Search Results

This work presents an application of the simulated annealing algorithm to redundant system reliability optimization, and its objective is to analyze this method and compare to other optimization methods, which are Genetic Algorithm, Evolution Strategy and Lagrange Multipliers. A hidrid optimization algorithm, composed by Simulated Annealing and Genetic Algorithm, is also developed in this work, to conciliate the advantages of both methods. The results presented by the Simulated Annealing and the hybrid algorithm are significant and validate these methods as robust tools for parameters optimization in engineering systems developments

A software for the management and test of rotors was developedand that is capable to set up models and to generate solutions as frequency response of a rotor on line with the design process. That system can be applied in the study of problems of rotordynamics, taking into account the interaction between the rotor and the supporting structure where it is set up. The Rotortest software is capable of make the management of the input data of each element that composes a rotor: shafts, hydrodynamic and roller bearings, mechanical seals and foundations. Besides, starting from the input data, it presents illustrative outputs as diagrams and schemes, to accomplish the assembly of the mathematical models, and to find the frequency response solution of the complete system, presenting the results graphically.

The study of rotative machines occupies an outstanding position in the context of machines and structures in view of the significant amount of typical phenomena in the operation of those equipments. The existence of a rotative component leaning by bearings and transmitting power and torque creates a family of problems that are found in the most several machines. Therefore, in the study of the dynamic behavior of those systems, it is necessary to be determined the interaction among all the components that affect in a significant way the dynamic behavior of the system. It is proposed, this way, a simplified mathematical model, through the method of finite elements, of the Rotor-Bearing-Coupling system, considering the support structure or foundation practically rigid, in way to allow the analysis of the transverse or flexional vibrations of the system.

The study of rotative machines occupies an outstanding position in the context of machines and structures in view of significant amount of typical phenomena in the operation of those equipments. The existence of a rotative component leaning by bearings and transmitting power and torque creates a family of problems that are found in the most several machines. Therefore, in the study of the dynamic behavior of those systems, it is necessary to be determined the interaction among all the components that affect in a significant way the dynamic behavior of the system. It is proposed, this way, to determine the dynamic behavior of Rotor-Bearing-Coupling system using a Multiobjective Genetic Algorithm coupled with Surface Response Method. The Rotor-Bearing-Coupling system has the support structure or foundation practically rigid, in way to allow the analysis of the transverse or flexional vibrations of the system.

This work has for purpose to develop a mathematical model for the hydrodynamic lubrication of bearings with oscillating motion. The motivation lies in the tribology of internal combustion engines, specifically the lubrication problem at the bearing existing between the connecting rod and the piston pin. This kind of bearing does not perform a complete rotation, thus characterizing a new class of oscillating motion bearings. For the analysis of lubrication, a viscous fluid flow with negligible inertia inside a narrow gap formed by two surfaces is considered. The surfaces can be considered flat and are inclined to each other. The inner surface is fixed whereas the outer one performs an oscillating motion which generates a combined Couette-Poiseuille flow inside the gap. Using the same assumptions as in classical Reynolds' lubrication equation, the simplified mass and momentum conservation equations are analytically solved to determine the velocity and pressure distributions.

The pressure generation within the lubricant fluid present in the clearance between a thrust bearing and the collar attached to the shaft has a fundamental importance to avoid contact between solid parts with axial relative motion. Any existing contact can lead to friction, wear and, as a consequence, failure of elements on a rotating machine. Therefore, in order to design an effective bearing, it is important to know how the pressure is generated within the oil film and the magnitude of the load capacity transmitted from the collar to the bearing throughout the fluid. Thus, it is necessary to solve the Reynolds' Equation to obtain the distribution of pressure on the sections under Hydrodynamic Lubrication. Then several operational parameters can be obtained, such as, the total load capacity, lubricant fluid flow, position of the maximum pressure and so on.

The present research proposes the use of linear model updating along with unrestricted optimization methods, in order to obtain a methodology, which allows the calibration of mathematical models of rotating systems. An experimental set up of a symmetric rotor, on a rigid foundation supported by two hydrodynamic cylindrical bearings and with a central disk of considerable mass, working as an unbalancing excitation force, was used for this purpose. Once the numeric and experimental values are obtained, error vectors are then defined, which serve as minimization parameters, through the variation of the numeric model parameters. The method presented satisfactory results, as it was able to calibrate the mathematical model and then from that, obtain reliable responses for the physical system studied.

The reliability optimization of redundant systems aims to improve some design parameters as reliability or cost, using optimization methods, in order to determine the optimum number of redundancies in each component of the system. This work compares some optimization methods (Lagrange Multipliers, Evolution Strategy and Genetic Algorithm) and its objective is to determine which method is more adjusted to the problem.

One of the great challenges of engineering teams nowadays is to overcome long and costly project experimentation phases. One effective way of decreasing such project demands is to come up with a firsthand prototype with high success probability. In order to do so, the project team should rely on robust numerical models, which can represent most of the real-life product behaviors, for instance system dynamics. For rolling element bearings, such dynamic models have to consider the dynamic interactions between its components, i.e., rolling elements and raceways. The only vibration transmitting points on rolling element bearings are the lubricated contacts. Therefore, in order to represent the full bearing dynamic behavior on a numerical model, an efficient transient contact model, which depicts the actual contact behavior, is fundamental.