Residential and District-Scale Ground-Coupled Heat Pump Performance with Fiber Optic Distributed Temperature Sensing

Abstract

Ground-coupled heat pumps (GCHP) have garnered considerable attention in recent years as an energy-saving alternative to conventional space heating and cooling systems, but there exists some debate about their effectiveness and efficiency. Some of this debate stems from the inherent complexity of the subsurface heat transfer problem; the rate of subsurface heat transfer ultimately drives GCHP system performance. A given GCHP site may have many geologically distinct layers of soil and rock?each with its own macro-scale thermal properties?which complicates the prediction of heat transfer. Even within a given layer, there exists variability in its properties due to depositional processes, weathering, and fracturing. This work opens with an introduction that prefaces the study and fundamentals of GCHPs, with particular attention paid to the interaction between site geology and heat transfer within the system (Chapter 1). The general concept of using heat pumps for heating and cooling is discussed, and some of the tools for GCHP design, like thermal response tests (TRTs), are briefly explained. The background (Chapter 2) introduces distributed temperature sensing (DTS), the history of its development, and the methods behind its use. The first paper presented in this thesis (Chapter 3) is a two-part study, the first focusing on recording thermal properties of specific lithofacies in a particular region (Wisconsin, USA), and the second part modeling how these measurements could change expected GCHP design and performance in a typical residential setting. Representative sedimentary, igneous, and metamorphic rocks from Wisconsin were characterized in previous work by Meyer (2013) using guarded-comparative-longitudinal heat flow experiments (ASTM E1225), calorimetry, and weight-volume assessments. The model was developed by coupling the rate of heat extraction in a heat pump to an analytical model of heat transfer through a pipe with constant wall temperature. The constant-wall temperature used in the pipe model was taken from an implementation of the infinite line source (ILS) model. This strategy allows the modeled heat extraction rate to be limited by the characteristics of the pipe, which is an advantage over using the ILS model on its own. However, coupling ILS to pipe conduction is still a major simplification in modeling subsurface heat transfer. The indexed thermophysical values from Meyer (2013) and supplemental laboratory measurements were used to construct a hypothetical vertical heat exchange loop penetrating vertically consecutive Paleozoic strata. The hypothetical loop was examined using the rate of system temperature drop under full load as a measure of performance during the heating season. Expected system performance was compared between commonly cited thermal conductivity values of their general lithologies (e.g., sandstone, dolomite, shale) and the laboratory-measured thermal properties of the specific formations. Thermal conductivity indexed by general lithology proved to be insufficient as a design parameter to generate accurate assessments of GSHP performance. The second paper of this study (Chapter 4) and its appendix (Appendix 1) describe the design and implementation of distributed temperature sensing (DTS) as a method of evaluating subsurface heat transfer. DTS is a measuring technique that utilizes fiber-optic cable to measure temperature along continuous profiles. DTS fibers were deployed permanently in an instrumented GCHP system at a residence located in Grand Marsh, Wisconsin, USA. The site was originally intended to be a test location for an unconventionally deep borehole described in Meyer (2013), but was instead converted to a more traditional system with three vertical ground heat exchangers (GHEs) of ~100 m depth each. However, the system was atypical in a number of ways. Air conditioning, space heating, and domestic hot water were supplied by a pair of heat pumps sharing the same GHE loop. A manifold separately controlling flow to each of the GHEs was constructed to allow the flow circuit to be configured as parallel, series, or single-GHE. Temperature, flow rate, and power sensors were installed on components of the heat pump system in a plan similar to that described in the appendix of Meyer (2013). This allowed the system?s coefficient of performance (COP) to be measured along with weather conditions on the site. The dataloggers collecting these measurements were connected to network-connected laptops for remote monitoring and configuration. Though monitoring on the site is in the early stages, daily COP observations were made and related to the electrical power consumption of the heat pumps and the weather conditions during the first half of the 2014-2015 heating season. The daily COP was correlated with which of the two heat pumps delivered the majority of heating to the residence on that day. The water-to-air heat pump was the more efficient of the two, so higher proportional use of the water-to-air heat pump was correlated positively to the daily COP. Heating COP was also positively correlated to the entering water temperature (EWT). This outcome was expected because the Carnot efficiency (i.e., the theoretical maximum efficiency for any heat pump or heat engine) is directly related to the temperature difference between the heat source and heat sink. Heat pump usage, as indicated by their electrical power consumption, was correlated with drops in outdoor temperature. Representative temperature profiles within the system?s GHEs were acquired with the DTS system during seasons with different space conditioning requirements. The profiles, which were taken in January of 2014, indicated that heat recovery was lower than desired. The 2013-2014 heating season was colder than normal, with 223 more heating degree days (Celsius) during the Dec-January time period than average for the central Wisconsin region (Wisconsin Department of Administration). The profiles also showed that heat transfer in the horizontal trenches leading to the GHEs was limited at best, and, during extreme conditions, the direction of heat flux within the trenches could be reversed, likely due to the depressed temperature at the trench?s shallow depth (2 m). Summer profiles showed that heat was easily transferred to the subsurface, due in part to relatively cool ambient temperatures underground (<12 ?C) and heat already extracted during the preceding heating season. The final component of this research was the installation of a DTS network at a large heat exchange borefield in Verona, Wisconsin, USA. Described in Appendix 1, this effort has already produced background DTS measurements and will begin to record the effect of activating a portion of the borefield of over 2500 150-m-deep GHEs early in 2015. The DTS array at this borefield features eight double-ended loops of fiber optic cable connecting 13 boreholes, each with a depth of 150 m. The total length of DTS fiber at this system will surpass 10 km.