Automatic Fault Detection in domestic heat pumps

Identifier:  TDX:3974

Handle: https://hdl.handle.net/20.500.11797/TDX3974

Authors:  Bellanco Bellanco, Ivan

Abstract:
In the current context of climate emergency, governments around the world are taking actions to reduce greenhouse gases emissions (GHG), which are the main agents of climate change. Buildings accounted for 32% of the total energy consumption worldwide, being water heating and space heating production the main use of energy in domestic buildings. Heat pumps could provide water heating and space heating and cooling with high efficiency and, coupled with renewable electric sources, with low GHG emissions. Despite their high efficiency, they could be working below their design efficiency in domestic installations or with undetected faults. Fault Detection and Diagnosis (FDD) monitors are focused on the detection of performance losses and the diagnosis of faults. The main objective of this doctoral thesis is the development of an FDD monitor for domestic heat pumps, which could work with variable speed compressors and diagnose different faults. The first part of this thesis is dedicated to the literature review of faults incidence, effects and its emulation or simulation. The next part is focused on the development and first validation of the FDD monitor. The monitor counts with several modules to increase the reliability of the solution. The monitor uses polynomial regressions for the generation of fault-free models and expert knowledge for the diagnosis. The solution is first validated with model data, achieving fault detection before a 5% of performance degradation and diagnosing several faults. Experimental faults test were performed in a variable-speed heat pump charged with propane. The procedure is described and the fault effects are quantified. Then, this data is used to validate again the monitor. Despite some modifications were required to adapt to real-time data, the monitor detected the faults before an 8% of degradation and diagnosing the majority of faults.