
elevators, which have more flexibility in the selection of
the anchor location, can help address the abovementioned
problems. Their major difference with the equatorial ele-
vators is that the tether is located entirely on one side of
the equatorial plane. Therefore, the tether would also be
inclined towards the equatorial plane owing to gravity. The
taper function for the tether/ribbon is still the foundation of
dynamic analyses. Gassend [13] presented the continuous
taper function of a uniform-stress tether in a general potential
field. However, the tether’s spatial configuration is difficult
to obtain by analytical means for the nonequatorial space
elevator. Cowan and Leonard [14] calculated the tether’s
taper profile on the basis of a finite segments model of the
tether, which did not need to consider spatial configuration.

2169-3536
2018 IEEE. Translations and content mining are permitted for academic research only.
74940 Personal use is also permitted, but republication/redistribution requires IEEE permission. VOLUME 6, 2018
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Z. Wang et al.: Modal and Dynamic Analysis of a Tether for a Nonequatorial Space Elevator
Indeed, it is recommended that the model used to calculate
the tether’s taper profile and spatial configuration should be
consistent with the dynamic analysis to avoid calculation
error. Therefore, the finite segments model will be simplified
as a lumped mass tether model that is sufficiently precise
and widely used in the dynamic analysis of tethered space
systems.
This paper is concerned with the dynamic properties of a
nonequatorial space elevator system with a moving elevator
on the tether. Thus, a universal model applicable to any
latitude is established and described in Section II. Then, these
equations are linearized by applying the cosine theorem and
the binomial theorem on the basis of the equilibrium posi-
tion to perform a modal analysis in Section III. Meanwhile,
the configurations and the maximum deployment latitude
of the tether are studied in this section. Section IV describes
the investigation of the effect of the latitudes on the system
dynamic response with a propelled elevator moving along
the tether at different velocities. Finally, the conclusions and
planned future work are presented in Section V.
II. MATHEMATICAL MODEL
Tether modeling is a fundamental aspect with extensive
supporting literature, pertaining in particular to the studies
of tethered satellites. From another perspective, the space
elevator is an extreme version of a tethered satellite [15].
In the modeling approaches for the tether, the lumped-mass
model has been used extensively for the tethered system
in the space environment [5], [16]–[20]. The lumped mass
approaches physically discretize the tether/cable into a sig-
nificant number of discrete masses. The adjacent masses are
connected by viscoelastic springs. This modeling approach
is employed herein. To ensure the model has high fidelity,
the full interaction between the elevator and the tether should
be considered [5]. In this study, the elevator is integrated into
the system as an independent lumped mass.
FIGURE 1. Lumped mass model of the nonequatorial space elevator.
A. MODEL DESCRIPTION AND TETHER DISTRIBUTION
The nonequatorial space elevator system is the generalized
form of the traditional one. Similarly, the model comprises a
tether, elevator and counterweight, as shown in Fig. 1. The
symbol θ0 is used to describe the anchored latitude of the
tether. The tether has a nominal (unstressed) length of L. With
the application of the lumped-mass model, n masses are used
to represent the tether. The masses are numbered beginning
from the Earth side and ending at the counterweight side.
VOLUME 6, 2018
The counterweight is also idealized as a point mass mc rel-
ative to the tether, which has a large length. An elevator
carrying cargo having a total mass me moves along the tether
into space. The entire system rotates around the Earth’s axis
with the Earth’s rotation angular velocity ωE .
Let EXYZ be the rotating
coordinate frame at the Earth
center. e = e1 , e2 , e3 is the base vector of the coor-
dinate frame. ri is the vector from the origin of EXYZ to
the i-th lumped mass. Then, the displacement of the lumped
mass i can be expressed as
ri = r Ti e (1)
where ri = xi , yi , zi is the projection of the lumped mass

in the Earth rotating coordinate frame.
FIGURE 2. Nonuniform mass distribution in the tether.
For the discrete model to be more reasonable, the tether
should be tapered because of the uniform stress. Hence,
the discrete masses that are distributed uniformly along the
tether are usually nonuniform masses. The lumped masses are
treated as if they are located in the center of each segment, and
the adjacent two masses are joined together with a viscoelas-
tic spring, as shown in Fig. 2. The mass of the i-th element is
calculated as follows:
γ li
mi = (Ai + Ai+1 ) (2)
2
where γ is the density of the tether material, Ai represents the
cross-sectional area on the left side of the i-th segment, and li
is the unstrained length of the i-th segment.
The relative displacement d i of two lumped masses rep-
resents the tensile state of the spring. The unstrained length
of each spring is defined as Li = (li−1 + li )/2. Note that
the spring tension force equals zero when the modulus of
the relative displacement is less that the unstrained length,
i.e., di < Li , because of the flexibility of the tether. The
relative displacement of the i-th mass from the (i−1)-th mass
is written as:
d i = ri − ri−1 (3)
where r0 is the displacement of the anchorage on the Earth
surface.
B. DYNAMICS EQUATIONS OF THE TETHER
The derivation of the model is presented in the rotating
coordinate system, as shown in Fig. 1. A straightforward
Newtonian approach is used to derive the equations of
motion. The displacement of each lumped mass containing
74941
 Z. Wang et al.: Modal and Dynamic Analysis of a Tether for a Nonequatorial Space Elevator
a counterweight was previously described in (1). Hence,
the inertial acceleration vector of any mass is the second-
order derivative of the displacement as
ai = r̈i + ω̇ × ri + 2ω × ṙi + ω × (ω × ri ) (4)
The angular velocity matrix, i.e., ω = 0, 0, ωE is

obtained from the transformation matrix of the coordinate
frame.
The dynamics equations of these masses are derived under
the assumption that only the gravity, tension and damping
forces are present. However, there are various disturbance
factors from the real environment, such as J2 gravitational
perturbation, lunisolar perturbations, and the wind load.
We plan to study these influences in the future. The dynamics
equations of the system (no elevator) can be written as:
g relationships of the tether length on one side of the eleva-
mi ai = Fi + Fsi + Fdi (5)
where the superscripts (g), (s) and (d) represent gravity,
spring, and damping forces, respectively. These forces are
defined in the following subsections.
1) GRAVITY FORCES
The force of gravity is calculated on the basis of the instan-
taneous displacement of the discrete mass. The expression is
written as:
g µmi
Fi = − 3 ri (6)
|ri |
where µ is the gravitational constant of the Earth.
2) TENSION FORCES
Tether tension produces a tether tensile state. In this case, the
corresponding tension can be determined via Hooke’s law
as Ti = Ki (|d i | − Li ), where Ki = EAi /Li is the spring
stiffness of the tether element and E is the Young’s modulus
of elasticity. However, no tension force exists when the tether
is in a state of relaxation. The tension forces acting on each
mass can be written as:
K (|d | − L ) d i , |d | ≥ L

i i i i i aspect is not addressed, the oscillation of the tether could
Ti = |d i | (7)
0, |d i | < Li

Note that Equation (7) must be modified because of the
presence of a moving elevator. The modified equation is
presented in (14). The resultant spring force on the i-th mass is
Fsi = T i+1 − T i (8)
3) DAMPING FORCES
The damping forces acting on the masses can help attenuate
the vibration energy and improve computational stability.
To calculate the damping force, it is assumed that there is
a relationship between the damping force and relative dis-
placement rate, i.e., Di = C ḋ i · d i /|d i |, where C is the
tension force in the case of tether relaxation. The expression
for the damping force is:

C ḋ i · d i d , |d | ≥ L

i i i
Di = |d i |2 (9)
|d i | < Li

0,
The total damping force acting on mass i is obtained as:
Fdi = Di+1 − Di (10)
Thus, the dynamic equations of the free-moving system
(no elevator) are considered established.
C. COUPLING OF ELEVATOR DYNAMICS
In many models of space-tethered systems, the elevator is
integrated into the tether dynamic equations via kinematic
tor [7]–[9], [21]–[24]. Thus, it is challenging to reflect the
local coupling oscillation between the elevator and tether
segment. To fully describe the dynamic characteristics of
the elevator, the coupling of the elevator with the tether
requires special consideration. The elevator is also regarded
as a lumped mass that contacts the tether at a point. It can be
imagined that the main coupling forces are the spring tension
forces and the friction forces. However, only gravity and the
propulsive forces are considered external forces. The motion
of the elevator is determined by the following equation:
me ae = FN f g p
e + Fe + Fe + Fe (11)
where the superscripts (N ), (f ) and (p) represent normal
force, friction force and propulsive force, respectively. Note
that the Coriolis forces will cause a bowing effect on the tether
when the elevator moving along the tether. A normal force is
produced in this case. The detailed derivation of these forces
is given in the following subsections.
1) DEFINITION OF DIRECTION OF FORCES
The elevator will contact various tether segments as it moves
along the tether. Thus, the external forces on the elevator
change at each instant with the position of the elevator. If this
adversely affect the motion of the elevator. Williams [5] com-
puted the forces on the basis of two vectors of the current posi-
tion of the elevator on the local tether segment, i.e., q1 and q2 ,
as shown in Fig. 3.
q1 = re − ri−1 (12)
q2 = ri − re (13)
A more important assumption is that the unstrained length
of the tether element is segmented via the current position of
the elevator. According to the principle of the series spring,
we can verify that the tension forces on both sides of the ele-
vator are equal. The value of the tension forces is calculated
as follows
Te = Ki∗ q1  + q2  − Li∗
   

damping constant. The damping force is the same as the (14)

74942 VOLUME 6, 2018

Z. Wang et al.: Modal and Dynamic Analysis of a Tether for a Nonequatorial Space Elevator

Here, a unit transition vector between mi and mi+1 is used

to define the other vectors of the local coordinate frame
ri+1 − ri−1
tb = (17)
|ri+1 − ri−1 |
The normal vector n at the point of contact is redefined as:
kN × t b
n= (18)
|kN × t b |
The tangent vector t at the point of contact is redefined as:
kN × n
t= (19)
|kN × n|

2) NORMAL FORCE
The normal forces acting on the elevator are the tensions
FIGURE 3. Local elevator tether segment. (a) Local coordinate frame and of the two subsegments of the local elevator tether segment
mutation of the vectors. (b) Modified local coordinate frame.
projected onto the direction of the normal vector.
" ! #
Equation (14) illustrates that the direction of the normal N q2 q1
Fe = Te   −   ·n n (20)
force is along the direction of a normal vector n that bisects q2  q1 
the angle formed by −q1 and q2 . The direction of the friction
This expression allows for two cases when the normal
force and the propulsive force is defined as the direction of
force exerted on the elevator is zero. In one case, the tensions
the tangent at the contact point. Hence, a plane composed of
in the local elevator tether segment equal zero. In the other
normal and tangential vectors is used to define the external
case, the subsegments in the local elevator tether segment are
forces. This local coordinate frame is formed by the unit
parallel.
vector kN normal to the plane, the normal vector n and the
tangent vector t, as shown in Fig. 3 a.
3) FRICTIONAL FORCE
Special attention should be paid to the emergence of sin-
In most elevator structures, the elevator is propelled by pinch
gularities when the elevator is close to a lumped mass [5].
rollers. Generally, a frictional force, i.e., traction, braking or
In such a case, the stiffness of the spring tends towards
parking, exists during the entire process of elevator operation.
infinity because of q1 or q2 approaching zero in (14). To solve
The kinetic frictional force is generated when relative slip
this problem, a proximity tolerance is defined. The value
occurs between the pinch rollers and the tether/ribbon. In this
of the proximity tolerance, Rmin , should be appropriate for
case, reactions of the tension forces are generated as the
the conditions. The vector q2 is redefined as the vector
result of ribbon deformation. The normal force was used to
between me and mi+1 when the elevator is within the scope
calculate the kinetic friction in [5]; however, this factor has
of the tolerance. The corresponding variables in (14) are
only a small impact on the coupling movement.
adjusted accordingly. For instance, the spring stiffness Ki∗ is
At this moment, the frictional forces depend on the gravity
adjusted as:
and the propulsive force as the elevator is parked in a certain
Ki Ki+1 position of the tether. The friction is generated tangentially to
Ki∗ = (15) the tether in the direction of the vector t, as defined by
Ki + Ki+1
Ffe = − Fge + Fpe · t t, |ve | = 0


The unstrained length Li∗ is also adjusted accordingly. (21)
This method of dealing with the singularities may lead to
4) EVALUATION CONDITIONS OF THE PASSING POINT MASS
the speed oscillation problem of the elevator in the simulation.
The cause of the oscillation is the direction mutation of As the elevator moves along the tether, the conditions under
vectors t and n. The phenomenon is more prominent with which the elevator arrives at the next tether segment should be
increasing elevator altitude along the tether. established. The strategy adopted in this paper is to compute
To reduce the oscillation problem generated, the vec- the modulus of the vectors q1 and q2 and the angle β between
tors t and n are modified in this section. The modified local the two vectors. The local elevator tether segment will be
coordinate system is shown in Fig. 3 b. The unit vector kN , updated as the following conditions are satisfied:
which is normal to the plane containing the vectors t and n, q1  > q2  and cos β < 0
   
(22)
remains unchanged; thus, the resulting expression is as
follows: This procedure avoids situations in which the tether may be
coiled. In particular, the local elevator segment may be retro-
q × q2
kN =  1  (16) gressed as a result of the oscillation of the lumped masses.
q1 × q2 
This procedure is sufficient for all situations.

VOLUME 6, 2018 74943

 Z. Wang et al.: Modal and Dynamic Analysis of a Tether for a Nonequatorial Space Elevator

III. MODAL ANALYSIS the nonlinear term 1/|ri |3 in the gravity equation can be
Equation (5) cannot be directly used to perform a modal linearized via the cosine theorem  and binomial theorem
analysis because of some nonlinear terms in the force of 1 1 3Ri · ρ i
= 1− (25)
equations. The cosine theorem and binomial theorem are |ri |3 |Ri |3 |Ri |2
used to linearize the kinetic equations around the equilibrium The gravity equation after linearization can be obtained as
position to obtain the Jacobian of the system (no elevator). follows by substituting (23) and (25) into (6), ignoring the
The equations of an equilibrium configuration can also be small high-order nonlinear terms, and simplifying:
achieved from the static equilibrium equations. Then, a modal µmi
3µmi
g
analysis can be performed by calculating the eigenvalues Fi = − 3
Ri + ρ i + Ri · ρ i (26)
|Ri | |Ri |5
and eigenvectors of the Jacobian, through which the stability
of the system for a particular tether configuration can be where Ri is the dyadic tensor of the static equilibrium dis-
simultaneously revealed. placement Ri , calculated as follows:
Ri = Ri Ri = eT Ri RTi e (27)
The equations of tether tension and damping force are
linearized in the same way. The nonlinear term 1/|d i |
in (7) and (9) can be simplified as
1Ri · 1ρ i
 
1 1
= 1− (28)
|d i | |1Ri | |1Ri |2
where
1Ri = Ri − Ri−1 (29)
1ρ i = ρ i − ρ i−1 (30)
The equations of the tension and damping forces after
linearization
 are obtained
 as  
Li Li
T i = Ki 1 − 1Ri + Ki 1 − 1ρ i
|1Ri | |1Ri |
Li
+ Ki 1Ri · 1ρ i , |d i | ≥ Li (31)
|1Ri |3
C 2C
Di = 2
1Ri · 1ρ̇ i + 1Ri · 1ρ i 1ρ̇ i , |d i | ≥ Li
|1Ri | |1Ri |2
FIGURE 4. Instantaneous displacement on the basis of the equilibrium (32)
position.
where
A. LINEARIZATION OF THE EQUATIONS OF MOTION 1Ri = eT 1Ri 1RTi e (33)
In this section, the equations are linearized on basis of the
equilibrium position and the relative displacement of the Once (24)–(32) are solved, the results can be substituted
lumped masses, as shown in Fig. 4. Ri is the static equilibrium into (5) for each mass. Then, the system (no elevator)
displacement of the lumped masses in the rotating coordinate dynamic equations can be simplified as
frame. ρ i is the relative displacement of mi between the mi ρ̈ i
µmi
 
instantaneous position and the static equilibrium position. Li+1
= −mi ω×(ω×Ri ) − Ri +Ki+1 1− 1Ri+1
Hence, the instantaneous displacement of mi in the rotating |R |3 |1Ri+1 |
coordinate system can be defined as   i
Li C
− Ki 1 − 1Ri − 2mi  × ρ̇ i − 1Ri
ri = Ri + ρ i (23) |1Ri | |1Ri |2
C 2C
· 1ρ̇ i + 1Ri+1 · 1ρ̇ i+1 − 1Ri ·1ρ i 1ρ̇ i
By substituting (23) into (4) and simplifying, the equations |1Ri+1 |2 |1Ri |2
of motion of the tether can be rewritten as 2C
1Ri+1 · 1ρ i+1 1ρ̇ i+1 − mi  ×  × ρ i


+ 2
ai = ρ̈ i + 2ω × ρ̇ i + ω × ω × ρ i + ω × (ω × Ri ) (24)

|1Ri+1 |
µmi
 
3µmi Li
However, the equations of gravity, spring, and damping − ρ i + Ri · ρ i − Ki 1 − 1ρ i
|Ri |3 |Ri |5 |1Ri |
forces are complex nonlinear equations in the tether dynamic Li

Li+1

equation. Hence, a premise condition is proposed in which − Ki 3
1Ri · 1ρ i + Ki+1 1 − 1ρ i+1
|1Ri | |1Ri+1 |
the mode of the relative displacement ρ i is significantly less
Li+1
than the mode of the staticequilibrium displacement Ri for + Ki+1 1Ri+1 · 1ρ i+1 (34)
the lumped mass mi (i.e., ρ i /|Ri |  1). In such a case, |1Ri+1 |3


74944 VOLUME 6, 2018

Z. Wang et al.: Modal and Dynamic Analysis of a Tether for a Nonequatorial Space Elevator

B. TETHER RADIAL EQUILIBRIUM CONFIGURATIONS

Care must be exercised during the implementation of modal
and dynamic analyses because the constant term, which
denotes only the static equilibrium displacement, should
equal zero, i.e.,
µmi
 
Li+1
−mi ω × (ω × Ri ) − Ri + Ki+1 1− 1Ri+1
|Ri |3 |1Ri+1 |
 
Li
− Ki 1 − 1Ri = 0 (35)
|1Ri |
The discrete approach described by (35) is used to study
the static characteristics of the nonequatorial space elevator,
which also correspond to the range of deployed latitudes in
the equilibrium configurations. For a uniform-stress tether,
the nominal strain can be written as ε0 = σ0 /E, where σ0
is the design stress of the tether. Thus, the length of the nomi-
nally stretched tether element is given by |1Ri | = Li (1 + ε0 ). FIGURE 6. The inclination of an earth space elevator varies with the
To describe the bending deformation of the tether, the droop anchored latitudes of the design stresses.

angle αi is introduced, as shown in Fig. 5.

FIGURE 5. Lumped masses under static equilibrium state.

Numerical simulations are conducted to calculate the

tether’s spatial configuration and the maximum deployment
latitude by solving (35). One parameter, typically the initial
cross-sectional area or the counterweight’s mass, should be
defined prior to solving (35). In the current cases, the ini-
tial cross-sectional area of a 5-mm2 tether is adopted. The
unstrained length of the total tether and each tether element
are 100,000 km and 100 km for the equatorial space elevator,
respectively. The density of the tether material is 1,100 kg/m3 .
The counterweights are located at the same height to ensure
that the ratios of the counterweight mass to the tether mass
are approximately equal. Note that this value is important to
the coupling of the counterweight motion to the tether modes,
which is largely a function of the tether tension and the mass
ratio. These equations can be solved using the least squares
method. MATLAB has a tool package called FSOLVE that
is available to implement the abovementioned method. The
simulation results are shown in Fig. 6 and Fig. 7.
Fig. 6 shows the results of the inclination (i.e., droop angle
at anchor α1 ) as a function of deployed latitude. The incli-
nation increases with the increase in the deployed latitudes.
Each line represents a different value of design stress, and
the red line denotes the limit case of the maximum reachable
latitude. The maximum reachable latitude occurs when the
tether and the ground are tangential at the anchorage point
(i.e., αmax ≤ 90◦ − θ0 ), implying that the system can only
support itself. The results show that the maximum reachable
latitude is sensitive to the effects of the design stress. Note that
FIGURE 7. System equilibrium configurations.
for a design stress of 20 GPa, the maximum reachable latitude
is approximately 21.5◦ . However, the value of the maximum
reachable latitude increases to 51.8◦ when the design stress
increases to 70 GPa. These results depend greatly on the
material of the tether.
Fig. 7 shows the equilibrium configurations for several
deployed latitudes. Note that the scale is different along the
x- and y-axes. The design stress is set to 60 GPa to achieve a
wider deployment range. The latitudes start in the range from
0◦ to 40◦ . The ratio of the counterweight mass to the tether
mass is approximately 0.742. From the above two diagrams,
the results suggest that avoiding the heavily occupied GEO
orbit is easy. However, the tether requires a higher design
stress to anchor at higher latitudes to avoid the most intense
areas of the radiation belts, which extend over 2,000 km above
the equatorial plane.
C. NUMERICAL SIMULATION OF MODAL ANALYSIS
A modal analysis is performed using (34). The system
Jacobian can easily be obtained when the equilibrium config-
uration is achieved using the methodology described earlier.
The stability information and modal data of the system are

VOLUME 6, 2018 74945

 Z. Wang et al.: Modal and Dynamic Analysis of a Tether for a Nonequatorial Space Elevator
revealed from the eigenvalues and eigenvectors of the system
Jacobian.
To study the influence of the deployment latitude to the
modal of the tethered system (no elevator), the parameters of
the system are consistent with Fig. 7. Three kinds of tether
spatial configuration are used to perform a modal analysis.
Note that the system is undamped, including the longitudinal
damping of the tether material.
Figs. 8–14 show the first seven modes of the system
anchored at various latitudes. The amplitude of vibration is
arbitrary due to the linear vibration hypothesis.
FIGURE 8. Mode 1 of the space elevator anchored at different latitudes:
(a) θ0 = 0◦ ; (b) θ0 = 15◦ ; and (c) θ0 = 30◦ .
FIGURE 9. Mode 2 of the space elevator anchored at different latitudes:
(a) θ0 = 0◦ ; (b) θ0 = 15◦ ; and (c) θ0 = 30◦ .
FIGURE 10. Mode 3 of the space elevator anchored at different latitudes:
(a) θ0 = 0◦ ; (b) θ0 = 15◦ ; and (c) θ0 = 30◦ .
FIGURE 11. Mode 4 of the space elevator anchored at different latitudes:
(a) θ0 = 0◦ ; (b) θ0 = 15◦ ; and (c) θ0 = 30◦ .
Fig. 8 shows the system fundamental mode, which is
a pendular mode centered on the anchorage point. In this
mode, the counterweight moves with the maximum ampli-
tude only in the XY plane of the rotating frame, irrespective
of the system inclination. The oscillation period is approxi-
mately 124 h, with a negative real part of the eigenvalue of
−1.974×10−10 for the equatorial space elevator, whereas the
period is approximately 139.5 h when θ0 = 30◦ . Similarly,
the second-order mode is also a pendular mode, except that it
74946
FIGURE 12. Mode 5 of the space elevator anchored at different latitudes:
(a) θ0 = 0◦ ; (b) θ0 = 15◦ ; and (c) θ0 = 30◦ .
FIGURE 13. Mode 6 of the space elevator anchored at different latitudes:
(a) θ0 = 0◦ ; (b) θ0 = 15◦ ; and (c) θ0 = 30◦ .
FIGURE 14. Mode 7 of the space elevator anchored at different latitudes:
(a) θ0 = 0◦ ; (b) θ0 = 15◦ ; and (c) θ0 = 30◦ .
is perpendicular to the equatorial plane shown in Fig. 9. The
period of this mode is not sensitive to the deployed latitudes,
and the value of the period is approximately 23.5 h with a
negative real part of the eigenvalue of −5.515×10−9 . This
period is slightly less than the one of the Earth’s rotation.
Fig. 10 shows the third-order mode of the system, which
is similar to the second bending mode of a cantilever beam
in the equatorial plane. In fact, the tethered system can be
regarded as a curved cantilever beam for the nonequato-
rial space elevator, in which one node is located adjacent
to the counterweight. The period is approximately 8.1 h
for the equatorial space elevator. There is a small increase
of 0.05 h in the period when the system is deployed in
a direction of 30◦ north or south. The fourth-order mode,
shown in Fig. 11, has a mode shape that is similar to that of
the third one. However, the vibration plane is perpendicular
to the equatorial plane. The periods in the three cases are
7.67 h, 7.7 h and 7.8 h. Hence, except for the case of the first
pendular mode, the modes are not sensitive to the deployed
latitude. The third and fourth modes have a negative real part
of the eigenvalue, which indicates that these modes are stable.
The fifth-order mode, shown in Fig. 12, demonstrates
some differences, mainly embodied in the tether longitudinal
direction of the movement for the equatorial space elevator,
whereas the mode contains the coupling movement on the
Z-axis direction for the nonequatorial space elevator. The
mode shapes, shown in Figs. 12 b and c, are similar to the
third-order bending mode of a cantilever beam. The position
of one node is near the position where the inclination of tether
is leveling off, whereas another node is near the position
VOLUME 6, 2018
Z. Wang et al.: Modal and Dynamic Analysis of a Tether for a Nonequatorial Space Elevator
adjacent to the counterweight. This mode is also stable, with
periods of approximately 6.44 h, 6.5 h and 6.7 h.
Figs. 13 and 14 show 6th and 7th modes respectively, which
correspond to the third bending mode of a cantilever beam in
the directions of two mutually vertical planes. It is obvious
that the node on the right is much closer to the counterweight
than the one of third-order and fourth-order modes. However,
another node is much closer to the middle of the tether than
it is for the fifth mode, as shown in Figs. 12 b and c. The
periods of the modes are 4.16 and 4.1 h for the equatorial
space elevator. It is evident that the period is not sensitive
to the deployed latitude. Notably, the peak of the amplitude
closer to the counterweight is larger than the one closer to the
Earth.
From the abovementioned linear modal analysis, the coun-
terweight apparently oscillates in conjunction with the tether
modes. Except for the first-order modes, the periods of the
modes are insensitive to the anchored latitudes. In addition,
the motion of the fifth-order mode couples with the motion
along the Z-axis for a nonequatorial space elevator on the
basis of an equatorial space elevator. It is notable that the
system is stable for all modes.
IV. DYNAMIC RESPONSE OF THE SYSTEM WITH AN
ELEVATOR CLIMBING TO GEO ORBIT
To study the dynamic response of the system as an eleva-
tor climbs to GEO orbit, various numerical simulations are
implemented, with consideration of the nonequatorial space
elevator in particular. To force the elevator to run at a specified
speed profile, a proportional-integrator (PI) control method
is used to correct the propulsive force. The speed profile of
the elevator is prescribed directly in the kinematic analysis.
Then, a parametric study of the effects of climbing speed and
deployed latitude on the tether response is carried out.
A. ELEVATOR MOTION AND CONTROL
In this section, the speed of the elevator V (t) is prescribed
using a particular type of ascension from its initial altitude Hi
to its final one Hf , and a PI controller is used to accommodate
the elevator moving with the prescribed speed profile when
the elevator is propelled by an external force.
The principle of designing the speed of the elevator is to
make it continuous and without mutation. The speed curve of
the elevator is divided into three phases of duration: Ta , Tc
and Td [8]. In phase 1 and phase 3, the acceleration varies
according to the half-sinusoid curve. Hence, the speed of the
elevator could change smoothly and continuously from zero
to cruise speed, or from cruise speed to zero. In phase 2,
the elevator maintains this cruise speed. The expression is
defined as:
πt
   
vc

 1 − cos , 0 ≤ t ≤ Ta


 2 Ta
v , T ≤ t ≤ T

V (t) = v 
c a  ac (36)
π (t − Ta − Tc + Td )




c
1 − cos , algorithm is not good, although it can meet the following
2 Td



Tac ≤ t ≤ Tad
VOLUME 6, 2018
where vc is the cruise speed, Tac = Ta +Tc and Tad = Tac +Td
are the time parameters.
The expressions for acceleration can be determined
through differentiation. Note that the acceleration initially
equals zero at both the breakpoints and the end of the motion.
In practice, the duration of the acceleration phase should
be reasonable after determining the operating power and air
resistance of the atmosphere in comprehensive consideration.
Therefore, the parameter of the atmosphere height Ha is
introduced to compute the duration of the acceleration phase,
i.e., Ta = 2Ha /vc .
To conduct a parametric analysis, an acceleration limit
amax is assigned to the elevator movement in phase 3. Hence,
the duration of the deceleration phase is Td = πvc /(2amax ).
Then, the prescribed motion is defined by only four parame-
ters: Ha , Hf , vc and Td . The motion can be written mathemat-
ically in terms of these parameters as:
Hf − Ha Td
Tc = − (37)
vc 2
In the dynamic model, the motion of the elevator is
achieved through an external force. For the elevator to be able
to track a specified velocity profile, a PI controller [11] is
used to produce a response that tracks the velocity curve accu-
rately. The expression of the propulsive force is as follows:
Z
Fe = KP [Ve (t) − V (t)] + KI [Ve (t) − V (t)] dt (38)
p
The prescribed elevator speed V (t) is the input. The error
signal is calculated from the tangential velocity of the elevator
on the tether, Ve (t). The proportional gain KP and the integral
time constant KI are used to calculate the required propulsive
force. Note that the propulsive force calculated from (38) is
tangential to the local tether.
At present, the existing methods to determine the opti-
mal proportional gain and integral time constant require the
whole closed-loop system to be stable. However, the work-
ing point changes with the movement of the elevator, and
the transfer function is difficult to give. Therefore, a large
number of numerical simulations are used to select these gain
parameters under different input signals, so that the control
system can respond quickly and track smoothly with small
overshoot.
Fig. 15 shows the response of the elevator to step input
commands of 50 m/s and 100 m/s. The elevator is 1000 kg
and starts climbing from the anchor point. The durations of
the three phases are set to 1000 s, 3600 s, and 1000 s. The gain
parameters are selected as KP = 0.1 and KI = 5. The results
illustrate that the control has an additional margin due to the
sufficient durations of phases 1 and 3. The elevator speed
tracks the specified velocity profile with little overshoot and
steady-state errors. Indeed, the robustness of the PI control
simulation accuracy requirements. Better control methods
can be adopted in future work [25], [26].
74947
 Z. Wang et al.: Modal and Dynamic Analysis of a Tether for a Nonequatorial Space Elevator

time, as shown in Fig. 16. The tether remains deflected in the

direction of the negative y-axis as the elevator climbs along
the tether to the GEO orbit, and the pendular oscillation is
always in the equatorial plane. The tether remains predomi-
nantly straight before the elevator climbs to a height of nearly
half the height to the GEO orbit. Then, the tether tends to
move to the static equilibrium position. After a period of time
in which it maintains this deflection angle, the tether gener-
ates higher-order oscillations. However, the response follows
mainly the first mode shape. The peak of the deflection at the
end is approximately 12 km, which is equivalent to a tether
deflection of 0.0065◦ .

FIGURE 15. Validation of elevator PI controller.

TABLE 1. Time required to climb to GEO orbit.

B. NUMERICAL SIMULATIONS
It is important to study the dynamic response of the system
to the elevator travelling along the tether because there is a
possibility for resonance and other dynamic effects to influ- FIGURE 16. Response of the system to an elevator climbing to GEO
ence the system. This section describes the conduction of two altitude at 50 m/s.

parametric studies: One involves investigating the effect of

the deployed latitudes on an elevator moving along the tether
from the anchor to the GEO orbit, and the other analyzes the
effect of the elevator speeds.
The first parametric study mainly simulates a space eleva-
tor for which the anchor is deployed at 0◦ –30◦ . The elevator
begins from rest, accelerates to a cruising speed of 50 m/s, and
then decelerates and comes to rest at approximately the GEO
altitude. Gain parameters of KP and KI are set to 0.1 and 5,
respectively, and the duration of phase 3 is set to 1000 s to
reduce the relative error. The other parameters are consistent
with these in Fig. 7. The total simulation time in each case
is equal to the time taken to come to rest at the GEO altitude
and an additional time of 140 h. The additional simulation
time is to enable the occurrence of at least one full cycle
of the pendular modes, i.e., a period of 139.5 h. A constant
simulation time step of 0.05 s is used for each elevator speed.
Note that the time step should be less than the minimum mode
period.
FIGURE 17. Projections of the tether shape along the
For a range of anchored latitudes, the times required to anchor–counterweight line for a 50 m/s transit to
travel to the GEO orbit are calculated and listed in Table 1. GEO (shown for every 1.5 h).
The initial altitude of the elevator is assumed to be 0.1 km.
From Table 1, the longest simulation is of the order of 343 h. Fig. 17 shows the lateral response of the tether, which is
The response of the system anchored at 0◦ is shown via obtained by projecting the tether motion along the line that
projections of the tether(s) and elevator at discrete points in connects the anchor point to the counterweight. For the results

74948 VOLUME 6, 2018

Z. Wang et al.: Modal and Dynamic Analysis of a Tether for a Nonequatorial Space Elevator
to be distinct, the projections are shown for every 1.5 h. Note
that the spacing between the individual tethers on the x-axis
is arbitrary. The results show that the lateral oscillations of
the tether are minimal relative to the length of the tether.
The effects of the first mode of the oscillation can be seen
at the anchor-end of the tether at early time points. The
higher-frequency modes have a coupled effect on the motion
of the elevator. As the elevator height increases, excitation of
the higher tether modes occurs, and higher mode amplitudes
are attained. The deformation of the tether exhibits fluctua-
tions as a result of the higher-order oscillations.
Fig. 18 highlights the effects of the lateral oscillations
by directly projecting the oscillations around a fixed axis.
The peak of the lateral oscillation is approximately 0.62 km
during the transition of the elevator to the GEO orbit. The low
amplitude of the oscillation caused by the elevator is largely
due to the very high tension present in the tether and the
relatively low elevator mass. A higher tension results in less
lateral deflection of the tether for a given lateral force.
FIGURE 18. Direct projections of the tether shape for a 50 m/s transit
(shown for every 1.5 h).
FIGURE 19. Response of the system to an elevator climbing to GEO
altitude at 50 m/s with an anchorage at (a) 10◦ , (b) 20◦ , and (c) 30◦
(shown for every 1 h).
The response of the system for anchored latitudes between
10◦ and 30◦ are shown in Fig. 19. The results show a
clear progression of the effects of higher anchored latitudes
on the lateral dynamics of the tether. The time evolution
plots suggest that higher anchored latitudes cause larger
lateral pendular oscillations. In contrast to in the case of
the equatorial space elevator, the pendular motion of the
tether initially deflects to the positive direction of the y-axis,
VOLUME 6, 2018
FIGURE 20. Projections of the tether shape relative to the static
equilibrium position for an elevator transit to GEO at 50 m/s with an
anchorage at (a) 10◦ , (b) 20◦ , and (c) 30◦ (shown for every 1 h).
FIGURE 21. Projections of the tether shape along the
anchor–counterweight line on the basis of Fig. 21 for an elevator transit
to GEO orbit at 50 m/s with an anchorage at (a) 10◦ , (b) 20◦ , and (c) 30◦
(shown for every 1 h).
which is consistent with the fact that the orbital velocity of the
lumped masses of the tether increases as a result of the tether
being inclined to the equatorial plane, affected by the moving
elevator. Fig. 20 summarizes the vertical response of the
tether by projecting the deflection along the static equilibrium
position. For the results to be legible, projections are shown
for every 1 h. The response closely follows the predicted
mode shape for the second mode. A similar conclusion can be
obtained that higher anchored latitudes cause larger vertical
pendular oscillations, although the amplitude of the vertical
pendular oscillation is much smaller than the lateral oscilla-
tion. Fig. 21 highlights the effects of the lateral and vertical
oscillations by directly projecting the oscillations around the
anchor-counterweight line on the basis of Fig. 20. The results
show that the peak of the lateral oscillation is slightly less than
that of the vertical oscillation and that the amplitudes of the
higher-order oscillations decrease obviously with increasing
anchored latitude.
Fig. 22 summarizes the resulting motion of the counter-
weight after the elevator has stopped moving. The timescale
is in hours and is not normalized. Several features are evident
from Fig. 22. First, the amplitude of the oscillation of the
counterweight generally increases with increasing anchorage
angle. The increase is most dramatic when the system is
anchored farther away from the equatorial. The period of
the transient oscillation is very close to the predicted value
between 124 h and 139 h in the lateral direction and 23 h
in the vertical direction, and the amplitude of the transient
oscillation does not appear to grow or decay over one cycle.
In other words, the low-frequency mode does not appear to
be destabilizing in the full nonlinear simulation.
The second parametric study simulates an elevator moving
at different cruising speeds along the tether from the anchor
to the GEO altitude. In these cases, the anchor is deployed
at 30◦ . The cruising speeds are set to 50, 80 and 100 m/s,
corresponding approximately to the running speeds of the
74949
 Z. Wang et al.: Modal and Dynamic Analysis of a Tether for a Nonequatorial Space Elevator
FIGURE 22. Residual oscillation of the counterweight following elevator
movement for different anchored latitudes. (a) Residual oscillation of the
counterweight in the lateral direction. (b) Residual oscillation of the
counterweight in the vertical direction.
TABLE 2. Time required for the climb to GEO orbit.
three types of high-speed railway operating in China. Higher
speeds, which are challenging to implement, are not studied.
The combinations of the projections over time and the
direct projections of the tether shape for speeds between
50 m/s and 100 m/s are shown in Fig. 23. The results show
a clear progression of the effects of higher speeds on the
lateral dynamics of the tether. The amplitudes of the lateral
pendular oscillations caused by higher transition speeds are
almost equal in the process of the elevator climbing to GEO
orbit. Note that the tether is in different states when the
elevator is climbing to the GEO orbit with different speeds.
74950
FIGURE 23. Response of the system with an anchorage at 30◦ to an
elevator climbing to GEO altitude at (a) 50 m/s, (b) 80 m/s, and
(c) 100 m/s (shown for every 1 h).
FIGURE 24. Projections of the tether shape relative to the static
equilibrium position for an elevator transit to GEO altitude at (a) 50 m/s,
(b) 80 m/s, and (c) 100 m/s (shown for every 1 h).
FIGURE 25. Projections of the tether shape along the anchor–
counterweight line on the basis of Fig. 24 for an elevator
transit to GEO altitude at (a) 50 m/s, (b) 80 m/s, and
(c) 100 m/s (shown for every 1 h).
Hence, the amplitudes of the residual oscillations correspond
to different results. Fig. 24 summarizes the vertical response
of the tether by projecting the deflection along the static
equilibrium position. Again, the projections are shown for
every 1 h to ensure legible results. The results show that
higher speeds have a small effect on the amplitudes of both
the vertical pendular oscillations and the lateral pendular
oscillations. In Fig. 25, the results show that the peaks of the
lateral oscillation and vertical oscillation demonstrate only
small changes with increasing elevator speeds.
Fig. 26 summarizes the resulting motion of the counter-
weight after the elevator has stopped moving, for speeds
between 50 and 100 m/s. Fig. 26 a shows that the ampli-
tude of the lateral oscillation of the counterweight increases
with increasing elevator speed. The growth is relatively
small between 50 and 80 m/s, whereas it is most dra-
matic between 80 and 100 m/s, wherein the amplitude
increases almost 25 km. The amplitude of the vertical oscil-
lation increases substantially with increasing elevator speed,
as shown in Fig. 26 b.
This section addresses the parametric studies conducted to
examine the effect of the deployed latitudes and the elevator
speeds on the system dynamic response. The results clearly
illustrate that the system is stable during the motion to the
GEO altitude at different anchored latitudes and transition
speeds. Unfortunately, one point of concern has not been
analyzed here, namely, the effect of the elevator mass on
VOLUME 6, 2018
Z. Wang et al.: Modal and Dynamic Analysis of a Tether for a Nonequatorial Space Elevator
FIGURE 26. R esidual oscillation of the counterweight following elevator
movement at different cruising speeds. (a) Residual oscillation of the
counterweight in the lateral direction. (b) Residual oscillation of the
counterweight in the vertical direction.
the tether response. In previous studies, a series of simu-
lations was performed considering changes in the elevator
mass for the equatorial space elevator. The results from these
simulations show that the maximum amplitude of the lateral
oscillation is directly proportional to the elevator mass and
that the amplitude increases linearly with increasing elevator
mass. However, the problem of interest here pertains to the
method of calculation of the payload capacity of the system
for a nonequatorial space elevator. Note that a linear rela-
tionship does not exist between the payload capacity and the
anchored latitudes. Again, the payload capacity of the system
is affected by the traveling waves created by the elevator.
V. CONCLUSIONS
This paper has proposed a universal model to analyze a
space elevator considering different anchored latitudes. The
maximum anchored latitude of the nonequatorial elevator
VOLUME 6, 2018
is affected by the design stress of the tether. Linearized
equations were formulated by assuming the displacement
relative to the static equilibrium position, and the modal
analysis was performed. The system is asymptotically stable,
as determined by analyzing the eigenvalues. The fundamental
period increases with an increase in the anchored latitudes,
and the modal shape of the fifth-order modal is influenced
when the anchored latitude is not equal to zero because this
modal shape of the traditional space elevator is along the
tether, whereas the tether of the nonequatorial elevator is
curvilinear. In a part of the dynamic analysis, the control
algorithm is improved to avoid the generation of the simu-
lation errors of the control system under the condition of a
small distance existing between the elevator and the lumped
mass. The system with a climbing elevator remains stable
with variation in the anchored latitudes and the speed of the
elevator. The vibration forms of the tether with a climbing
elevator are composed mainly of the first- and second-order
modal shapes. The amplitude of the residual oscillation in the
vertical or horizontal direction is magnified with an increase
in the anchored latitudes or in the speeds of a traveling
elevator.
The residual vibration suppression of space elevator sys-
tems has always been an important research direction.
At present, the main way to restrain the tether oscillation of
the space elevator system is to optimize elevator movement.
However, the elevator motion does not have enough control
authority to damp the out-of-plane motion. Our future work
will focus on an idea of energy collection to suppress the
vibration. For this method, a piezoelectric material is attached
to the surface of the tether. Piezoelectric materials will absorb
energy when the tether is deformed due to vibration. The
absorbed energy will provide a part of the electric energy of
the elevator motion. This method can not only dissipate the
vibration energy of the tether but also provide energy for the
elevator movement.
REFERENCES
[1] J. Pearson, ‘‘The orbital tower: A spacecraft launcher using the Earth’s
rotational energy,’’ Acta Astronautica, vol. 2, nos. 9–10, pp. 785–799,
Sep./Oct. 1975.
[2] B. C. Edwards, ‘‘Design and deployment of a space elevator,’’ Acta Astro-
nautica, vol. 47, no. 10, pp. 735–744, Nov. 2000.
[3] B. C. Edwards, ‘‘The space elevator: A new tool for space studies,’’ Grav-
itational Space Biol. Bull., Amer. Soc. Gravitational Space Biol., vol. 16,
no. 2, pp. 101–105, Jun. 2003.
[4] S. S. Cohen and A. K. Misra, ‘‘Elastic oscillations of the space elevator
ribbon,’’ J. Guid. Control Dyn., vol. 30, no. 6, pp. 1711–1717, Nov. 2007.
[5] P. Williams, ‘‘Dynamic multibody modeling for tethered space elevators,’’
Acta Astronautica, vol. 65, nos. 3–4, pp. 399–422, Aug. 2009.
[6] C. R. McInnes, ‘‘Dynamics of a particle moving along an orbital tower,’’
J. Guid. Control Dyn., vol. 28, no. 2, pp. 380–382, Mar. 2005.
[7] S. S. Cohen and A. K. Misra, ‘‘The effect of climber transit on the space
elevator dynamics,’’ Acta Astronautica, vol. 64, nos. 5–6, pp. 538–553,
Mar./Apr. 2009.
[8] P. Woo and A. K. Misra, ‘‘Dynamics of a partial space elevator with
multiple climbers,’’ Acta Astronautica, vol. 67, nos. 7–8, pp. 753–763,
Oct./Nov. 2010.
[9] P. Williams and W. Ockels, ‘‘Climber motion optimization for the tethered
space elevator,’’ Acta Astronautica, vol. 66, nos. 9–10, pp. 1458–1467,
May/Jun. 2010.
74951
 Z. Wang et al.: Modal and Dynamic Analysis of a Tether for a Nonequatorial Space Elevator
[10] P. Swan, Space Elevators: An Assessment of the Technological Feasibility
and the Way Forward. Lynchburg, VA, USA: Virginia Edition Publishing
Company, 2013, pp. 291–292.
[11] V. S. Aslanov, A. S. Ledkov, A. K. Misra, and A. D. Guerman, ‘‘Dynamics
of space elevator after tether rupture,’’ J. Guid. Control Dyn., vol. 36, no. 4,
pp. 986–992, Jul. 2013.
[12] A. M. Jorgensen, B. Gassend, S. E. Patamia, R. H. W. Friedel, and
T. Cayton, ‘‘Radiation hazards on a space elevator and ways of miti-
gating them,’’ in Proc. 55th IAC IAF, Vancouver, BC, Canada, 2004,
pp. 3038–3049.
[13] B. Gassend, ‘‘Non-equatorial uniform-stress space elevators,’’ in Proc. 3rd
ISEC, Washington, DC, USA, 2004, pp. 1–10.
[14] F. Cowan and J. Leonard, ‘‘A Mars colonization approach,’’ in Proc. AIAA
Space Conf. Expo., Long Beach, CA, USA, 2007, pp. 1–17.
[15] J. A. Carroll, ‘‘Tether applications in space transportation,’’ Acta Astronau-
tica, vol. 13, no. 4, pp. 165–174, 1986.
[16] P. Williams and P. Trivailo, ‘‘Performance optimization for a towed cable
system with attached wind sock,’’ J. Guid. Control Dyn., vol. 29, no. 6,
pp. 1415–1417, Nov. 2006.
[17] H. A. Fujii, T. Watanabe, T. Kusagaya, D. Sato, and M. Ohta, ‘‘Dynamics
of a flexible space tether equipped with a crawler mass,’’ J. Guid., Control
Dyn., vol. 31, no. 2, pp. 436–440, Mar. 2008.
[18] B. Yu, D. Jin, and H. Wen, ‘‘Nonlinear dynamics of flexible tethered
satellite system subject to space environment,’’ Appl. Math. Mech. -Engl.
Ed., vol. 37, no. 4, pp. 485–500, 2016.
[19] X. Sun, M. Xu, and R. Zhong, ‘‘Dynamic analysis of the tether transporta-
tion system using absolute nodal coordinate formulation,’’ Acta Astronau-
tica, vol. 139, pp. 266–277, Oct. 2017.
[20] P. Williams, C. Blanksby, and P. Trivailo, ‘‘Optimal control of flexible
tethers,’’ in Proc. 54th IAC IAF, Bremen, Germany, 2003, pp. 1–11.
[21] C. R. McInnes, ‘‘Dynamics of a particle moving along an orbital tower,’’
J. Guid. Control Dyn., vol. 28, no. 2, pp. 381–382, Nov. 2005.
[22] N. Takeichi, ‘‘Geostationary station keeping control of a space elevator
during initial cable deployment,’’ Acta Astronautica, vol. 70, pp. 85–94,
Jan./Feb. 2012.
[23] H. Kojima, K. Fukatsu, and P. M. Trivailo, ‘‘Mission-function control of
tethered satellite/climber system,’’ Acta Astronautica, vol. 106, pp. 24–32,
Jan./Feb. 2015.
[24] A. Steindl, ‘‘Time optimal control for the deployment of a tethered satel-
lite allowing for a massive tether,’’ Meccanica, vol. 51, pp. 2741–2751,
Sep. 2016.
[25] M. Yu and C. Li, ‘‘Robust adaptive iterative learning control for discrete-
time nonlinear systems with time-iteration-varying parameters,’’ IEEE
Trans. Syst., Man, Cybern. Syst., vol. 47, no. 7, pp. 1737–1745, Jul. 2017.
[26] J. Peng, C. Li, and X. Ye, ‘‘Cooperative control of high-order nonlinear
systems with unknown control directions,’’ Syst. Control Lett., vol. 113,
pp. 101–108, Mar. 2018.
ZHENKUN WANG received the B.S. degree in
mechanical design manufacturing and automa-
tion from Guizhou University, Guiyang, China,
in 2010, and the M.S. degree in mechanics from the
Harbin Institute of Technology, Shenzhen, China,
in 2012, where he is currently pursuing the Ph.D.
degree in mechanics.
From 2012 to 2016, he was an Assistant Teacher
with the Department of Aerospace Engineering,
Harbin Institute of Technology, Harbin, China. His
research interests include the dynamic analysis of space-tethered system and
vibration suppression and control.
74952
NAIGANG CUI received the B.S. degree in vehicle
system engineering from the National University
of Defense Technology, Changsha, China, in 1986,
and the M.S. degree in flight mechanics and the
Ph.D. degree in guidance and control and sim-
ulation from the Harbin Institute of Technology,
Harbin, China, in 1989 and 1996, respectively.
He has been a Professor with the School
of Astronautics, Harbin Institute of Technology,
since 2000. His research interests include flight
mechanics and control of spacecraft, filtering theory and application, and
spacecraft dynamics and control.
Dr. Cui is a Committee Member of the China Space Institute of Aerody-
namics and Flight Mechanics of the Specialized Committee.
YOUHUA FAN received the B.S. degree in nuclear
physics from the University of South China,
Hengyang, China, in 1996, and the M.S. degree
in flight vehicle design and the Ph.D. degree
in general and fundamental mechanics from the
Harbin Institute of Technology, Harbin, China,
in 1998 and 2001, respectively.
He has been a Professor with the Harbin Insti-
tute of Technology, Shenzhen, since 2010. He has
authored over 50 research papers. His research
interests include elastodynamics and its application and structural dynamics
and applications.
Dr. Fan is a Lifetime Member of the China Geophysical Society and the
Committee of Heilongjiang Industrial and Applied Mathematics Society.
DUN LIU was born in Beijing, China, in 1932.
He received the degree from the Harbin Institute
of Technology, Harbin, China, in 1956.
From 1960 to 1962, he was the Director of the
Mechanical Technology Specialized Committee,
Department of Mechanical Engineering, China.
From 1962 to 1972, he was the Director of the
Teaching and Research Office of Aircraft Design,
Department of Engineering Mechanics, Harbin
Institute of Technology. From 1972 to 1977,
he was a Teacher with the Automatic Control Teaching and Research Office,
Harbin Institute of Technology. From 1977 to 2002, he was a Professor and
the Director of the Automatic Flight and Dynamics Research Office, Harbin
Institute of Technology. His research interests include static pressure gas
lubrication and dynamics of space vehicle.
Mr. Liu is a Foreign Academician of the Russian Academy of
Astronautics.
VOLUME 6, 2018