Feature-level Data Fusion for Energy Consumption Analytics in Additive Manufacturing

Published

September 1, 2020

Author(s)

Paul Witherell

Abstract

The issue of Additive Manufacturing (AM) energy consumption attracts attention in both industry and academia, as the increasing trend of AM technologies being employed in the manufacturing industry. It is crucial to analyze, understand, and manage the energy consumption of AM for better efficiency and sustainability. The energy consumption of AM systems is related to various correlated attributes in different phases of an AM process. Existing studies focus mainly on analyzing the impacts of different processing and material attributes, while factors related to design and working environment have not been paid enough attention to. Such factors involve features with various dimensions and nested structures that are difficult to handle in the analysis. To tackle these issues, a feature-level data fusion approach is proposed to integrate heterogeneous data in order to build an AM energy consumption model to uncover energy-relevant information and knowledge. A case study using real-world data collected from a selective laser sintering (SLS) system is presented to validate the proposed approach, and the results indicate that the fusion strategy achieves better performances on energy consumption prediction than the individual ones. Based on the analysis of feature importance, the geometry relevant features are found to have significant impacts on AM energy consumption.

Proceedings Title

Proc. of IEEE Conference on Automation Science and Engineering 2020

Conference Dates

August 20-21, 2020

Conference Location

Hong Kong, HK

Pub Type

Conferences

Keywords

additive manufacturing

Citation

Witherell, P. (2020), Feature-level Data Fusion for Energy Consumption Analytics in Additive Manufacturing, Proc. of IEEE Conference on Automation Science and Engineering 2020, Hong Kong, HK (Accessed December 30, 2024)