Thèse de doctorat en Génie mécanique - procédés de fabrication
Sous la direction de Jean-Yves Dantan et de Alain Etienne.
Thèses en préparation à Paris, ENSAM , dans le cadre de École doctorale Sciences des métiers de l'ingénieur (Paris) , en partenariat avec Laboratoire de Conception Fabrication Commande (laboratoire) depuis le 05-03-2015 .
L’objectif de cette thèse est de prendre en compte les paramètres du procédé de FA choisi afin d’optimiser la conception des produits dans le but d’assurer la qualité du produit et de minimiser le cout de fabrication. Dans un premier temps il convient donc d’étudier les caractéristiques du procédé afin de proposer un modèle d’estimation des couts. Dans un deuxième temps un travail sur la qualité géométrique obtenue par un procédé de FA doit être mené afin de proposer un modèle de prédiction des incertitudes de fabrication. Enfin, Une méthode regroupant ces 2 approches doit nous permettre de proposer une (re)conception de produit permettant l’atteinte de la qualité visée dans un cout maitrisé.
Models and Methods for variation and uncertainty management in Micro-manufacturing during the design phase
This proposal focuses on design for variation and uncertainty management of the micro-manufactured product and its micro-manufacturing process. At the most basic level, “design” is a process that is used to generate, evaluate and select a solution for a given problem. Uncertainty and risk are unavoidable when designing in the absence of full knowledge. Uncertainty stems from the lack of knowledge required to model performance. The risk that is associated with a particular design may be ascertained if one is able to use models to define a probability of success. In short, it is better to be in a position to ascertain risk than to be in a position of uncertainty. Throughout this assessment, little evidence has been found to suggest that stochastic methods have been employed during the design of NLBMM parts. Given the nascent state of the NLBMM technology, one finds variation in fabrication processes, metrology tools, material properties, and other aspects which must to be considered during the design process. The ability to model these variations and use such models to assess risk is important for two reasons: 1. Knowing the risks may help to prevent the stifling of design concepts in the face of erroneously perceived high risks. 2. A good understanding of risks may be used to guide the distribution of resources toward designs that have a higher probability of success. The scientific main objective is to improve methods for variation and uncertainty management during this design phase. This proposal will focus on a contribution in developing: • Models to predict the quality of micro-manufactured part (Prediction of surface roughness and dimensional deviation). Prediction of surface finish and dimensional deviation is an essential prerequisite for developing a micro-manufacturing process. Two main attributes of job quality are surface roughness and dimensional deviation. Surface finish has a great influence on the reliable functioning of two mating parts. A reasonably good surface finish is desired for improving the tribological properties, fatigue strength, corrosion resistance, aesthetic appeal of the product, ... Excessively better surface finish may involve more cost of manufacturing. • Methods for risk assessment. There is a need for modeling and simulation tools that link the performance of components to variations in the characteristics of the part and the variations in the performance of a system to variations in the performance of its parts. In many cases, the preceding may be addressed by custom-made, stand-alone simulations (e.g., Monte Carlo simulations). Unfortunately, many of these tools are difficult to learn and they are not integrated with existing geometric and behavioral modeling tools. This is perhaps the reason that stochastic design tools are not widely used in engineering design. • Methods for parts tolerancing. Micro-manufacturing processes are characterized by high process variability and an increased significance of measurement uncertainty in relation to tight tolerance specifications. Therefore, it is important to take into account not only the micro manufacturing process capabilities but the measurement uncertainty.