Enhancing long-term motion prediction for floating energy platforms: A physics-informed framework multi-scale recalibrated fusion transformer model

2026-02-06

PRE

Yang Chen

Lihao Yuan *

Baicheng Lyu

Mingyang Guo

Zhi Zhou

Zhongming Li

DOI：10.1016/j.energy.2026.140325

Original

Original request for help

Save

Abstract

En 中文

• Propose a physics-informed Transformer model (MRF-Transformer) for long-term motion prediction of floating energy platforms. • Achieve significant accuracy improvements in motion forecasts under operational and typhoon conditions. • Develop an environment-adaptive mechanism to dynamically recalibrate multi-scale ocean features. • Design dual-residual structures to mitigate long-term prediction drift and error accumulation. • Enable enhanced operational safety and CO2 reduction for offshore renewable energy systems.

Keywords:

long-term motion prediction

floating energy platforms

physics-informed Transformer

multi-scale feature fusion

error accumulation mitigation

Enhancing long-term motion prediction for floating energy platforms: A physics-informed framework multi-scale recalibrated fusion transformer model

Abstract

Journal

Organization