A Hybrid Machine Learning Approach for Energy Consumption Prediction in Additive Manufacturing

Additive manufacturing (AM), as a fast-developing technology for rapid manufacturing, offers paradigm shift in terms of process flexibility and product customisation, showing great potential widespread adoption the industry. In recent years, energy consumption has increasingly attracted attention both academia industry due to increasing demands applications AM systems production. However, are considered highly complex, consisting several subsystems, where is related various correlated factors. These factors stem from different sources typically contain features with types dimensions, posing challenges integration analysing modelling. To tackle this issue, hybrid machine learning (ML) approach that integrates extreme gradient boosting (XGBoost) decision tree density-based spatial clustering noise (DBSCAN) technique, proposed handle such multi-source data granularities structures prediction. paper, four sources, including design, process, working environment, material, taken into account. The unstructured clustered by DBSCAN so reduce dimensionality combined handcrafted XGBoost model A case study was conducted, focusing on real-world SLS system demonstrate effectiveness method.