Development and Demonstration of a Method to Determine Residential Air-Conditioning System Fault Prevalence

Air conditioning system efficiency and effectiveness can be reduced by undetected installation faults. Many fault detection and diagnosis (FDD) methods have been proposed during the past three decades, but they are not widely applied in residential systems. Almost all methods can only handle single faults, rather than multiple simultaneous faults. An additional challenge is that prevalence of each fault is uncertain, so that the cost-effectiveness of FDD is unclear to homeowners or decision-makers, slowing adoption of FDD and investment into further development. This dissertation describes part of a project to address the need for fault prevalence data in residential air-conditioning systems. It discusses existing FDD methods that were compared and selected for a large fault prevalence field study. To evaluate the effectiveness of those methods, it is necessary to have test data with combinations of simultaneous installation faults, including improper refrigerant charge (CH), improper evaporator airflow (EA), liquid line restrictions (LL), and the presence of non-condensable gas (NC). However, no such data previously existed. Therefore, an extensive laboratory investigation was conducted, in which these faults were imposed both individually and simultaneously on a 4-ton split heat pump equipped with TXVs and a 3-ton split air conditioner equipped with an FXO in cooling mode. The impacts of single, double, triple, and quadruple faults on the performance and characteristic fault detection features of each system were investigated. The superposition principle for combined faults (up to 4) was tested with the collected data to characterize synergistic or cancelling effects. Results and analysis of this novel testing are provided in the dissertation. The FDD methods for the fault prevalence field study – used to detect CH, LL, and NC faults – were validated, modified, and trained by using the experimental data, along with data from previous studies. A case study of air conditioning systems in the field was conducted to demonstrate the method for two faults – NC and EA. The COVID 19 pandemic prevented further field testing, which was initially planned