Climate-adaptive control of a hybrid heat pump-desiccant wheel cooling system

Conventional air conditioning systems are energy intensive and have a limited ability to control humidity independently, especially in hot and humid climates. Existing hybrid air-conditioning systems often lack the flexibility and adaptability required to operate effectively under dynamic, real-world environmental conditions. Therefore, this proposes a novel, climate-adaptive control framework for a hybrid heat pump-desiccant wheel system designed to optimize energy consumption and indoor thermal comfort for an office building. The integrated system consists of an energy recovery ventilation unit, a desiccant wheel, and a heat pump. The methodology integrates an artificial neural network surrogate model for the desiccant wheel with a physics-based heat pump model, governed by a multi-objective optimization framework that adaptively selects operational modes and continuously optimizes wheel speed, face velocity, regeneration temperature, and bypass fraction to maintain temperature and humidity setpoints. The ANN surrogate model accurately predicted desiccant wheel performance (R2 = 0.998), and the adaptive control system maintained humidity tracking errors within ±0.1 g·kg−1 by dynamically adjusting operational parameters under varying climate conditions. Furthermore, compared to a conventional direct expansion vapor compression air conditioning system, the hybrid system achieved energy savings, ranging from over 45% in moderate humidity to 15% in extreme latent load conditions, with a system equivalent Coefficient of Performance peaking at 2.6. This study demonstrates that a machine learning-assisted, multi-objective control strategy can significantly improve the energy efficiency and operational performance of hybrid cooling systems under challenging climatic conditions.