Search Results

Technological systems are designed to carry out very specific functions. Because of that, their components should have measurements that can guarantee their operability within the range of precision. Furthermore, the current systems are inherent parts of design involving multi-disciplinary aspects. Their development and analysis expose the designer to a series of unknown parameters from several sources such as material properties, environmental and operational conditions. Therefore, the qualification and quantification of these inherent sources of design uncertainties become very important in several aspects in the context of design development and so, a system is reliable and robust if it allows a certain range of uncertainties before the first failure occurs. With this in mind, we propose here the development of a methodology that can identified the sources of uncertainties and parameters that largely influence the whole design.

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.

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.

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

This paper proposes a elaboration of a Fault Tree Analysis (FTA) for gearboxes to study reliability of statistical models more adequate for components which may be able to generate failures. Therefore there was generated a fault tree with scale and shape factors for each component. Once obtained the statistical models for each failure, a study about critical levels of the system components will be valid by reduction of the FTA models. The aim of the actions in relation to the improvements will increase the reliability of the gearboxes. This method can be applied in various vehicles systems allowing a complete statistic modeling. Thus, a general map of the product can be made, in which there are components that generate or will be able to generate failures in specific kilometer. With this, it is possible to obtain a model for management of the actions of warranty and descriptions for the areas of product development and manufacturing.

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.

In a large part of the mechanical elements used in machines and equipment, the preponderant failure mode is not that of fatigue of the element itself, but certainly the fatigue of a small point where the contact occurs. The prime example of this are the roller bearings, that they fail not by “breaking”, but by surface fatigue at contact points or on the tracks where there is contact between the rings and the rolling bodies. The optimization of the contact geometry, the material and the lubrication used can allow us to have larger admissible loads or lower system's costs where there is great influence of the contact fatigue. To make this optimization easier, a software was developed for a Windows platform, including the whole contact theory, the life calculations under surface fatigue and the lubricating fluid film thickness. In an interactive way, the user can change the data entrances such as material or geometry until an ideal solution is found for its problem.