An integrated approach for sustainable machining processes: Assessment, performance analysis, and optimization

Abstract

Abstract Nowadays, the implementation of sustainable techniques is an essential requirement in machining processes; however, there is a need to establish a detailed and reliable platform for sustainable machining processes. In this work, a novel approach for sustainable machining processes is presented and discussed to provide the machining process designer with the optimal parameters through an exhaustive assessment process. This approach is mainly based on two main pillars. The first pillar presents a comprehensive guideline for the assessment of machining processes considering the life cycle assessment impacts as well as the existing sustainable machining elements (i.e. energy consumption, machining costs, personal health and safety, and impacts of waste management). Besides, the product quality aspects (i.e. surface integrity indicators) are integrated into the sustainable machining metrics. This proposed integration establishes a new performance analysis model to accurately evaluate the sustainable machining process. Material extraction and manufacturing stages of different inputs (e.g. cutting tool and cooling strategy) as well as the obtained machining characteristics and output materials are covered by the proposed assessment process. The second pillar presents a novel algorithm to model and optimizes the sustainable machining processes by providing the optimal solutions using multi-objective optimization of all metrics of the proposed performance analysis model for multiple input alternatives (i.e. different machining strategies). Applying the non-dominated sorting process on the obtained optimal solutions shows the superiority range for each machining input and provides the capability to compare different machining alternatives. To validate the effectiveness of the proposed approach, two case studies have been used and the results are discussed. It should be stated that the proposed model offered an overall accuracy of 84% for the first case and 90.3% for the second case.