Abstract
With global demand for energy and the pressing need for climate change, the development of an energy-efficient and sustainable cooling technology has become a significant research priority. The traditional mechanical vapour compression (MVC) refrigeration systems contribute greatly to greenhouse gas emissions through their high consumption of electricity and utilisation of high global warming potential refrigerants. Adsorption chillers (ADCs) are an environmentally benign alternative cooling system driven by low-grade waste heat (e.g., from industrial processes or solar collectors), thereby transforming a wasted energy stream into a valuable cooling output. Despite their potential and many advantages over the conventional MVC systems, ADCs usually suffer from lower Coefficients of Performance (COP) and larger physical footprints, which hinder their commercial deployment. The review of existing literature highlighted that although experimental and component-level studies are significant, there is limited application and relative under-exploration of advanced, multi-objective metaheuristic optimisation algorithms to improve the performance of ADC. A key research gap was the lack of explicit treatment of waste heat recovery efficiency ππ as a primary optimisation objective alongside traditional metrics like COP and πππ. This study develops a robust computational ADC framework to identify the maximum performance potential of a silica gelβwater ADC, starting from the single-objective analysis to explore the complex, multi-dimensional trade-offs between three performance indicators.
The primary aim of this study was to develop, apply, and compare a multi-objective optimisation framework based on three distinct metaheuristic algorithms to maximise the performance of a single-stage, dual-bed silica gelβwater adsorption chiller. It started by developing statistically validated surrogate regression models for three performance indicators: coefficient of performance (COP), cooling capacity (πππ) and waste heat recovery efficiency ππ). ππ was introduced as a novel co-equal objective to provide a "source-aware" perspective that is seldom reported in ADC literature. The study continued by applying, solving and comparing the tri-objective optimisation problem using three metaheuristic algorithms, namely the Multi-Objective Grey Wolf Optimiser (MOGWO), Multi-Objective Antlion Optimiser (MOALO), and Multi-Objective Particle Swarm Optimisation (MOPSO). The study also identified Pareto-optimal solution fronts, analysed the fundamental trade-offs between the three KPIs to translate the abstract optimisation results into a practical, implementation-ready Operational Set-Point Matrix for system designers and operators. The performance objective equations were modelled using surrogate regression models derived from a statistically...