Date of Award

Spring 1-1-2013

Document Type

Thesis

Degree Name

Master of Science (MS)

First Advisor

Michael Brandemuehl

Second Advisor

Moncef Krarti

Third Advisor

Jon Winkler

Abstract

This thesis provides the capability for improved predictions of the annual energy use of residential air conditioners and heat pumps.The annual simulation engine - EnergyPlus - was used to evaluate the effects that manufacturers data has on predicted annual energy consumption. A new approach to modeling residential air conditioners and heat pumps (inside of the framework of EnergyPlus) was developed and is presented here. The new approach has updated system information and involves a reduced number of user inputs. This approach was also adjusted for use with the simulation engine, DOE-2.

To inform this research, manufacturer data was collected for the majority of currently available air conditioners and heat pumps. This data was parsed and written into a SQLite database that was used to generate full sets of model inputs for over 450 heat pumps and air conditioners. The predicted performance of these different units was evaluated through annual simulation in EnergyPlus. Inspection of the simulation results led to questions regarding the quality of data coming from the manufacturers. The combination of questionable data and the benefits of simplified model inputs led to a sensitivity analysis on the various model inputs.

The results of this sensitivity analysis showed limited impact from the complex regression curve-fits on the expanded performance tables, especially for single speed units. Representative curves were selected and a parametric study was performed in various climates that demonstrated the impact of using the selected curves in place of individual unit curves. Rated values were the cause of most of the variability observed in the initial simulations. Rated values were evaluated through a comparison with AHRI published data. Results from the sensitivity study and the rated value comparison led to recommendations for best modeling practices. Information on how the manufacturers' data impacted predicted energy use can be used to inform future reporting standards.

Share

COinS