title   
  

Rapid modelling of multi-stepped rotor dynamics

Keskin, Hakan (2004) Rapid modelling of multi-stepped rotor dynamics. [Thesis]

[img]PDF - Registered users only - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
1263Kb

Official URL: http://risc01.sabanciuniv.edu/record=b1124916 (Table of Contents)

Abstract

The major objective of the studies on the dynamic behaviour of rotors is to allow development of rotating machinery that will be free from vibrational problems. Rotor dynamics prediction has several important consequences for a large group of machinery. First of all, there are many machines with rotating components, i.e. rotors, such as turbines, compressors and pumps, machine tools, helicopters, dentist's tooth grinders etc. The dynamics of these systems are very critical for their proper functioning, stability, efficiency and health. For example, if one of the natural frequencies is in the operation region of the system, i.e. one of the excitation frequencies during the operation is close to one of the natural frequencies, then the system may resonate and the resulting large amplitude vibrations will cause loss of accuracy and efficiency in the operation, and may also damage the bearings and the rest of the machinery. There are two main approaches to rotor dynamics analysis. For existing rotors, the dynamics can be measured using experimental techniques such as modal analysis. Although this can be a fast approach, it requires the experimental set up be available. Also, if there are many different components added to the rotor, such as different tool holders and tools on a machine tool spindle, then the measurements must be repeated for each combination which may result in high number of tests and waste of productive time on the machine. Prediction of the rotor dynamics during the design is another critical case where dynamics analysis is required. Obviously, testing is not a possible technique for this case. FEA can be used for the prediction of the dynamics. This may be a viable solution in many cases, however the bearing contact parameters, i.e. stiffness and damping, must be known in order to develop the model. Also, many simulations have to be performed for optimal design resulting in highest dynamic rigidity with smallest possible rotary inertia. This is usually very time consuming, and the optimal configuration may never be obtained. The objective of this study is to develop a fast method for the dynamic analysis of rotor-bearing systems so that they can be used during the design and the operation of the machinery. The method can be used to determine the frequency response function of the system at critical locations for different geometric configurations very fast so that the design could be optimized. The effects and optimal values of some internal system parameters such as bearing preloads could also be determined using the method. Also, the dynamic response of the system could be updated after new components are added to the rotor.

Item Type:Thesis
Uncontrolled Keywords:Prediction of protein contact maps. -- Disulfide bond prediction. -- Neural networks. -- Suppart vector machines
Subjects:T Technology > TJ Mechanical engineering and machinery
ID Code:8222
Deposited By:IC-Cataloging
Deposited On:16 Apr 2008 16:51
Last Modified:26 Dec 2008 13:15

Repository Staff Only: item control page