Search Results

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.

The development area of the bearing's industry constantly searches for a better efficiency and a lower film oil thickness in surfaces with high roughness under relative motion. In these cases, operational conditions like high loads and temperatures, as well as low safety margins for weight and size, considering the lubricant viscosity, should be taken into account as fundamental design parameters. In order to know better the elastohydrodinamic lubrication effect, firstly, it is necessary to understand deeply and accurately the applied loads on the ball element bearings. For this purpose, the accurate analysis and study of the performance of these machine components is carried out, using analytical methods and giving special focus on the velocities and accelerations involved, as well as different types of loads applied on the ball element and their distribution and consequences during the ball motion inside the bearing rings.

The study about the dynamic characteristics of a great number of mechanical parts has been promoted by the necessity of decreasing the vibrational effects in mechanical systems, as the reduction of superficial fatigue. In this way, the research around, even, a simple part like a hydrodynamic bearing is very important, especially in the automotive industry. In this case, the lubricant acts like a flexible liking element between the journal-bearing surfaces. The lubrication is essential for the engine, because it reduces the wear between the internal parts and prevents the metal contact. Due to the shear stresses present in the lubricant, the temperature rises and, consequently, it changes the lubricant properties. The viscosity is strongly dependent on the temperature and it is the parameter that characterizes the fluid flow and its dynamic behavior. Any temperature change induces a consequent modification in the lubricant behavior.

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 productivity of a system can be measured by a combination of parameters as the reliability, mantenability, availability and dependability, which are influenced by the time of functioning and facilities in the repair of the equipment. This work is resulted of the development of a software in the Department of Mechanical Design of the Faculty of Mechanical Engineering of the State University of Campinas. This software carries out, basically, Analyses of Failure, Maintenance and Availability. The Analyses of Availability and Dependability as well as a previous analysis of the Economic Viability have been the subjects treated in this work.

The purpose of this software we developed was to make na analyses of the quality of mechanical components using reliability calculus. The computer program is a powerful tool, for it deals with failure tests, commonly used in practice to obtain the mean life, which is a fundamental parameter in advanced techniques of maintenance. Moreover, the work intends to make accessible concepts related to the reliability of systems and components in Mechanic Engineering, making possible a better understanding and familiarity. The subject is of great application in the great industries and projects with high levels of precision and security. What will determine the satisfaction of the customer after a component, and in consequence, its retention is the performance of its technique characteristics in the period of its life.

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.

The dependence among the components and the data associated with a repair politic, that can consists of the number of available teams and the components repair priority, rarely evaluate the measures of reliability. Thus, to evalute the reliability and availability values with these conditions, this article develops a methodology using the theory of Markovian stochastic processes. A system of ordinary differential equations is obtained using the concepts of component failure and repair rates, constants or variable in the time. Solving it, these values are obtained, allowing the calculation of the MTTF of the analyzed system, for an allowed error.

To analyze a complete real machine, it can be convenient to divide the system into sub-systems, analyzing each sub-system individually, and then, assembling them together in the whole system. Many of these sub-systems can be found in an automotive engine, being the hydrodynamic bearing one of the most common mechanical components present in all kinds of power generation systems. Journal bearings are linking elements between parts with relative motion, and these linking elements must work to support radial loads with minimal friction and power loss. In 1925, Stodola realized that the bearing is not a rigid support, but it works like a set of springs and dashpots whose characteristics have an expressive effect on the dynamical behavior of the supported rotating shaft.

This paper presents a software developed to apply the Hertz contact formulation to bearings with rolling elements, estimating the occurrence of surface fatigue failure in this kind of parts. Considering the wide range of designs in the automotive industry using this kind of bearings, a dedicated software, using a simple language, becomes a useful tool to reduce the design time, predicting in a satisfactory manner the failure of the bearing. The software developed shows a high level of accuracy, when compared to data provided by rolling bearing manufacturers.

This work has as objective the construction of a reduced fault tree to gearboxes in commercial vehicles. In this reduction, the representative statistical model of reliability is studied for all the components that generate failures, where it is complemented with an application of factorial planning with two replicates to identify the components most sensible to the variation of the parameters in the reliability distributions. Thus, fault tree was constructed from the factors of scale (α) and form (β) for each component. From the data statistical models are obtained to individual failures, and the study of the degree of sensitivity of these components into system becomes viable by the reduction of the model of the FTA and applications of design of experiments. With these techniques, the FTA can be reduced as well as the number of failures around 30%, which contributes in the direction of the improvement of product's reliability, in this case, gearboxes.

In the current scenery, to be competitive, the companies face the challenge of developing new products in a short period of time, with improved technology with respect to the previous one and with reduced costs to guarantee the business survival. The success is directly connected to the customers requirements, where Quality, Reliability, Supply and Price are the minimum desired. In this work, methodologies are considered, which are used for the Planning and Warranty of the Quality, which should already be used in the phase of the product design conception. In the design conception phase the quality born and significantly improves the performance, the reliability and the final price of the product.

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 comparative study involving failure data of life tests of two bearings types is presented. In a first stage, the statistical distribution more adapted for the modeling of the failure process is selected. Soon after, the parameters of the distribution and the confidence intervals are estimate. The approximationed sample size is supplied in function of an interval with a magnitude and specific confidence level. Tests of hypotheses are used for several significant levels, rejecting or accepting the hypothesis of equality among the adjusted curves of accumulated failure to each bearing.

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.

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.