Deep learning models for straddle carriers: Predictive maintenance

The secret to decreasing downtime, guaranteeing smooth operations, and raising productivity in machine maintenance is predictive maintenance. With predictive maintenance, the need for emergency maintenance decreases. The goal of this study was to forecast spreader problems with the straddle carriers that Cargotec (Kalmar) uses. Machines called "straddle carriers" are used to pick and place shipping containers. The pick and ground action is carried out by the spreader, which is a part of the straddle carrier. The investigation was conducted using straddle carrier logs from their on-board automation systems. With different training times, all four of the advanced deep learning models were able to minimize false positives and false negatives and accurately forecast failures. This study gives a thorough overview of different deep learning models in the context of predictive maintenance, as well as a comprehension of the advantages and disadvantages of the models that were employed. • This paper discusses the use of advanced deep learning algorithms for predictive maintenance of straddle carriers. • Some forms of recurrent neural networks and convolutional neural networks are used and compared. • The data preprocessing for sequence-based modeling is described along with hyper parameter tuning and training times. • This is groundwork for improved maintenance procedures driven by data-driven strategies in container terminal operations.