Track: Business Analytics

Abstract

Additive manufacturing attains supply chain resilience by enabling localized production, reducing prototyping time through swift design iterations, and achieving on-demand manufacturing by crafting customized products when required. These advantages stem from its interconnected layer-by-layer fabrication process, redefining traditional manufacturing approaches. Like other industries, AM exhibits intricate parameter interdependencies, necessitating novel predictive models. In this study we used a dataset obtained from PLA-based Additive manufacturing process and used data-driven methodologies on the dataset including measurements of input and output parameters.

The study aims to predict energy consumption, print duration, and object weight based on input variables - layer thickness, printing speed, and line infill pattern. We used advanced machine learning techniques including linear regression, decision trees and random forests to create a predictive model for data driven decision making.

We segregated the collected data into two distinct sets: a training set which was used to train the machine learning model, and a test set to assess the model's performance on unseen data. This approach ensures the model's robustness and ability to generalize, allowing it to make accurate predictions for real-world additive manufacturing scenarios.

Results highlight machine learning's efficacy in forecasting output parameters. The research emphasizes the potential of computational models for real-time process control. It enhances resource efficiency by minimizing material waste and energy consumption. Collectively, these outcomes empower companies to streamline processes, make informed choices, and ultimately deliver high-quality products efficiently and competitively echoing the trend of data-driven transformations in manufacturing.

Additive manufacturing attains supply chain resilience by enabling localized production, reducing prototyping time through swift design iterations, and achieving on-demand manufacturing by crafting customized products when required. These advantages stem from its interconnected layer-by-layer fabrication process, redefining traditional manufacturing approaches. Like other industries, AM exhibits intricate parameter interdependencies, necessitating novel predictive models. In this study we used a dataset obtained from PLA-based Additive manufacturing process and used data-driven methodologies on the dataset including measurements of input and output parameters.

The study aims to predict energy consumption, print duration, and object weight based on input variables - layer thickness, printing speed, and line infill pattern. We used advanced machine learning techniques including linear regression, decision trees and random forests to create a predictive model for data driven decision making. We segregated the collected data into two distinct sets: a training set which was used to train the machine learning model, and a test set to assess the model's performance on unseen data. This approach ensures the model's robustness and ability to generalize, allowing it to make accurate predictions for real-world additive manufacturing scenarios. 

Results highlight machine learning's efficacy in forecasting output parameters. The research emphasizes the potential of computational models for real-time process control. It enhances resource efficiency by minimizing material waste and energy consumption. Collectively, these outcomes empower companies to streamline processes, make informed choices, and ultimately deliver high-quality products efficiently and competitively echoing the trend of data-driven transformations in manufacturing.