Abstract

PINTO SANTOS, Jorge Adolfo et al. Survival Analysis of a Tool Using a Bayesian Network Model-CPH. RIDE. Rev. Iberoam. Investig. Desarro. Educ [online]. 2022, vol.13, n.25, e061.  Epub June 12, 2023. ISSN 2007-7467.  https://doi.org/10.23913/ride.v13i25.1361.

One of the major problems in the industry is maintenance, and specifically the replacement of typical wear tools. Currently, companies manage total productive maintenance, which focuses maintenance technicians mainly on corrective maintenance and, to a lesser degree, on preventive maintenance. However, very few companies formally analyze predictive maintenance. Establishing a methodology for predictive maintenance requires analyzing the different degradation models of the tool by relating it to the probability distribution function it develops. This paper contemplates the analysis of an electrode for contact (ultrasonic) welding through the behavior of its degradation, thanks to which it is possible to obtain the probability density function that best fits the behavior of tool wear. In addition, the factors influencing non-compliance in the tensile strength of welded parts are determined. The Cox proportional hazard model and design of experiments techniques are used, which is considered as the basis for implementing the predictive maintenance program.

Keywords : degradation analysis; survival analysis; predictive maintenance; Cox proportional hazard model; Bayesian network.