Applied Mechanics and Materials Vols 1-2 (2004) pp 51-58

© (2004) Trans Tech Publications, Switzerland

doi:10.4028/www.scientific.net/AMM.1-2.51

On Non-Ideal Simple Portal Frame Structural Model: Experimental

Results under a Non-Ideal Excitation
J.M. Balthazar1,a, R.M.L.R.F. Brasil2,b, F.J. Garzeri3, c
1
State University of São Paulo at Rio Claro, PO Box 178, 13500-230, Rio Claro, SP, Brazil.
2
University of São Paulo, PO Box61546, 05424-970, São Paulo, Brazil.
3
General Motors of Brazil
a
jmbaltha@unesp.rc.unesp.br, b rmlrdfb@usp.br, c garzeri@terra.com.br

Keywords: Non-ideal motors, Non-linear dynamics, Experimental analysis

Abstract. We present measurements of the non-linear oscillations of a portal frame foundation for
a non-ideal motor. We consider a three-time redundant structure with two columns, clamped in their
bases and a horizontal beam. An electrical unbalanced motor is mounted at mid span of the beam.
Two non-linear phenomena are studied: a) mode saturation and energy transfer between modes; b)
interaction between high amplitude motions of the structure and the rotation regime of a real limited
power motor. The dynamic characteristics of the structure were chosen to have one-to-two internal
resonance between the anti-symmetrical mode (sway motions) and the first symmetrical mode
natural frequencies. As the excitation frequency reaches near resonance conditions with the 2nd
natural frequency, the amplitude of this mode grows up to a certain level and then it saturates. The
surplus energy pumped into the system is transferred to the sway mode, which experiences a sudden
increase in its amplitude. Energy is transformed from low amplitude high frequency motion into
high amplitude low frequency motion. Such a transformation is potentially dangerous.
We consider the fact that real motors, such as the one used in this study, have limited power
output. In this case, this energy source is said to be non-ideal, in contrast to the ideal source whose
amplitude and frequency are independent of the motion of the structure. Our experimental research
detected the Sommerfeld Effect: as the motor accelerates to reach near resonant conditions, a
considerable part of its output energy is consumed to generate large amplitude motions of the
structure and not to increase its own angular speed. For certain parameters of the system, the motor
can get stuck at resonance not having enough power to reach higher rotation regimes. If some more
power is available, jump phenomena may occur from near resonance to considerably higher motor
speed regimes, no stable motions being possible between these two.

Introduction
Usually, an experimental structural analysis supposes that the excitation devices are ideal or, in
other words, non-dependent of the dynamic response of that structure. Electric motors are an
example of non-ideal excitation devices, because the mechanical output power depends on the
motion of its armature and on the dynamic characteristics of its rotor, among other factors. When
motors are attached to structures that need excitation power levels similar to the power capacity of
those motors, interesting non-linear phenomena may happen, such as: modal saturation and
Sommeferld effect. There are excellent theoretical and experimental researches about this subject;
the well known ones, but not recent, are: [1], [2], [3], [4] and [5]. Some researchers have carried out
theoretical and experimental analyses of portal frames excited by non-ideal excitation devices [6],
[7] and [8]. A first announcement of this work was done in [8].

A non-linear experimental analysis of a metallic portal frame excited by a DC motor is presented

in this work. The two columns of the structure were clamped in their bases to a large seismic mass
and the motor was attached to the center of the horizontal beam. In order to allow for non-linear
phenomena observation possible, the dimensions of the structure were chosen in such a way that
there was a ratio of one-to-two between the first two frequencies of the first two natural modes
(anti-symmetrical or sway mode and symmetrical or bending mode); for the same reason, all joints

All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of Trans
Tech Publications, www.ttp.net. (ID: 128.42.202.150, Rice University, Fondren Library, Houston, USA-14/04/15,01:42:40)
52 Advances in Experimental Mechanics

between the columns and the horizontal beam were welded, in order to reduce the structural
damping. Modal saturation, Sommerfeld effect and other interesting phenomena were observed by
using modern experimental techniques.

Fig. 1 Multiple flash exposure of modal saturation phenomenon in the portal frame analyzed.
Angular motor speed: 1298rpm; lateral acceleration in the middle of the horizontal beam: 23m/s2;
vertical acceleration: 99.8m/s2; lateral sway frequency: 10.9Hz; vertical frequency: 21.8Hz.

Experimental set-up
All tests were performed with the columns of the structure clamped to a seismic mass of 4000kg, in
a temperature-controlled room (See Fig. 1). First, the structural system was analyzed with its motor
turned off, in order to confirm the 1:2 ratio between the first two natural modes. Modal analysis was
used to confirm the frequencies values as well as the mode shapes (nine low-weight accelerometers,
a small electrodynamics shaker and a force transducer were attached to the structure; a digital data
acquisition system recorded the dynamic signals). Free vibration tests were also performed. Vertical
and horizontal static displacements were independently set and, right after, suddenly released by a
device specially built for that purpose. Vertical and horizontal acceleration components in the
middle of the horizontal beam were recorded during the motion dissipation. A third set of tests was
carried out with the DC motor running. During those tests, the vertical and horizontal acceleration
components in the middle of the horizontal beam were also obtained, as well as the electrical power
characteristics delivered to the DC motor (current and voltage) and its angular velocity. Basically,
the test procedure was the following: with the motor stopped, a certain voltage was set to the
electric power supply and, at the same time, data acquisition was started. The test was considered
ended when the motor reached a constant angular velocity. Finally, some tests were carried out with
the DC motor running, passing through the second system resonance. The same physical parameters
obtained in the previous tests were recorded (accelerations, angular velocity of the motor, voltage
and electric current). The characteristics of the instrumentation used during the tests played an
important role in the acquisition of useful dynamic data. Some details are mentioned below. ICP
accelerometers were used in order to reduce low frequency noise, usually found when the
piezoelectric type is used. Besides, pre-amplifiers with filters guaranteed a good signal-to-noise
ratio between 1 to 1000Hz. A fast optical encoder was used to get the angular velocity of the DC
motor. By this sensor, it was possible to observe small variations of rpm every 1/12 of turn of the
rotor. A low noise, fast, simultaneous multi-channel data acquisition was fundamental to get reliable
dynamic data. The PC-based equipment used had four channels, 85 dB signal-to-noise ratios and a
variable sampling rate (up to 44kHz).
 Applied Mechanics and Materials Vols. 1-2 53

Results for free vibrations

Fig. 2 shows the horizontal and vertical acceleration component measured in the middle of the
horizontal beam for an initial horizontal displacement of 4.0 ± 0.1 mm.

Fig. 2 Acceleration peaks measured in the middle of the horizontal beam during free vibrations of
the system. Initial horizontal static displacement: 4.0 ± 0.1 mm

20

15 4

H o rizo n ta l a cce le ra tio n

Horizontal Acceleration (m/s2 )

10

Vertical Acceleration(m/s2 )
2
5

0 0

-5

-2
-1 0

V e rtica l a cce le ra tio n

-1 5
-4

0 20 40 60 80 100 120 1 40
T im e (s )

Results for forced vibrations

Figs. 3, 4 and 5 present the motor angular velocity and the acceleration peaks in the middle of the
horizontal beam when a constant voltage is supplied to the motor. Three quite different dynamic
behaviors can be observed, depending on how close the angular velocity of the motor is to the 2nd
system resonance.
54 Advances in Experimental Mechanics

Horizontal
20 2nd Resonance

Vertical
1300
2 Angular Velocity

Vertical Acceleration
1100
1 10
Acceleration (m/s2 )

900

RPM
0
0 1st Resonance 700
Angular Velocity

-1 500

-10
300
Horizontal Acceleration
-2 Rotor Rotation

100

8 2 3 4 5 6 7 8 2 3 4 5 6 7 8
10 0 10 1 10 2
Time (s)
Fig. 3 Acceleration peaks in the middle of the horizontal beam during acceleration of the DC
motor. Power Supply Voltage 2.70±0.05V

100
1300
4 2nd Resonance Angular Velocity
Horizontal

Vertical

1100
50 Horizontal Acceleration
2

900

RPM
Acceleration (m/s2 )

0
0
700
1st Resonance
Angular Velocity
-2 -50 500

Vertical Acceleration 300

-4 -100
Angular Velocity
100
10 30 50 70 90
Tim e (s)

Fig. 4 Acceleration peaks measured in the middle of the horizontal beam during acceleration of the
motor. Power supply voltage: 2.83±0.05V.
 Applied Mechanics and Materials Vols. 1-2 55

20

Horizontal
2nd Resonance

Vertical
2 1300
Angular Velocity Vertical Acceleration
10
1
Acceleration (m/s2 )

900
0 0

RPM
1st Resonance
-1 Angular Velocity
-10
500

-2 Horizontal Acceleration
-20
Angular Velocity

100
7 8 2 3 4 5 6 7 8 2 3 4 5 6 7 8
10 0 10 1 10 2
Time (s)

Fig. 5 Accelerations peaks measured in the middle of the horizontal beam during acceleration of the
motor. Power supply voltage: 3,00±0.05V

Fig. 6 shows the occurrence of the modal saturation phenomenon. The sudden increase of the
horizontal acceleration peak measured in the middle of the beam is observed; in the same instant,
occurs a short variation of the vertical acceleration peak; during the phenomenon, the vertical
acceleration stays stable. In the same figure, the rotation of the electric motor deserves special
attention. Some variation is observed around the value corresponding to 2nd resonance during
elapsing of the modal saturation phenomenon.

100 2nd R esonance

1300
A n g u la r V e lo c ity
A n g u la r V e lo c ity
Angular Velocity (rpm)
Acceleration (m/s )

50
2

1200
H o riz o n ta l A c c e le ra tio n
0

1100
-5 0

-1 0 0 1000
V e rtic a l A c c e le ra tio n

-1 5 0
900
5 30 55 80 105 130 155 180
T im e (s )

Fig. 6 Acceleration peaks measured at the middle of the beam during occurrence of the modal
saturation phenomenon
 56 Advances in Experimental Mechanics

Fig. 7 and 8 shows the behavior of the measured behavior of the system during the passage
through the 2nd resonance with increasing rotation of the motor. There is an instability zone near the
resonance (shown shaded) where abrupt accelerations of the angular velocity of the electric motor
were observed. When raising continually the electric tension delivered to the motor (when its
angular velocity is close to the 2nd resonance of the system), its rotation is increased, until the
moment that it is stabilized around a certain value (around the 2nd natural frequency of the system).
Successive increments of the tension only elevate the electric power absorbed by the motor; all
energy is used by the structural system to increase its vibration amplitudes. The process possesses a
limit: when the structure is not more able to absorb the energy, a sudden increase of the angular
velocity is observed. After a certain period, the rotation of the motor stabilizes at a proportional
value to the feeding tension.

Transient 3.0V
18 Zone Vertical

Transient
120 Component
Electric Power (W)

2.80V

Zone
Acceleration (m/s2 RMS)
16

80

14

2.70V
2nd Resonance

40
Horizontal
12
Component
2.5V
0
10
1150 1200 1250 1300 1350 1400
19 20 21 22 23
Angular Velocity (rpm)
Frequency (Hz)

Fig. 7 Electric power absorbed during motor

acceleration through 2nd mode resonance Fig. 8 Acceleration peaks in the middle
of the beam during motor acceleration
through 2nd mode resonance

Conclusions
Considering the results presented and several other experiences performed with the structure of
Fig.8, some conclusions are drawn: modal saturation phenomenon can happen in structures subject
to small displacements; in the analyzed case, the manifestation of that phenomenon was observed
with displacements of the order of 1:200; changes of energy between the first two natural modes
depends on the relationship between their frequencies and on the structural damping; the transients
observed before and after the modal saturation phenomenon are related to those factors;
Sommerfeld effect depends on the electric and dynamic characteristics of the motor; kinetic energy
(stored in the flywheels and in the rotor) and torque (related to the feeding current, motor
temperature and conditions of the brushes and of the collector) dictate the way as the motor will be
captured by the resonance of the system, as well as the fast acceleration of the angular speed of the
rotor after the modal saturation phenomenon; electric brush motors are not very well appropriate for
the analysis of dynamic events as those presented in this work.; the commutator temperature and the
incapacity of the system that presses the brushes against the commutator segments in keeping a
constant pressure (when the structural system is in resonance) affects the observation of the
Sommerfeld effect and of the modal saturation phenomenon. We also remarked that, in the general,
the success of the kind of analysis presented here is due to three factors: to the judicious choices of
 Applied Mechanics and Materials Vols. 1-2 57

the structural parameters, which allowed only two natural vibration modes to characterize the
dynamic response of the system; to the constructive characteristics, which contributed to the
reduction of the structural damping; to the instrumentation, which made it possible to record low
amplitude transients and time-consuming phenomena. Although not presented here, many other
non-linear phenomena were observed, such as: transfers of energy between the electric motor and
the structure, transients during the deceleration of the motor, the dependence of non-linear
phenomena with the excitation magnitude and the changes of energy between the first two vibration
modes when the excitation frequency was close to the first natural mode.

Acknowledgements
The authors thank the support given by FAPESP, Fundação de Apoio à Pesquisa do Estado de São
Paulo, by CNPq, Conselho Nacional de Desenvolvimento Científico e Tecnológico and by General
Motors of Brazil.

References
[1] A . Sommerfeld. Beitrage zum dynamischen ausbau de festigkeitslehre”. Physikalische
Zeitschr, Vol. 4, (1904), p. 631-636.
[2] V. O. Kononenko: Vibrating Systems with a Limited Power Supply London (Illife Books
1969)
[3] JM Balthazar, DT Mook, DT, RMLRF Brasil, HI Weber, A. Fenili, D. Belato, JLP Felix: An
Overview on non-ideal vibrations, Meccanica, 38, (2003), p. 613-621.
[4] Balachandran B., A Theoretical nd Experimental Study of Modal Interacions in Resonantly
Forced Structures, (1990), Phd Dissertation, Vrginia Tech, Blacksburg, Virginia, USA.
[5] Mattos MC, J M Balthazar, DT Mook, S. Wieczorek. Experimental Study of Vibrations of
Non-ideal Systems. (1997), (Proceedings of DETEC’97. CD ROM ASME Design Engineering
Technical Conferences, Sacramento, CA, 1997), 10 pages.
[6] Garzeri, F.J.: On The Dynamical and Experimental Analyzis of a Non-Ideal Portal Frame
Structure( In Portuguese), Phd Dissertation, SP, Brazil.
[7] Felix JLF: On Dynamics and Control of a Non-Ideal Portal Frame Strcuture (In Portuguese),
(2002), Phd Dissertation, Campinas, SP, Brazil.
[8] MRF Brasil, J M Balthazar, F J Garzeri, Experimental Validation of a Non-Linear Model
Under Non-Ideal Excitation, (2003), (Proceedings of DETEC’03. CD ROM ASME Design
Engineering Technical Conferences, Chicago, Illinois, USA, 2003), 10 pages.
Advances in Experimental Mechanics
10.4028/www.scientific.net/AMM.1-2

On Non-Ideal Simple Portal Frame Structural Model: Experimental Results under a Non-Ideal
Excitation
10.4028/www.scientific.net/AMM.1-2.51

DOI References
[3] JM Balthazar, DT Mook, DT, RMLRF Brasil, HI Weber, A. Fenili, D. Belato, JLP Felix: An verview on
non-ideal vibrations, Meccanica, 38, (2003), p. 613-621.
doi:10.1023/A:1025877308510
[3] JM Balthazar, DT Mook, DT, RMLRF Brasil, HI Weber, A. Fenili, D. Belato, JLP Felix: An Overview on
non-ideal vibrations, Meccanica, 38, (2003), p. 613-621.
doi:10.1023/A:1025877308510