Optimization of Energy Consumption of Conveyor Belts in Self-Unloading Ships Using Machine Learning Models

The energy consumption of conveyor belts in self-unloading ships is a significant concern in the maritime industry due to its impact on costs and sustainability. This study demonstrates the potential of machine learning in managing conveyor belt energy consumption, resulting in substantial savings and improved operational efficiency. Introducing intelligence in this process changes operational conditions, achieves significant energy savings, and improves efficiency and environmental sustainability by reducing greenhouse gas emissions associated with bulk cargo transportation. This study is designed in collaboration with our industrial partner based in North America. For this, a comparative study of different analytical approaches, such as Decision Trees (DT), Support Vector Regressor (SVR), and Random Forest (RF), was conducted to choose the most efficient to optimize conveyor belt energy consumption. Through careful data pre-processing and hyperparameter tuning, we demonstrate that RF yielded the best results, with an average train R-squared of 0.930 and a test R-squared of 0.89.