Challenges in cable burial installation: new opportunities with machine learning

This study investigates submarine cable burial via trench ploughing, where accurate prediction of tow force is critical for design, scheduling, and cost. Traditional analytical models are often oversimplified and inaccurate, while advanced numerical methods are computationally costly and require advanced laboratory testing for calibration. This paper explores Machine Learning (ML) to improve predictions, using field data and geotechnical parameters from standard site investigations. Benchmarking with real data shows the limited predictive capabilities of conventional models, while the ML model achieves higher accuracy when blind-tested. Results highlight ML potential to deliver more reliable, efficient, and cost-effective submarine cable installation design.