Artificial Neural Networks Approach for Ductility Prediction of Copper Cold Spray Laser Heat Treated Coatings

点击翻译按钮获取中文摘要

There is over 3 million nuclear used fuel bundles (UFBs) in Canada as by-products of electricity production by nuclear power plants and over 120,000 UFBs will be generated per year going forward. The goal of the partnership is to use in a synergetic approach existing knowledge and resources available at uOttawa and NRC to support The Nuclear Waste Management Organization (NWMO) in its quest to use cold spray (CS) additive manufacturing (CSAM) in the development of innovative used fuel container (UFC). The project is part of the NWMO extensive research and development program that will ensure the safe storage of UFBs in UFCs in deep geological repositories, replacing short-term above ground storage sites. Copper coating are to be used to cover the weld closure zone (welded junction between the main UFC body and the closing end cap) for corrosion protection of the UFCs. The project will allow achieving the coating ductility required to comply with external forces that it could encounter over the years without resulting in cracks and leaks of the radioactive material outside the UFCs. This will be achieved through the development of a laser heat treatment that will be designed using a trained neural network model able to correlate the laser parameters to the resulting ductility (input-output). In addition to the creation of new fundamental knowledge/training of highly qualified personnel, the project specific outcome will be the development of a new fast laser heat treatment (HT) on thick CSAM copper coatings (used on UFCs) to ensure a minimum ductility of 10% for safe usage. The work will have a direct impact on the UFC design viability proposed by NWMO for deep ground repository. Furthermore, it will create new knowledge on laser HT that can be adapted to industrial fields important to Canada such as aerospace and automotive.