Concluding Remarks and Future Outlook
Simulation-based process qualification has become an active area of research, with metal AM being a prime candidate for application of such techniques. The increasing availability of computational resources and their incorporation in industrial/research environments further promotes the potential for tailoring properties and processes for metal AM using modelling techniques. Thus, several areas for continued development can easily be identified for modelling of metal AM.
The current techniques for model reduction for macro-scale modelling can be complemented by application of coupled global-local models, where thermo-mechanical and microstructural responses are calculated at different length and time scales within the same simulation. More optimally designed experimental results would then be necessary in the near future to better capture process parameter uncertainties and correlations for macro-, meso- and micro-scale models, eventually leading to better predictability of corresponding outputs. Such studies would pre-requisite application of appropriate sensitivity and identifiability analysis techniques that can indicate the adequacy of experiments towards calibrating the model parameters. Having such calibrated and well-characterised models of the metal AM process would open the possibilities for optimising the process/ process-chain for newer materials and applications at significantly accelerated rates.
This chapter has highlighted the various models of interest for the metal AM process and discussed their applicability towards simulating known defects and behaviour of metal AM components. The high-fidelity models discussed in this chapter are relevant when critical components are to be produced in known materials as well as when specific defects and/or process windows need to be evaluated for new metallic alloys. On the other hand, the implementation of the highest fidelity models for daily continuous evalu- ation/control of the metal AM process is not feasible due to the long computation time at high computational resource requirements. Instead, simplified modelling approaches need to be adopted and calibrated according to the specific metal AM machine and process - the existing approach in industries venturing into metal AM for production of their specific products. The faster models, when supported with monitoring systems, can enable effective feed-forward control of the metal AM process. More specifically, these faster models can be used for quality control of the metal AM process within four major areas: (a) process window identification, (b) hotspot/porosity indicator, (c) local phase fraction/morphology prediction and (d) warpage/deformation estimation. The relevance of developing both approaches (highest fidelity modelling and faster modelling) remains high nonetheless due to the rapid progress in the field of industrial metal AM, which results in newer machines every year for the same core metal AM process. Leveraging the progress in either of the development pathways is the way to ensure that the various simulation possibilities are exploited with the highest efficiency for enabling precision in metal AM.
