Single-pressure absorption refrigeration systems for low-source-temperature applications
Rattner, Alexander S.
MetadataShow full item record
The diffusion absorption refrigeration (DAR) cycle is a promising technology for fully thermally driven cooling. It is well suited to applications in medicine refrigeration and air-conditioning in off-grid settings. However, design and engineering knowhow for the technology is limited; therefore, system development has historically been an iterative and expensive process. Additionally, conventional system designs require high-grade energy input for operation, and are unsuitable for low-temperature solar- or waste-heat activated applications. In the present effort, component- and system-level DAR engineering analyses are performed. Detailed bubble-pump generator (BPG) component models are developed, and are validated experimentally and with direct simulations. Investigations into the BPG focus on the Taylor flow pattern in the intermediate Bond number regime, which has not yet been thoroughly characterized in the literature, and has numerous industry applications, including nuclear fuel processing and well dewatering. A coupling-fluid heated BPG design is also investigated experimentally for low-source-temperature operation. Phase-change simulation methodologies are developed to rigorously study the continuously developing flow pattern in this BPG configuration. Detailed component-level models are also formulated for all of the other DAR heat and mass exchangers, and are integrated to yield a complete system-level model. Results from these modeling studies are applied to develop a novel fully passive low-source-temperature (110 - 130°C) DAR system that delivers refrigeration grade cooling. This design achieves operation at target conditions through the use of alternate working fluids (NH3-NaSCN-He), the coupling-fluid heated BPG, and a novel absorber configuration. The complete DAR system is demonstrated experimentally, and evaluated over a range of operating conditions. Experimental results are applied to assess and refine component- and system- level models.