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

net/publication/325717124

An enhanced simulation model for DC motor belt drive conveyor system

control

Conference Paper · May 2018

DOI: 10.1109/MOCAST.2018.8376636

CITATIONS READS

4 1,026

3 authors:

Andromachi G. Katsioula Yannis L Karnavas

Democritus University of Thrace Democritus University of Thrace
1 PUBLICATION   4 CITATIONS    61 PUBLICATIONS   326 CITATIONS   

SEE PROFILE SEE PROFILE

Yiannis Boutalis
Democritus University of Thrace
186 PUBLICATIONS   3,098 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Fuzzy Cognitive Networks on Fault Detection, Prognostics and Health Management Applications View project

NeuroFuzzy adaptive control View project

All content following this page was uploaded by Yannis L Karnavas on 11 September 2018.

The user has requested enhancement of the downloaded file.

 2018 7th International Conference on Modern Circuits and Systems Technologies (MOCAST)

An Enhanced Simulation Model for DC Motor

Belt Drive Conveyor System Control

Andromachi G. Katsioula Yannis L. Karnavas Yiannis S. Boutalis

Electrical Machines Laboratory Electrical Machines Laboratory Automatic Control Systems & Robotics Lab.
Dept. of Electrical & Computer Engineering Dept. of Electrical & Computer Engineering Dept. of Electrical & Computer Engineering
Democritus University of Thrace Democritus University of Thrace Democritus University of Thrace
Xanthi, Hellas (GR) Xanthi, Hellas (GR) Xanthi, Hellas (GR)
[email protected] [email protected] [email protected]

Abstract—Belt drive systems provide freedom to position the have changed or environment disturbances have occurred.
motor relative to the load and this phenomenon enables Adaptive fuzzy control [8], sliding mode control [2-5] as well
reduction of the conveyor inertia. It also facilitates quick as disturbance observer techniques [9] have been utilized to
response when employed in robotics applications. Unfortunately, improve performance. Also, neuro-fuzzy and genetic
the flexible dynamics deteriorates the positioning accuracy.
Therefore, there exists a trade-off between the simplicity of the
algorithms have been proposed in [10-12] in order to improve
control strategy to reject time varying disturbance caused by the PID control strategy. System efficiency was the focus of
belt's flexibility and the precision in performance. Further, another intelligent control strategy in [13], whereas self-
resonance of the system leads to vibrations and poor accuracy in learning approach shown in [14] achieved satisfactory speed
positioning. In this paper, an overall model of belt drive conveyor control. More relevant accurate results were obtained in [15].
system is presented which is enhanced by the dc motor model However, the main drawback of the aforementioned efforts
and an realistic friction model. This leads to more accurate was again the use of a rather simple BDCS model, rather than a
positioning of the belt driven mechanism. Thus, reliable more representative to the actual system. Therefore, the need
simulations can be performed prior to any controller design. The for more accurate models along with advanced control methods
effectiveness of the proposed model was proven by means of
simulation -where actual belt driven system parameters used-
is emerged. In the above context, an attempt is made in the
which carried out with and without these enhancements. present paper, to provide an enhanced simulation BDCS model
which takes into account system non-linearities due to a)
Keywords—DC motor, belt drive system, non-linear system friction, vibration and belt flexibility components and b) the dc
modeling, simulation, industrial applications, controller design motor model driving the system. The validity of the procedure
followed and the subsequent proposed model is examined by
I. INTRODUCTION comparing its response results over its simpler counterparts.
A belt drive conveyor system (BDCS) is a typical energy II. PROBLEM DESCRIPTION
conversion system from electrical to mechanical energy. It is
widely used in agricultural machines, electric generators, A. Brief Theoretical Background
robotic arms, machine tools, and textile machines. It is a The structure of a belt drive system is presented in Fig. 1.
simple, lightweight, low cost power transmission system and This system consists of a belt, a driving and a driven pulley, a
provides flexibility in situating the motor relative to the load, motor, a speed reducer and a cart. The driving and driven
allowing for high speed positioning [1]. However, the non- pulleys are used to stretch the belt so that the cart can be
linear friction, vibration and belt flexibility compose a set of regulated to the desired position precisely. The motor and the
negative factors for a precision control which may in turn lead speed reducer are coupled to the driving pulley so that the belt-
to limit reliability of the controlled system. Problems pulley sub-system can be rotated. This rotation is converted
associated with precision control of such systems, have gained into linear motion of the cart, which represents the load side of
the focus of the researchers over the last decades. In [2-7], the system. In Fig. 1, the system layout, its physical
control algorithms achieved vibration suppression and representation as well as some of its parameters are depicted.
improved servo performance under system uncertainties.
A BDCS is difficult to be controlled accurately due to the
abovementioned problems. In other words, its control
constitutes of the utmost importance so that the parameters can
better meet the requirements of the process. Several control
techniques to improve the precision of a BDCS have been
proposed so far. Firstly, the conventional PI controller had
been used for simplicity, but as a technique depends on the
precise mathematical model of the plant. Thus, it cannot
provide the high dynamic performances as required by variable (a) (b)
speed regulation of belt conveyor, when the system variables Fig. 1. Belt drive system, a) layout representation, b) physical model.

XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE

978-1-5386-4788-2/18/$31.00 ©2018 IEEE

 2018 7th International Conference on Modern Circuits and Systems Technologies (MOCAST)

The mathematical model of this system can be described by the III. MODEL ENHANCEMENT
following set of second order differential equations [2]:
A. Pulleys geometry
( J1 + G 2 ( J G + J M ))q&&1 + τ f 1 = All studies which have been done so far, have used the
= GT − R[ K1 ( x)( Rq1 − x) − K3 ( Rq2 − Rq1 )] model described by Eq. (1), assuming that the pulleys' radii are
(1) identical (R1=R2=R). In the present application radii will be
J 2 q&&2 + τ f 2 = R[ K 2 ( x)( x − Rq2 ) − K3 ( Rq2 − Rq1 )]
regarded as unequal, so a generic model is derived as follows:
x + Ff = K1 ( x)( Rq1 − x) − K 2 ( x)( x − Rq2 )
M c &&
( J1 + G 2 ( J G + J M ))q&&1 + τ f 1 =
where J1, J2 are the inertia moments of driving and driven
pulley, G is the speed reducer ratio and JG, JM are the inertia = GT − [ R12 K1 ( x)q1 − R1K1 ( x) x − R22 K 3q2 + R2 K 3 R1q1 ]
(6)
moments of speed reducer and servomotor, respectively. J 2 q&&2 + τ f 2 = R2 [ K 2 ( x) x − R2 K 2 ( x)q2 − R2 K 3 q2 + K 3 R1q1 ]
Further, q1 and q2 are the angular positions of driving and
driven pulley, τf1 and τf2 are the corresponding friction x + Ff = K1 ( x) R1q1 − K1( x) x − K 2 ( x) x + K 2 ( x ) R2 q2
M c &&
moments, T represents the torque developed by the motor, R is
the outer radius of the pulleys, K1, K2 and K3 denote belt’s B. DC motor model
elasticity coefficients, Mc is the cart mass, x is the cart position Models described either by Eq. (1) or Eq. (6), consider the
and Ff is the friction force which acts to the cart. driving torque T provided by the motor to be constant w.r.t.
overall system dynamics. Since this is not the case in the real
B. Flexibility of the belt world problem -because a motor presents its own dynamics-, a
Since belt flexibility is one of the main disadvantages of a motor model should be adopted and implemented appropriately
BDCS, its modeling is of great importance. In model of Eq. to complete the accurate modeling of the system. A typical DC
(1), the contribution of this factor is presented from the motor is described by the following well-known set of
coefficients Ki (i=1,2,3) which can be calculated as: equations (both for electrical and mechanical parts),
Fl ∆l Va = Ea + I a Ra + La ( dI a dt ) , where Ea = kt Φωm
Ki = , where ε = (2)
ε (li − x) li Tem = TL + T f + Tout = kt ΦI a , where ωm = dθ dt (7)
Eq. (2) implies that coefficient Ki is dependent on an external T f = Bmωm , and Tout = J M ( d ωm dt )
force (Fl) applied along the belt's length, the ratio (ε) of the
change in length to the total belt length, and the cart's relative where Va, Ea, Ia, Ra, La are the armature voltage, the back-EMF,
position (li-x), determining the extent of its tension. It is the armature current, the armature resistance and armature
assumed that there is no slip between the belt and the pulleys inductance respectively. Also, Tem, Tf, Tout and TL are the
of this system. electromagnetic torque, friction torque, output torque and load
torque respectively. Finally, Bm, Φ and ωm are the viscous
C. System inertia friction coefficient, the magnetic flux and the rotor’s
mechanical speed. Eqs. (7) can be modeled as shown in Fig. 2
For effective modeling, the system components’ inertia and this motor model is used here (upper right block in Fig. 3).
calculation is necessary, including all its rotational parts.
Specifically, these parts can be the pulleys, the motor's rotor,
C. Friction model
the coupling which connects the motor shaft with the gearbox,
the gearbox and a speed encoder. Thus, the total inertia is: Finally, in order to complete the accurate modeling of the
system, the friction contribution should be studied and
J tot = J1 + J 2 + J G + J M + J coupling + J encoder (3) implemented. Friction is a very complex phenomenon affecting
the response of the system to a great extent. For simplicity, it is
In our case, it is assumed that the last two (inertia of the considered that friction is applied only to the load side. The
coupling and the encoder) are included in the inertia of the Tustin’s friction model [17] was chosen for implementation in
motor, and the driving/driven pulleys inertia are equal (J1=J2). the BDCS and is shown in Eq. (8). It is dependent on cart’s
In general, the inertia for any component taken into account position and describes satisfactorily the friction at low speeds.
which presents a mass (m), a radius (r) and a corresponding Also, it may help the system to prevent (damping) possible
mass per volume unit (ρ) can be calculated by, [16]: “sharp” oscillations in the beginning of the movement.
∆m dm x&
F ( x& ) = Fc sgn( x& ) + ( Fs − Fc ) exp[−( )] + Fv x& (8)
J i = ∫ ρi ri 2 dV , where ρi = lim = , dm = ρ dV (4) x&s
∆V dV
As shown in Eq. (8), friction model is composed of three types
Indicatively, the driving pulley’s, inertia can be calculated as: of friction; static friction, Coulomb friction and viscous
friction. Each one has its own particularity and use. Static
πρ L( R14 − r14 )
r2

J1 = ∫ r 2 dm = 2πρ L ∫ r 3dr = (5) friction opposes to all motion until the applied force magnitude
r1
2 is less than the maximum stiction force Fs. Coulomb friction Fc
where r1 and R1 are the inner and outer radius respectively and opposes the relative motion. Finally, viscous friction Fv is
L is the axial length of the pulley. proportional to the velocity and produced by the viscosity of
lubricants.
 2018 7th International Conference on Modern Circuits and Systems Technologies (MOCAST)

Fig. 2. Analytical DC motor simulation model implemented and used here.

Fig. 3. Overall proposed DC motor BDCS simulation model.

The reference cart position profile is set at 0.15 m at t=0. It

IV. SIMULATION RESULTS
should be also noted that in all cases, the gains of the involved
Fig. 3 depicts the overall developed BDCS simulation PI controllers were optimally tuned through genetic algorithm.
model and its performance is presented next. In order to Also, Model C uses two PI controllers (for motor's speed and
validate its robustness, numerous simulations were conducted torque). In Fig. 4 the time response of the cart position with
(not all shown here) under several operating conditions and the respect to the reference profile is depicted. It can be seen that
results were compared with those of its simpler counterparts. Model A presents higher settling time and is much more
The examined models are: oscillatory than Models B and C. Model B however, presents
Model A. BDCS model without Tustin's friction model and competitive settling time with this of Model C, though exhibits
without DC motor model, higher overshoot. The respective angular velocities time
Model B. BDCS with Tustin's friction model without DC responses can be observed in Fig. 5. It can be seen from this
motor model, figure that for Model A, the driven pulley is subjected to very
Model C. BDCS with Tustin's friction model and with DC high acceleration/deceleration while trying to follow driving
motor model (proposed), pulley. On the other hand, in Model B, presents very good
positioning speeds but, as it was mentioned, there are no any
 2018 7th International Conference on Modern Circuits and Systems Technologies (MOCAST)

Indicative simulation results from models’ operation revealed

its reliable and proper functionality. Future work includes an
undergoing effort towards different control strategies adoption
including fuzzy logic and bio-inspired optimization techniques.

REFERENCES
[1] T. S. S. Jayawardene, M. Nakamura, S. Goto, “Accurate control position
of belt drives under acceleration and velocity constraints”, International
Journal of Control, Automation, and Systems, vol. 1, no. 4, December,
Fig. 4. Time responses of cart position (for Model A, Model B & Model C). 2003, pp. 474-483.
[2] A. Hace, K. Jezernik and A.Sabanovic, “SMC with disturbance observer
for a linear belt drive”, IEEE Transactions on Industrial Electronics,
vol. 54, no. 6, November 19, 2007, pp. 3402–3412.
[3] A. Hace, K. Jezernik and A.Sabanovic, “Improved design of VSS
controller for a linear belt-driven servomechanism”, in Proc. of the 8th
IEEE International Workshop on Advanced Motion Control, Kawasaki,
Japan, March 25-28, 2004.
[4] A. Hace, K. Jezernik and A. Sabanovic, “A new robust position control
algorithm for a linear belt-drive”, in Proc. of the IEEE International
Conference on Mechatronics, Istanbul, Turkey, June 3-5, 2004.
Fig. 5. Time responses of angular velocities of driven pulley of the BDCS [5] A. Sabanovic, O. Sozbilir, G. Goktug and N. Sabanovic, “Sliding mode
control of timing-belt servosystem”, in Proc. of IEEE International
(for Model A, Model B and Model C).
Symposium on Industrial Electronics, Rio de Janeiro, Brazil, June 9-11,
2003.
[6] T. F. Lee and A. C.Huang, “Vibration suppression in belt-driven servo
systems containing uncertain nonlinear dynamics”, Journal of Sound
and Vibration, vol. 330, no. 1, January 3, 2011, pp. 17-26.
[7] J. Wu, G. Xu and Z. Yin, “Residual vibration reduction for a flexible
positioning system using input shaping technique”, in Proc. of the 7th
World Congress onIntelligent Control and Automation, Chongqing,
China, June 25-27, 2008.
[8] M. A. El-Sharkawi and Y. Guo, “Adaptive fuzzy control of a belt-driven
precision positioning table”, in Proc. of the IEEE International Electric
Fig. 6. Time response of the dc motor's electromagnetic torque (Model C). Machines and Drives Conference, Madison, WI, USA, June 1-4, 2003.
[9] C. Mitsantisuk, S. Katsura and Kiyoshi Ohishi, “Sensorless interaction
TABLE I. COMPARISON OF SIMULATION MODELS force control based on modal space design for twin belt-driven system”,
in Proc. of SICE Annual Conference, Takamatsu, Japan, September 17-
Model Controllers' Gains Settling time Max. Overshoot 20, 2007.
[10] O. F. Lutfy, S. B. Mohd Noor, M. H. Marhaban and K. A. Abbas, “Non-
“A” Kp = 128.36458, Ki = -130.36340 >> 1 sec 196.53 %
linear modelling and control of a conveyor-beltgrain dryer utilizing
“B” Kp = 15.03437, Ki = -57.05871 1.82 sec 11.86 % neuro-fuzzy systems”, Journal of Systems and Control Engineering, vol.
225, no. 5, July 2011, pp. 611-622.
Kpw = -0.41649, Kiw = 42.21471
“C” 0.78 sec 0.00 % [11] X. Chen, Md. L. Rahman, Y. Zhang, “Adaptive fuzzy PID speed control
Kpt = 24.66191, Kit = -766.75432
of DC belt conveyor system”, in Proc. of the 17th IEEE/ACIS
International Conference on Software Engineering, Artificial
motor dynamics included in the system in this case. The latter Intelligence, Networking and Parallel/Distributed Computing (SNPD),
is presented in the same figure for Model C where a more Shanghai, China, May-June 30-1, 2016.
realistic driven pulley speed is achieved. Especially for Model [12] Ma Xian-Min, “Genetic Taboo hybrid strategy for PID regulator
C, the motor's electromagnetic torque is shown in Fig. 6, parameter adaptation in belt conveyor”, in Proc. of International
whereas the relevant dynamics are exhibited and justify the Conference on Machine Learning and Cybernetics, Dalian, China,
August 13-16, 2006.
enhancement proposed here. Finally, the corresponding PI
controller gains are shown in Table I, along with each model’s [13] L. B. Ristic and B. I. Jeftenic, “Implementation of fuzzy control to
improve energy efficiency of variable speed bulk material
settling time and maximum overshoot characteristics. It is clear transportation”, IEEE Transactions on Industrial Electronics, vol. 59,
for this Table that Model C presents better response no. 7, September 26, 2011, pp. 2959-2969.
characteristics compared to those of the other two models. [14] J. Li, G. You and S. Lei, “Fuzzy and frequency control system study on
belt conveyor”, in Proc. of the 4th International Conference on
Bioinformatics and Biomedical Engineering, Chengdu, China, June 18-
V. CONCLUSIONS 20, 2010.
A simulation model for a belt drive conveyor system is [15] A. S. Kulkarni and M.A. El-Sharkawi, “Intelligent precision position
developed and presented in this paper. The model is able to control of elastic drive systems”, IEEE Transactions on Energy
take into account the inherent high non-linearities imposed by Conversion, vol. 16, no. 1, March 2001 pp. 26-31.
the real world’s system such as friction, vibration and [16] R. A. Serway, Physics: For scientists and engineers, 1982, ISBN 0-03-
057903-1.
resonance components. Furthermore, the model proposed was
enhanced by the dc motor model driving the system, which [17] S. S. Ge, T. H. Lee and S. X. Ren, “Adaptive friction compensation of
servo mechanisms”, International Journal of Systems Science, vol. 32,
allows the motor’s dynamics to be taken into account as well. no. 4, 2001, pp. 523-532.

View publication stats