See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/371782966

MATHEMATICAL MODELING AND SIMULATION OF A HYBRID COMPOUND

MECHANICAL SYSTEM

Conference Paper · June 2023

CITATIONS READS

0 141

4 authors, including:

Milica Tufegdžić Violeta Đorđević

Academy of Professional Studies Sumadija 10 PUBLICATIONS 10 CITATIONS
35 PUBLICATIONS 47 CITATIONS
SEE PROFILE
SEE PROFILE

All content following this page was uploaded by Milica Tufegdžić on 22 June 2023.

The user has requested enhancement of the downloaded file.

 13th International Scientific Conference
Science and Higher Education in Function of Sustainable Development –
SED 2023

MATHEMATICALMODELING AND SIMULATION OF A HYBRID

COMPOUND MECHANICAL SYSTEM
Milica Tufegdžić1, PhD; Zvonko Petrović2; MSc; Violeta Đorđević3, MSc
1
Academy of professional studies Šumadija, Trstenik, SERBIA, [email protected],
2
Academy of professional studies Šumadija, Trstenik, SERBIA, [email protected],
3
Academy of professional studies Šumadija, Trstenik, SERBIA, [email protected]

Abstract: Analysis of complex dynamics of compound mechanical systems requires different approaches, methods, and tools, such as
mathematical modeling and simulation. This study is conducted at the representative example of compound mechanical system with
the aim to present possibilities for the application of modern tools in solving problems that include different movements, namely
translation and rotation. Mathematical model is derived in the form of system of differential equations, as well as the state space model.
In order to facilitate motion transmission analysis and visualize results an appropriate script in Matlab is written. The simulation
model is created based on the derived mathematical model in the Simulink graphical environment. After the simulation process, the
influences of individual inputs on the outputs from the system are plotted and analyzed. The changes of the output values in the form
of translational and rotational displacement over time are also observed.

Keywords: differential equation, translational displacement, rotational displacement

1. INTRODUCTION

Mathematical model of a system is a set of differential and algebraic equations that describe the behavior of the physical
system with a certain degree of approximation, based on knowledge of the physical properties and laws of the system
itself. Such are, for example, Newton's laws for modeling mechanical systems or Kirchhoff's laws for modeling electrical
systems. These models usually describe a system through relations between variable quantities that represent certain
properties of the system and establish relationships between input and output of the system. This enables forecasting of
system behavior under specific conditions [1,2,3]. Models are suitable for improving the system design and designing
optimal control policy, with the aim to achieve the desired operational performance [4].
Differential equations in the form of ordinary differential equations (ODE) or system of linear differential equations are
usually used to describe the dynamics of the system [5,6]. For more complex systems nonlinear and/or partial differential
equations (PDEs) are better [2]. Once the mathematical model is established, adequate analytical or numerical algorithms
are used to find solutions by the methods of direct integration or direct and inverse Laplace transforms [6]. In cases where
complex calculations are required appropriate software is used [4,7], such as Matlab with its ODE suite, with several
ODE solvers, and graphical environment Simulink for dynamic system modeling, simulation and interactive analyses [7].
Mechanical systems generally fall into two broad categories, translational and rotational, but in practice, we often have
to deal with their combination [8]. Such hybrid systems contain building blocks from both categories, such as mass, spring
and dashpot as the elements of translation system, and the moment of inertia, the torsion spring, and the rotary damper,
as the elements of rotational system [2, 8, 9]. The second Newton's law of motion in appropriate forms for translational
and rotational motion provides the base for mathematical modelling of these systems:
d2 x
m = m‫ ∑ = ̈ݔ‬Xi (1)
dt2
d2 θ
J = mθ̈ = ∑ τi (2)
dt2
2
d x
where m states for mass, represents acceleration, ∑ Xi states for the sum of forces acting on material point, J is
and ‫ݔ‬ሷ
dt2
2
d θ
the moment of inertia, and θ̈ represents angular acceleration, and ∑ τi states for sum of moments or torques [6,7]. The
dt2
inputs are the force F, which is characteristic of translational systems, and/or moment or torque , which is characteristic

1-84
 13th International Scientific Conference
Science and Higher Education in Function of Sustainable Development – SED 2023

of rotational system. Those inputs provide outputs in the form of linear displacement x, and rotational displacement, or
angle  [8].
Unified modeling methodology has been proposed for modeling and simulating complex dynamic mechanical systems,
based on similarity between mechanical system and electric circuit composed of inductance (L), capacitance (C) and
resistance (R) [9]. Transfer function modeling was done using Node-Voltage analysis, by force/torque-voltage analogy
[10]. Mechanical connections between the driving motor and the working mechanism were analyzed, and the equivalent
electrical circuits were defined for the mechanical systems [11]. Lumped-system approach was used for obtaining the
mathematical model of a mechanical system in the form of ODEs [4]. Nonhomogeneous system of differential equations
of motion for mechanical system with two degrees of freedom, derived from Newton's second law, Lagrange‘s equations
and the Hamilton‘s equations, was solved in Matlab/Simulink using State-Space and Transfer Fcn blocks [12]. Dynamical
equations obtained for oscillating system composed of car seat with driver were solved with the aid of ODE solver in
Matlab, and in Simulink, using different model representations [13]. Hybrid systems of differential equations composed
from ODEs and PDEs has been derived for studying systems of interconnected rigid bodies mounted on an Euler-
Bernoulli beam with the aid of elastic joints. Based on these systems the proper software was developed with the aim to
analyze free vibrations of mechanical systems [14]. Some techniques for simulating dynamical systems based on neural
networks were provided [5].
For the purpose of this study, an example of hybrid compound system is presented. System is comprised of mass, springs,
dashpot, two racks, and a double gear, covering translation and rotational movement. The external force provides the
input, and the outputs in the form of linear and rotational displacement fully reflect the nature of the chosen system. This
simple mechanism, which enables motion transmission due to adjacent mechanical elements, is studied through several
steps. Mathematical model is derived in the form of two differential equations which describe translation and rotation
movement respectively. The system of differential equations is solved via Simulink with the aid of a Matlab script and
visualization and analysis of the results is conducted as well. The model is represented in a state-space and should be used
as a basis for obtaining the matrix of transfer functions. Visualization of the results of the study in the form of plots
indicates forced and underdamped oscillatory translational and rotational movements.

2. MODELING METHODOLOGY

The focus of this study is a hybrid compound mechanical system which contains a two-disc rigid pinion with moment of
inertia JC and radii R1 and R2 (double gear), two racks of masses m1 and m2, a body of mass m3, two pairs of linear springs
of stiffnesses c1 and c2, and a damper of viscous damping coefficient b (Figure 1). Pinion with fixed rotation passes
through point C meshes with racks (without slippage) which perform translational and parallel motions. External force F
acting on on the body of mass m3 has a dead zone that affects the force over the 1 to -1 N range.

Figure 1: Hybrid compound mechanical system [adapted from 6]

In order to derive a mathematical model for the proposed mechanical system it was necessary to decompose the system
into components and create free body diagrams (presented in Figure 2). The reaction forces originating from racks are
marked with FA and FB, while xA and xB denote the racks movements in the tangential direction. The elastic (spring) forces
are c1 xA , c1 (xA -x3 ), and c1 xB respectively, while the damping force is bẋ 3 , where ẋ 3 stands for velocity of body of mass
m3. The translational displacement of this body is denoted with x3.

Figure 2: Free-body diagrams

1-85
 13th International Scientific Conference
Science and Higher Education in Function of Sustainable Development – SED 2023

Applying Newton’s second law for translational and rotational motion (Equations (1) and (2)) yields a system of
differential equations:
JC θ̈ = FA R2 -FB R1
m1 ẍ A = -c1 xA -FA -c1 (xA -x3 )
m3 ẍ 3 = c1 (xA -x3 )+F-bẋ 3
0 = FB - 2c2 xB . (3)
The input in the system is the applied force F, while the outputs are translational displacement x3, and rotational
displacement . The coordinates xA and xB are connected with  as xA =R2 θ, and xB =R1 θ. After eliminating FA and FB,
using previously defined relations, mathematical model of the studied mechanical system is written in the following form:
θ̈ = a11 θ + a12 x3
ẍ 3 = a21 ẋ 3 +a22 x3 +a23 θ+a24 F, (4)
2c1 (R21 +R22 ) b c1 c1 R2 1
with: a11 = - , a12 = c1 R2 , a21 = - , a22 = - , a23 = , a24 = .
JC+m1 R22 m3 m3 m3 m3
Organizing the above differential equations into a state-space model is done by introducing the following state variables
to form the state vector {x}:
x1 = θ; x2 = θ̇ ; x3 = x3 ; x4 = ẋ 3 ; {x} = {x1 x2 x3 x4 }T . (5)
The independent variable t (time) has been dropped to simplify the notation, and all variables in Equations (4) and (5)
are functions of time. The second form of Equation (4) and the particular choice of the state variables of Equation (5)
indicate that
ẋ 2 = a11 x1 +a12 x3
ẋ 4 = a23 x1 +a22 x3 + a21 x4 +a24 F, (6)
where the obvious relations such as ẋ 2 = θ̈ , and ‫ݔ‬ሶ ସ = ‫ݔ‬ሷ
ଷ, which arise from Equation (5) are used.
Equations (6) are collected in vector-matrix form and represent a differential equation in the state-space model
ẋ 1 0 0 0 0 x1 0
ẋ 2 a11 0 a12 0 x2 0
൞ ൢ=൦ ൪൞x ൢ + ൦ ൪{F}. (7)
ẋ 3 0 0 0 0 3 0
ẋ 4 a23 0 a22 a21 x4 a24
The output equation is easily obtained from the first and the second expression in Equation (5) in the form
x1
θ 1 0 0 0 x2
൜ ൠ= ቂ ቃ൞ ൢ +{0}. (8)
x3 0 0 1 0 x3
x4
The last two equations form the state-space model. This model could be further used for calculating the transfer function
matrix [G(s)] using the expression presented in [6]
G(s) = C(sI-A)-1 B+D, (9)
where A, B, C, and D represent matrix from state space model, I stands for identity matrix, and s is the complex variable.

3. RESULTS AND DISCUSSION

Assigned values m3 = 0.6 kg, JC = 0.05 kgm2, c1 = 120 N/m, c2 = 100 N/m, b = 32 Ns/m, R2 = 2R1 = 0.03 m,
and expressions for calculating constants are entered into Matlab script, along with proper functions for plotting
translational displacement x3 versus time, and rotational displacement  versus time, under zero initial conditions (Figure
3). External force is sinusoidal with amplitude of 7 N and frequency of 2 rad/s.

Figure 3: Part of Matlab script

1-86
 13th International Scientific Conference
Science and Higher Education in Function of Sustainable Development – SED 2023

Differential equations (Equation (4)) for both cases are integrated in the Simulink model (Figure 4), when the input is
regular and when the input is with a dead zone.

Figure 4: Simulink model for hybrid mechanical system

The model contains two identical structures representing the left sides of the equations in system (Equation (4)). With the
aim to plot the values of the angle  in degrees, the output values are converted from radians using the Gain block. These
values, together with the values of translational displacement x3, are exported to the Matlab workspace using Simout
blocks in order to plot translational and rotational displacement versus time. Blocks thr and x3r are used to export the
output values in the case when the input is regular, while the x3dz and thdz blocks are used for exporting the output values
in the case when the input is with a dead zone. Two additional Simout blocks are added into the model (for regular input
Fr, and for input with dead zone Frdz) with the aim of creating plots that depict the change of force over time. After
runing the Matlab script, plots for input values of force are obtained in the case of a regular input (Fr, Figure 5), as well
as in the case of an input with a dead zone (Fdz, Figure 5).

Figure 5: Regular and dead zone input signal time-domain plots

1-87
 13th International Scientific Conference
Science and Higher Education in Function of Sustainable Development – SED 2023

Plots that represent pinion rotational displacement  versus time, and body translational displacement x3 versus time, are
presented in Figures 6 and 7.

Figure 6: Output time-domain plots for pinion's rotation angle in the cases of regular and dead zone input signal

Figure 7: Output time-domain plots for body's translational displacement in the cases of regular and dead zone input
signal
The output plots, presented in Figures 6 and 7, indicate oscillatory movements, and highlight the differences between the
effects of the two inputs. Dead zone affects the values of amplitude in the case of translational, as well in the case of
rotational movement, by reducing them. Both oscillatory motions are forced and underdamped, due to the external force
acting on the body, and can be treated as complex oscillations. The body is moving right and left in an alternating fashion
which causes the pinion to rotate clockwise and counterclockwise respectively. The graphs presented above are similar
as in the case of superposition of two sinusoidal with close amplitudes and slightly different frequencies, showing the so-

1-88
 13th International Scientific Conference
Science and Higher Education in Function of Sustainable Development – SED 2023

called beats with the difference frequency, and with slowly changing amplitudes. The period of the beat is easy to
determine from the plots and its value is 21.5 s (see Figure 6).

4. CONCLUSION

An efficient and accurate dynamic analysis method is required to address the problem of complex dynamic analysis
processes in hybrid mechanical systems. Presented methodology is easy to implement and should be used for similar
problems in mechanics, in order to predict how a real system evolves over time under given conditions, as well as for
teaching purposes in subjects that focus on solving dynamic behavior of mechanical systems. The results are visualized
in such manner that allows monitoring the system behavior over time. Inputs to the system are easy to modify by changing
their parameters or even their nature. The obtained outputs can be compared in order to find the optimal solution for the
set of initial conditions.

REFERENCES

[1] S. K. Sharma, and U. K. Gupta, "An approach to mathematical modelling of mechanical and electrical systems and
engineering", International Journal of Statistics and Applied Mathematics, vol. 2, no. 4, pp. 52-56, 2017.
[2] A. Patil, and S. More, "Mathematical Modeling of Physical System", IOSR Journal of Electrical and Electronics
Engineering (IOSR-JEEE), vol. 9, no. 3 Ver. II, pp. 57-64, May - Jun 2014.
[3] K. A. Craig, "Introduction to Dynamic Systems and Control," in Dynamic Systems: Modeling, Simulation, and
Control, John Wiley & Sons, Inc., 2015, Chapter 1, pp. 1-6.
[4] C. M. Legaard, T. Schranz, G. Schweiger, J, Drgoňa, B. Falay, C. Gomes, A. Iosifidis, M. Abkar, and P. G. Larsen,
"Constructing Neural Network-Based Models for Simulating Dynamical Systems", ACM Comput. Surv., vol. 1, no. 1,
Article, pp 1-35, July 2022. Accessed: Feb. 05, 2023. doi: 10.48550/arXiv.2111.01495. [Online].
Available: https://arxiv.org/abs/2111.01495
[5] N. M. Bhondge, M. V. Sarode, and A. D. Chaudhari, "Mathematical Modeling of Mechanical and Electrical System
using Differential Equations", International Journal of Artificial Intelligence and Mechatronics, vol. 4, no. 3, pp. 100-
102, 2015.
[6] N. Lobontiu, System Dynamics for Engineering Students, Concepts and Applications, Second Edition, Elsevier, 2018.
[7] M. Tufegdžić, V. Jevremović, Z. Petrović, and V. Bojović, "Modeling material point's motion under the action of air
resistance", presented at 28. IKT konferencija YUINFO 2022, Kopaonik, Serbia, 13 .- 16. mart, 2022., str. 303-307.
[8] D. K. Chaturvedi, Modeling and Simulation of Systems Using MATLAB and Simulink, 1st Edition, CRC Press, 2010.
[9] M. Akbaba, "Modeling and simulation of dynamic mechanical systems using electric circuit analogy", Turkish Journal
of Engineering, vol. 5, no. 3. pp. 111-117, 2021. doi: 10.31127/tuje.695769.
[10] A. V. Nair, and K. Madhavan, "Mathematical Modelling of Linear Complex Mechanical Systems by Inspection
Method based on Force/Torque-Voltage Analogy", International Journal of Current Engineering and Technology, vol.
8, no. 2, pp. 450-452, March/April 2018. Accessed: Feb. 06, 2023. doi: 10.14741/ijcet/v.8.2.39. Available:
http://inpressco.com/category/ijcet
[11] A. Popenda, "Mathematical modelling of real transmission shafts and mechanical connections with clearances",
Przegląd Elektrotechniczny, R. 93, NR 1, pp. 189-192, 2017. doi:10.15199/48.2017.01.46
[12] P. Siváka, and D. Hroncová, "State-Space model of a mechanical system in MATLAB/Simulink", Procedia
Engineering, vol. 48, pp. 629 – 635, 2012. doi: 10.1016/j.proeng.2012.09.563
[13] D. Hroncová, A. Gmiterko, T. Lipták, and M. Bruchánek, "Computer simulation of the mechanical system with one
degree of freedom", Acta Mechatronica - International Scientific Journal about Mechatronics, vol. 1, no. 4, pp. 13-18,
2016.
[14] A. D. Mizhidon, "A generalized mathematical model for a class of mechanical systems with lumped and distributed
parameters", AIMS Mathematics, vol. 4, no. 3, pp. 751–762.,2019. doi: 10.3934/math.2019.3.751

1-89

View publication stats