Performance optimization of advanced vapor compression systems working with low-GWP refrigerants using numerical and experimental methods

Haider, Muhammad

Permalink

https://hdl.handle.net/2142/124484 Copy

Description

Title

Performance optimization of advanced vapor compression systems working with low-GWP refrigerants using numerical and experimental methods

Author(s)

Haider, Muhammad

Issue Date

2024-03-01

Director of Research (if dissertation) or Advisor (if thesis)

Elbel, Stefan

Doctoral Committee Chair(s)

Miljkovic, Nenad

Committee Member(s)

• Wang, Sophie
• Kozlowski, Tomasz

Department of Study

Mechanical Sci & Engineering

Discipline

Mechanical Engineering

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• Advanced Vapor Compression Systems
• Steady-state System Modeling
• Co2
• Ejector Performance Map
• Heat Exchanger Modeling
• Low-gwp Zeotropic Blend
• Circulation Composition Measurement

Language

eng

Abstract

Recent research has focused on reducing the environmental impact of HVAC&R systems while enhancing their performance. Current trends indicate that low-GWP pure refrigerants and blends hold significance for future applications. Moreover, the emergence of new applications suggests that future vapor compression systems (VCS) will be more complex than their predecessors. Therefore, advanced design tools are necessary to handle complex VCS using pure refrigerants with ease, along with analyzing the intricacies of refrigerant blends. One objective of this research is to develop robust, accurate and computationally efficient steady-state system solvers using gradient methods for both pure and mixture refrigerants. Experimental methods are utilized not only for validating different components and system models, but also to develop novel techniques when essential knowledge gaps need filling. The research has focused on ejector systems, as they represent a family of complex VCS and are a good representative example of advanced VCS. The research develops robust and computationally efficient simultaneous system solvers by modeling a transcritical CO2 air conditioning system operating on a standard ejector cycle with an internal heat exchanger. Computationally efficient and accurate component models are developed. A novel ejector performance map is developed from experimental data to predict the performance of a fixed-geometry ejector across a wide range of operating conditions. For heat exchangers, both finite volume (FV) and artificial neural network (ANN) approaches are developed and compared. It is found that the advanced VCS transforms into a non-convex and ill-conditioned problem, which makes gradient methods struggle to find convergence, particularly when the number of solver variables is increased. The work proposes that the robustness of simultaneous solvers can be improved by appropriate scaling factors in the residual equations. Furthermore, using constrained algorithms instead of unconstrained algorithms improves the solver’s robustness even with a higher number of solver variables. The simultaneous solvers are extended to model vapor compression systems working with zeotropic blends. The focus is to identify an ejector cycle whose performance can be enhanced with blends. Three different cycle architectures, one conventional system and two ejector cycles, namely, the standard ejector cycle and the COS cycle, are modeled using low-GWP mixtures of R1234yf/R32. The conventional system and the COS ejector cycle have the same circulation composition as the system is fed from the liquid port of the receiver, whereas the standard ejector cycle has two circulation compositions for high and low-side due to fractionation inside the separator. The numerical models provide insight that the performance of systems with the same circulation composition is expected to improve upon the addition of the more-volatile substance, whereas the performance of the standard ejector cycle will decrease. The decrease is attributed to fractionation as it causes most of the added more-volatile substance to move towards the high-side of the cycle, thus leading to an increase in compressor power. Based on this analysis, experiments are conducted on a chiller facility to test the performance of the conventional system and the COS ejector cycle using R134a/R32 mixtures. It is found that the COP of the conventional system with an internal heat exchanger is increased by 27% under matched capacity conditions, whereas the COP of COS ejector cycle increases by almost 23%. Thus, this proves that gain in the ejector system with blends is possible and requires careful investigation of the storage vessel’s location inside the cycle. Another important aspect in a blend system is to measure the circulation composition accurately, which could be different from the charged composition. For measuring circulation composition, a novel PTD (gas) method is developed. Furthermore, three different in-situ estimation techniques are evaluated for their effectiveness. It is recommended that a calibration procedure should be used before using any of the in-situ estimation methods to avoid a 2-5% error in reporting the system performance. The present research can be useful in analyzing advanced VCS with relative ease. The findings can help improve the finite volume heat exchanger model and integrate complex closure equations like charge modeling. In addition, it also presents a few experimental techniques like a novel ejector performance map that can help with ejector selection decision, and a novel PTD (gas) method to measure a blend’s circulation composition, making experimental studies of blends relatively an easier endeavor.

Graduation Semester

2024-05

Type of Resource

Text

Handle URL

https://hdl.handle.net/2142/124484

Copyright and License Information

Copyright 2024 Muhammad Haider

Owning Collections