WANA (Oct 06) – Iranian Researchers at Amirkabir University of Technology have developed an innovative machine-learning model capable of predicting significant wave height with an accuracy rate of 93 to 97 percent. The model, which requires no specialized expertise, is designed to be accessible to all users through an intelligent, AI-based graphical user interface (GUI).

 

According to the research team, this development aims to support the operation and planning of wave energy converters (WECs) and facilitate decision-making in marine renewable energy projects.

 

The model’s standout feature is its ability to operate effectively with limited and scattered data — a challenge often faced in marine environments where data collection is costly and time-consuming. Its user-friendly interface enables individuals with varying technical backgrounds to utilize it for the optimal planning of wave energy systems.

 

Amirhossein Shahbazbegian, a graduate in Marine Engineering under the supervision of Dr. Mahmoud Ghiassi, Associate Professor at the Faculty of Marine Engineering, led the project and co-authored a paper titled “Development of a Machine-Learning-Based Graphical User Interface for Predicting Significant Wave Height to Support Wave Energy Converter Operations Planning.” The paper was recently published in Renewable Energy, a reputable journal by Elsevier.

 

Speaking about the research, Shahbazbegian said: “Together with Dr. Ghiassi, we developed an innovative and user-friendly artificial intelligence model to predict significant wave height — a key parameter in designing and operating marine energy converters. Even small prediction errors can lead to lower efficiency, higher maintenance costs, or even irreversible damage.”

 

He explained that one of the main challenges in this field is the heavy reliance of conventional machine-learning models on large datasets — something often unavailable in many coastal regions, especially in developing countries.

 

“Our research directly addresses this challenge,” he added. “We developed an optimized AI model capable of learning from small and scattered datasets — an important advancement for marine applications.”

 

The team trained the model using data from four offshore buoys located along various parts of Australia’s coastline, deliberately chosen to represent a wide range of climatic conditions and wave patterns. The results were described as “remarkable,” with the model outperforming many existing machine-learning algorithms.

 

“But the real innovation,” Shahbazbegian emphasized, “was bridging the gap between academia and real-world application. We implemented the model into an intelligent and intuitive graphical user interface that allows any user — from coastal engineers to energy planners — to input local data and receive precise wave-height forecasts for timeframes ranging from one month to a year, displayed through easy-to-understand charts and reports.”

 

He noted that this approach uniquely combines advanced artificial intelligence with user-centered design — a feature still rare in the field of marine renewable energy. “With this tool, users can identify optimal locations for wave energy converters, significantly reduce initial study costs, and obtain reliable predictions even in data-scarce regions,” he said.

 

Shahbazbegian concluded by highlighting the project’s broader impact: “This research is not just a scientific achievement; it’s a practical step toward a cleaner, more sustainable future. At a time when the world faces the urgent challenge of climate change and the need to reduce dependence on fossil fuels, intelligent tools like this can play a crucial role in advancing marine renewable energy and supporting the transition to a low-carbon world.”

Artificial intelligence (AI)