Search Results

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.

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

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 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 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 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.

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 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.

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.