MODELLING AND OPTIMIZATION OF PHYSICAL SYSTEMS 8, pp.

73-76, Gliwice 2009

MODELLING TORSIONAL VIBRATIONS IN SIMMECHANICS KRZYSZTOF KAWLEWSKI EUGENIUSZ WITOSKI

Department of Applied Mechanics, Faculty of Mechanical Engineering, Silesian University of Technology krzysztof.kawlewski@polsl.pl, eugeniusz.switonski@polsl.pl

Abstract: This paper deals with modeling of torsional vibrations of mechanical system in Matlab/Simmechanics. Model of mechanical system was created in Matlab/Simulink Simmechanics toolbox. Results of simulation and test results on the real object were compared and presented in this article 1. INTRODUCTION In among groups of mechanical systems and mechanisms can be separated group of systems which behavior can be modeled using 2 or 3 mass/inertial system connect by spring and damper elements. Devices like rolling mills, robotic arms, transmission systems, flexible shafts in still rising demands for higher motion speed, provide to possibility of excitation of resonance frequencies. In addition, the increasing demands for better dynamics generate a constant need for new solutions. To model and simulate mechanical systems can be used a large group of software. Simmechanics with Simscape language and Matlab/Simulink environment is often chosen for modeling, simulating and development of various mechanical systems. Simmechanics software provides diagram block modeling and can be a reliable tool for modeling mechanical systems using rigid finite elements or even FEM [2]. Simmechanics block library contains groups of basic elements, parameters or functions like bodies, sensors, actuators, joints, springs and dampers. By connecting blocks and setting parameters of some of the blocks can be modeled almost each mechanical system. 2. MODEL OF MECHANICAL SYSTEM Laboratory stand (Fig. 1) consists of AC motor drive and shaft of mechanical device with connected flexible shaft. On shaft were mounted two disks which position can be changed by moving them along shaft. This mechanical system was modeled in simulink environment as three inertial system connect with elastic rotational spring elements. Each inertia element was modeled as body with two parameter, matrix of inertia and dimension. Inertia of shafts and clutches was neglected. To model rotational spring it was used the Joint Sensor and the Joint Actuator block and Gain block from simulink (Fig 3), gain value was calculated using shear modulus of elasticity. Fig. 2 shows model created in simulink environment with Simmechanics block diagrams. Visualization of modeled system can be displayed in time of simulation and each body can be represented by ellipsoid of inertia (Fig 4)

74

K. KAWLEWSKI, E. WITOSKI

Figure 1. View of the laboratory stand and dynamical model of mechanical system Between AC motor shaft and flexible shaft with disks is located the torque sensor for measure torque generated by motor and transmitted to mechanical system. Measurements from torque sensor can be visualized on the screen of oscilloscope or saved on the computer hard disk drive via analog digital converter with usb interface. Measurements are saved as values with specified sample rate, on saved values can be performed FFT transformation. At simulations performed on laboratory stand sample rate of measurements were five times greater than second torsional vibration frequency of system.
B Ground1 F C S1 C S4 Revolute1 C S2 C S3 Revolute2 B F C S1 C S4 C S2 C S3 Revolute3 B F C S1 C S2 C S3 Disc2_body D2_b_sensor av
Pulse Generator ACM_a

AC Motor_body

Disc1_body

Jr2s

AC M_b_sensor

1
1 s

p a

Jr2a D1_b_sensor

p a

Jr3s

Jr3a

3
-12760 shaft_stiffness 1 s

TE

-1410 ts_stiffness

1 s

Clock FFT_acm To Workspace

Scope

Figure 2. Model of mechanical system created in simmechanics Stiffness coefficient value of each elastic element was obtained from its geometry and material modulus of elasticity in shear. Geometry and moment of inertia of each element of mechanical system were combined and modeled in simmechanics, AC motor was considered

MODELLING TORSIONAL VIBRATIONS IN SIMMECHANICS

75

only from mechanical point of view, electric parameters were neglected and only rotors inertia was included in model.

Figure 3. Diagram of blocks modelling shaft stiffness

Figure 4. View of Simmechanics visualization window with model in time of simulation

3. SIMULATION RESULTS Model was analyzed in two cases, first one when disks were close together (50mm) and second one when disks were far from each other (300mm). Resonance frequency of torsional vibrations of mechanical system was obtained from signal of torque sensor measurements using FFT algorithm (Fig. 5,6) and from angular position of each body in simulation (Fig 7,8). Results have been compared in Table 1. Distance between disks 300 mm 50 mm Table 1. Vibration frequencies obtained from simulations and measurements Frequencies calculated Frequencies measured Frequencies measured from frequency response of from frequency domain from eigenvalues of signal from torque sensor in Simmechanics model stiffness matrix 124,66 Hz 129,87 Hz 123,3 Hz 211,45 Hz 210,28 Hz 210 Hz 123 Hz 127,18 Hz 124,8 Hz 516,8 Hz --508,9 Hz
129,87Hz 127,18Hz

210,28Hz

Figure 5. Angular position of disk in frequency domain response measured on laboratory stand (300 mm between disks)

Figure 6. Angular position of disk in frequency domain response measured on laboratory stand (50 mm between disks)

76
0 5

K. KAWLEWSKI, E. WITOSKI
123,3 Hz 210,0Hz
Power/frequency Power/frequency -100

-50

124,8Hz 508,9Hz

-50

-100

-150

-150

-200

-200

-250 0

10

150

200

250 Frequency

300

350

400

450

500

-250 0

100

200

300 Frequency

400

500

600

Figure 7. Angular position of bodies in domain frequency (distance between disks 300mm) 3. CONCLUSIONS

Figure 8. Angular position of bodies in domain frequency (distance between disks 50mm)

Measurements from tests and simulations were similar in both cases and values of torsional frequencies were comparable. Even for that simplification and neglects in model creation, Simmechanics is a suitable tool for modeling mechanical systems. Strong point of modeling in Simmechanics is possibility to use Matlab m-files and Simulink models. With connection and use of the other toolboxes in environment of Matlab/Simulink it is easy to model physical systems which cover multiple domains like electrical, mechanical, pneumatic etc. REFERENCES 1. E. witoski: Modeling of mechatronic drive systems, ISBN 83-7335-287-2, Gliwice 2004 2. S. Duda: Flexible multibody systems in simmechanics, In: Modelling and Optimization of Physical Systems ISBN 978-83-60102-46-6, pp. 15-18, 2008 3. K. Kawlewski: Modelling torsional vibrations of mechanical system in Matlab in statespace approach, In: 10th International Scientific Conference Applied Mechanics 2008, ISBN 978-83-60102-49-7, 2008

MODELOWANIE DRGA SKRTNYCH W SIMMECHANICS

Streszczenie: W artykule przedstawiono sposb modelowania drga skrtnych w rodowisku Matlab/Simulink w programie Simmechanics. Dla ukadu elektromechanicznego skadajcego si z silnika prdu przemiennego z doczonym do niego podatnym wakiem z dwiema tarczami zosta utworzony model drga skrtnych w programie Simmechanics. Utworzony model mia na celu okrelenie czstotliwoci drga skrtnych powstajcych w ukadzie jak i wpyw pooenia tarcz na powstajce czstotliwoci. Wyniki symulacji oraz wyniki bada na obiekcie rzeczywistym zostay porwnane i przedstawione w artykule.