Predictive Maintenance For Straddle Carriers Using Advanced Deep Learning Models: A Comparative Exploration

Abstract

Predictive analysis for machine maintenance is the key to reducing downtimes and ensuring smooth operations while increasing productivity. Emergency repairs reduce while the lifespan of machinery increases. In this study, we focused on predicting spreader issues of straddle carriers used at Cargotec (Kalmar). Straddle carriers are machines to pick and place shipping containers. The spreader, a component of the straddle carrier, performs the pick and ground operation. On-board automation system logs of straddle carriers were used for the analysis. All four advanced deep learning models were able to effectively predict failures while minimizing false positives and false negatives with varying training times.

This study offers a nuanced understanding of the strengths and weaknesses of the models that were used and provides a comprehensive and practical understanding of various deep learning models in the context of predictive maintenance.