Machine learning algorithm functional on environmental sustainability assessment in turbomachinery sector: Application on centrifugal compressors

Heliyon. 2024 Jun 24;10(13):e33480. doi: 10.1016/j.heliyon.2024.e33480. eCollection 2024 Jul 15.

Abstract

The current government directives have focused industries' attention on environmental sustainability issues in products and processes. There is indeed a growing demand from customers to conduct environmental impact assessments of the products they purchase. This work presents the implementation of a predictive model developed in an industrial context to evaluate the environmental sustainability of a centrifugal compressor rotor assembly. The development of the predictive model arises from the objective of overcoming the limitations of the traditional Life Cycle Assessment approach, which is based on a retrospective evaluation of the product life cycle. The functionality of predictive models is to estimate product environmental sustainability to meet customer demands and guide them toward choices that aim for carbon neutrality. The implementation of the model has been conducted in parallel with a tailored measurement campaign of the primary inventory flows involved in various manufacturing operations. The article details the methodological approach that led to the development of the predictive models and their respective functionality in supporting the design engineer in evaluating the eco-profile of the assembly. In addition to the methodological aspect, the work also includes a case study through which the functionality of the models can be illustrated.

Keywords: Life cycle assessment; Predictive life cycle assessment; Sustainable manufacturing; Sustainable turbomachinery design.