A Build-Time Estimator for Additive Manufactured Objects

Abstract

Additive Manufacturing is a very time consuming technology. An estimation of the build time is fundamental to: Evaluate the production cost in budgeting process.Make use of optimization methods, which use as parameter the build time, for determining optimal build direction. In both these cases a fast and valid build time estimator, which can work with a few input data deducible from geometric model, is required. In the proposed paper a reliable parametric-based method to determine the build time for additive manufactured objects is provided. The implemented method is based on a back-propagation artificial neural network, which gives the possibility to implement the complex functions that elapse some driving build-time factors and the build time. The neural network training is based on data provided by a properly developed analyzer of the list of commands given to AM machines, which performs an analytical estimation of the build time. The implementation of the proposed methodology is illustrated and some comparisons between the real and estimated build-time are provided, then the results are critically analyzed.