Research progress on intelligent optimization techniques for energy-efficient design of ship hull forms

Shuwei Zhu, Ning Sun, Siying Lv, Kaifeng Chen, Wei Fang, Leilei Cao

Introduction

The maritime industry is currently undergoing a significant transformation driven by the need to meet aggressive decarbonization targets. With the International Maritime Organization (IMO) aiming for a 50% reduction in greenhouse gas emissions by 2050, ship efficiency has become a critical research priority. Consequently, the field has seen a dramatic increase in research activity focused on intelligent optimization techniques for hull design.

The increasing trend in publications related to intelligent ship hull optimization from 2014 to 2023.

The Computational Bottleneck

Traditionally, ship hull design relied on empirical "rule-of-thumb" methods and physical model testing. The advent of Computational Fluid Dynamics (CFD) allowed for high-fidelity simulations of water flow, providing a "gold standard" for evaluating hull efficiency. However, CFD is computationally expensive; a single simulation can take hours or even days. This creates a bottleneck when optimization algorithms require thousands of iterations to identify an optimal shape. To bridge this gap, researchers are turning to "intelligent" shortcuts—algorithms and mathematical models that maintain accuracy while drastically reducing time.

Simulation-Based Design (SBD) and Intelligent Optimization

To overcome these computational hurdles, researchers have shifted toward Simulation-Based Design (SBD) powered by intelligent optimization. This framework integrates automated geometry reconstruction, high-fidelity CFD, and mathematical "shortcuts" to explore the design space efficiently.

The SBD technique framework, illustrating the integration of intelligent optimization methods, approximation (surrogate) models, and CFD calculations.

The core of this "intelligent" toolkit includes:

  • Hull Geometry Reconstruction: Techniques such as Free-Form Deformation (FFD) and Radial Basis Functions (RBF) allow for the automatic "morphing" of hull shapes using a limited set of control variables.
  • Surrogate Models (Metamodels): These are mathematical approximations—such as Kriging, Neural Networks, or Support Vector Machines—trained on a small dataset of CFD results. They can predict the performance of new hull forms almost instantaneously, acting as "cheap" stand-ins for full simulations.
  • Intelligent Optimization Algorithms: The industry is moving away from simple gradient-based searches toward nature-inspired, derivative-free algorithms like Particle Swarm Optimization (PSO) and Genetic Algorithms (GA), which are better at finding global optima in complex design spaces.

Benchmarking and Performance Gains

The effectiveness of these techniques is typically validated using industry-standard benchmarks such as the DTMB-5415 (a U.S. Navy destroyer design), the Japan Bulk Carrier (JBC), and the KCS container ship.

Research has demonstrated that surrogate-assisted optimization can achieve significant improvements. For instance, studies on the DTMB-5415 hull using Co-Kriging models reported total drag reductions ranging from 5.13% to 10.37%. Furthermore, Surrogate-Assisted Evolutionary Algorithms (SAEAs) have been shown to find optimal designs using 90% fewer high-fidelity simulations than traditional methods.

Future Directions: Beyond Calm Water

Modern design is evolving from single-objective optimization (e.g., minimizing resistance at one speed) to Multi-Objective Evolutionary Algorithms (MOEAs). These allow designers to find "Pareto optimal" solutions that perform efficiently across various speeds and sea states simultaneously.

Looking forward, the field is moving toward:

  1. Holistic Optimization: Treating the hull, propeller, and engine as a single coupled system rather than optimizing components in isolation.
  2. Robust Design: Accounting for real-world uncertainties such as stochastic weather conditions and manufacturing tolerances to ensure performance beyond calm water.
  3. Physics-Aware AI: Combining the pattern-recognition capabilities of deep learning with the fundamental physical laws of fluid mechanics to create more reliable hybrid models.

By bridging the gap between high-fidelity simulation and practical engineering timelines, intelligent optimization is paving the way for a new generation of energy-efficient maritime vessels.

Summary updated

BibTeX
@article{zhu2024research,
  author = {Shuwei Zhu and Ning Sun and Siying Lv and Kaifeng Chen and Wei Fang and Leilei Cao},
  title = {{Research progress on intelligent optimization techniques for energy-efficient design of ship hull forms}},
  journal = {J. Membr. Comput.},
  year = {2024},
  doi = {10.1007/s41965-024-00169-6},
}
Figure