Strojniški vestnik - Journal of Mechanical Engineering 67(2021)6, 287-301 Received for review: 2021-03-21

© 2021 Journal of Mechanical Engineering. All rights reserved. Received revised form: 2021-05-06
DOI:10.5545/sv-jme.2020.7179 Original Scientific Paper Accepted for publication: 2021-05-07

Dynamic Modelling, Experimental Identification and Computer

Simulations of Non-Stationary Vibration in High-Speed Elevators
Đokić, R. – Vladić, J. – Kljajin, M. – Jovanović, V. – Marković, G. – Karakašić, M.
Radomir Đokić1,* – Jovan Vladić1 – Milan Kljajin2 – Vesna Jovanović3 – Goran Marković4 – Mirko Karakašić5
1 University
of Novi Sad, Faculty of Technical Sciences, Serbia
2 University
North, University Center Varaždin, Croatia
3 University of Niš, Faculty of Mechanical Engineering, Serbia
4 University of Kragujevac, Faculty of Mechanical and Civil Engineering in Kraljevo, Serbia
5 University of Slavonski Brod, Mechanical Engineering Faculty, Croatia

Modelling the dynamic behaviour of elevators with high lifting velocities (contemporary elevators in building construction and mine elevators)
is a complex task and an important step in the design process and creating conditions for safe and reliable exploitation of these machines.
Due to high heights and lifting velocities, the standard procedures for dynamic exploitation are not adequate. The study presents the method
of forming a dynamic model to analyse nonstationary vibrations of a rope with time-varying length with nonholonomic boundary conditions
in the position where the rope is connected with the cabin (cage) and in the upcoming point of its winding onto the pulley (drum). A unique
method was applied to identify the basic parameters of the dynamic model (stiffness and damping) based on experimental measures for a
concrete elevator. Due to the verification of this procedure, the experiment was conducted on a mine elevator in RTB Bor, Serbia. Using the
obtained computer-experimental results, the simulations of the dynamic behaviour of an empty and loaded cage were shown. In addition, the
study shows the specific method as the basis for forming a control program that would enable the decrease in vertical vibrations during an
elevator starting and braking mode.
Keywords: high-speed elevators, dynamic analysis, a rope with time-varying length, mechanical characteristics of steel ropes,
longitudinal oscillations, control program

Highlights
• The complexity of the dynamic analysis of elevators is because these are systems for lifting (lowering) people and load to great
heights (depths) with high velocities and variable parameters.
• Determining the critical hoisting velocity of the elevator car can be performed in the function with mechanical characteristics of
ropes, such as the elasticity modulus and damping and loads in steel ropes.
• Based on the theory of free harmonic damping oscillations, the mechanical characteristics of steel ropes can be determined
through the oscillation diagrams obtained by measurement.
• By defining the basis for the driving mechanism control program, it is possible to provide minimum dynamic loads of elevators
based on adequate models and simulations of their operation in real conditions.

0 INTRODUCTION values. The first part is a driving mechanism (engine,

reductor, brake, and couplings), while the second
Mines with underground exploitation and operating part is a cabin lifting system, mostly made of steel
levels up to 2500 m, along with the rising number of ropes for lifting the counterweights on one end and
exceptionally tall buildings, with heights reaching up cabin (cage) on the other end and their guide rails.
to 850 m nowadays - such situations require electric The driving mechanism comprises elements that are
elevators with specific characteristics, whose velocity much more rigid (c1, c2, Fig. 1) and have a smaller
reaches up to 20 m/s, and load capacity up to over mass than the cabin lifting system (c, Fig. 1), which
ten tons. The elevator quality is estimated according in turn causes the oscillations in smaller amplitudes
to several important indicators. Safety, comfort, and higher frequencies. As well as that, bearing in
and reliability are especially important features [1]. mind that the oscillations are indirectly transferred
These indicators depend on, first of all, vibrations to the cabin (cage) via ropes, it can be deduced that
occurring while the elevator is in motion. Vibrations the lifting system has a much bigger influence on the
are a consequence of driving parameters, inertial comfort during the motion than driving mechanism
characteristics and elasticity of the binding elevator elements does.
elements. Fig. 1 shows the elements with the greatest Deep shaft mines require special mining ropes to
impact. hoist personnel and materials safely and efficiently.
An elevator can be divided into two basic parts They are made of round wires that must be either
according to the dynamic impact on the vibration bright or galvanized. The values of the rope safety
*Corr. Author’s Address: University of Novi Sad, Faculty of Technical Sciences, Trg D. Obradovića 6, 21000 Novi Sad, Serbia, [email protected] 287
 Strojniški vestnik - Journal of Mechanical Engineering 67(2021)6, 287-301

factors for hoisting in mines depend on shaft depth should be pointed towards innovative methods for
and rope number and are higher in the case of analysing the longitudinal oscillations with variable
personnel hoisting. According to ISO standards for parameters [9].
the area of mine elevators, safety factors vary between
4 and 8 for new ropes and between 3.6 and 6.4 for 1 DYNAMIC MODELS FOR THE ANALYSIS
ropes to be discarded, depending on shaft depth. The OF ELEVATOR LONGITUDINAL OSCILLATIONS
highest value of 13 is required by special regulations
for hoisting people to depths of up to 600 m. 1.1 Standard Models

Many researchers are interested in studying

longitudinal oscillations, and their studies have been
based on the general theory about the application of
oscillation of elastic bars with constant parameters
(mass, stiffness, and damping). These are the so-
called standard models [8] and [10]. Those models
are acceptable for analysing elevators with low lifting
velocities and heights. Figs. 2a and b show models
with one and two degrees of freedom and a rope of
constant lengths, represented here as Hook’s, i.e.,
Calvin‘s body.
A certain improvement has been made with
the analysis of high-lift elevators (≥ 35 m) and low
velocities (till 3 m/s) by using the model represented
in Fig. 2c. The model represents a bar of a constant
Fig. 1. Elements with the biggest impact
length with an equally spread mass q (kg/m), i.e., it is
on vibration occurrence in elevators
a model of an elastic body with an unlimited number
of degrees of freedom and a concentrated mass at the
While the elevator is in operation mode, the
bottom end as the boundary condition.
hoist ropes increase and decrease their free length, so
Based on the analysis which was shown in detail
the parameters, such as rope stiffness and damping,
in [11] to [14], in the case when the free rope length
are constantly changing [2] to [4]. In high-speed
is small compared to the cabin mass, it is possible to
elevators, dynamic instability may occur during
significantly simplify the dynamic model analysis.
lifting (reducing free length) due to increased relative
deformation. This instability seriously impacts the
safety of the passengers. Since classic models are
based on elastic body (rope) oscillations with constant
dynamic parameters (mass, stiffness and damping), it
is necessary to form dynamic models that will enable
the analysis and definition of the dynamic behaviour
of elevators with variable parameters [5] and [6].
Because special attention has to be paid to the
accuracy of installing and making of cabin guide
rails and counterweight in high-speed elevators, the
following conclusion arises. Without the addition
of external influences, it can be concluded that
longitudinal oscillations are dominant, as opposed to Fig. 2. Standard dynamic models; а) with one degree of freedom;
transversal oscillations [7] and [8]. b) with two degrees of freedom; c) a “heavy” constant length bar
Taking into consideration that up until this
moment, the problems in driving mechanism Fig. 3 shows an oscillation diagram for the first
vibrations were the subject of a great number of three harmonics. Due to the very small oscillation
scientific and research papers, with standard analyses amplitudes in higher harmonics, their influence can
as the most frequently applied method, it seems logical be neglected. Thus, the total oscillation process with
that the main focus of dynamic research of elevators infinite degrees of freedom, whose total oscillation
288 Đokić, R. – Vladić, J. – Kljajin, M. – Jovanović, V. – Marković, G. – Karakašić, M.
 Strojniški vestnik - Journal of Mechanical Engineering 67(2021)6, 287-301

form is shown in Fig. 3 with a dashed line (d), can when the motion velocity depended a great deal on the
be replaced with a straight line (a), with satisfactory driving electromotor’s mechanical characteristics and
accuracy. In other words, it is replaced with a system brake system (the dynamic equilibrium condition).
with one degree of freedom and constant dilatation (ε) In the earlier periods, replacing one-speed engines
along the free rope end. with two-speed engines was observed as a significant
improvement. This improved the motion comfort in
braking instances and aided the accuracy of stopping
the cabin. As for the process of a regular elevator, in
cases in which there is no slipping of the steel rope on
the driving pulley and when the driving characteristic
is represented via a rope velocity at the meeting point
of the rope and pulley, the elevator model can be
Fig. 3. Oscillation shapes (forms) of the first three harmonics (a, b simplified and represented in the form shown in Fig.
and c), and the summary oscillation form for α = 0.1 5.
Upon observing just the upcoming rope end,
In accordance with the above, regarding high-lift the model can be represented as a system with an
elevators (≥ 35 m) and low velocities (to 3 m/s), it is unlimited number of degrees of freedom; at one end
roughly possible to create load oscillation models with it is rolling onto the pulley at a v(t) velocity, while on
one degree of freedom, with a “heavy” spring, which the other it is burdened with concentrated mass. Due
was studied in general literature. Also, it is necessary to the variable rope length during the motion, the
to replace the total mass (of both load and rope) with stiffness (c = EA/L) changes. This is a characteristic of
an equivalent mass Me = M + (1/3) ∙ qL, reduced in the parametric oscillations and contributes to the possible
cabin place, [15] and [16]. occurrence of resonance. To this end, it is necessary
to complete certain steps in the analysis of dynamic
1.2 Dynamic Elevator Models with Dynamic Variable behaviour. The critical lifting velocity, during which
Parameters the unstable motion occurs, i.e., the rope strain is
increased when its free length is reduced, needs to be
Fig. 4a shows the most common solutions of the determined.
lifting systems for high-speed elevators with a driving
pulley, while the corresponding dynamic model is
shown in Fig. 4b.

Fig. 5. Elevator model with a rope of a variable length

with boundary conditions
Fig. 4. Elevator models; a) Köepe system,
b) a dynamic model of a high-speed elevator The deformation of an arbitrary cross-section
represents the function of the position x, and the time
In order to secure comfort during the motion, t, i.e.:
control programs are used in contemporary elevators. u = f ( x, t ). (1)
They define the circumferential velocity of the pulley.
Thus, they also define the cabin motion velocity (a Upon observing the equilibrium of the elementary
kinematic condition), as opposed to the previous period part (dx), it can be deduced that:
Dynamic Modelling, Experimental Identification and Computer Simulations of Non-Stationary Vibration in High-Speed Elevators 289
 Strojniški vestnik - Journal of Mechanical Engineering 67(2021)6, 287-301

q  dx  2 u ( x, t ) 1.3 Boundary Conditions

   S ( x, t )  S ( x, t )
g t 2
In order to solve a partial differential equation, (Eq.
S q  dx (7)), i.e., an equation system (Eqs. (5) to (7)), it is
 dx  q  dx   a. (2)
x g necessary to define the boundary conditions in the
By representing the rope as Calvin’s model, incoming point of the rope to the pulley in the meeting
where the influence of internal friction can be taken point of hoist ropes and the elevator cabin.
Boundary conditions at point C, where the rope
into consideration via the so-called rope resistance
makes the first contact with the pulley, are dependent
force factor (bf), the dependence of the inner force in
on whether the winding is with or without rope
the rope on its deformation can be noted in this form:
slipping. Fig. 6a shows the distribution of the forces
  u ( x, t )  on the wrap angle of the pulley (α) for a quasistatic
S ( x, t )  E  A   u ( x, t )  bf  . (3)
x  t  case of elevator operation (without the influence of
dynamic forces). The regular elevator operations
If the Eq. (2) is divided with (q·dx) / g and a must not allow for the rope to slip in the whole wrap
replacement for S(x,t), this is obtained: angle (α), which is regulated with a safety degree
against slipping, i.e., with the existence of a suitable
 2 u ( x, t ) g  E  A  2  u ( x, t ) 
  2  u ( x, t )  bf   angle (αM) with the so-called relative abeyance of
t 2 q x  t  the rope on a driving pulley. It should be noted that,
 g  a. (4) in quasistatic conditions, the zone with the elastic
slipping of the rope on a driving pulley (αK), Fig.
In Eqs. (2) and (3), (E) is the rope elasticity 6a, always occurs on the descending side. However,
modulus whose magnitude depends on the elasticity due to the oscillation, the force in the incoming
modulus of wires (Er = 2.1 ∙ 105 MPa) and the end changes, so slipping is probable in the pulley’s
construction of the rope [17] and [18], whose value can incoming zone. The change of force in the wound rope
be twice smaller (stranded wire ropes). part can only be maintained if this change is smaller
By using the differential equation, (Eq. (4)) and than the adhesive force making it possible. Fig. 6d
the equilibrium condition of moments on the driving shows different cases of force distribution over the
pulley, it is possible to form a system of equations that wound rope length as a function of elevator velocity.
describes a dynamic equilibrium on the driving pulley With high-speed elevators, it should be expected that
in the case of a model shown in Fig. 4b in this form the force change in the wound part of the rope is
smaller than the friction force change in the incoming
[12] and [19]:
zone (curve v3, Fig. 6d). Basically, in this case, it can
be accepted that there is no slipping in the rope and
q  2 u1 ( x, t ) 2  u ( x, t ) 
  E  A   u1 ( x, t )  bf  1 rope pulley meeting point. The problem is analysed in
g t 2 x 2  t  great detail in [12] and briefly summarized in [20].
 a
 q  1   , (5)
 g

q  2 u 2 ( x, t ) 2  u ( x, t ) 
  E  A 2  u2 ( x, t )  bf  2
g t 2
x  t 
 a
 q  1   , (6)
 g

R 
Mm   E  A
i  x
  
 u1 (l1 , t )  u2 (l2 , t )  bf u1 (l1 , t )  u2 (l2 , t ) 
 t 
a i Fig. 6. Boundary conditions; a) force distribution on the wrap
Jr  . (7) angle of the pulley, b) driving pulley without slipping,
R
c) cage (cabin), d) different cases of the force distribution

290 Đokić, R. – Vladić, J. – Kljajin, M. – Jovanović, V. – Marković, G. – Karakašić, M.

 Strojniški vestnik - Journal of Mechanical Engineering 67(2021)6, 287-301

The boundary condition at the meeting place

of the rope and the pulley without slipping, Fig. 6b
(point C), is presented in this form:
t ∂u (l , t )  dl 
u (l , t ) = ∫0 ∂x  dt 
  dt . (8)

The boundary condition in the meeting point of

the rope and elevator cabin (mine elevator cage) or
counterweight is given in this form:

  u ( L, t ) 
Q  E  A  u ( L, t )  bf  
x  t 
 Q  2 u ( L, t ) 
   a   Ff , (9)
 g  t 2

where is Ff     FN  sign (v) friction force between
sliding guide shoes and guide rails. Fig. 7. Deformation of a rope

1.3.1 Estimation of Critical Velocity By applying this procedure to the differential

equation (Eq. (5)), replacing the floating argument
The non-integrated boundary condition (Eq. (8)) with (s), multiplying by the function K(x,s,l)
prevents the solution of the partial differential equation and performing the necessary mathematical
(Eq. (5)) so that the solution can be sought through transformations, the rope deformation in the form is
the formation of integral equations, which contain obtained [20]:
both the differential equations and the corresponding
L   2u 
boundary conditions. Simplifying the mathematical u ( x , t )    K ( x , s , l )  q ( s )   2  g  a  ds
model, friction within the boundary condition is
l
 t 
omitted in Eq. (9). t u (l , t ) L  3u
 v(t )dt  E  A K ( x, s, l ) 2 ds
In Fig. 7, a weightless fibre is shown, loaded at 0 x l s t
point (S) by the force Si, where the displacement of  u ( L, t )
2

the rope point (without sliding over the drive pulley) bf  E  A   K ( x, L, l ). (13)
xt
in the region s<x, shows a linear increase from zero to
the boundary value u(s), while the displacement below Using the method of particular integrals, taking
point S is stable and equal to the boundary value u(s). into account only the first form of oscillations of the
Therefore, for the elemental force (Si = 1), there is a rope and that the length of the free rope (L – l) changes
deformation: “slowly” with time, and if a new variable v(x,t) is
introduced, so that:
 xl
 for s  x u ( x, t )  v( x, t )  u (l , t ) ,
K  x, s    E  A , (10)
 s l for s  x
the deformation of rope free side in the phase of load-
 E  A lifting will be:
and displacement u(x) = K(x,s)∙Si.
  (l)dt 
1

For the case with several acting forces with acting  L  l0  4 m t

v( x, t )  ( x  l )  h0     e  cos
points at x=si there is a displacement:  Ll  0

x l  q  (2  L  x  l )   a 
n
 Q    1   , (14)
u ( x)   K ( x, si )  Si . (11) EA  2   g
i 1

For the load case with evenly distributed rope where:

load (rope’s weight), there is the expression: Q a
h0    ,
L
u ( x)   K ( x, s )  q  ds. (12) EA g
l
gEA
 2 (l ) of Non-Stationary Vibration in High-Speed
Dynamic Modelling, Experimental Identification and Computer Simulations , Elevators 291
 q  (L  l) 
 Q    ( L  l )
3
 Strojniški vestnik - Journal of Mechanical Engineering 67(2021)6, 287-301
Q a
h0    ,
EA g
gEA
 2 (l )  ,
 q  (L  l) 
 Q    ( L  l )
3
b t
m   f    2 (l )  dt.
2 0
Analysing Eq. (7), it can be concluded that during
the reduction of the free length of the rope (load-
lifting), its deformation can increase under certain
conditions, which causes a permanent increase in the
dynamic load of the hoist rope. Such a phenomenon,
which is usually described as unstable lifting, will
not occur if the lifting velocity is less than the critical Fig. 8. Driving machine of a mine elevator and the connecting
velocity, which is defined by the expression: tools for connecting the hoist ropes and the cage

gEA EA Table 1. Mine elevator technical characteristics

vkr  2  bf   2  bf  , (15)
q  (L  l) Me
Q Power: 1500/2860 kW,
3 Driving electric
number of revolutions: 122.2 rpm,
motor
torque: 117.2/233.4 kNm
 q  (L  l) 
where M e   Q   g is the reduced mass Mass: 13 t (it includes the cage and the connecting
 3  Cage (cabin)
tools for connecting the hoist ropes), Fig. 8
of the cabin and the rope part as a “heavy spring”. Counterweight Mass: 21 t
The level of critical velocity depends on the z = 6 pieces,
damping characteristic of the wire rope and the d = 27 mm (150 wires per cross-section).
elevator-s basic characteristics. When it comes to Lang’s lay ropes, three right-hand Lang’s lay ropes,
Hoist ropes
elevators with high load capacity, high velocity and three left-hand Lang’s lay ropes.
Breaking force: 561 kN, tensile strength of the
and lifting height (high-speed elevators and mine wires: 1700 MPa.
elevators), the lifting velocity can exceed the critical Unit mass: 3.02 kg/m.
velocity, which is why it is necessary to check the z = 2 pieces,
stability of the lifting process already in the design Compensation d = 50 mm (222 wires per cross-section).
phase. It should be taken into account that Eq. (15) ropes Connected with the cage via rotating hooks to
prevent unwinding. Unit mass: 9.64 kg/m.
does not contain the friction in guide rails, which
adds to the oscillation reduction, and simultaneously
increases the critical velocity value.
2.1 Equipment Used and the Method for Measuring Data
Acquisition
2 EXPERIMENTAL DETERMINATION
OF DYNAMIC PARAMETERS
The following measuring equipment was used for the
mine elevator experiment:
Experimental studies shown in this paper represent the
sequel to the research in [13]. The measuring was done • Universal 8-channel measuring amplifier (2
on a mine elevator in RTB Bor, Serbia; the maximum pieces), QUANTUM X MX480B, pos. 10a and
projected cage lifting velocity was 16 m/s, and the 10b (Fig. 9),
projected lifting height was 523 m in the first phase • A computer for storing the measuring signals,
and 763 m in the second phase of mining. The driving pos. 13 (Fig. 9),
mechanism is shown in Fig. 8. The other significant • Antennas for the wireless transfer of the
characteristics of this exploitation machine can be measuring signal (2 pieces) NanoStation loco
seen in Tab. 1. NS2L, made by IBIQUITI NETWORKS; 5 km
A mine shaft is like a round cross-section with range with the antennas optical visibility, pos. 12a
a 10 m diameter. The transfer of the driving moment and 12b (Fig. 9),
to the hoist ropes is done through friction (Köepe • Incremental encoder - the number of revolutions
system) from the grooved drum with a 2.5 m diameter. sensor AINS 41, made by Meyer Industrie-
Electronic GmbH – MEYLE; measuring range up
292 Đokić, R. – Vladić, J. – Kljajin, M. – Jovanović, V. – Marković, G. – Karakašić, M.
 Strojniški vestnik - Journal of Mechanical Engineering 67(2021)6, 287-301

to 6000 rpm and with an allowed axial/radial load 9), set on the connecting tools of the cage through
at the output shaft of 30/20 N, pos. 9 (Fig. 9), a magnetic holder, Fig. 11a. The force changes in
• Optical measuring device for the number of the hoist ropes were monitored by measuring the
revolutions ROS type, made by Monarch deformation of the connecting tools (pos. 3 to 8,
Instrument; measuring range varies from 1 to 250 Fig. 9). The deformations were measured on each
000 rpm, pos. 1 (Fig. 9), connecting tool (out of 6 in total) by using the HBM
• Accelerometer HBM B12, pos. 2 (Fig. 9), LY41-6/120 strain gauges, Fig. 11b.
• Strain gauges HBM LY41-6/120, pos. 3 to 8 (Fig.
9),
• Software for acquisition and processing of
measuring signals, HBM catmanEasy-AP.

a) b)
Fig. 10. Measuring places; a) the incremental encoder connected
with the measuring wheel positioned on the rim of the brake disc,
and b) the optical sensor (number of revolutions sensor) positioned
on the cage

a) b)
Fig. 11. Measurements on the cage connecting tools; a) cage
acceleration, and b) deformations
Fig. 9. Schematic layout of the measuring places
All measuring signals were led to the other eight-
The experiment was done according to the channel measuring amplifier (pos. 10a, Fig. 9), and
principle of summary data acquisition via a wireless they were recorded on the computer (pos. 13, Fig. 9)
data transfer by using the NanoStation loco NS2L via a LAN switch (pos. 11a, Fig. 9).
antennas (pos. 12a and 12b, Fig. 9). In this way, a
complete synchronization of the collected data on the 2.2 Protocols and Results of the Experiment
cage roof and machine room was achieved. The drum
number of revolutions was measured with an AINS 41 In Fig. 12, a schematic layout of a mine shaft elevator
incremental encoder (pos. 9, Fig. 9). It was set on the is shown with labelled levels, whose positions were
adaptive carrier with a 127 mm diameter measuring set with altitude, using plus and minus. The ground
wheel, which was directly leaned onto the rim of the level was at an altitude of +436 m, while the machine
brake disc, Fig. 10a. The encoder measuring signal room was placed at +465 m.
was led to the amplifier (pos. 10b, Fig. 9), and then Experiment results and determination of the
via a LAN switch (pos. 11b, Fig. 9) and the antenna dynamic model parameters in this paper are shown in
(pos. 12b, Fig. 9) wirelessly connected to the other four characteristic motion cases (lifting and lowering)
radio antenna positioned on the cage (pos. 12a, Fig. of the cage, with and without load. A significantly
9). The measuring signal was stored in a common higher number of measurements was completed for
computer file (pos. 13, Fig. 9) placed on the cage. It different cases (different positions for starting and
was done via another LAN switch (pos. 11a, Fig. 9). stopping the cage motion etc.), and the results are
The lifting and lowering of the velocity of the presented in [11]. Here, the results of the lowering
cage was measured via a roller guide, whose change and lifting of an empty cage and a loaded cage
in the number of revolutions was registered with an (traction machine) are presented. According to those,
optical sensor (pos. 1, Fig. 9), positioned on a small the system oscillation parameters are defined. The
magnet table, Fig. 10b. The change in the cage’s characteristics of experimental cases are as follows:
acceleration during the measuring process was I. Lowering an empty cage from position
registered by an HBM B12 accelerometer (pos. 2, Fig. +104 m to position +56 m (48 m). The action of
Dynamic Modelling, Experimental Identification and Computer Simulations of Non-Stationary Vibration in High-Speed Elevators 293
 Strojniški vestnik - Journal of Mechanical Engineering 67(2021)6, 287-301

stopping the cage was performed by suddenly

switching off the driving electromotor (at a high
velocity, Fig. 13a).
II. Lifting of an empty cage from position
+56 m to a position +199 m (143 m). Stopping
the cage was done just like in the case above,
with the sudden switching off the power to the
driving electromotor, at a high velocity, Fig. 13b.
III. Lifting a loaded cage (the traction machine
for wagons, mass ~9.35 t) from position a
-71 m, to position +52 m (123 m). Stopping the
cage was done at a low speed, Fig. 13c.
IV. Lowering the loaded cage (the traction machine
for wagons, mass ~9.35 t) from position +52 m to
-71 m (123 m). Stopping the cage was done at a
low speed, Fig. 13d.
The results of the conducted experimental
research can be seen in the following figures, which
are showing the changes in the winding velocities
of hoist ropes that were wound onto the drum; cage
acceleration and changes in the forces in the hoist
ropes, i.e., on the elements for connecting the ropes
to the cage. The results are valid for all four chosen
examples of motion. The diagram shows the character
of the changes of the said parameters. It shows the
cage oscillation amplitudes that occur after stopping
the driving machine. This part of the diagram (after
stopping the driving machine) is used to set the
Fig. 12. Schematic layout of a vertical mine shaft in “Jama” mine mechanical characteristics of steel ropes (harmonic
of RTB Bor oscillations).

Fig. 14. Diagram with the results obtained during lowering an

empty cage and sudden stopping, at a “high” velocity
(I motion case)

Fig. 13. Mine elevator parameters that are relevant for the Fig. 15. Diagram with the results obtained during lifting the empty
analysis; a) lowering an empty cage, b) lifting of an empty cage, c), cage and sudden stopping at a “high” velocity (II motion case)
lifting a loaded cage, d) lowering the loaded cage

294 Đokić, R. – Vladić, J. – Kljajin, M. – Jovanović, V. – Marković, G. – Karakašić, M.

 Strojniški vestnik - Journal of Mechanical Engineering 67(2021)6, 287-301

variability in the elasticity modulus magnitude can be

observed, and significant differences occur in loading
and unloading the rope. The average (median) value
of elasticity modulus can be set for boundary working
stresses - ES (σlower, σupper).

Fig. 16. Diagram with the results obtained during the loaded cage
lifting process (III motion case)

Fig. 18. Tangent and average rope elasticity modulus

Fig. 17. Diagram with the results obtained during loaded cage
lowering process (IV motion case)
Damping. In many practical systems, oscillation
energy is gradually transformed into heat or sound,
2.3 Determination of Dynamic Parameters Based on the effect mostly known as damping. Although
the quantity of such energy is relatively small, it is
Measurement Results
important to dedicate some attention to damping to
The stiffness and damping parameters can be predict the oscillatory system’s response, such as the
elevator systems considered in this paper.
determined, and the elasticity modulus as mechanical
Damping at the elevator is complex, and it
characteristics of a steel rope can be indirectly
happens because of the rope’s inner friction (viscous
determined too, by using the above diagrams and
and hysteretic damping), Coulomb’s friction on guide
determining the oscillation period, i.e., the frequency
rails and damping due to the airflow around the cabin
and logarithm decrement of the damping. In [13], and
(cage) in the elevator shaft [13] and [15].
especially in [17], the said characteristics are described
Since the presented experiment results (Tab.
in detail, so this study only mentions the important
2) refer to the moment of stopping the cage (v = 0),
parts for explaining the results obtained with new
damping as a consequence of airflow around the cage
measurements.
can be neglected.
Stiffness and elasticity modulus. Rope stiffness
is defined with an expression [13]:
EA
c . (16)
L  l (t )
The stiffness depends on the change in the rope’s
free length, but it also depends on the elasticity
modulus, usually taken as a constant parameter in
dynamic behaviour analyses. However, due to the
complex construction of a steel rope (consisting of
wires, strands, and a core), its change depending on Fig. 19. An oscillating system with Coulomb’s friction;
the strain magnitude and loading character (loading- a) motion with the load, b) a model with Coulomb’s damping
unloading) cannot be neglected. According to [18], a
distinction can be made between the so-called tangent The friction on the elevator guide rails creates the
elasticity modulus (Et) and an average (medium) damping force (Coulomb’s friction), which is constant
elasticity modulus between two stresses (ES), Fig. 18. in its magnitude, but it is of the opposite direction
The tangent elasticity modulus represents a theoretical compared to the motion of the oscillating load. Since
angle of the inclination of the tangent to the σ-ε this type of elevator requires special attention to be
curve for the current stress value (σZ). The constant paid to the guiding accuracy and reducing the guide
Dynamic Modelling, Experimental Identification and Computer Simulations of Non-Stationary Vibration in High-Speed Elevators 295
 Strojniški vestnik - Journal of Mechanical Engineering 67(2021)6, 287-301

rails friction, in the example with centrical cabin seen in the diagrams shown in Figs. 14 to 17, which
load, the friction forces can be neglected in relation show the cumulative influence of rope’s internal
to the total load. Nevertheless, the influence can be friction and the guide rails’ friction is generally of a
of importance for cabin oscillation analyses [21]. viscous damping character.
A typical diagram of an oscillator with Coulomb’s The size of the impact caused by rope damping
friction is presented in Fig. 19a. The declining of and damping due to guide rails friction can be defined
amplitudes, as opposed to inner friction oscillation, is via the “overlapping” of the diagram measurement
linear in its nature [15] and [17]. results and simulation results, as presented in Fig. 22.
As it was explained in [13], the total force of Based on the theory about free harmonic
Coulomb’s friction on guide rails is:
damping oscillations (Fig. 21a), it is concluded that
Fc  nv  nt    FN . (17) for determining the dynamic parameters, one should
take the part of the cage oscillation diagram after the
For the most part, the literature presents the driving machine is stopped.
inner friction in the rope as viscous friction, like in
homogenous bodies, Figs. 2a and b. Nevertheless,
while the rope is being deformed, the energy
dissipation also appears due to the friction between
wires and strands, which slide against one another
during the deformation process. This creates the
hysteresis loop (Fig. 20a). The effect causes the
damping known as hysteretic or structural damping.
The loss of energy per unit of material volume in one
loading and unloading cycle is equal to the area closed
by the hysteresis loop. Experiments have confirmed
that the energy loss per cycle is approximately Fig. 21. Oscillating system with viscous integral damping;
a) damped oscillations b) a lifting system model
proportional to the square of the oscillating amplitude.
A model with hysteretic damping is presented in
Fig. 20b. Similarly to the above, the rope’s inner
friction (br) is a combination of viscous and hysteretic
damping, Fig. 20c.

Fig. 22. The impact of inner and Coulomb’s friction on oscillating

systems when b = 5161 Ns/m (Table 2)

By measuring the amplitudes and oscillating

periods of free damping oscillations (changes in cage
acceleration), the damping coefficient (δ) can be
determined via a logarithmic decrement, and based on
Fig. 20. Hysteretic damping; a) hysteresis loop, this, a resistance force coefficient (b):
b) a model with hysteretic damping, and
c) an equivalent model with internal rope damping xi 1 x D
D  ln  ln i    T    , (18)
xi 1 n xi  n T
The total damping of the considered system is
an overall influence of the damping due to the rope’s b  2   M e . (19)
internal friction and the damping caused by the friction
in guide rails, so it can be represented by the parallel Determining the stiffness coefficient (c) and
connections in the oscillator model, Fig. 21b. This is elasticity modulus (E) of hoist ropes can be performed
296 Đokić, R. – Vladić, J. – Kljajin, M. – Jovanović, V. – Marković, G. – Karakašić, M.
 Strojniški vestnik - Journal of Mechanical Engineering 67(2021)6, 287-301

Table 2. Values obtained by measuring and parameters of hoist ropes

L [m] Me [kg] δ [s–1] ω [rad/s] c [N/m] b [Ns/m] E [MPa] δ/ω
I (Fig. 13a) ↓ 409 18 797 0.33 4.67 409 163 12 222 98 440 0.070
II (Fig. 13b) ↑ 266 20 525 0.21 5.69 663 329 8718 103 792 0.037
III (Fig. 13c) ↑ 413 28 099 0.09 4.26 510 803 5161 124 095 0.022
IV (Fig. 13d) ↓ 536 26 613 0.18 3.88 401 206 9639 126 498 0.047

with the measured oscillation parameters (oscillating 2.4 Computer Simulations of Mine Elevator Dynamic
period, oscillating amplitudes, etc.): Behaviour and Result Correlation

c  M e 2 , (20) In order to simulate the mine elevator operation

cL process, the study has set the change in hoisting rope
E . (21) velocity at the point of winding onto the driving drum
A
Fig. 23 is an illustration of determining the said as the change (diagram), obtained by direct measuring
mechanical characteristics of hoist ropes based on the on the driving machine via an incremental encoder
cage acceleration changes diagram (after the driving (pos. 9, Fig. 9).
machine was stopped) in the III motion case. An oscillation diagram was separated with the
purpose of verifying the dynamic model during the
whole period while the cage was moving, as well
as after the complete stop in the driving machine
moving. The acceleration diagram in the III motion
case is shown, “overlapping” the diagram with the
measured results, Fig. 24.
Judging by the diagram in Fig. 24, it can be
concluded that the dynamic model with dynamic
parameters, determined by the measuring with
satisfactory accuracy, shows the real behaviour of
a mine elevator. As a result, this makes computer
Fig. 23. The measured values of cage acceleration simulations possible, which will analyse the real loads
for III motion case of the exploitation facility [11] and [22]. The described
method is a new approach that will ensure the analysis
Based on the diagram in Figs. 14 to 17 and of dynamic behaviour in systems for vertical hoisting,
the expressions (Eq. (18) to (21)), mechanical which can be found in exploitation.
characteristics of steel ropes are defined, as shown in
Table 2. Upon analysing the measured results, it can
be concluded that the data for elasticity modulus are in
agreement with the data from the literature [18]. The
higher values of elasticity modulus for loaded cabins
are seen as a consequence of the rope constructed
by putting the lays of wires in strands and strands
into a rope. This confirms the validity of the applied
procedure, enabling us to define the real (exploitation)
values in mine elevators. The damping coefficient Fig. 24. Cage acceleration diagrams in the motion case III
values, for which there is no significant comparative (Fig. 13c)
data, are not constant in size but appear to be different
in the analysed experiments. Regarding the limitations After comparing the results of numerical analysis
in conducting the experiments in working conditions, with measurement results, a conclusion arises that
it was concluded that the experiment should be it would be useful to have some form of numerical
prepared in a laboratory setting in the future. value for result correlation. In the literature, this value
is known as Pearson’s correlation coefficient. This
Dynamic Modelling, Experimental Identification and Computer Simulations of Non-Stationary Vibration in High-Speed Elevators 297
 Strojniški vestnik - Journal of Mechanical Engineering 67(2021)6, 287-301

study did not determine the correlation coefficient for

individual motions, so that it could be taken as one of
the ideas for future research.
As an illustration of the possibility of dynamic
analysis, the following section shows diagrams that
represent changes in individual values. They were
created in specialized software for dynamic analysis
using the experimental data shown in Table 2. Fig. 29. Diagram showing the changes in cage position, velocity,
and acceleration, for III motion case

Fig. 25. Diagram showing the changes in the cage position,

velocity, and acceleration, for I motion case Fig. 30. Comparative diagrams showing the changes in the
elongation, stiffness coefficient, and resistance force coefficient of
hoist ropes for III motion case

Fig. 26. Comparative diagrams showing the changes in the

elongation, stiffness coefficient, and resistance force coefficient Fig. 31. Diagram showing the changes in cage position, velocity,
of hoist ropes, for I motion case and acceleration for IV motion case

Fig. 27. Diagram showing the changes in cage position, velocity, Fig. 32. Comparative diagrams showing the changes in the
and acceleration, for II motion case elongation, stiffness coefficient, and resistance force coefficient of
hoist ropes for IV motion case

3 AN ANALYSIS OF THE HOISTING VELOCITY IMPACT AS THE

BASIS FOR CONTROL SYSTEM

The material of this paper confirms that there are

different analyses of individual parameters for vertical
hoist systems. One part is of special interest, and that
Fig. 28. Comparative diagrams showing the changes in the is the research on the changes in hoist velocities,
elongation, stiffness coefficient, and resistance force coefficient of i.e., defining the control system [23]. The above
hoist ropes, for II motion case method can be used for determining an optimal form
298 Đokić, R. – Vladić, J. – Kljajin, M. – Jovanović, V. – Marković, G. – Karakašić, M.
 Strojniški vestnik - Journal of Mechanical Engineering 67(2021)6, 287-301

for changing the hoist velocity, to reach maximum The following figures present some of the possible
capacity, i.e., the shortest time for the motion, along ways of impacting driving kinematic characteristics
with securing a high level of motion, especially in through a control program. Fig. 33 shows how the
modern passenger elevators, whose velocities reach changes in velocity form characteristics can impact
up to 20 m/s. As stated above, the dynamic parameters the cabin (cage) oscillating amplitudes, as well as
during the elevator or exploitation facility design the motion comfort. As it appears, it is important to
process can only be determined with limited accuracy, note that by controlling the drive in the acceleration
so the exploitation deviation can be significant. The period, and especially by choosing the right moment
method shown in this paper can be used as a basis, to switch from acceleration to stationary velocity, the
which shows that it is possible to apply straightforward system can be significantly “relieved”, i.e., the motion
methods and form an optimal elevator facility. comfort can be bettered (the amplitudes can be quite
The application of the illustrated method makes it decreased), pos. 4 in Fig. 33.
possible to determine real parameters and perform Installing an acceleration transducer
dynamic analysis even after the installation process. (accelerometer) on the points where ropes connect
Therefore, the method makes it possible to adjust with the cabin (cage) has acceleration change as a
and correct the exploitation facility and improve the response at any moment. This is vital with high-speed
previously defined optimal operating conditions. This elevators, in cases in which nominal velocity is not
particularly concerns defining the appropriate control reached between two neighbouring floors, but the
program, which ensures the shortest motion time with acceleration is followed by braking, Fig. 34.
the maximum level of comfort, i.e., with the control of The right control is key for braking at the most
the dynamic loads. convenient moment, as shown by line (2) in the

Fig. 33. The influence of driving system kinematic characteristics on the motion comfort;
a) change of velocity, b) change of acceleration

Fig. 34. The layout of the influence of controlling the magnitude of oscillating system amplitudes;
a) change of velocity, b) change of acceleration

Dynamic Modelling, Experimental Identification and Computer Simulations of Non-Stationary Vibration in High-Speed Elevators 299
 Strojniški vestnik - Journal of Mechanical Engineering 67(2021)6, 287-301

velocity change diagram, Fig. 34a, which causes much for comfort during the motion, while simultaneously
smaller oscillating amplitudes, Fig. 34b. The distance securing the optimal motion time, i.e., the efficiency
the cabin (cage) has passed and the accuracy of station of high-speed passenger elevators, especially in
landing must be taken into account when creating a transient operation regimes.
control system. As for mine elevators, strict safety conditions are
required in order for the facility to operate, especially
4 CONCLUSION regarding elevators for the transport of people. It is
necessary to secure special conditions at the moment
The dynamic analysis and scientific approach shown of power transfers and friction motion to avoid the
in this paper are of special importance taking into slipping of the rope as a whole along the driving
account the nature of the systems; they are meant for pulley. Finally, this is vital from the point of view of
lifting (lowering) people and load to great heights defining suitable boundary conditions for a dynamic
(depths) at velocities up to 20 m/s. The complexity of model at the points where the ropes get on or off the
the dynamic analysis can also be seen because these driving pulley.
are oscillating processes with variable parameters and
boundary conditions at the incoming/outgoing points 5 NOMENCLATURES
of the steel rope and the driving pulley. Furthermore,
the stiffness and elasticity modulus of steel ropes u1, u2 elastic deformation of the rope on the
are not constant in magnitude but rather change incoming and outgoing rope end, [mm]
depending on the changes in cage position, as well as E rope elasticity modulus, [MPa]
construction, stress, and rope exploitation time. A rope’s cross-section area, [mm2]
The dynamic model that is suitable for the a acceleration of the driving mechanism, [m/s2]
analysis can be formed by observing a particular
Mm driving motor torque, [Nm]
system and by simplifying (i.e., omitting) the small
i gear ratio, [-]
values of higher order.
η driving mechanism efficiency, [-]
The presented approach on forming the
Jr moment of inertia of rotational masses, reduced
differential motion equation for the incoming end
to the shaft of a driving pulley, [kgm2]
of the rope onto the driving pulley helps determine
R driving pulley radius, [m]
the critical hoisting velocity in the function with
q rope weight per meter, [kg/m]
mechanical characteristics (elasticity modulus and
l wound rope length, [m]
damping) and stress (load) in steel ropes. Moreover,
v (v = dl / dt) circumferential velocity, [m/s]
this method also makes it possible to execute
computer simulations and dynamic behaviour analysis Ff friction force between sliding guide shoes and
during the motion by using the appropriate specialized guide rails, [N]
software for the dynamic analysis of mechanical l0 length of winded rope at time t = 0, [m]
systems. nv (nv = 4) number of the roller guide groups, [-]
The paper presents an experimental method that nt (nt = 3) number of rollers in the guide group, [-]
can be used to determine dynamic parameters such as µ rolling resistance of the roller guide on the guide
stiffness, elasticity modulus, and damping in exploited rail, [-]
elevator steel ropes. Specific elasticity modulus values FN force of the roller guide pressure on the guide
presented in Table 2 reveal the noticeable dependence rail for centric load, which depends on the spring
on the load, i.e., rope stress. In addition, the damping tightening during installation, [N]
coefficient in ropes is not a constant size, but it xi (also, xi+1 and xi+n) measured oscillating
depends on the position of a cage. It can be deduced amplitudes, [mm]
that ropes experience a combination of viscous T measured oscillating period of free damping
damping and hysteretic damping, which should be oscillations, [s]
further investigated in these systems. Me reduced oscillating mass (Me = M + (q · L(t)) / 3),
By forming adequate dynamic models, [kg]
determining real values for dynamic parameters, and M total mass hanging on hoist ropes (loaded cage
simulations of dynamic behaviour for real facilities, it and compensation ropes), [kg]
is possible to define the basis for a program that would L(t) hoist ropes’ free length ( L(t )  L   v(t )dt ), [m]
control the driving mechanism. That would ensure v(t) circumferential velocity of the pulley (drum),
minimal dynamic loads, which is especially relevant [m/s]
300 Đokić, R. – Vladić, J. – Kljajin, M. – Jovanović, V. – Marković, G. – Karakašić, M.
 Strojniški vestnik - Journal of Mechanical Engineering 67(2021)6, 287-301

ω circular frequency of free oscillations [10] Li, C., Hua, C., Qin, J., Zhu, Z. (2019). Research on the Dynamic
(    2   2 ), [rad/s] Characteristics of High-Speed Elevator System. Advances in
ω damping oscillation frequency (  2 T ), Engineering Research, vol. 181, p. 105-109, DOI:10.2991/
[rad/s] ice2me-19.2019.23.
L hoist ropes length, [m] [11] Đokić, R. (2016). Dynamics researching and development of
vertical transport machines using numerical-experimental
δ damping coefficient, [s‒1]
procedures. PhD thesis, University of Novi Sad, Faculty of
c ropes stiffness coefficient, [N/m] Technical Sciences, Novi Sad.
b resistance force coefficient, [Ns/m] [12] Vladić, J. (1982). Contribution to the determination of the
safety degree against slipping of a dynamically loaded rope in
6 REFERENCES the power transmission system by a driving pulley. MSc thesis,
University of Novi Sad, Faculty of Technical Sciences, Novi Sad.
[1] Peng, Q., Xu, P., Yuan, H., Ma, H., Xue, J., He, Z., Li, S. (2020). [13] Vladić, J., Jovanović, M., Đokić, R., Kljajin, M., Karakašić, M.
Analysis of Vibration Monitoring Data of Flexible Suspension (2015). Theoretical and experimental analysis of elevator
Lifting Structure Based on Time-Varying Theory. Sensors, vol. dynamic characteristics. Tehnički vjesnik / Technical
20, no. 22, 6586, DOI:10.3390/s20226586. Gazette, vol. 22, no. 4, p. 1011-1020, DOI:10.17559/TV-
[2] Zhang, Q., Yang, Y., Hou, T., Zhang, R. (2019). Dynamic 20150107175453.
analysis of high-speed traction elevator and traction car–rope [14] Vladić, J., Đokić, R., Kljajin, M., Karakašić, M. (2011).
time-varying system. Noise & Vibration Worldwide, vol. 50, no. Modelling and simulations of elevator dynamic behaviour.
2, p. 37-45, DOI:10.1177/0957456519827929. Tehnički vjesnik / Technical Gazette, vol. 18, no. 3, p. 423-
[3] Bao, J., Zhang, P., Zhu, C. (2011). Modeling of Rope 434.
Longitudinal Vibration on Flexible Hoisting System with [15] Rao, S. S. (2011). Mechanical vibrations. Prentice Hall, Upper
Time-Varying Length. Applied Mechanics and Materials, vol. Saddle River.
130-134, p. 2783-2788, DOI:10.4028/www.scientific.net/ [16] Sinha, A. (2010). Vibration of Mechanical Systems. Cambridge
AMM.130-134.2783.
University Press, Cambridge.
[4] Zhang, Y., Pota, H.R., Agrawal, S.K. (2002). Modification of
[17] Vladić, J., Đokić, R., Jovanović, M. (2019). Steel wire rope and
residual vibrations in elevators with time-varying cable lengths.
computational-experimental procedures for the analysis of
Proceedings of the American Control Conference Anchorage,
specific transport machines. University of Novi Sad, Faculty of
vol. 6, p. 4962-4966, DOI:10.1109/ACC.2002.1025450.
Technical Sciences, Novi Sad.
[5] Arrasate, X., Kaczmarczyk, S., Almandoz, G., Abete, J.M., Isasa,
[18] Feyrer, K. (2007). Wire Ropes. University of Stuttgart, Stuttgart.
I. (2014). The modelling, simulation and experimental testing
[19] Goroshko, O.A. (2007). Evolution of the dynamic theory of
of the dynamic responses of an elevator system. Mechanical
hoist ropes. International Applied Mechanics, vol. 43, no. 1, p.
Systems and Signal Processing, vol. 42, no. 1-2, p. 258-282,
DOI:10.1016/j.ymssp.2013.05.021. 64-67, DOI:10.1007/s10778-007-0007-9.
[6] Herrera, I., Su, H., Kaczmarczyk, S. (2010). Investigation into [20] Vladić, J., Malešev, P., Šostakov, R., Brkljač, N. (2008).
the damping and stiffness characteristics of an elevator car Dynamic analysis of the load lifting mechanisms. Strojniški
system. Applied Mechanics and Materials, vol. 24-25, p. 77- Vestnik - Journal of Mechanical Engineering, vol. 54, no. 10,
82, DOI:10.4028/www.scientific.net/AMM.24-25.77. p. 655-661.
[7] Kimura, H., Ito, H., Nakagawa, T. (2007). Vibration analysis [21] Zhang, X., Li, H., Meng, G. (2008). Effect of Friction on
of elevator rope (Forced vibration of rope with time-varying the Slide Guide in an Elevator System. Journal of Physics:
length). Journal of Environment and Engineering, vol. 2, no. 1, Conference Series, vol. 96, no. 012074, DOI:10.1088/1742-
p. 87-96, DOI:10.1299/jee.2.87. 6596/96/1/012074.
[8] Guo, Y., Zhang, D., Chen, K., Feng, C., Ge, S. (2018). [22] Vladić, J., Đokić, R. (2017). Characteristics of mathematical
Longitudinal dynamic characteristics of steel wire rope in a methods and specialized software systems for dynamic
friction hoisting system and its coupling effect with friction analysis of elevators and mining elevators. IMK-14 - Research
transmission. Tribology International, vol. 119, p. 731-743, and development in Heavy Machinery, vol. 23, no. 2, p. 31-38,
DOI:10.1016/j.triboint.2017.12.014. DOI:10.5937/imk1702031v.
[9] Watanabe, S., Okawa, T. (2018). Vertical vibration of elevator [23] Knezevic, B.Z., Blanusa, B., Marcetic, D.P. (2017). A synergistic
compensating sheave due to brake activation of traction method for vibration suppression of an elevator mechatronic
machine. Journal of Physics: Conference Series, vol. 1048, no. system. Journal of Sound and Vibration, vol. 406, p. 29-50,
012012, DOI:10.1088/1742-6596/1048/1/012012. DOI:10.1016/j.jsv.2017.06.006.

Dynamic Modelling, Experimental Identification and Computer Simulations of Non-Stationary Vibration in High-Speed Elevators 301