Structural Optimization of Disk Diffusers Based on Radial Basis Functions and Multi-Island Genetic Algorithms

HVAC systems are major energy consumers in buildings, accounting for 30% to 50% of total energy usage. Supply outlets, as the terminal devices of HVAC systems, significantly influence the system’s operational performance, energy consumption, and indoor environmental quality. However, current designs and performance optimizations of supply outlets still encounter challenges such as high resistance and low energy efficiency, posing substantial obstacles to building energy conservation, indoor air quality, and fan energy consumption. To address the optimization of resistance reduction and efficiency enhancement for local components (specifically the supply outlets) in HVAC systems, this study focuses on the commonly used disk diffusers. Utilizing a combined research methodology that integrates theoretical analysis, numerical simulation, full-scale experiments, surrogate model prediction, and multi-island genetic algorithm optimization, this study investigates both fluid flow and resistance distribution characteristics. The ADEI comprehensive evaluation index is employed to assess the operational energy efficiency of the disk diffusers. Based on an optimized Latin hypercube sampling method and incorporating RBF surrogate models, surrogate models relating the structural parameters of disk diffusers to their resistance and range are developed. A multi-island genetic algorithm is then applied to optimize the RBF surrogate models. The optimization results demonstrate that the new type of disk diffuser achieves a 33.07% reduction in ADEI compared to traditional disk diffusers, while resistance decreases by 23.10% and jet length increases by 7.19%.