As a new passive cooling technology, space radiation cooling has great potential for development because the cooling itself has no energy consumption, and the radiation heat exchanger does not affect the appearance, with low noise and low cost. Several rectangular stainless steel plates coated with RLHY-2 material are used as the transmitter for the field test. The experimental results show that, in the case of no windscreen, the increase of outdoor humidity will reduce the cooling effect, and the greater the humidity, the more pronounced the reduction effect. Significantly when the humidity increases from 78% to 90%, the cooling power of the cooler reduces from 102 to 67 W/m2. The thickness of the cloud layer also affects the cooling effect of the space radiative cooler. Compared with the clear weather, the cooling power of the cooler is reduced by 11.65 W/m2 on average under foggy weather conditions. Compared with the force-1 wind and the force-3 wind, the cooling effect of the cooler is the worst under the condition of the force-2 wind, and the average cooling power is only 49.76 W/m2. In addition, laying polyethylene (PE) film as a windscreen is beneficial to improving the radiative cooling effect, and the difference in surface temperature between the two is up to 3 °C. This research provides a theoretical basis and practical reference for applying radiative cooling technology in different regions and seasons and adjusting and improving its effects.
Accurate and real-time prediction of ship infrared (IR) radiation characteristics is crucial for maritime search-and-rescue (SAR) operations and naval stealth management, but traditional methods are either computationally inefficient (e.g., Computational Fluid Dynamics requiring hours per scenario) or lack cross-scenario generalization (e.g., pure data-driven models). This study proposes a Kernel-Attention Mechanisms U-Net (KEU-Net) framework to address these issues. The framework integrates U-Net’s robust spatial feature extraction—via skip connections that preserve edge details of small IR targets (e.g., exhaust plumes, hull thermal gradients)—with a kernel-attention module. This module dynamically adjusts feature weighting based on maritime environmental factors (solar angles, sea states) and ship operational parameters (heading, speed), avoiding parameter redundancy of multi-layer perceptions (MLPs). Trained on three datasets covering seasons, weather, and thermal suppression scenarios, KEU-Net achieves 97.3% accuracy against finite element benchmarks, maintains ≤1.7°C temperature error across Beaufort 3–8 sea states, and completes full-azimuth IR pattern prediction in 12 seconds per scenario (32× faster than response surface models). It supports real-time SAR and stealth planning, with strong compatibility with naval systems.