Isochrone-Based Voyage Optimization Methods to Increase Shipping Energy Efficiency

Licentiate thesis, 2024

The International Maritime Organization (IMO) mandates improved energy efficiency in shipping, with voyage optimization systems being a key measure. The systems rely on ship models to esti-mate energy costs, and optimization algorithms to find the optimal route. However, models using various modelling techniques have been arbitrarily applied for voyage optimization in existing research. Their effectiveness needs systematic evaluation. Optimization algorithms plan the opti-mal voyage before departure, but uncertainties en route, such as weather and market changes, also require the algorithms to update voyages in real time to ensure optimal operations.

The main objectives of this thesis are to 1) investigate the sensitivity of various ship energy cost models in evaluating energy efficiency for ship voyage optimization, 2) develop strategies to im-prove the Isochrone voyage optimization algorithm, which is well-known for its computational efficiency, 3) propose an Isochrone-based predictive optimization algorithm for real-time ship voyage planning and execution for energy efficiency shipping.

Firstly, five ship models are integrated into a voyage optimization algorithm to identify their im-pacts on optimization. It is found that machine learning models present better reliability than theo-retical models in diverse sailing environments, and considering specific fuel oil consumption (SFOC) variation in actual operation is essential to estimate accurate fuel costs. Secondly, five strategies are proposed to improve the traditional Isochrone algorithm. It is found that one pro-posed strategy, named ‘Isochrone A*’, can effectively resolve its drawbacks, such as unrealistic routes and locally optimized results. Meanwhile, it provides the most fuel savings. Based on these two findings, an Isochrone-based predictive optimization (IPO) method is finally proposed, with enhanced performance while remaining computational efficient. The proposed method has been validated using six case study voyages with full-scale measurements, and compared with four voy-age optimization algorithms. It is found that, the proposed IPO method demonstrates smoother voyages with on-time arrivals and an average fuel reduction of 5% across diverse sailing condi-tions. Its runtime of around 40 seconds is also suitable for real-time usages.