No abstract

Metal Additive Manufacturing (AM) has been attracting a continuously increasing attention due to its great advantages compared to traditional subtractive manufacturing in terms of higher design flexibility, shorter development time, lower tooling cost, and fewer production wastes. However, the lack of process robustness, stability and repeatability caused by the unsolved complex relationships between material properties, product design, process parameters, process signatures, post AM processes and product quality has significantly impeded its broad acceptance in the industry. To facilitate efficient implementation of advanced data analytics in metal AM, which would support the development of intelligent process monitoring, control and optimisation, this paper proposes a novel Digital Twin (DT)-enabled collaborative data management framework for metal AM systems, where a Cloud DT communicates with distributed Edge DTs in different product lifecycle stages. A metal AM product data model that contains a comprehensive list of specific product lifecycle data is developed to support the collaborative data management. The feasibility and advantages of the proposed framework are validated through the practical implementation in a distributed metal AM system developed in the project MANUELA. A representative application scenario of cloud-based and deep learning-enabled metal AM layer defect analysis is also presented. The proposed DT-enabled collaborative data management has shown great potential in enhancing fundamental understanding of metal AM processes, developing simulation and prediction models, reducing development times and costs, and improving product quality and production efficiency.

This paper presents R U LES-5, a new induction algorithm for effectively handling problems involving continuous attributes. R U LES-5 is a 'covering' algorithm that extracts IF -TH EN rules from examples presented to it. The paper rst reviews existing methods of rule extraction and dealing with continuous attributes. It then describes the techniques adopted for R U LES-5 and gives a step-by-step example to illustrate their operation. The paper nally gives the results of applying R U LES-5 and other algorithms to benchmark problems. These clearly show that R U LES-5 generates rule sets that are more accurate than those produced by its immediate predecessor R U LES-3 Plus and by a well-known commercially available divide-and-conquer machine learning algorithm.

No abstract