A new machine learning based geometry feature extraction approach for energy consumption estimation in mask image projection stereolithography

Abstract

Listen

Translate article

The sustainability issues have become increasingly critical due to the limited capacity of the environment and non-renewable natural resources. Manufacturing, as the main foundation of human society and civilization, plays a significant role in sustainability. As a new production approach, additive manufacturing fabricates products layer-by-layer. and has become a promising alternative to traditional subtractive manufacturing. Nevertheless, the sustainability performance of additive manufacturing has not been sufficiently estimated and evaluated. In current literature, the majority of the energy consumption studies on additive manufacturing aim to establish the relationships between process parameters and power consumption. While these studies can facilitate the joint consideration of process planning and environmental sustainability, they fail to relate the product geometry with the sustainability performance, and therefore lack predictive ability. Hence, they cannot be directly used to support product design and redesign. In addition, bridging the product geometry with the required power consumption can aid the establishment of the life cycle inventory database for additive manufacturing processes. In this paper, a new machine learning-based approach is adopted to extract the geometry-related features in order to estimate and predict the energy consumption of mask image projection stereolithography process. By bridging the product geometry and process energy consumption, the research outcomes will serve as a critical part of the unit manufacturing process models and contribute to the life cycle inventory of additive manufacturing.