Development and reduction of a fault tree for gearboxes of heavy commercial vehicles based on identification of critical components

• Published in: Quality and reliability engineering international Vol. 24; no. 2; pp. 183 - 198
• Main Authors: Morello, Marcelo Guimarães, Cavalca, Katia Lucchesi, Silveira, Zilda de Castro
• Format: Journal Article
• Language: English
• Published: Chichester, UK John Wiley & Sons, Ltd 01.03.2008
• Subjects: commercial vehicles; design of experiments; fail tree analysis; transmission
• ISSN: 0748-8017; 1099-1638
• DOI: 10.1002/qre.879

This work deals with the development and reduction of a fault tree applied to gearboxes of heavy commercial vehicles. The improvement with respect to the classical failure tree analysis (FTA) is to reduce the number of FTA components based on the sensitivity of the system reliability to the statistical parameters of the components failure models during a certain lifetime. These parameters determine the shape and slope of the decreasing reliability curve. The lifetime plays an important role because it is the base for any decision about guarantee time and periodic inspection in vehicle maintenance. As the components are usually in series configuration in this kind of mechanical system, the proposal is to consider only the components to which the system presents a significant sensitivity. A factorial planning with two replicates is applied to identify the system sensitivity with respect to these parameters taking into account the confidence interval in each case, as the parameters are evaluated from a sample with a specific size, which has a significant influence on the confidence limits. Therefore, the lifetime is defined in terms of kilometers and it must be economically viable for the company with respect to guarantee period, vehicle stops and maintenance costs. Following this objective, it is important to highlight that the reliability function sensitivity of the system changes depending on the lifetime and the behavior of each FTA component's reliability (failure model parameters). Then, the fault tree is assembled and successively reduced considering the sensitivity to scale (α) and shape (β) factors for components with Weibull distribution, as well as mean (µ) and standard deviation (σ) for Gaussian and Lognormal distributions, in specific kilometer intervals. This methodology allows a reliability model conception for management of the actions focused on products' guarantee and provides design descriptions for the development areas and manufacturing. In this model, it is possible to obtain information about lifetime to assist in activities of performance studies and optimization in design engineering as well as the identification of problems related to design and manufacturing for several operation intervals. In the present case study, the methodology led to a 30% reduction on the number of components considered in the complete fault tree, contributing to guide future improvements to gearboxes reliability. Copyright © 2007 John Wiley & Sons, Ltd.