A hybrid mechanism-based and data-driven approach to forecast energy consumption of fused deposition modelling
Zhiqiang Yan,Jizhuang Hui,Jingxiang Lv,Donald Huisingh,Jian Huang,Kai Ding,Hao Zhang,Qingtao Liu
DOI: https://doi.org/10.1016/j.jclepro.2023.137500
IF: 11.1
2023-08-01
Journal of Cleaner Production
Abstract:Additive manufacturing (AM) has been recognized as a promising technology for achieving more secure and resilient supply chains to support social transitions to sustainable post-fossil carbon societies. AM is considered to be efficient and sustainable and its environmental impact is mainly from energy consumption. Therefore, it is of great significance to accurate predict and thus reduce the energy consumption of AM. However, current models to predict AM energy consumption are not sufficiently accurate. In this paper, a hybrid mechanism-based and data-driven approach is proposed to forecast the AM energy consumption. This approach accurately modelled the working time and power of each energy-consuming component of the FDM machine. Firstly, the physical model was used to establish the energy consumption prediction framework which documents the total energy into energy consumed by each component. Secondly, the data-driven method was leveraged via data fitting and deep learning to obtain the power and temporal model of the nozzle and the platform, considering their working states and interactions between them and the cooling fan. Thirdly, experimental measurements and G-code parsing were utilized to obtain data of the power model and the temporal model of the remaining components. In this work, the different types of parts were fabricated by the fused deposition modelling (FDM) printer and the energy consumption was measured and compared with the predicted values to verify the accuracy and effectiveness of the proposed method. Finally, the proposed method was compared with other methods to document the benefits of the proposed method. The results showed that the prediction accuracy of the proposed method for these four experiments was more than 94% accurate and it outperformed other methods. The proposed method can be used to guide process design to more accurately predict and thus reduce the energy consumption of AM parts during their manufacturing processes.
environmental sciences,green & sustainable science & technology,engineering, environmental