The cost of the wells is usually one of the larger items in the overall cost of a geothermal system, and increases with resource depth. Compared to many geothermal areas worldwide, well depth requirements in the western United States are relatively shallow; most larger geothermal systems there operate with production wells of less than 2000 ft, and many at less than 1000 ft.

Direct use of geothermal energy must occur near the resource. The reason is primarily economic; although geothermal (or secondary) fluid could be transmitted over moderately long distances (greater than 60 miles) without great temperature loss, such transmission is generally not economically feasible. Most existing geothermal projects have transmission distances of less than 1 mile.

Energy output from a production well varies directly with the fluid flow rate. The energy cost at the wellhead varies inversely with the well flow rate. A typical good resource has a production rate of 400 to 800 gpm per production well; however, geothermal direct-use wells have been designed to produce up to 2000 gpm.

The available temperature is fixed by the prevailing resource. The temperature can restrict applications. It often requires a reevaluation of accepted application temperatures, which were developed for uses served by conventional fuels for which the application temperature could be selected at any value in a relatively broad range. Most existing direct-use projects use fluids in the 130 to 230°F range.

Because well flow is limited, power output from a geothermal well is directly proportional to the temperature drop of the geothermal fluid connected to the system. Consequently, a larger temperature drop reduces operating (pumping) and capital (well and production pump) costs.

Cascading geothermal fluid to uses with lower temperature requirements can help achieve a large temperature difference (Δt). Most geothermal systems are designed for a Δt between 30 and 50°F, although one system was designed for a Δt of 100°F with a 190°F resource temperature.

Defined as the ratio of the average load to the design capacity of the system, the load factor effectively reflects the fraction of time that the initial investment in the system is working. Again, because geothermal cost is primarily initial rather than operating cost, this factor significantly affects a geothermal system’s viability. As the load factor increases, so does the economy of using geothermal energy. The two main ways of increasing the load factor are (1) to select applications where it is naturally high, and (2) to use peaking equipment so that the geothermal design load is not the application peak load, but rather a reduced load that occurs over a longer period.

The quality of the produced fluid is site specific and may vary from less than 1000 ppm TDS to heavily brined. Fluid quality influences two aspects of the design: (1) material selection to avoid corrosion and scaling effects, and (2) disposal or ultimate end use of the fluid.

The costs associated with disposal, particularly when injection is involved, can substantially affect development costs. Historically, most geothermal effluent was disposed of on the surface, including discharge to irrigation, rivers, and lakes. This method of disposal is considerably less expensive than constructing injection wells.

Geothermal fluids sometimes contain chemical constituents that make surface disposal problematic. Some of these constituents are listed in Table 1.

Most new, large geothermal systems use injection for disposal to minimize environmental concerns and ensure long-term resource reliability. If injection is chosen, the depth at which the fluid can be injected affects well cost substantially. Many jurisdictions require the fluid be returned to the same or similar aquifers; thus, it may be necessary to bore the injection well to the same depth as the production well. Direct-use injection wells are considered Class V wells under the U.S. Environmental Protection Agency’s Underground Injection Control (UIC) program. Water wells, along with terminology relating to the technology, are discussed in the section on Groundwater Heat Pumps.

Low-temperature geothermal fluids commonly contain seven key chemical species that can significantly corrode standard materials of construction (Ellis 1989). These include

The principal effects of these species are summarized in Table 2. Except as noted, the described effects are for carbon steel. Kindle and Woodruff (1981) present recommended procedures for complete chemical analysis of geothermal well water.

Two of these species are not reliably detected by standard water chemistry tests and deserve special mention. Dissolved oxygen does not occur naturally in low-temperature (120 to 220°F) geothermal fluids that contain traces of hydrogen sulfide. However, because of slow-reaction kinetics, oxygen from air inleakage may persist for some minutes. Once the geothermal fluid is produced, it is extremely difficult to prevent contamination, especially if pumps used to move the fluid are not downhole-submersible or lineshaft turbine pumps. Even if fluid systems are maintained at positive pressure, air inleakage at pump seals is likely, particularly with the low level of maintenance in many installations.

Hydrogen sulfide is ubiquitous in extremely low concentrations in geothermal fluids above 120°F. This corrosive species also occurs naturally in many cooler groundwaters. For strongly affected alloys, such as cupronickel, hydrogen sulfide concentrations in the low parts per billion (10⁹) range may have a serious detrimental effect, especially if oxygen is also present. At these levels, the characteristic rotten egg odor of hydrogen sulfide may be absent, so field testing may be required for detection. Hydrogen sulfide levels down to 50 ppb can be detected using a simple field kit; however, absence of hydrogen sulfide at this low level may not preclude damage by this species.

Two other key species that should be measured in the field are pH and carbon dioxide concentrations. This is necessary because most geothermal fluids release carbon dioxide rapidly, causing a rise in pH.

Production of suspended solids (sand) from a well should be addressed during well construction with gravel pack, screen, or both. Proper selection of the screen/gravel pack is based on sieve analysis of cutting samples from drilling. Surface separation is less desirable because it requires sand to pass first through the pump, reducing its useful life.

Biological fouling is largely a phenomenon of low-temperature (<90°F) wells. The most prominent organisms are various strains (Galionella, Crenothrix) of what are commonly referred to as iron bacteria. These organisms typically inhabit water with a pH range of 6.0 to 8.0, dissolved oxygen content of less than 5 ppm, ferrous iron content of less than 0.2 ppm, and a temperature of 46 to 61°F (Hackett and Lehr 1985). Iron bacteria can be identified microscopically. The most common treatment for iron bacteria infestation is chlorination, surging, and flushing; success depends on maintaining proper pH (less than 8.5), dosage, free residual chlorine content (200 to 500 ppm), contact time (24 h minimum), and agitation or surging. Precleaning (wire brushing) of the screen and redevelopment of the well after treatment are key to effectiveness. Hackett and Lehr (1985) provide additional detail on treatment.

Carbon Steel. The Ryznar index has traditionally been used to estimate the corrosivity and scaling tendencies of potable water supplies. However, one study found no significant correlation (at the 95% confidence level) between carbon steel corrosion and the Ryznar index (Ellis and Smith 1983). Therefore, the Ryznar and other indices based on calcium carbonate saturation should not be used to predict corrosion in geothermal systems, though they remain valid for scaling prediction.

In Class Va geothermal fluids [as described by Ellis (1989); <5000 ppm total key species (TKS), total alkalinity 207 to 1329 ppm as CaCO₃, pH 6.7 to 7.6], corrosion rates of about 5 to 20 mil/yr can be expected, often with severe pitting.

In Class Vb geothermal fluids [as described by Ellis (1989); <5000 ppm TKS, total alkalinity <210 ppm as CaCO₃, pH 7.8 to 9.85], carbon steel piping has given good service in a number of systems, as long as system design rigorously excluded oxygen. However, introduction of 30 ppb oxygen under turbulent flow conditions causes a fourfold increase in uniform corrosion. Saturation with air often increases the corrosion rate by at least 15 times. Oxygen contamination at the 50 ppb level often causes severe pitting. Chronic oxygen contamination causes rapid failure.

External surfaces of buried steel pipe must be protected from contact with groundwater. Groundwater is aerated and has caused pipe failures by external corrosion. Required external protection can be obtained by coatings, pipe-wrap, or preinsulated piping, provided the selected material resists the system operating temperature and thermal stress.

At temperatures above 135°F, galvanizing (zinc coating) does not reliably protect steel from either geothermal fluid or groundwater. Hydrogen blistering can be prevented by using void-free (killed) steels.

Low-alloy steels (steels containing not more than 4% alloying elements) have corrosion resistance similar, in most respects, to carbon steels. As with carbon steels, sulfide promotes entry of atomic hydrogen into the metal lattice. If the steel exceeds a hardness of Rockwell C22, sulfide stress cracking may occur.

Copper and Copper Alloys. Copper-tubed fan-coil units and heat exchangers have consistently poor performance because of traces of sulfide species found in geothermal fluids in the United States. Copper tubing rapidly becomes fouled with cuprous sulfide films more than 1 mm thick. Serious crevice corrosion occurs at cracks in the film, and uniform corrosion rates of 2 to 6 mil/yr appear typical, based on failure analyses.

Experience in Iceland also indicates that copper is unsatisfactory for heat exchange service and that most brasses (Cu-Zn) and bronzes (Cu-Sn) are even less suitable. Cupronickel often performs more poorly than copper in low-temperature geothermal service because of trace sulfide.

Much less information is available regarding copper and copper alloys in non-heat-transfer service. Copper pipe shows corrosion behavior similar to copper heat exchange tubes under conditions of moderate turbulence (Reynolds numbers of 40,000 to 70,000). An internal inspection of yellow brass valves showed no significant corrosion. However, silicon bronze CA 875 (12-16Cr, 3-5Si, <0.05Pb, <0.05P), an alloy normally resistant to dealloying, failed in less than three years when used as a pump impeller. Leaded red brass (CA 836 or 838) and leaded red bronze (SAE 67) appear viable as pump internal parts. Based on a few tests at Class Va sites, aluminum bronzes have shown potential for corrosion in heavy-walled components (Ellis 1989).

Solder is yet another problem area for copper equipment. Lead-tin solder (50Pb, 50Sn) was observed to fail by dealloying after a few years’ exposure. Silver solder (1Ag, 7P, Cu) was completely removed from joints in under two years. If the designer elects to accept this risk, solders containing at least 70% tin should be used.

Stainless Steel. Unlike copper and cupronickel, stainless steels are not affected by traces of hydrogen sulfide. Their most likely application is heat exchange surfaces. For economic reasons, most heat exchangers are probably of the plate-and-frame type, most of which are fabricated with one of two standard alloys, Type 304 and Type 316 stainless steel. Some pump and valve trim also are fabricated from these or other stainless steels.

These alloys are subject to pitting and crevice corrosion above a threshold chloride level, which depends on the chromium and molybdenum content of the alloy and on the temperature of the geothermal fluid. Above this temperature, the passivation film, which gives the stainless steel its corrosion resistance, is ruptured, and local pitting and crevice corrosion occur. Figure 4 shows the relationship between temperature, chloride level, and occurrence of localized corrosion of Type 304 and Type 316 stainless steel. This figure indicates, for example, that localized corrosion of Type 304 may occur in 80°F geothermal fluid if the chloride level exceeds approximately 210 ppm; Type 316 is resistant at that temperature until the chloride level reaches approximately 510 ppm. Because of its 2 to 3% molybdenum content, Type 316 is always more resistant to chlorides than is Type 304.

Figure 4. Chloride Concentration Required to Produce Localized Corrosion of Stainless Steel as Function of Temperature
(Efrid and Moeller 1978)

Aluminum. Aluminum alloys are not acceptable in most cases because of catastrophic pitting.

Titanium. This material has extremely good corrosion resistance and could be used for heat exchanger plates in any low-temperature geothermal fluid, regardless of dissolved oxygen content. Great care is required if acid cleaning is to be performed. The vendor’s instructions must be followed. The titanium should not be scratched with iron or steel tools; this can cause pitting.

Chlorinated Polyvinyl Chloride (CPVC) and Fiber-Reinforced Plastic (FRP). These materials are easily fabricated and are not adversely affected by oxygen intrusion. External protection against groundwater is not required. The mechanical properties of these materials at higher temperatures may vary greatly from those at ambient temperature, and the materials’ mechanical limits should not be exceeded. The usual mode of failure is creep rupture: strength decays with time. Manufacturer’s directions for joining should be followed to avoid premature failure of joints.

Elastomeric Seals. Tests on O-ring materials in a low-temperature system in Texas indicated that a fluoroelastomer is the best material for piping of this nature; Buna-N is also acceptable (Ellis 1989). Neoprene, which developed extreme compression set, was a failure. Natural rubber and Buna-S should also be avoided. Ethylene-propylene terpolymer (EPDM) has been used successfully in gasket, O-ring, and valve seats in many systems. EPDM materials have swollen in some systems using oil-lubricated turbine pumps.

Production well pumps are among the most critical components in a geothermal system and have been the source of much system downtime. Therefore, proper selection and design of the production well pump is extremely important. Well pumps are available for larger systems in two general configurations: lineshaft and submersible. The lineshaft type is most often used for direct-use systems (Rafferty 1989b).

Lineshaft Pumps. Lineshaft pumps are similar to those typically used in irrigation applications. An aboveground driver, typically an electric motor, rotates a vertical shaft extending down the well to the pump. The shaft rotates pump impellers in the pump bowl assembly, which is positioned at such a depth in the wellbore that adequate net positive suction head pressure (NPSH) is available when the unit is operating. Two designs for the shaft/bearing portion of the pump are available: open and enclosed.

In the open lineshaft pump, the shaft bearings are supported in “spiders,” which are anchored to the pump column pipe at 5 to 10 ft intervals. The shaft and bearings are lubricated by the fluid flowing up the pump column. In geothermal applications, bearing materials for open lineshaft designs are typically elastomer compounds. The shaft material is typically stainless steel. Experience with this design in geothermal applications has been mixed. It appears that the open lineshaft design is most successful in applications with high (<50 ft) static water levels or flowing artesian conditions. Open lineshaft pumps are generally less expensive than enclosed lineshaft pumps for the same application.

In an enclosed lineshaft pump, an enclosing tube protects the shaft and bearings from exposure to the pumped fluid. A lubricating fluid is admitted to the enclosed tube at the wellhead. It flows down the tube, lubricates the bearings, and exits where the column attaches to the bowl assembly. The bowl shaft and bearings are lubricated by the pumped fluid. Oil-lubricated, enclosed lineshaft pumps have the longest service life in low-temperature, direct-use applications.

These pumps typically include carbon or stainless steel shafts and bronze bearings in the lineshaft assembly, and stainless steel shafts and leaded red bronze bearings in the bowl assembly. Keyed-type impeller connections (to the pump shaft) are superior to collet-type connections (Rafferty 1989b).

Because of the lineshaft bearings, lineshaft pump reliability decreases as pump-setting depth increases. Nichols (1978) indicates that, below about 800 ft, reliability is questionable, even under good pumping conditions.

Submersible Pumps. The electrical submersible pump consists of three primary components located downhole: the pump, the drive motor, and the motor protector. The pump is a vertical multistage centrifugal type. The motor is usually a three-phase induction type that is filled with oil for cooling and lubrication; it is cooled by heat transfer to the pumped fluid moving up the well. The motor protector is located between the pump and the motor and isolates the motor from the well fluid while allowing pressure equalization between the pump intake and the motor cavity.

The electrical submersible pump has several advantages over lineshaft pumps, particularly for wells requiring greater pump bowl setting depths. The deeper the well, the greater the economic advantage of the submersible pump. Moreover, it is more versatile, adapting more easily to different depths.

Submersible pumps have not demonstrated acceptable lifetimes in most geothermal applications. Although they are commonly used in high-temperature, downhole applications in the oil and gas industry, the acceptable overhaul interval in that industry is much shorter than in a geothermal application. In addition, most submersibles operate at 3600 rpm, resulting in greater susceptibility to erosion in aquifers that produce moderate amounts of sand. They have, however, been applied in geothermal projects where an existing well of relatively small diameter must be used. At 3600 rpm, they provide greater flow capacity for a given bowl size than an equivalent 1750 rpm lineshaft pump.

Standard cold-water submersible motors can be used at temperatures up to approximately 120°F with adequate precautions. These consist primarily of ensuring adequate water velocity past the motor (minimum 3 ft/s), which may require the use of a sleeve, and a small degree of motor oversizing (Franklin Electric 2001).

Well Pump Control. Well pumps serving variable loads are often controlled using variable-speed drives (VSDs). Submersible pumps can also be controlled using VSDs, but special precautions are required. Drive-rated motors are not commonly available for these applications, so external electronic protection should be used to prevent premature motor failure. In addition, the motor manufacturer must be aware that the motor will applied in a variable-speed application. Finally, because of the large static head in many well pump applications, controls should be configured to prevent the pump from operating at no-flow conditions.

Geothermal fluids can be isolated with large central heat exchangers, as in the case of a district heating system, or with exchangers at individual buildings or loads. In both cases, the principle is to isolate the geothermal fluid from complicated systems or those that cannot readily be designed to be compatible with the geothermal fluid. The main types of heat exchangers used in transferring energy from the geothermal fluid are plate and downhole.

Plate Heat Exchangers. For all but the very smallest applications, plate-and-frame heat exchangers are the most commonly used design. Available in corrosion-resistant materials, easily cleanable, and able to accommodate increased loads by adding plates, these exchangers are well suited to geothermal applications. Their high performance is also an asset in many system designs. Because geothermal resource temperatures are often less than those used in conventional hot-water heating system design, minimizing temperature loss at the heat exchanger is frequently a design issue. Approach temperatures of 5°F and less are common.

Materials for plate heat exchangers in direct-use applications normally include Buna-N or EPDM gaskets and 316 or titanium plates. Plate selection is often a function of temperature and chloride content of the water. For applications characterized by chloride contents of >50 ppm at 200°F, titanium would be used. At lower temperatures, much higher chloride exposure can be tolerated (see Figure 4).

Downhole Heat Exchangers. The downhole heat exchanger (DHE) is an arrangement of pipes or tubes suspended in a wellbore (Culver and Reistad 1978). A secondary fluid circulates from the load through the exchanger and back to the load in a closed loop. The primary advantage of a DHE is that only heat is extracted from the earth, which eliminates the need to dispose of spent fluids. Other advantages are the elimination of (1) well pumps with their initial operating and maintenance costs, (2) the potential for depletion of groundwater, and (3) environmental and institutional restrictions on surface disposal. One disadvantage of a DHE is the limited amount of heat that can be extracted from or rejected to the well. The amount of heat extracted depends on the hydraulic conductivity of the aquifer and well design. Because of the limitations of natural convection, only about 10% of the heat output of the well is available from a DHE in comparison to pumping and using surface heat exchange (Reistad et al. 1979). With wells of approximately 200°F and depths of 500 ft, output under favorable conditions is sufficient to serve the needs of up to five homes.

The DHE in low- to moderate-temperature geothermal wells is installed in a casing, as shown in Figure 5.

Downhole heat exchangers with higher outputs rely on water circulation within the well, whereas lower-output DHEs rely on earth conduction. Circulation in the well can be accomplished by two methods: (1) undersized casing and (2) convection tube. Both methods rely on the difference in density between the water surrounding the DHE and that in the aquifer.

Circulation provides the following advantages:

Figure 5. Typical Connection of Downhole Heat Exchanger for Space and Domestic Hot-Water Heating
(Reistad et al. 1979)

Figure 5 shows well construction in competent formation (i.e., where the wellbore will stand open without a casing). An undersized casing with perforations at the lowest producing zone (usually near the bottom) and just below the static water level is installed. A packer near the top of the competent formation allows installation of an annular seal between it and the surface. When the DHE is installed and heat extracted, thermosiphoning causes cooler water inside the casing to move to the bottom, and hotter water moves up the annulus outside the casing.

Because most DHEs are used for space heating (an intermittent operation), heated rocks in the upper portion of the well store heat for the next cycle.

Where the well will not stand open without casing, a convection tube can be used. This is a pipe one-half the diameter of the casing either hung with its lower end above the well bottom and its upper end below the surface or set on the bottom with perforations at the bottom and below the static water level. If a U-bend DHE is used, it can be either inside or outside the convection tube. DHEs operate best in aquifers with a high hydraulic conductivity and that provide water movement for heat and mass transfer.

In large (>2.5 in.) pipe sizes, resilient-lined butterfly valves are preferred for geothermal applications. The lining material protects the valve body from exposure to the geothermal fluid. The rotary rather than reciprocating motion of the stem makes the valve less susceptible to leakage and build-up of scale deposits. For many direct-use applications, these valves are composed of Buna-N or EPDM seats, stainless steel shafts, and bronze or stainless steel disks. Where oil-lubricated well pumps are used, a seat material of oil-resistant material is recommended. Gate valves have been used in some larger geothermal systems but have been subject to stem leakage and seizure. After several years of use, they are no longer capable of 100% shutoff.

Piping in geothermal systems can be divided into two broad groups: pipes used inside buildings and those used outside, typically buried. Indoor piping carrying geothermal water is usually limited to the mechanical room. Carbon steel with grooved end joining is the most common material.

For buried piping, many existing systems use some form of nonmetallic piping, particularly asbestos cement (which is no longer available) and glass fiber. With the cost of glass fiber for larger sizes (>6 in.) sometimes prohibitive, ductile iron is frequently used. Available in sizes >2 in., ductile iron offers several positive characteristics: low cost, familiarity to installation crews, and wide availability. It requires no allowances for thermal expansion if push-on fittings are used.

Most larger-diameter buried piping is preinsulated. The basic ductile iron pipe is surrounded by a layer of insulation (typically polyurethane), which is protected by an outer jacket of PVC or high-density polyethylene (HDPE).

Standard ductile iron used for municipal water systems is sometimes modified for geothermal use. The seal coat used to protect the cement lining of the pipe is not suitable for the temperature of most geothermal applications; in applications where the geothermal water is especially soft or low in pH, the cement lining should be omitted, as well. Special high-temperature gaskets (usually EPDM) are used in geothermal applications. Few problems have been encountered in using ferrous piping with low-temperature geothermal fluids unless high chloride concentration, low (<7.0) pH, or oxygen is present in the fluid. Most cases of corrosion failure have resulted from external attack by soil moisture in buried applications. For more information on piping systems for these applications, refer to Chapter 12 in the 2016 ASHRAE Handbook—HVAC Systems and Equipment and to ASHRAE (2013a, 2013b).

Figure 6 illustrates a system that uses geothermal fluid at 170°F (Austin 1978). The geothermal fluid is used in two main equipment components for heating the buildings: (1) a plate heat exchanger that supplies energy to a closed heating loop previously heated by a natural gas boiler (the boiler remains as a standby unit) and (2) a water-to-air coil used for preheating ventilation air. In this system, proper control is crucial for economical operation.

The average temperature of discharged fluid is 120 to 130°F. Geothermal fluid is used directly in the preheat coils in the buildings, which would probably not be the case if the system were designed today (Lienau 1979).

Figure 7 shows a geothermal district heating system that has a unique feature: its design is based on a peak load Δt of 100°F using a 190°F resource. It is of closed-loop design with central heat exchangers. The production well has an artesian shut-in pressure of 25 psi, so the system operates with no production pump for most of the year. During colder weather, a surface centrifugal pump located at the wellhead boosts the pressure.

Geothermal flow from the production well is initially controlled by a throttling valve on the supply line to the main heat exchanger, which responds to a temperature signal from the supply water on the closed-loop side of the heat exchanger. When the throttling valve reaches the full-open position, the production booster pump is enabled. The pump is controlled through a variable-speed drive that responds to the same supply-water signal as the throttling valve. The booster pump is designed for a peak flow rate of 300 gpm of 190°F water.

A few district heating systems have also been installed using an open distribution system. In this design, central heat exchangers (as in Figure 7) are eliminated and the geothermal water is delivered to individual building heat exchangers. When more than a few buildings are connected to the system, using central heat exchangers is normally more cost-effective.

Terminal equipment used in geothermal systems is the same as that used in nongeothermal heating systems. However, certain types of equipment are better suited to geothermal design than others.

In many cases, buildings heated by low-temperature geothermal sources operate at lower supply water temperatures than conventional hydronic designs. Because many geothermal sources are designed to take advantage of a large Δt, proper selection of equipment for low flow and low temperature is important.

Finned-coil, forced-air systems generally function best in this low-temperature/high-Δt situation. One or two additional rows of coil depth compensate for the lower supply water temperature. Although an increased Δt affects coil circuiting, it improves controllability. This type of system should be able to use a supply water temperature as low as 110°F.

Radiant floor panels are well suited to very low water temperatures, particularly in industrial applications with little or no floor covering. In industrial settings, with a bare floor and a relatively low space-temperature requirement, the average water temperature could be as low as 95°F. For a higher space temperature and/or thick floor coverings, a higher water temperature may be required.

Baseboard convectors and similar equipment are the least capable of operating at low supply-water temperature. At 150°F average water temperatures, derating factors for this design load may be affected. This type of equipment can be operated at low temperatures from the geothermal source to provide base-load heating, with peak load supplied by a conventional boiler. Ensure the boiler does not supply a higher fraction of the load than intended by the designer.

Figure 6. Heating System Schematic

Domestic water heating in a district space-heating system is beneficial because it increases the overall size of the energy load, energy demand density, and load factor. For those resources that cannot heat water to the required temperature, preheating is usually possible. Whenever possible, the domestic hot-water load should be placed in series with the space-heating load to reduce system flow rates and increase Δt.

Figure 7. Closed Geothermal District Heating System
(Rafferty 1989a)

Geothermal energy has seldom been used for cooling, although emphasis on solar energy and waste heat has created interest in cooling with thermal energy. The absorption cycle is most often used, and lithium bromide/water absorption machines are available in a wide range of capacities. Temperature and flow requirements for absorption chillers run counter to the general design philosophy for geothermal systems: they require high supply water temperatures and a small Δt on the hot-water side. Figure 8 illustrates the effect of reduced supply water temperature on machine performance. The machine is rated at a 240°F input temperature, so derating factors must be applied if the machine is operated below this temperature. For example, operation at a 200°F supply water temperature results in a 50% decrease in capacity, which seriously affects the economics of absorption cooling at a low resource temperature.

Coefficient of performance (COP) is less seriously affected by reduced supply water temperature. The nominal COP of a single-stage machine at 240°F is 0.65 to 0.70; that is, for each ton of cooling output, a heat input of 12,000 Btu/h divided by 0.65, or 18,460 Btu/h, is required.

Most absorption equipment is designed for steam input (an isothermal process) to the generator section. When this equipment is operated from a hot-water source, a relatively small Δt must be used. This creates a mismatch between building flow requirements for space heating and cooling. For example, assume a 200,000 ft² building is to use a geothermal resource for heating and cooling. At 25 Btu/h · ft² and a design Δt of 40°F, the flow requirement for heating is 250 gpm. At 30 Btu/h · ft², a Δt of 15°F, and a COP of 0.65, the flow requirement for cooling is 1230 gpm.

Figure 8. Typical Lithium Bromide Absorption Chiller Performance Versus Temperature
(Christen 1977)

Some small-capacity (3 to 25 ton) absorption equipment has been optimized for low-temperature operation in conjunction with solar heat. Although this equipment could be applied to geothermal resources, the prospects are questionable. Small absorption equipment generally competes with packaged direct-expansion units in this range; absorption equipment requires a great deal more mechanical auxiliary equipment for a given capacity. The cost of the chilled-water piping, pump, and coil; cooling-water piping, pump, and tower; and hot-water piping raises the capital cost of the absorption equipment substantially. Only in large sizes (>10 tons) and in areas with high electric rates and high cooling requirements (>2000 full-load hours) would this type of equipment offer an attractive investment to the owner (Rafferty 1989a).

The GCHP is a subset of the GSHP and is often called a closed-loop heat pump. A GCHP system consists of a reversible vapor compression cycle that is linked to a closed ground heat exchanger (also called a ground loop) buried in soil (Figure 9). The most widely used unit is a water-to-air heat pump, which circulates a water or a water/antifreeze solution through a liquid-to-refrigerant heat exchanger and a buried thermoplastic piping network. Heat pump units often include desuperheater heat exchangers (shown on the left in Figure 9). These devices use hot refrigerant at the compressor outlet to heat water. A second type of GCHP is the direct-expansion (DX) GCHP, which uses a buried copper piping network through which refrigerant is circulated.

The GCHP is further subdivided according to ground heat exchanger design: vertical and horizontal. Vertical GCHPs (Figure 10) generally consist of two small-diameter, high-density polyethylene (HDPE) tubes placed in a vertical borehole that is subsequently filled with a solid medium. The tubes are thermally fused at the bottom of the bore to a close return U-bend. Vertical tubes range from 0.75 to 1.5 in. nominal diameter. Bore depths normally range from 50 to 400 ft depending on local drilling conditions and available equipment, but can go to 600 ft or more if procedures for deep boreholes are followed (see the section on Pump and Piping System Options). Boreholes are typically 4 to 6 in. in diameter.

To reduce thermal interference between individual bores, a minimum borehole separation distance of 20 ft is recommended when loops are placed in a grid pattern. This distance may be reduced when bores are placed in a single row, the annual ground load is balanced (i.e., energy released in the ground is approximately equal to the energy extracted on an annual basis), or water movement or evaporation and subsequent recharge mitigates the effect of heat build-up in the loop field.

Figure 9. Vertical Closed-Loop Ground-Coupled Heat Pump System
(Kavanaugh 1985)

Advantages of the vertical GCHP are that it (1) requires relatively small plots of ground, (2) is in contact with soil that varies very little in temperature and thermal properties, (3) requires the smallest amount of pipe and pumping energy, and (4) can yield the most efficient GCHP system performance. Disadvantages are (1) typically higher cost because of expensive equipment needed to drill the borehole and (2) the limited availability of contractors to perform such work.

Hybrid systems are a variation of ground-coupled systems in which a smaller ground heat exchanger is used, augmented in cooling mode by a fluid cooler or a cooling tower. This approach can have merit in large cooling-dominated applications. The ground heat exchanger is sized to meet the heating requirements. The downsized loop is used in conjunction with the fluid cooler or cooling tower with an isolation heat exchanger to meet the heat rejection load. Using the cooler reduces the capital cost of the ground heat exchanger in such applications, but somewhat increases maintenance requirements. For heavily heating-dominant applications, a downsized loop also can be augmented with an auxiliary heat source such as electric resistance, solar collectors, or fossil fuel.

Horizontal GCHPs (Figure 11) can be divided into several subgroups, including single-pipe, multiple-pipe, spiral, and horizontally bored. Single-pipe horizontal GCHPs were initially placed in narrow trenches at least 4 ft deep. These designs require the greatest amount of ground area. Multiple pipes (usually two, four, or six), placed in a single trench, can reduce the amount of required ground area. Trench length is reduced with multiple-pipe GCHPs, but total pipe length must be increased to overcome thermal interference from adjacent pipes. The spiral coil is reported to further reduce required ground area. These horizontal ground heat exchangers are made by stretching small-diameter polyethylene tubing from the tight coil in which it is shipped into an extended coil that can be placed vertically in a narrow trench or laid flat at the bottom of a wide trench. Recommended trench lengths are much shorter than those of single-pipe horizontal GCHPs, but pipe lengths must be much longer to achieve equivalent thermal performance. When horizontally bored loops are grouted and placed in the deep earth, as shown in the bottom of Figure 11, design lengths are near those for vertical systems, because annual temperature and moisture content variations approach deep-earth values.

Figure 10. Vertical Ground-Coupled Heat Pump Piping

Advantages of horizontal GCHPs are that (1) they are typically less expensive than vertical GCHPs because relatively low-cost installation equipment is widely available, (2) many residential applications have adequate ground area, and (3) trained equipment operators are more widely available. Disadvantages include, in addition to a larger ground area requirement, (1) greater adverse variations in performance because ground temperatures and thermal properties fluctuate with season, rainfall, and burial depth; (2) slightly higher pumping-energy requirements; and (3) lower system efficiencies. OSU (1988a, 1988b), Remund and Carda (2014), and Svec (1990) cover the design and installation of horizontal GCHPs.

The second subset of GSHPs is groundwater heat pumps (Figure 12). Until the development of GCHPs, they were the most widely used type of GSHP. In the commercial sector, GWHPs can be an attractive alternative because large quantities of water can be delivered from and returned to relatively inexpensive wells that require very little ground area. Whereas the cost per unit capacity of the ground heat exchanger is relatively constant for GCHPs, the cost per unit capacity of a well water system is much lower for a large GWHP system. A single pair of high-volume wells can serve an entire building. Properly designed groundwater loops with correctly developed water wells require no more maintenance than conventional air and water central HVAC. When groundwater is injected back into the aquifer by a second well, net water use is zero.

One widely used design places a central water-to-water heat exchanger between the groundwater and a closed water loop, which is connected to water-to-air heat pumps in the building. A second possibility is to circulate groundwater through a heat recovery chiller (isolated with a heat exchanger), and to heat and cool the building with a distributed hydronic loop.

Figure 11. Trenched Horizontal (top) and Horizontally Bored (bottom) Ground-Coupled Heat Pump Piping

Figure 12. Unitary Groundwater Heat Pump System

Both types and other variations may be suited for direct preconditioning in much of the United States. Groundwater below 60°F can be circulated directly through hydronic coils in series or in parallel with heat pumps. The cool groundwater can displace a large amount of energy that would otherwise have to be generated by mechanical refrigeration.

Advantages of GWHPs under suitable conditions are (1) they cost less than GCHP equipment, (2) the space required for the water well is very compact, (3) water well contractors are widely available, and (4) the technology has been used for decades in some of the largest commercial systems.

Disadvantages are that (1) local environmental regulations may be restrictive, (2) water availability may be limited, (3) fouling precautions may be necessary if groundwater is used directly in the heat pumps and water quality is poor, and (4) pumping energy may be high if the system is poorly designed or draws from a deep aquifer.

Surface water heat pumps are included as a subset of GSHPs because of the similarities in applications and installation methods. SWHPs can be either closed-loop systems similar to GCHPs or open-loop systems similar to GWHPs. However, the thermal characteristics of surface water bodies are quite different than those of the ground or groundwater. Some unique applications are possible, though special precautions may be warranted.

Closed-loop SWHPs (Figures 13 and 34) consist of water-to-air or water-to-water heat pumps connected to a piping network (also called a surface water loop) placed in a lake, river, or other open body of water. A pump circulates water or a water/antifreeze solution through the heat pump water-to-refrigerant heat exchanger and the submerged piping loop, which transfers heat to or from the body of water. The recommended piping material is thermally fused HDPE tubing with ultraviolet (UV) radiation protection.

Advantages of closed-loop SWHPs are (1) relatively low cost (compared to GCHPs) because of reduced excavation costs, (2) low pumping-energy requirements, (3) low maintenance requirements, and (4) low operating cost. Disadvantages are (1) the possibility of coil damage in public lakes and (2) wide variation in water temperature with outdoor conditions if lakes are small and/or shallow. Such variation in water temperature would cause undesirable variations in efficiency and capacity, though not as severe as with air-source heat pumps.

Figure 13. Lake Loop Piping

Open-loop SWHPs can use surface water bodies the way cooling towers are used, but without the need for fan energy or frequent maintenance. In warm climates, lakes can also serve as heat sources during winter heating mode, but in colder climates where water temperatures drop below 45°F, closed-loop systems are the only viable option for heating.

Lake water can be pumped directly to water-to-air or water-to-water heat pumps or through an intermediate heat exchanger that is connected to the units with a closed piping loop. Direct systems tend to be smaller, having only a few heat pumps. In deep lakes (40 ft or more), there is often enough thermal stratification throughout the year that direct cooling or precooling is possible. Water can be pumped from the bottom of deep lakes through a coil in the return air duct. Total cooling is possible if water is 50°F or below. Precooling is possible with warmer water, which can then be circulated through the heat pump units. Large-scale cooling-only systems have been deployed successfully in some locations, including Cornell University and the city of Toronto [Cornell University 2006; Enwave (no date)].

Site characteristics influence the type of GSHP system most suitable for a particular location. Site characterization is the evaluation of a site’s geology and hydrogeology with respect to its effect on GSHP system design. Important issues include presence or absence of water, depth to water, water (or soil/rock) temperature, groundwater quality, depth to rock, rock type, and the nature and thickness of unconsolidated materials overlying the rock. Information about the nature of water resources at the site helps to determine whether an open-loop system may be possible. Depth to water affects pumping energy for an open-loop system and possibly the type of rig used for drilling closed-loop boreholes. Groundwater temperature in most locations is the same as the undisturbed ground temperature. These temperatures are key inputs to the design of GSHP systems. The types of soil and rock allow a preliminary evaluation of the range of thermal conductivity/diffusivity that might be expected. The thickness and nature of the unconsolidated (soil, gravel, sand, clay, etc.) materials overlying the rock influence whether casing is required in the upper portion of boreholes for closed-loop systems, a factor which increases drilling cost.

After the GSHP system type has been decided, specific details about the subsurface materials’ (rock/soil) thermal conductivity and diffusivity, water well static and pumping levels, drawdown, etc., are necessary to design the system. Ways to obtain these more detailed data are discussed in other parts of this chapter.

There are many sources for gathering site characterization information, such as geologic and hydrologic maps, state geology and water regulatory agencies, the U.S. Geological Survey (USGS 2000), and any information that may be available from geotechnical studies of the site. Among the best sources of information are completion reports for nearby water wells. These reports are filed by the driller upon completion of a water well and provide a great deal of information of interest for both open- and closed-loop designs. The most thorough versions of well completion reports (level of detail varies by state) cover all of the issues of interest listed at the beginning of this section. Information about access to and interpretation of these reports and other sources of information for site characterization is included in Rafferty (2000a) and Sachs (2002).

Once the type of system has been selected, more site-specific tests (ground thermal properties test for GCHP or well flow test for GWHP) can be used to determine the parameters necessary for system design. In many areas, ground heat exchangers are regulated by the state or other jurisdictions and under the jurisdiction of a state water rights authority, department of natural resources or environmental quality, or possibly a federal agency such as the U.S. Army Corps of Engineers. The regulation scope may include any type of ground-coupled system. The engineer or designer should be aware of and versed in regulatory issues affecting the project site. More recently, the U.S. Department of Homeland Security has required that certain activities, especially those close to drinking water sources, be excluded or stringently regulated. For security reasons, resource protection zones may not be found in public records archives. The simplest solution may be to contact the permitting authority.

The design phase of GSHP commissioning requires a thorough site survey and characterization, accurate load modeling, and ensuring that the design chosen (and its documentation) meets the design intent.

The construction phase is dominated by observation of installation and verification of prefunctional checks and tests. It also involves planning, training development, and other activities to help future building operators understand the HVAC system.

The acceptance phase starts with functional tests and verification of all test results. It continues with full documentation: completing the commission report to include records of design changes and all as-built plans and documents, and completing the operations and maintenance manual and system manual. Finally, after system testing and balancing is complete, the owner’s operating staff are trained. The acceptance phase ends when “substantial completion” is reached. The warranty period begins from this date.

Table 3 provides information on tasks and participants involved in the GSHP commissioning process. Additional details on this topic, along with preventive maintenance and troubleshooting information, are included in Caneta Research (2001).

This section provides an overview of a suggested design procedure for vertical, ground-coupled systems; related information and equations are discussed in more detail in Kavanaugh and Rafferty (1997). Several public software programs are available for performing the repetitive computations necessary for system optimization. Shonder et al. (1999, 2000) tested the accuracy of these programs, and agreement was attained with several programs in subsequent evaluations.

A more recent publication (Kavanaugh 2008) updates the design recommendations for GCHP systems as follows:

Deliverables from this process that are necessary to adequately describe a GCHP installation include, as a minimum,

Thermal Property Testing. In the design of vertical GCHPs, accurate knowledge of soil/rock formation thermal properties is critical. These properties can be estimated in the field by installing a loop of approximately the same size and depth as the heat exchangers planned for the site. The test loop location should be chosen with care, and designed to be used for the eventual full borefield, especially if geothermal is a likely final system choice (this may require the test loop to meet all local ground heat exchanger standards). Heat is added in a water loop at a constant rate, and data are collected as shown in Figure 14. Inverse methods are applied to find thermal conductivity, diffusivity, and temperature of the formation. These methods are based on the either the line source (Gehlin 1998; Mogensen 1983; Witte et al. 2002), the cylindrical heat source (Ingersoll and Zobel 1954), or a numerical algorithm (Austin et al. 2000; Shonder and Beck 1999; Spitler et al. 1999). More than one of these methods should be applied, when possible, to enhance reported accuracy. Recommended test specifications are as follows (Kavanaugh 2000, 2001):

Figure 14. Thermal Properties Test Apparatus

Ground Heat Exchanger Sizing. This is perhaps the most critical step in design of a vertical GCHP. Ground-loop design methods must proceed with limited information; a major missing component is long-term, field-monitored data, which are needed to further validate the design method to address effects of water movement and long-term heat storage more fully. The conservative designer can assume no benefit from water movement; designers who assume maximum benefit must ignore annual imbalances in heat rejection and absorption.

Two design methods are presented in the following section. Both methods have been implemented in design software tools, and are based on the assumption that heat transfer in the ground is governed by conduction only. The first method is based on the solution of the equation for heat transfer from a cylinder buried in the earth. This equation was developed and evaluated by Carslaw and Jaeger (1947) and was suggested by Ingersoll and Zobel (1954) as an appropriate method of sizing ground heat exchangers. Kavanaugh (1985) adjusted the method to account for the U-bend arrangement and hourly heat rate variations. Alternative design methods are described by Eskilson (1987), Morrison (1997), Philippe et al. (2010), Spitler (2000), and Spitler et al. (2000). A second method, attributed to Eskilson, is presented after this first method.

Vertical Ground Heat Exchanger Sizing. The method of Ingersoll and Zobel (1954), based on the following steady-state heat transfer equation, can be used to size vertical ground heat exchangers:

(1)

where

The heat rate delivered to the ground in the cooling mode by the condenser includes the heat of the heat pump and auxiliary equipment. Thus, q_cond can be calculated to be

(2)

where

However, the heat of the heat pump and auxiliary equipment in the heating mode is delivered to the building. Thus the heat removed from the ground by the evaporator is

(3)

where q_evap is the heat pump evaporator heat extraction rate from ground, and q_lh is the building design heating block load, both of which are in Btu/h.

The equation is rearranged to solve for the required bore length L. The steady-state equation is modified to represent the variable heat rate of a ground heat exchanger by using a series of constant-heat-rate “pulses.” Thermal resistance of the ground per unit length is calculated as a function of time corresponding to the time span over which a particular heat pulse occurs. The term R_b is also included to account for thermal resistance of the borehole, which includes the thermal resistance of the pipe wall and interfaces between the pipe and fluid and the pipe and the ground. The resulting equation takes the following form for required length to satisfy cooling loads:

(4)

The required length for heating is

(5)

where

Note: Heat transfer rate, building loads, and temperature penalties are positive for heating and negative for cooling.

Equations (4) and (5) consider three different pulses of heat to account for long-term heat imbalances, average monthly heat rates during the design month, and maximum heat rates for a short-term period during a design day. This period could be as short as 1 h, but a 4 to 6 h block is recommended.

The required total bore length is the larger of the two lengths L_c and L_h calculated with Equations (4) and (5). If L_c is larger than L_h, an oversized coil could be beneficial during the heating season. A second option is to install the smaller heating length along with a cooling tower to compensate for the undersized coil (see the section on Hybrid System Design). If L_h is larger, the designer should install this length, and during cooling mode the efficiency benefits of an oversized ground coil could be used to compensate for the higher first cost.

Thermal resistance of the ground is calculated from ground properties, pipe dimensions, and operating periods of the representative heat rate pulses. Table 4 lists typical thermal properties for soils and fills for the annular region of the boreholes. Type of fill material depends on thermal, regulatory, and economic considerations. Historically, a relatively low-thermal-conductivity bentonite grout common in the water well industry had been used. More recently, thermally enhanced materials have been developed to supplement or replace conventional grouts. In some cases, drill cuttings or a manufactured sand/gravel mix have been placed instead. Note that placing such fill material from the surface may cause the borehole to bridge, leaving voids or ungrouted sections of the borehole, so using a tremie pipe to inject fill from the bottom upward is recommended. Additionally, using any pumped grout in a formation with large voids, karst, or caverns can also result in voids through grout settling or flowing out of the borehole. Nutter et al. (2001) contains a detailed evaluation of potential fills and grouts for vertical boreholes. Also, Jenkins (2009), Sachs (2002), and Skouby (2011) have additional recommendations regarding grout and grout placement.

Table 5 summarizes potential completion methods for various geological conditions. “Two-fill” refers to the practice of placing a low-permeability material in the upper portion of the hole and/or in intervals where it is required to separate individual aquifers, and a more thermally advantageous material in the remaining intervals.

The thermal resistance of the ground heat exchanger vertical bore considers the effects of pipe resistance R_p and bore annulus grout resistance R_grt.

(6)

Pipe resistance includes the convective film resistance of the fluid and the conductive resistance of the pipe walls. Contact resistances are negligible compared to the high resistance of plastic pipe walls and annular grouts. For a single U-tube (two tubes) the pipe resistance is

(7)

where h_conv is the convection coefficient inside the pipes and k_p is pipe thermal conductivity.

A correlation for the grout’s thermal resistance has been developed using shape factor correlations (Remund 1999):

(8)

where k_grt is grout thermal conductivity. Coefficients β₀ and β₁ in Equation (8) have been developed for three locations of the tubes, as shown in Figure 15: centered in the bore and in contact each other (A), centered and spaced evenly in the bore (B), and centered and in contact with the bore wall (C). However, the most likely location of the U-tubes is BC, and coefficients for this location are unavailable. A similar but slightly more detailed solution has been developed by Hellström (1991) and applied to a few design and simulation tools (Liu 2008; Philippe et. al. 2010).

Because locations of U-tubes cannot be determined even when spacers are installed, exact computation of bore resistance values is somewhat uncertain. It is possible to apply the results from thermal property tests to calculate the bore resistance if the U-tube dimensions, grout conductivity and borehole diameter are known (Kavanaugh 2010). Thermal property tests were conducted at 15 installations where these values were known and the bore resistance was calculated. The field calculated bore resistances best matched the values computed with Equations (6) to (8) using

Table 6 provides the bore resistances computed using Equations (6) to (8) for three different grout conductivities, three different fluid flow regimes (laminar, transition, and fully turbulent), three different U-tubes sizes, and three different bore diameters for location B, and location C. The resistance is also computed a double U-tube in a bore. These values were calculated with a value of k_p equal to 0.24 Btu/h · ft · °F. Designers can choose to use the values of location B (conservative), BC (average), C (risky), or Double.

The most difficult parameters to evaluate in Equations (4) and (5) are the equivalent thermal resistances of the ground. The solutions of Carslaw and Jaeger (1947) require that the time of operation, bore diameter, and thermal diffusivity of the ground be related in the dimensionless Fourier number (Fo):

(9)

where

The method may be modified to permit calculation of equivalent thermal resistances for varying heat pulses. A system can be modeled by three heat pulses: a 10 year (3650 day) pulse of q_a, a 1 month (30 day) pulse of q_m, and a 4 h (0.167 day) pulse of q_st. Three times are defined as

The Fourier number is then computed with the following values:

An intermediate step in computing the ground’s thermal resistance using the methods of Ingersoll and Zobel (1954) is to identify a G-factor, which is then determined from Figure 16 for each Fourier value. [The algorithm proposed by Cooper (1976) provides an alternative to the use of Figure 15.]

(10)

(11)

(12)

Correlations for the values of R_ga, R_gm, and R_gst were presented by Philippe et al. (2010) for a wide range of conditions.

Ranges of the ground thermal conductivity k_g are given in Table 6. State geological surveys are a good source of soil and rock data. However, geotechnical site surveys are highly recommended to determine load soil, rock types, and drilling conditions.

Figure 15. Coefficients for Equation (8)

Performance degrades somewhat because of short-circuiting heat losses between the upward- and downward-flowing legs of a conventional U-bend loop. This degradation can be accounted for by introducing the short-circuit heat loss factor [F_sc in Equations (4) and (5)] in the following table. Normally U-tubes are piped in parallel to the supply and return headers. Occasionally, when bore depths are shallow, two or three loops can be piped in series. In these cases, short-circuit heat loss is reduced; thus, the values for F_sc are smaller than that for a single bore piped in parallel.

Figure 16. Fourier/G-Factor Graph for Ground Thermal Resistance
(Kavanaugh and Rafferty 1997)

The remaining terms in Equations (4) and (5) are temperatures. The local deep-ground temperature t_g can best be obtained from local water well logs and geological surveys. A second, less accurate source is a temperature contour map, similar to Figure 17, prepared by state geological surveys. A third source, which can yield ground temperatures within 4°F, is a map with contours, such as Figure 18. Comparison of Figures 17 and 18 indicates the complex variations that would not be accounted for without detailed contour maps.

Selecting the temperature t_wi of water entering the unit is critical in the design process. Choosing a value close to ground temperature results in higher system efficiency, but makes the required ground coil length very long and thus unreasonably expensive. Choosing a value far from t_g allows selection of a small, inexpensive ground coil, but the system’s heat pumps will have both greatly reduced capacity and increased electric demand. Selecting t_wi to be 20 to 30°F higher than t_g in cooling and 10 to 20°F lower than t_g in heating is a good compromise between first cost and efficiency in many regions of the United States.

A final temperature to consider is the temperature penalty t_p resulting from thermal interferences from adjacent bores. The designer must select a reasonable separation distance to minimize required land area without causing large increases in the required bore length (L_c, L_h). Table 7 presents the temperature penalty for a 10 by 10 vertical grid of bores for various operating conditions after 10 years of operation in a nonporous soil where cooling effects from moisture evaporation or water movement do not mitigate temperature change (Kavanaugh 2003; Kavanaugh and Rafferty 1997). Correction factors are included to find the temperature penalty for four other grid patterns. Note that the higher the number of internal bores, the larger the correction factor.

Figure 17. Water and Ground Temperatures in Alabama at 50 to 100 ft Depth
(Chandler 1987)

In the table, adjustments are made to the number of equivalent full-load hours (EFLHs) in cooling and heating that correspond to values consistent with the local ground temperature. To mitigate long-term heat build-up for small separation distances, the required heat exchanger length is extended to maintain good system efficiencies. Larger separation distances result in shorter required lengths and smaller temperature changes, because there is greater thermal capacity available and greater area to diffuse heat to the far field.

The table applies only to a limited number of specific cases and is not intended for application to actual designs. It is intended to demonstrate trends for various ground temperatures, hours of operation in heating and cooling, and bore separation distances. Note that values of t_p in the table are significantly different from those obtained using the approach presented by Bernier et al. (2008).

Smaller bore lengths per ton of peak block load result in larger temperature changes; the relationship between bore length and temperature change is inverse and linear.

Table 7 applies only to a limited number of specific cases and is not intended for application to actual designs. It is intended to demonstrate trends for various ground temperatures, hours of operation in heating and cooling, and bore separation distances. Note that values of t_p in the table are significantly different from those obtained using the approach presented by Bernier et al. (2008). Values represent worst-case scenarios, and the temperature change is usually mitigated by groundwater recharge (vertical flow), groundwater movement (horizontal flow), and evaporation (and condensation) of water in the soil.

Groundwater movement strongly affects the long-term temperature change in a densely packed bore field (Chiasson et al. 2000a). A related factor is the evaporative cooling effect experienced with heat addition to the ground. Although thermal conductivity is somewhat reduced with lower moisture content (see Table 5), the net effect is beneficial in porous soils when water movement recharges the ground to original moisture levels. A similar effect may be experienced in cold climates when soil moisture freezes and the heat of solidification mitigates excessive temperature decline. Because these effects have not been thoroughly studied, the design engineer must establish a range of design lengths between one based on minimal groundwater movement, as in very tight clay soils with poor percolation rates, and a second based on the higher rates characteristic of porous aquifers.

Kavanaugh and Kavanaugh (2012) examined ground heat exchanger performance in 40 commercial buildings with vertical ground heat exchangers and between 5 and 25 years of operation. They calculated maximum approach temperature (difference between average loop temperatures t_wi and t_wo and initial ground temperature t_g) for all of the buildings; higher approach temperatures as years of operation increased would indicate an increase in ground temperature and raise concern about the expected life of ground heat exchangers with imbalanced cooling loads compared to heating loads.

In fact, the data suggested that older GSHP systems had lower approach temperatures. Results were not adjusted for many important factors such as vertical bore length, ground thermal properties, and vertical bore separation distance. Newer systems tended to have slightly shorter ground heat exchangers, but this was offset somewhat by the older systems’ tendency to have smaller vertical bore separation distances and lower conductivity grout and fill. Of the loops with the largest approach, three of the newer systems had vertical bore lengths less than 120 ft/ton. Two systems with long loops but large approach temperatures had low thermal conductivity grout (0.38 Btu/h · ft · °F), 15 ft bore separation, and indoor air temperatures below 70°F.

Figure 18. Approximate Groundwater Temperature (°C) in the Continental United States

The study’s data set is small, and significant long-term temperature change cannot be excluded at this point. Although much more field study is desirable, the absence of any significant trend of increased ground temperature (noted by elevation of maximum approach temperature) with increased years of GSHP operation suggests that long-term ground temperature change is not prevalent in properly designed GSHPs.

Results from this project cannot be applied to long-term temperature decline in which the amount of heat removed from the ground in heating far exceeds the heat rejected in cooling. In cold climates the heat capacity available at the freeze point of water is significant but the impact on grout thermal and physical properties also needs further field study.

The long-term temperature change table and a commonly used software for GCHP system design both use EFLH calculations. To the extent that annual loads are proportional to peak loads, the equivalent full-load hours method provides a simple estimate of annual loads from peak loads. The EFLHs in Table 8 provide a quick means to estimate annual loads needed to size ground heat exchangers at the initial feasibility study phase of a project. Because EFLHs vary with changes in both annual loads and peak loads, not all building parameters effects are included in EFLHs. For instance, building operating hours change annual loads by increasing the amount of time that internal gains are at elevated levels, but they do not change the peak load. Occupancy hours can add load without increasing the installed capacity, thereby changing the EFLHs. Furthermore, changes in other parameters, such as internal gains, do not necessarily scale with system capacity in the same proportion as annual load, again leading to changing EFLHs. Potential users of EFLHs must understand these sources of variability to use them effectively (Carlson 2001).

Example 1.

Size a vertical ground-coupled heat pump system for a six-zone classroom addition in Atlanta, GA. The addition has a peak cooling (block) load q_lc of 247,000 Btu/h and a peak heating (block) load q_lh of 160,000 Btu/h. The design monthly part-load factor PLF_m is 0.28.

Ground temperature t_g = 65°F

Ground conductivity k_g and diffusivity α_g = 1.4 Btu/h · ft · °F, 1.0 ft²/day

Bore fill conductivity k_b = 1.0 Btu/h · ft · °F

Vertical U-tube = 1.0 in. nominal, DR11, HDPE, 5 in. borehole diameter

2 × 10 grid (20 vertical bores) with 20 ft separation

Heat pump inlet and outlet temperatures t_wi and t_wo = 85 and 95°F

Heat pump cooling and heating efficiency (EER, COP) = 14.9 Btu/Wh, 4.4

10 year (3650 day), 1 month (30 day), and 4 h (0.167 day) heat pulse analysis:

EFLH_c = 760 h, EFLH_h = 245 h (Table 8, Average Atlanta values for school)

R_b = 0.185 h · ft · °F/Btu (average of Locations B and C, k_grout = 1.0 Btu/h · ft · °F, transition flow).

The required total bore length, assuming no long-term ground temperature change (t_p = 0) caused by moisture evaporation and subsequent groundwater recharge is

For nonporous soils, a 2 × 10 grid on 20 ft bore centers with EFLH_c = 1000 h and EFLH_h = 500 h (i.e., the example building operates 515 h more in cooling than in heating), the temperature penalty from Table 7 is

Repeat the process using Equation (5) to find the bore length for heating L_h. The design bore length is the larger value of L_c and L_h.

Alternative Sizing Method. The preceding traditional design method has several advantages. For instance, the effective ground thermal resistances R_ga, R_gm, and R_gst and the temperature penalty t_p are assumed to be independent of borehole length. This implies that the design length can be determined by solving a relatively simple equation. In contrast, the following method, used in some software design tools, is more complex and involves an iterative process to evaluate the design length.

Figure 19. Typical g-Function Curves for 3 × 2 Bore Field

Thermal response factors, also known as g-functions (not to be confused with the G-factor) may be used as an alternative to calculate the ground thermal resistance required in Equations (4) and (5). The concept was first introduced by Eskilson (1987) and extended to short-time steps by Yavuzturk and Spitler (1999).The g-functions give a relation between the heat extracted (or rejected) from the ground per unit borehole length q_L and the borehole wall temperature T_b. The borehole wall temperature is given by

(13)

where g represents the g-function. As shown in Equation (13), which depend on three non-dimensional parameters: B/H, the ratio of the borehole spacing over the borehole length; r_b/H, the ratio of the borehole radius over the borehole length; and t/t_s, a nondimensional time where t_s is a characteristic time (= H²/9α_g). Typical g-functions curves are presented in Figure 19 for a 3 × 2 bore field.

The g-function curves are presented graphically in Figure 19 as a function of ln (t/t_s) for six bore field spacings (B/H) and for a value of r_b/H = 0.0005. The curve for B/H= ∞ corresponds to the g-function of a single borehole. One of the major advantages of these nondimensional curves is that they apply to any 3 × 2 bore field.

Eskilson (1987) provides g-function curves for a number of bore field geometries. Design software tools that use the g-function concept have a relatively large data set of g-function curves to choose from. Eskilson (1987) calculated g-functions using two-dimensional transient finite-difference equations on a radial-axial coordinate system for a single borehole in homogeneous ground. The temperature fields from a single borehole were superimposed in space to obtain the response from a borehole field with a certain configuration.

The g-functions can be used to determine the design length of a bore field. One possible approach is to use Equations (4) and (5) but with two modifications. First, when g-functions are used, thermal interference among boreholes is implicitly accounted for and t_p can be eliminated. Second, the values of R_ga, R_gm, and R_gst are now based on g-functions. Hence, Equations (10) to (12) take the following forms:

(14)

(15)

(16)

where g_{(ts – ty)} is the g-function evaluated at ln[(t_x – t_y)/t_s] for a given bore field and B/H ratio. Note that determining L (i.e., n_b × H, where n_b is the number of boreholes) is an iterative process because R_ga, R_gm, and R_gst depend on H, which is unknown beforehand. Thus, software tools are often required to accomplish this task. Because g-functions account for 3D heat transfer in the borefield, they are considered to be more accurate than the G-factors, which derive from a radial-only heat transfer model.

Example 3.

A building has a cooling block load of 15 tons with a corresponding value of q_cond = −225,200 Btu/h. The annual ground imbalance q_a = −10,236 Btu/h, PLF_m = 0.30, and F_SC = 1.0. The 3 × 2 bore field has the following characteristics: r_b = 2 in., B = 16.4 ft, and R_b = 0.173 h · ft · °F/Btu. The undisturbed ground temperature is 50°F, the thermal conductivity is 1.93 Btu/h · ft · °F, and the thermal diffusivity is 1.04 ft²/day × 10⁻⁶. Calculate the equivalent thermal resistances R_ga, R_gm, and R_gst for three consecutive heat pulses of 10 years, 1 month, and 6 hours and the total required length if the maximum mean fluid temperature in the borehole is to be kept below 95°F.

From the problem statement, t_f = 3680.25 days, t₂ = 3680 days, and t₁ = 3650 days. After iterations, this leads to g_(tf) = 12.34, g_{(tf – t1)} = 3.99, and g_{(tf – t2)} = 1.55; and R_ga = 0.689 h · ft · °F/Btu, R_gm = 0.201 h · ft · °F/Btu, and R_gst = 0.128 h · ft · °F/Btu; and

Thus, 321 ft per bore is required with a borehole spacing of 16.4 ft. This represents a length of 131 ft per ton.

After the design length has been determined, it is often necessary to evaluate the outlet fluid temperature of a bore field as a function of time, generally on an hourly basis, and estimate the annual heat pump energy consumption. Energy simulation can be used to compute this temperature (they can also be used iteratively to assist in sizing the ground heat exchanger). Some energy simulation programs use the duct ground storage (DST) model introduced by Hellström (1989) to evaluate the outlet fluid temperature of a bore field as a function of time. Yavuzturk and Spitler (1999) describe the calculation method behind the DST model.

The DST model calculates the transient thermal process in densely packed borehole fields. The boreholes are assumed to be evenly distributed within a cylindrical storage region in the ground. Although the DST model was originally intended to simulate borehole thermal energy storage (BTES) systems, it has been used to simulate ground source heat pump systems.

Other energy simulation programs have a g-function-based routine to evaluate the outlet fluid temperature of a bore field as a function of time (Fisher et al. 2006; Liu 2008). The following analysis is intended to give only the salient features of an hourly simulation based on g-functions. As an example, assuming that F_SC = 1 and that the borehole length and the inlet fluid temperature are known, and that the heat transfer rates for three consecutive time intervals (0 to t₁, t₁ to t₂, and t₂ to t₃) are given by Q₁, Q₂, and Q₃, then, using temporal superposition, the mean fluid temperature at the end of the third time interval is given by

(17)

with T_m = (T_wi + T_wo)/2.

Based on the work of Yavuzturk and Spitler (1999), Equation (17) can be generalized for n time steps as follows:

(18)

Solving Equation (18) can be computationally intensive if the number of time steps is large, because there is no recurrence in the summation term. In other words, the calculations performed at time step n – 1 cannot be used at time step n, and the ground loop loading history must be updated at each time step. Load aggregation is typically used to reduce the number of terms in the summation without sacrificing accuracy. It is based on the fact that recent ground loads have a more significant effect on the current mean fluid temperature than distant ground loads. For example, in the case of hourly simulations, the determination of T_m at the end of a year would require a summation of 8760 hourly terms according to Equation (18). One possible alternative is to aggregate (i.e., average) the ground loads of the first 8000 hours, then aggregate the next 730 hours and keep intact the last 30 hours. The summation term would then be reduced to 32 terms. Other aggregation schemes have been proposed by Bernier et al. (2004), Liu (2005), and Yavuzturk and Spitler (1999).

The design methods described previously size the ground loop for the larger of the heating or cooling loads, including a temperature penalty for the amount of imbalance (which can be large in severe climates). An alternative approach for imbalanced buildings is to partially balance the load on the ground, both at peak and annual scale, by adding a supplemental device to help meet the larger of the two peak loads. This is a hybrid geothermal (or hybrid ground-coupled) system. Hybrids can provide several benefits for buildings with a load imbalance. The biggest economic effect is in decreasing the ground heat exchanger size/cost. First-cost savings have been reported of 6 to 16% of total HVAC system cost, with little consequence reported on operating cost (Hackel and Pertzborn 2011; Singh and Foster 1998) because the HVAC systems operates for the vast majority of the year at a fraction of peak design. Table 7 shows that the more balanced loads resulting from hybrids also significantly affect long-term performance too.

In most U.S. commercial buildings, the cooling load is dominant both annually and at the peak because of high internal loads, ventilation heat recovery, and good building envelopes. Heat from compressors, pumps, and fans also plays a factor; in heating mode, this heat is delivered to the building, so less heat is required from the ground. As a result, achieving annual thermal balance requires heat pumps in a geothermal system to operate in heating mode 1.6 to 1.8 h for every hour in cooling.

The ideal configuration of the ground heat exchanger and supplemental cooling device in a hybrid depends on many factors, such as climate, building peak load, and building annual loads. Carefully analyze which approach may work best for a specific building. One common configuration for cooling-dominated systems, a series hybrid, is shown in Figure 20A. This approach could also be taken with a closed-circuit cooling tower (i.e., fluid cooler) downstream of the ground heat exchanger (GHX). In general, it is most effective to place the lower-temperature heat sink downstream; an energy model can help determine which order most often results in this scenario throughout the year. As a rule of thumb, in drier climates with warmer ground (e.g., desert southwestern United States) the tower is almost always the lower-temperature sink, whereas in humid climates with moderate-temperature ground (e.g. southeastern United States), the ground is often the lower-temperature sink. The hybrid can also be configured in parallel, as shown in Figure 20B, which is especially desirable if the ground heat exchanger is small in comparison to the building peak cooling load (a series system in this example would require a more complex partial GHX bypass). In either case, there are two guidelines for design and operation of hybrid systems:

Control of a cooling-dominated hybrid depends on the configuration. If equipment is placed as shown in Figure 20A, the temperature downstream of the tower can be used to control the use and speed of the tower based on a high limit (some additional savings are possible if the tower is controlled by the Δt between entering fluid and ambient wet-bulb temperatures, though this method depends on a difficult measurement of wet bulb). If the tower is located upstream of the ground heat exchanger, the temperature exiting the tower and the ground heat exchanger should both be used in tower control (to ensure the ground is not being cooled). For parallel configurations (Figure 20B), one practical tower control sequence bases tower operation on a calculation of the average of fluid temperature entering the heat pumps over the previous week (Xu 2007). Xu also suggests a strategy for controlling the parallel three-way valve.

Figure 20. Hybrid System Configuration Options, (A) Series and (B) Parallel

A heating-dominated hybrid with a boiler instead of a cooling tower can use a series configuration, with the boiler downstream of the loop (because of the boiler’s high temperature output). The boiler is ideally controlled based on the temperature leaving the heat pumps.

Sizing hybrid components is a bit more complex than standard systems. For cooling dominated hybrids, Kavanaugh and Rafferty (1997) suggest that heat exchanger length for heating L_h be determined using Equation (5) with heating-mode loop temperatures t_wi and t_wo as low as possible to minimize L_c. A tower with an isolation heat exchanger is sized to meet the capacity difference between the required cooling length L_c from Equation (4) and the heating length L_h. Kavanaugh (1998) revised this method to include an additional iteration to size the ground heat exchanger only after estimating the annual heat rejection from the tower: q_tower(rated gpm) = gpm_system (L_c – L_h)/L_c, where L_c is calculated from Equation (5) but based on reduced EFLH_c to account for tower operation rejecting an estimated amount of the annual load. The strategy suggests eliminating long-term ground temperature change with additional tower operation.

A more detailed study (Hackel et al. 2009) included assumptions about typical installation and operating costs to demonstrate an optimized design strategy for cooling dominated hybrids. Based on life-cycle cost, this approach was roughly attractive whenever the peak heating load was less than 80% of the cooling load; savings increased logarithmically as the ratio decreased below 80%. A variety of cases were modeled, and the simplified best-fit regression for the hybrid ground heat exchanger length L_hyb in a cooling-dominated scenario was found to be proportional to heating load:

(19)

where C₁ = 254 ft · h · °F/kBtu, at k =1.4 Btu/ft · h · °F. For other ground conductivities, the change in ground heat exchanger size is approximately inversely proportional to the change in conductivity. In choosing t_wo, Hackel et al. also suggest in cooler climates it is often economical to include antifreeze in the system and allow the entering fluid temperature to drop to 35°F or lower. The supplemental cooling device (closed-circuit tower) should then be sized to meet the fraction of the cooling load that this smaller hybridized ground heat exchanger cannot. The study suggests that the tower should even be oversized slightly and its fan put on variable-speed control, to achieve optimal performance. Furthermore, tower sizing should be completed using the local peak wet-bulb and design entering fluid temperature for the hybrid.

This basic strategy of sizing hybrids was found to be valid for a wide range of economic scenarios. This economically justifies the general concept laid out in Kavanaugh (1998); however, results showed that it is generally not economically optimal to balance the load on the ground, and some increase in ground temperature can be accepted. Preliminary research showed that running the tower at night or in winter before there is significant cooling load (precooling the ground) to balance the ground load is not always necessary, and if done without care (i.e., in-depth energy analysis) could possibly lead to increased energy consumption.

Regardless of the calculation method used, detailed building load calculation is critical when sizing and configuring a hybrid, to determine the impacts of heating and cooling loads across time scales from peak to annual. Further refinement of hybrid sizing (and control) can be done through energy simulation software, including some of the design software mentioned earlier, as well as a tool created as a result of ASHRAE research project RP-1384 (Hackel et al. 2009). Building energy simulation can estimate loads at every hour of the year, establishing better understanding of annual ground loads as well as the load sharing between ground heat exchanger and supplemental device.

Heating-Dominated Hybrids. These hybrids are needed in only very cold climates. In heating-dominated buildings in such climates (i.e., with approximately twice as much annual heating load as cooling load), however, a heating-dominated hybrid could decrease the size of the ground heat exchanger. Optimally, the ground heat exchanger is sized to meet the cooling load and a supplemental device meets the remaining heating load through a supplemental heat coil on the ground-coupled fluid loop. Coil energy could be supplied by gas boiler, solar panels, electric resistance, or another source. Several things need to be considered if using this approach:

An alternative strategy for heating dominated buildings that poses some advantages is to simply remove some loads in the building from the ground-coupled loop and serve them by boilers, solar panels, or some other source which provides high enough temperatures to heat without added heat pump energy. Baseboard, unit heaters, and preheat coils can be good applications for this approach. This may alleviate some of the complications with boiler hybrids discussed previously.

Loop design can have a substantial effect on both pumping power requirements and system installed cost. A GSHP survey (Caneta Research 1995) reported that installed pumping power varied from 0.04 to 0.21 hp/ton of heat pump power. This represents 4 to 21% of the total demand of typical GSHP systems and up to 50% of the total energy for some pump control schemes. Table 9 gives a recommended set of guidelines for minimizing the power of closed-loop GSHPs and maximizing system efficiency.

Good Table 9 grades can be obtained by minimizing extensive piping arrangements with long interior and exterior piping runs, high-head-loss fittings, valves, and control devices. Designers must compare the costs and advantages of large central piping loops and larger pumps with those of multiple smaller loops and smaller pumps. Pumping rates greater than 3.0 gpm/ton in closed-loop systems result in marginal equipment capacity gains in modern water-to-air heat pumps, and typically decrease overall system efficiency.

Kavanaugh et al. (2002) found the total cost of vertical GCHP ground-loop systems (including headers) ranged from $6.00 to $25.60 per foot of vertical bore. The cost of headers is a significant portion of the total and in many cases exceeded the cost of the vertical bore. The savings in vertical loop costs because of central systems’ load diversity often is not warranted because of the increased cost of large-diameter piping networks connecting equipment inside the building and below-grade circuits connecting exterior ground heat exchangers.

In low-rise buildings with large footprints, such as a school, multiple unitary loop systems (Figure 21) are an effective option to offset the high cost of central interior piping and ground-loop header costs. Although the total length of vertical bore for unitary systems is greater than for central loop systems, the high cost of interior piping, exterior headers, and valve vaults often offsets the bore cost savings. Additionally, pump demand is substantially reduced in the unitary system because of the short header runs, so low-wattage on/off circulator pumps are suggested.

A compromise in applications with significant load diversity is to group ground heat exchangers into multiple smaller subcentral loops in different areas of the building (Figure 22). Subcentral loops can be served by on/off circulator pumps located on each heat pump if a check valve is installed on each heat pump to prevent reverse water circulation through idle units.

Figure 23 is an example of a central loop that can effectively reduce the cost of the required vertical bore in buildings with higher load diversities. The central ground heat exchanger consists of several subheader sets, each having 6 to 20 vertical U-tube heat exchangers. The subheaders are gathered into a valve manifold located either near the center of the loop field in a below-grade vault or in the building equipment room. Each subheader set has isolation valves for independent purging of air and debris. Interior piping is similar to conventional water-source heat pump systems in which interior piping is routed to individual water-to-air heat pumps in each zone and/or heat pump water heaters and water-to-water heat pumps.

Variable-speed drives (VSDs) are recommended for central loop systems because they offer substantial energy savings compared to primary/secondary pump schemes in GSHP applications. However, in buildings with primary occupancy of less than 60 h per week, measures should be incorporated to turn off the main VSD pump and provide some alternative means of pumping water to critical building zones during low-occupancy or unoccupied periods.

Table 10 is a summary of Tables 9 and 10 from Chapter 32 of the 2007 ASHRAE Handbook—HVAC Applications, which compared a central GCHP loop and a set of eight subcentral loops for a 40,000 ft², four-story office with a 100 ton cooling load. The total cost of the single central loop was higher ($2114 per ton, versus $1956 per ton). Although the vertical portion of the subcentral loop cost slightly more ($1343 per ton, versus $1327 per ton), the total system cost was projected to be lower because of the lower header cost.

Table 10 also compares a central loop with multiple unitary loops for a 57,000 ft² elementary school. The projected 17% reduction in total cost ($2099 per ton versus $2530 per ton) of the unitary system is more significant compared to the reduction in the four-story office because of the building’s layout. The single-story, large-footprint design increases the cost of a central loop because of the long interior piping runs, even though the vertical portion of the central loop system is smaller. The unitary loop system benefits because the large footprint provides ample space for loops to be located near heat pump units.

Figure 21. Unitary GCHP Loops with On/Off Circulator Pumps

Figure 22. Subcentral GCHP Loop with On/Off Circulator Pumps

GSHP Piping Materials. It is an industry standard that the buried ground-loop piping materials be fusion welded. For this reason, the current International Ground Source Heat Pump Association (IGSHPA) design and installation standards lists only high-density polyethylene (HDPE), and most recently an exception for cross-linked polyethylene (PEX-a) as approved materials for buried pipe. (IGSHPA 2011). In most cases, this pipe is buried without insulation.

Figure 23. Central Loop GCHP

Distribution systems in a building need to be carefully chosen and not solely driven by cost or ease of installation. As with any system, piping materials must be compatible with system design temperatures and with the fluids conveyed. Closed-loop systems typically use potable water and may also include an antifreeze solution and/or water treatment chemicals. With newer refrigerants, care is needed to ensure the piping materials are compatible with the oils used in the refrigeration system and that equipment seals are compatible with the heat transfer fluids conveyed; for instance, polyolester (POE) oil used with R-410A is not compatible with PVC pipe. Consult chemical resistance charts provided by materials manufacturers.

Whether designing a new system or retrofitting an existing one, it is the design engineer’s responsibility to size and select the proper piping materials for each application and to select only those materials allowed by code (ICC 2012). In large building systems, the aboveground piping specifications commonly include steel, iron, copper or PVC. Where the hydronic mains are 4 in. or larger, it is often more cost effective to specify steel pipe. Specifying steel pipe may require some type of water treatment to inhibit general corrosion and dielectric isolation when connected to geothermal heat pumps.

Water Quality, Heat Transfer Fluids, and Water Treatment. When engineers introduce dissimilar metals into closed-loop hydronic piping systems, improperly address water treatment requirements for their designs, or are not properly informed of the local groundwater regulations, water quality problems may result. Poor water quality contributes to a decline in mechanical system performance, increased maintenance, and can reduce the useful lifetime of mechanical system components.

The standard working fluid in small residential closed-loop piping systems is either potable water or, in colder regions, an antifreeze solution. Where required, the antifreeze solution is introduced after the ground loop is properly pressure tested, flushed, and purged of debris and air. There has been little problem with or concern about water quality in these GSHP systems, because piping materials have historically been HDPE, copper, or stainless steel hose kits for final connection at the heat pumps.

Quality of potable water can vary, depending upon the source, and rules on its use vary from state to state. The chemistry of this water is one of the contributors to corrosion and water quality problems in closed-loop piping systems, so engineers need to be precise with hydronic piping and water treatment system specifications. Chapter 49 offers some guidance on understanding the consequences of water quality on both open- and closed-loop hydronic systems.

The type of chemical used for water treatment in closed-loop systems is based on several criteria, including the materials being protected, the quality of the water in the piping system, local regulations, and cost. These systems are typically specified to include a rigorous process for cleaning the distribution piping, flushing the piping systems of air and debris, and then adjusting the water quality of the final local water to meet the long-term performance requirements of the building systems. Chemicals used to adjust water quality for these systems are often not acceptable for use when the circulating fluid is also connected to a ground loop. This is not a problem unless there is a pipe failure or problem with the ground loop, but a potential leak is a concern of the regulatory agencies that protect groundwater in each state.

Flush and Purge. To ensure that a geothermal heat pump system provides trouble-free operation, all ground-loop systems must be properly flushed and purged prior to connection to the building piping system (IGSHPA 2011). The current standard of care is defined by IGSHPA as providing a minimum velocity of 2 ft/s through each piping system, and flushing and purging each supply and return circuit in the forward and reverse directions for long enough to remove all debris and air from the system.

Recommendations for Good GSHP Piping System Design.

Pressure Considerations In Deeper Vertical Boreholes. Vertical systems have be designed and installed to depths greater than 400 ft with mixed results. More field data are needed to extend this range without prominent warning. Pay special attention to the effect of both the internal and external fluid(s) differential pressure on the HDPE pipe, regardless of the specified U-bend type or insertion depth.

In most cases, the only design compensation for increased hydrostatic pressure on the U-bend assembly at these depths was to specify heavier-walled HDPE pipe [e.g., DR-9 (Dimension Ratio)]. Below 350 ft, in a borehole full of fluid, pipe collapse is more probable than burst due to the fact that thermal grout is denser than water. This results in an imbalance of the hydrostatic head in the annulus of the borehole versus the hydrostatic pressure exerted by the fluid inside of the pipe.

Internal working pressure (IWP) and external pressure resistance (EPR) of HDPE material decreases with time of exposure and elevated temperature. The 10 h IWP and EPR for HDPE 3406/3408 and 4710 at 73.4°F for DR-11 and DR-9 are shown in Table 11. The designer should check the manufacturer’s material specifications for all buried piping. Should the lower pipe DR (thicker wall) be selected, the EPR pressure increases at the expense of an increase in friction loss at any given flow. HDPE material selection, DR selection, grout selection, and the short term use of additional pressure (air or fluid) on the U-bend assembly may be remedies to deal with excessive differential pressure on the U-bend assembly.

Recent developments in clay-based thermally enhanced grout include the addition of graphite along with, or in place of, silica sand. Graphite-based-grout can have a thermal conductivity of up to 1.60 Btu/h · ft · °F. In addition, graphite-based grout has a significantly lower density and viscosity than clay/silica sand mixtures at a given thermal conductivity (Table 4). Graphite-based products are more expensive than clay/silica sand thermally enhanced product, but economics, geology, and site conditions may warrant the additional expense. Contact the grout manufacturer for additional information. In the long term, the annular fluid in a properly grouted borehole is designed to gel using a clay-based grout, or harden (set) using a cementatious grout, thus mitigating part of the differential pressure. In general, the maximum working time for most grouts is 30 to 45 min, depending on factors such as makeup water temperature and chemistry, borehole temperature, etc. After this amount of time, the grout begins to gel or set and become unpumpable. Note that grouting is not required in some jurisdictions and not addressed in others; this has given rise to substituting a manufactured fill material, which may be acceptable in some cases.

In deeper boreholes it is especially important for the designer to be aware of conditions that may lead to pipe failure and design accordingly. Equation (19) may be used to calculate differential pressure between the internal and external pressures on the HDPE U-bend assembly at any depth:

(20)

where

The following is a design example for a deeper borehole pipe design. A borehole is designed to be 500 ft deep. The thermal conductivity test indicates that the average deep earth temperature is 73°F. The selected U-bend assembly is HDPE 3408, DR-11. The clay-based thermal grout selected weighs 15.1 ppg at a thermal conductivity of 1.2 Btu/h · ft · °F. The manufacturer’s data sheet indicates that the grout will gel within 30 to 45 min of mixing and placement. The U-bend assembly is full of water (8.34 ppg). Using Equation (19), the differential pressure is 0.052 lb_f/in² × (15.1 ppg − 8.34 ppg) × 500 ft, or 175.8 psi on the exterior of the pipe at 500 ft. Note: The 10 h EPR for 3408 DR-11 at 73°F is 60 psi. Safety and best practice dictate that a design change may be warranted. The design would need to include direction to either stage pressure on the U-bend assembly to offset the hydrostatic head of the grout, or use a lower-density grout (see Table 4 for more information on grout options).

The ground heat exchanger must absorb the heat of compression and heat from auxiliary equipment (e.g., fans, pumps) in cooling mode. In heating mode, heat from auxiliary equipment reduces the amount of heat required from the ground. Therefore, the cooling- and heating-mode power values W_c and W_h in Equations (4) and (5) must include the auxiliary input.

Rated values published in compliance with ANSI/ARI/ASHRAE/ISO Standard 13256-1 do not include the auxiliary power required to circulate air and water through the distribution systems. Furthermore, the auxiliary power required to distribute chilled air (and water) can have a substantial negative effect on the equipment’s cooling capacity.

Table 12 summarizes the air and water temperatures used to generate the rated performance of water-to-air heat pumps.

Table 13 summarizes the conditions for ANSI/ARI/ASHRAE/ISO Standard 13256-2, the rating standard for water-tower heat pumps.

Actual heat pump performance can be substantially different from rated conditions. Designers must convert rated performance to design conditions by accounting for the effect of auxiliary power input and for design ELTs and EWTs. When water-to-water heat pumps or chillers are used in this application, corrections should include power for the pump(s) of the source/sink loop and the chilled-/hot-water loop, and power for fans in the air distribution system. Table 14 demonstrates the difference between rated GSHP efficiency and actual system efficiencies for various options when the effect of auxiliary components is considered. Note that using a high-static-pressure air handler for air distribution significantly reduces cooling efficiency. In heating, 120°F hot water also lowers heating COP compared to direct condensation and hydronic systems (e.g., in-floor heating) that use lower-temperature water.

The buried pipe of a closed-loop GSHP may theoretically produce a change in temperature in the ground up to 16 ft away. For all practical purposes, however, the ground temperature is essentially unchanged beyond about 3 ft from the pipe loop. For that reason, the pipe can be buried relatively near the ground surface and still benefit from the moderating temperatures that the earth provides. Because the ground temperature may fluctuate as much as ±10°F at a depth of 6 ft, an antifreeze solution must be used in most heating-dominated regions. The critical design aspect of horizontal applications is to have enough buried pipe loop in the available land area to serve the equipment. The design guidelines for residential horizontal loop installations can be found in OSU (1988b).

A horizontal loop design has several advantages over a vertical loop design for a closed-loop ground source system:

Figure 24. Horizontal Ground Heat Exchanger Configurations

Some limitations on selecting a horizontal loop design include the following:

An overlapping spiral configuration, shown in Figure 25, has also been used with some success. However, it requires special attention during backfilling to ensure that soil fills all the pockets formed by the overlapping pipe. Large quantities of water must be added to compact the soil around the overlapping pipes. The backfilling must be performed in stages to guarantee complete filling around the pipes and good soil contact. The high pipe density (up to 10 ft of pipe per linear foot of trench) may cause problems in prolonged extreme weather conditions, either from soil drying during cooling or from freezing during heating. This spiral design has been used in vertical trenches cut with a chain trencher as well as in laying the coil flat on the bottom of a large pit excavated with a bulldozer. Installations using the horizontal spiral coil on the bottom of a pit have generally performed better than those with spiral coils that were stood upright in a vertical trench.

The extra time needed to backfill and the extra pipe length required make spiral configurations nearly as expensive to install as straight pipe configurations. However, the reduced land area needed for the more compact design may allow use on smaller residential lots that are too small for conventional horizontal-pipe ground-loop designs. The spiral pipe configuration laid flat in a horizontal pit arrangement is used commonly in the northern Midwest of the United States, where sandy soil causes vertical trenches to collapse. A large open pit is excavated by a bulldozer, and then the overlapping pipes laid flat on the bottom of the pit. The bulldozer is also used to cover the pipe, being careful to not run over them with the bulldozer tread.

Figure 25. General Layout of Spiral Earth Coil

Although most horizontal closed-loop systems are installed with either a chain trencher or a backhoe, horizontal boring machines are also now available for this application. Developed for buried utility applications such as electric or potable water service, these devices simply bore through the ground parallel to the ground surface. A detector at the surface can show the exact point where the boring head is located underground so that the bore does not penetrate other known utilities or cross over into a neighbor’s lot.

Most horizontal loop installations place the pipe loops in a parallel rather than a single (series) loop to reduce pumping power (Figure 26). Parallel loops may require slightly more pipe, but may use smaller pipe and thus have smaller internal volumes, requiring less antifreeze (if needed). Also, smaller pipe is typically much less expensive for a given length, so total pipe cost should be less for parallel loops. An added benefit is that parallel loops can be flushed out with a smaller purge pump than is required for a larger single-pipe loop. A disadvantage of parallel loops is the potential for unequal flow in the loops and thus nonuniform heat exchange efficiency.

The time required to install a horizontal loop is not much different from that for a vertical system. For the arrangements described, a two-person crew can typically install the ground heat exchanger for an average house in a single day.

Soil characteristics are an important concern for any ground heat exchanger design. With horizontal loops, the soil type can be more easily determined because the excavated soil can be inspected and tested. EPRI (1990) lists criteria and simple test procedures that can be used to classify soil and rock for horizontal ground-loop design.

Figure 26. Parallel and Series Ground Heat Exchanger Configurations

Although soil type and moisture content are important considerations in sizing the ground heat exchanger, some design guidelines have been developed based on extensive analysis of monitored systems in mostly southern climates (Kavanaugh and Calvert 1995). Table 15 gives recommended trench lengths for the various types of commonly used excavation methods. Heating-mode run times approaching 100% on a daily basis would be the norm at heating design conditions in heating-dominated climates. In contrast, daily run times of no more than 50% would be encountered at design cooling conditions in cooling-dominated climates. The combination of long run times and ice formation around the pipes makes performance of horizontal systems dependent on both the loop field design and how the system is matched to the building load. Though many thousands of these systems have been installed in heating climates, no comparable analysis has been performed to determine proper design guidelines. The loop length data in Table 15 for soil temperatures below 56°F are based on nominal heat pump capacity and use of supplemental resistance heat at design conditions. If installing such a system for the first time, contact several experienced contractors in the area to determine successful design lengths for the local climate and soil types.

Trench lengths in Table 15 are based on a minimum trench separation of 10 ft and minimum horizontal loop average burial depth of the greater of 5 ft or 2 ft below the frost line. Bore lengths are based on a vertical bore separation of 20 ft. Design ground heat exchanger temperatures are a maximum of 90°F return and 100°F entering in warm climates and a minimum of 28°F return and 22°F entering in cold climates.

Additional considerations for horizontal loop systems in colder climates arise from the potential for ice formation around the pipe loop. The loop should not pass within 2 ft of any buried water line (potable, sewer, or rainwater). If such proximity cannot be avoided, the GCHP loop can be insulated in that area. Horizontal loops should not be placed closer than 6 ft from a basement or crawl space wall when buried parallel to the wall. Heaving from ice formation could cause structural damage if placed in close proximity to the wall.

Leaks in heat-fused plastic pipe are rare when attention is paid to pipe cleanliness and proper fusion techniques. Should a leak occur, it is usually best to try to isolate the leaking parallel loop and abandon it in place. The effort required to find the source of the leak usually far outweighs the cost of replacing the defective loop. Because the loss of as little as 0.25 gal of water from the system causes the system to lose pressure and shut down, leaks cannot be located by looking for wet soil, as is commonly done with water lines.

Although leaks should be rare with properly thermally fused pipe, some states have adopted restrictions against the use of certain types of antifreeze mixtures in GCHP systems; check local water-quality regulations before selecting a mixture. Methanol has been used extensively because of its low cost and good physical properties when cold. Comprehensive studies by Heinonen and Tapscott (1996) and Heinonen et al. (1997) showed that propylene glycol is a good alternative when issues of flammability or environmental safety are important considerations. A more thorough discussion of antifreeze solutions is given in the Antifreeze Requirements section of this chapter.

Fluid Flow and Loop Circuiting. Residential systems, like commercial applications, sometimes have excessive pumping power. This trend may result from undersized piping, excessive amounts of viscous antifreeze solutions, or conservative pump sizing. Because a 3 ton heat pump with an energy efficiency ratio (EER) of 15 requires a total power (compressor and fan) of 2400 W, the addition of a second 1/6 hp pump (which draws 245 W) reduces system efficiency by 10%. Table 16 provides a guideline to ensure adequate liquid flow rate with the least possible number of pumps. It should be used in conjunction with Table 15 and applies to loops with 0 to 15% propylene glycol solutions (by volume; note that caution and additional treatment may be needed for solutions lower than 20%, because of risks of corrosion and biological growth below this level). This solution has the reputation of being the most difficult of the commonly used solutions to pump when cold. However, it is no more difficult to pump than ethyl alcohol, and pumping penalties can be mitigated by adding only the required amount of antifreeze. Shorter loops may require higher levels of antifreeze solutions. See the section on Antifreeze Requirements for more details. Any exposed piping above the frost line must be insulated with closed-cell insulation with ultraviolet (UV) protection (paint or wrap).

Central plant GCHP systems use central water-to-water equipment (e.g., a water-cooled chiller) to move thermal energy between a source loop (the ground coupled heat exchanger), a chilled-water loop, and a hot-water loop. Here, the term central plant implies the mechanical equipment is in one centralized location; it does not imply a campus is served, and single buildings are a common application. The central plant source loop is commonly a vertical closed-loop heat exchanger, but any combination of GCHP types may be used. The chilled- and hot-water loops may serve any HVAC distribution equipment (fan coils or radiant systems are efficient options) as long as design hot-water supply temperatures do not exceed heat pump temperature ranges (typically 130°F). Best practice in these applications is largely theoretical or anecdotal; research is needed on the efficacy of the various central plant GCHP applications and designers considering central plant GCHPs must be cautious.

Compared to traditional unitary (water-to-air) GCHP systems, central plant systems offer the potential advantages of (1) incorporating direct heat recovery from hot-water loads to chilled-water loads, (2) sometimes allowing waterside economizing, (3) centralizing equipment maintenance, and (4) expanding retrofit opportunities. However, when such plants are connected to conventional distribution such as variable air volume (VAV), the loss of zone level heating and cooling results in large fan and reheat energy penalties. Pumping power is often higher than equivalent unitary heat pump systems. Central plants also have significantly more complex design, controls, commissioning, operational training, and maintenance. Most existing tools and methods for sizing ground heat exchangers and evaluating energy performance do not accurately represent central plant GCHP operation.

As with unitary systems, central plant systems may take many forms. Figure 28 illustrates the basic building blocks of a central plant GCHP system. Any number of heat recovery chillers and heating or cooling heat pumps may be installed. Either or both of the load-side loops may be connected to the source loop to achieve direct heat transfer (water economizing). The loops may be separated by control valves or heat exchangers (as shown in Figure 28), but loops should be linked only by control valves where fluid type and pressure are compatible.

Most central plant heat pump equipment is designed for use in one of three basic strategies:

Figure 27. Residential Design Example

Figure 28. Central Plant GCHP System

To provide stable temperature control and part load operation, central plant GCHP systems can benefit from additional thermal capacitance (buffer tanks) on the load side loops. Any loop flow control methodology may be used; however, variable flow should be used to minimize pumping energy penalties where equipment allows. Consider using water as the working fluid wherever possible, because of the energy and capital cost associated with antifreeze solutions.

Central plant designers have the option to connect hybrid heating and cooling equipment to the load side of the heat pumps instead of the source/ground side, which adds redundancy and reduces pump and compressor energy at the expense of increased controls complexity. Hybrid air-cooled chillers may serve the chilled-water loop directly, whereas evaporative cooling equipment is best placed on the source loop in series or in parallel with the ground heat exchanger. Connecting a hybrid boiler to the hot-water loop reduces thermal stress for boilers designed to operate at higher temperatures and provides direct emergency heat; connecting a hybrid boiler to the source loop reduces control interconnection/complexity and reduces heat pump faults caused by low entering water temperatures.

An open-loop system design must balance well pumping power with heat pump performance. As groundwater flow increases through a system, more favorable average temperatures are produced for the heat pumps. Higher groundwater flow rates, to a point, increase system EER or COP: increased well pump power is outweighed by decreased heat pump power requirements (because of the more favorable temperatures). At some point, additional increases in groundwater flow result in a greater increase in well pump power than the resulting decrease in heat pump power. The key strategy in open-loop system design is identifying the point of maximum system performance with respect to heat pump and well pump power requirements. Once this optimum relationship has been established for the design condition, the method of controlling the well pump determines the extent to which the relationship is preserved at off-peak conditions. This optimization process involves evaluating the performance of the heat pumps and well pump(s) over a range of groundwater flows. Key data necessary to make this calculation include well performance (flow and drawdown at various groundwater flows) and heat pump performance versus entering water temperatures at different flow rates. Well information is generally derived from well pump test results. Heat pump performance data are available from the manufacturer.

GWHP systems employ the same type of extended-range unitary heat pumps as GCHP systems. Building loop pumping guidelines (see Table 9) in the GCHP portion of this chapter also apply to GWHP systems. In large commercial applications, the head loss associated with the isolation heat exchanger in a GWHP system is typically lower than that of an equivalently sized ground heat exchanger in a GCHP system. A guideline for building loop head loss in a GWHP system can be described as follows:

where d = pipeline distance ft from plate heat exchanger outlet to most distant heat pump unit inlet.

This calculation assumes a maximum head loss of 4 ft/100 ft fittings at 25% of total head loss and a heat pump unit head loss of 12 ft. Because of more extensive fittings, retrofits can sometimes exceed this value.

For moderate-efficiency heat pumps (EER of 14.2), efficient loop pump design (7.5 hp/100 tons), and a heat exchanger approach of 3°F, Figure 29 provides curves for two different groundwater temperatures (GWT = 50 and 65°F) and two well pump situations (SWL 75 ft/specific capacity 10 gpm/ft and SWL 300 ft/specific capacity 3 gpm/ft). The curves are plotted for constant well pump head, a situation which does not occur in practice. In reality, well pump head rises with flow but at a rate typically less than that in friction head applications.

Although the four curves show a clear optimum flow, sometimes operating at a lower groundwater flow reduces well/pump capital cost and the problem of fluid disposal. These considerations are highly project specific, but do afford the designer some latitude in flow selection. Generally, an optimum design results in a groundwater flow rate that is less than the building loop flow rate.

The exception to this occurs in the case of groundwater temperatures of less than 47°F and greater than 72°F. In these situations the groundwater flow requirement is influenced more by avoiding excessive heat pump EWT in the cooling mode (groundwater temperatures above 72°F) and heat pump LWTs that could result in freezing conditions in the heating mode (groundwater temperatures less than 47°F). In the case of low water temperatures, some designers have found it advantageous to use antifreeze in the building loop to slightly broaden the allowable loop temperature range.

Table 17 provides design data for a specific example system.

When possible, well testing should be completed before mechanical design. Only with actual flow test data and water chemical analysis information can accurate design proceed.

Flow testing can be divided into three different types of tests: rig, short-term, and long-term (Stiger et al. 1989). Rig tests are generally very short and are accomplished while the drilling rig is on site. The primary purpose of this test is to purge the well of remaining drilling fluids and cuttings and to get a preliminary indication of yield. The length of the test is generally governed by the time required for the water to run clean. The rate is determined by the available pumping equipment. Frequently, the well is “blown” or pumped with the drilling rig’s air compressor. As a result, limited information about the well’s production characteristics is available from a rig test. If the well is air lifted, it may not be useful to collect water samples for chemical analysis because certain chemical constituents may be oxidized by the compressed air.

Properly conducted, short-term, single-well tests lasting 4 to 24 h yield information about well flow rate, temperature, drawdown, and recovery. These tests are used most frequently for direct-use and GWHP applications. The test is generally run with a temporary electric submersible pump or lineshaft turbine pump driven by an internal combustion engine and are often performed by a well pump contractor.

A step test (Table 19), the most common type, involves at least three production rates, the largest being equal to the design flow rate for the system served. The three points are the minimum required to determine a productivity curve for the well that relates production to drawdown (Stiger et al. 1989). The key parameters monitored during these tests are well water level and water flow. Water level and pumping rate should be stabilized at each point before flow is increased. In many cases, water level is monitored with a bubbler or an electric sounder, and flow is measured using an orifice meter. More sophisticated instrumentation (e.g., pressure transducers for water level, magnetic flow meters, data loggers) can also be used. Short-term testing is generally used for small projects and provides information on yield, drawdown, and specific capacity.

Test results should reflect stable flows rates and individual flow “steps” are extended until water level readings stabilize. In many cases brief intervals of turbidity may occur at flow changes but extensive periods of turbidity indicate instability in the near well formation.

Long-term tests of up to 30 days provide information on the reservoir. Normally these tests involve monitoring nearby wells to evaluate interference effects. The data are useful in calculating transmissivity and storage coefficient, reservoir boundaries, and recharge areas (Stiger et al. 1989) but are rarely used for direct-use and GWHP systems.

It is also important to collect background information before the test, and water level recovery data after pumping has ceased. Recovery data in particular can be used to evaluate skin effect, which is a type of well flow resistance caused by residual drilling fluids, insufficient screen or slotted liner area, or improper filter pack.

The importance of groundwater quality depends on the system design. Systems using isolation heat exchangers commonly encounter no water quality issues (other than iron bacteria) that would prevent a GWHP system from operating under reasonable maintenance levels.

For systems that use groundwater directly in heat pump units (e.g., standing-column systems and small residential GWHP systems), several issues are of concern. The primary water quality problem in the United States is scaling, usually of calcium carbonate (lime). Because this type of scaling is partially temperature driven, the temperature of surfaces that groundwater contacts determines the extent to which scaling will occur. In these systems, peak temperatures in the refrigerant-to-water exchanger in cooling mode are likely to be over 160°F. For the same system using an isolation plate heat exchanger, the groundwater is unlikely to encounter temperatures over 90°F. Using the plate heat exchanger reduces the propensity for scaling and limits any scale that does occur to a single heat exchanger. Rafferty (2000b) provides information on water scaling potential on a state-by-state basis.

Although cupronickel, originally developed for seawater applications, is an excellent material in that role, it provides few benefits for most groundwater applications. It has been shown to perform more poorly than pure copper with hydrogen-sulfide-bearing waters.

Excessive iron, particularly ferrous iron, in the water can result in coating of heat transfer surfaces if the water is exposed to air (allowing the iron to oxidize to the ferric state, a form with much lower solubility in water). Periodically removing this iron from the plates of a single heat exchanger is much less labor-intensive than removing it from tens or hundreds of individual heat pump heat exchangers. Table 20 summarizes the minimum parameters that should be evaluated for a GWHP application.

Particulate matter (e.g., sand) in the groundwater stream, although usually not a problem in the mechanical system, can effectively plug injection wells. Sand production should be addressed in construction of the production well (screen/gravel pack/development). If it must be dealt with on the surface, a screen or strainer is preferable to a centrifugal separator, which can be ineffective at start-up and shutdown and can experience variable flow (Kavanaugh and Rafferty 1997). Perforation size selection is critical to a strainer’s effectiveness. Sizing should be based on 90 to 100% removal of the particulate material. Particle size information can be based on the sieve analysis results of drill cuttings (used to size the well screen) or of a sample taken during well flow testing. In applications with very fine sand, multiple strainers in parallel may be necessary to control pressure drop (Rafferty 2008).

Submersible pumps have not performed well in higher-temperature, direct-use projects. However, the submersible pump is a cost-effective option with normal groundwater temperatures, as encountered in heat pump applications. The low temperature eliminates the need to specify an industrial design for the motor/protector, thereby greatly reducing the first cost relative to direct use. Caution should still be used for wells that are expected to produce moderate amounts of sand. The high speed (nominal 3600 rpm) of most submersible pumps makes them susceptible to erosion damage.

Small groundwater systems have frequently been identified with excessive well pump energy consumption. The reasons for excessive pump energy consumption (high water flow rate, coupling to the domestic pressure tank, and low efficiency of small submersible pumps) are generally not present in large, commercial groundwater systems. In large systems, the groundwater flow per unit capacity is frequently less than half that of residential systems. Pressure at the wellhead is not the 30 to 50 psi typical of domestic systems, but is rather a function only of pressure losses through the groundwater loop. Finally, large well pumps have efficiencies of up to 83% compared to the 35 to 40% range for small submersible pumps.

In GWHP system design, the control method for the well pump determines the extent to which the optimum relationship between well pump power and heat pump power is preserved at off-peak conditions. There are several ways the pump can be controlled. Multiple pumps can be staged to meet system loads, either with multiple wells or with multiple pumps installed in a single well. A dual set-point control similar to that used in boiler/tower systems energizes the well pump above a given temperature in cooling mode and below a given temperature in heating mode. Between those temperatures, the building loop floats without the addition of groundwater. To control well pump cycling, it is necessary to establish a temperature range (difference between pump-on and pump-off temperatures) over which the pump operates in both the heating and cooling modes. The size of this range is primarily a function of the building loop water volume in terms of gallons per peak per ton of peak block system load (Rafferty 2000c). Table 21 summarizes these data. In the example in Table 17, the optimum system building loop return temperature (at peak system EER) is 80.6°F. If this system had a water volume of 8 gal/ton, from Table 21, a range of 8°F in cooling mode would be required. This range would result in a well pump start temperature of 80.6 + (8/2) = 84.6°F and a well pump stop temperature of 80.6 − (8/2) = 76.6°F. A similar calculation can be made for heating mode. From Table 21, for systems with very low thermal mass, the dual set-point method of control becomes impractical because of the very large temperature range required. For these applications, an alternative method of control (variable speed, staging, etc.) is required.

Well pumps may also be controlled using a variable-speed drive, which responds to building loop return temperature by varying groundwater flow to the exchanger to maintain the cooling or heating mode set point. Submersible-motor variable-speed applications are somewhat different than surface motor applications. Most manufacturers limit speed reduction to 50%, and other issues such as minimum water velocity for motor cooling, switching frequency, reactor requirement, and motor protection must be addressed. Additional information on VFD applications for submersible motors is available in Rafferty (2008).

Design of a plate-and-frame heat exchanger is largely a trade-off between pressure drop, which influences pumping (operating cost), and overall heat transfer coefficient, which influences surface area (capital cost). In general, exchangers in GWHP systems can be economically selected for approach temperatures (between loop return and groundwater leaving temperatures) as low as 3°F. Most selections involve an approach of between 3 and 7°F and a pressure drop of less than 10 psi on the building loop side. Excessive fouling factors (>0.0002 h · ft² · °F/Btu) should not be specified when selecting plate heat exchangers, which can be easily disassembled and cleaned.

Heat exchanger cost may be reduced for groundwater applications by using Type 304 stainless steel plates rather than the Type 316 or titanium plates common in direct-use projects. The low temperature and generally low chloride content of heat pump fluids frequently make the less expensive Type 304 material acceptable. Chloride content of the groundwater, particularly in coastal areas, should always be compared to values in Figure 4 to determine plate material acceptability. Exchanger performance should be checked at minimum system flow rates to ensure adequate heat transfer. In some cases, very low design pressure drop selections can encounter inadequate heat transfer at minimum flows.

Residential groundwater heat pump systems have the same design considerations as commercial groundwater heat pump systems, but differ on three main items: typically they (1) are integrated with a household domestic water system, (2) are single-zone systems, and (3) do not isolate the groundwater from the heat pump unit(s).

Groundwater heat pumps are a prudent choice in residential buildings on well water if the groundwater is of good quality. As such, the heat pump can be integrated into the domestic water system and considered another water-using appliance. Design care must be taken to ensure that the well and pressure tank have adequate capacity to handle the additional flow demand of the heat pump. Well pumps may be of the submersible or jet type, and the design groundwater flow rate should be chosen based on its temperature such that the system COP or EER is maximized. Flow control valves are recommended in the discharge line to ensure that the well is not over-pumped. Placement of a slow-closing motorized valve also on the discharge line ensures positive pressure on the heat pump water coil, and stops the flow of water when the heat pump is not operating (Figure 31).

Figure 31. Motorized Valve Placement

The pressure tank provides water at pressure on demand without short-cycling the well pump. A prepressurized bladder tank is preferred, and it should be large enough that filling it with the well pump takes at least 1 min.

Residential groundwater heat pump systems are small enough that the additional cost of an isolation heat exchanger is typically not economically justified. Raw groundwater is generally used as the heat transfer fluid, but provisions must be made (such as hose bibs or boiler drains) to allow for flushing and descaling of the heat pump water/refrigerant coil if necessary.

After exiting the heat pump, groundwater should be returned to a point of discharge in accordance with best practices and/or local codes. Surface discharge to a pond or wetland, or infiltration in to a dry well may be more of an option in these smaller systems than with larger commercial systems due to the correspondingly lower flow rates.

Central plant systems, in which a conventional or heat recovery central chiller is connected to a four-pipe system, are the oldest type of open-loop design, having first been installed in the late 1940s. Because of the cost and energy requirements of the central plant design, these systems typically do not result in the same level of energy efficiency as unitary GWHP systems.

For central plant groundwater systems, two heat exchangers are normally used: one in the chilled-water loop and one in the condenser water loop (Figure 32). The evaporator-loop exchanger provides a heat source for heating-dominated operation and the condenser-loop exchanger provides a heat sink for cooling-dominated operation.

Sizing the condenser-loop exchanger is based on providing sufficient capacity to reject the condenser load in the absence of any building heating requirement.

Sizing the chilled-water-loop exchanger must consider two loads. The primary criterion is the load required during heating-dominant operation. The exchanger must transfer sufficient heat (when combined with compressor heat) from the groundwater to the chilled-water loop to meet the building’s space heating requirement. Depending on the relative groundwater and chilled-water temperatures and on the design temperature rise, exchangers may also provide some free cooling during cooling-dominant operation. If groundwater temperature is lower than that of chilled water returning to the exchanger, some chilled-water load can be met by the exchanger. This mode is most likely available in regions with groundwater temperatures below 60°F.

Figure 32. Central Plant Groundwater System

Central plant chiller controls must also allow for the unique operation with a groundwater source. Controls can be similar to those on a heat-recovery chiller with a tower, with one important difference. In a conventional heat-recovery chiller, waste heat is available only when there is a building chilled-water (or conditioning) load. In a groundwater system, a heat source (the groundwater) is available year round. To take advantage of this source during the heating season, the chiller must be loaded in response to the heating load instead of the chilled-water load. That is, the control must include a heating-dominant mode and a cooling-dominant mode. Two general designs are available for this:

For buildings with a significant heating load, the former may be more attractive, whereas the latter may be appropriate for conventional buildings in moderate to warm climates.

In practice, standing-column wells (SCWs) are a trade-off between groundwater systems and ground-coupled systems. They do not require well flow testing of the sort necessary for groundwater systems, and conductive heat transfer can be based on existing closed-loop theory. Though standing-column systems have been applied mostly in the northeastern United States, approximately 60% of the country is underlain by near-surface (<150 ft) bedrock suitable for the systems.

A heating capacity of 350,000 to 420,000 Btu/h or cooling capacity of 30 to 35 tons can be expected from a 1500 ft deep standing-column well. Ideal spacing between SCWs is 50 to 75 ft (Orio et al. 2005) Typically, spacing between standing-column wells is greater than vertical closed-loop (GCHP) boreholes. These estimated capacities assume a 10% advective bleed flow, discussed later in this section. Closer spacing affects well field performance and can be evaluated with design software. Additional information on standing-column systems can be found in Spitler (2002).

The standing-column well combines supply and injection wells into one, and does not depend on the presence or flow of groundwater, beyond that of the typical bleed rate of 10 to 20% of total pumped flow. Standing-column wells are always augmented with a bleed circuit, to monitor the entering water temperature. Well water temperatures that are below or above design limits because of variations in rock conductivity, building anomalies, or nonstandard weather patterns can be restabilized (i.e., brought back to far-field temperatures by overflowing smaller amounts of water on command). This advective flow is a powerful short-term method of warming and cooling well columns that are beyond design limits.

Standing-column wells (Figure 33) consist of a borehole cased in steel or other material until competent bedrock is reached. The casing must be driven 25 to 50 ft into and sealed in competent bedrock. Bedrock sealing requirements vary by state. The remaining depth of the well is then self-supporting through bedrock. For deep commercial SCWs, a tail pipe (porter shroud) is inserted to form a conduit to draw up water, and an annulus to return water downward. This tail pipe is perforated at the bottom to form a diffuser. Water is drawn into the diffuser and up the central riser pipe to the submersible pump. The well pump must be located below the water table in line with the central riser pipe. The tail pipe allows a shorter, reduced-power wire size as well as more accessible well pump service.

The U.S. Environmental Protection Agency (EPA) Underground Injection Control program considers standing-column reinjection well water a Class V water use, type 5A7, noncontact cooling water for geothermal heating and cooling. The EPA and equivalent state agencies regard SCW reinjection as a beneficial use. Permitting or notice may be required, depending on average daily water flow rates. SCWs are serviced by qualified well contractors with minimal familiarization training.

Figure 33. Commercial Standing-Column Well

Heat is transferred to lakes by three primary modes: radiant energy from the sun, convective heat transfer from the surrounding air (when the air is warmer than the water), and conduction from the ground. Solar radiation, which can exceed 300 Btu/h per square foot of lake area, is the dominant heating mechanism, but it occurs primarily in the upper portion of the lake unless the lake is very clear. About 40% of solar radiation is absorbed at the surface (Pezent and Kavanaugh 1990). Approximately 93% of the remaining energy is absorbed at depths visible to the human eye.

Convection transfers heat to the lake when the lake surface is cooler than the air. Wind speed increases the rate at which heat is transferred to the lake, but maximum heat gain by convection is usually only 10 to 20% of maximum solar heat gain. Conduction gain from the ground is even less than convection gain (Pezent and Kavanaugh 1990).

Lakes are cooled primarily by evaporative heat transfer at the surface. Convective cooling or heating in warmer months contributes only a small percentage of the total because of the relatively small temperature difference between the air and lake surface. At night when the sky is clear, longwave radiation can account for significant amount of cooling. The relatively warm water surface radiates heat to the cooler sky. For example, on a clear night, a cooling rate of up to 50 Btu/h · ft² is possible from a lake 25°F warmer than the sky. The last mode of heat transfer, conduction to the ground, does not play a major role in lake cooling (Pezent and Kavanaugh 1990), though it does provide significant heating under winter conditions (Gu and Stefan 1990) when the surface of the lake is frozen.

To put these heat transfer rates in perspective, consider a 1 acre (43,560 ft²) lake used in connection with a 10 ton (120,000 Btu/h) heat pump. In cooling mode, the unit rejects approximately 150,000 Btu/h to the lake. This is 3.4 Btu/h · ft², or approximately 1% of the maximum heat gain from solar radiation in the summer. In winter, a 10 ton heat pump absorbs only about 90,000 Btu/h, or 2.1 Btu/h · ft², from the lake.

The maximum density of water occurs at 39.2°F, not at the freezing point of 32°F. This phenomenon, in combination with the normal modes of heat transfer to and from lakes, produces temperature profiles advantageous to efficient heat pump operation. In the winter, the coldest water is at the surface. It tends to remain at the surface and freeze. The bottom of a deep lake stays 5 to 10°F warmer than the surface. This condition is referred to as winter stagnation. The warmer water is a better heat source than the colder water at the surface.

As spring approaches, the surface water warms until the temperature approaches the maximum density point of 39.2°F. The winter stratification becomes unstable, and circulation loops begin to develop from top to bottom. This condition of spring overturn (Peirce 1964) causes the lake temperature to become fairly uniform.

Later in the spring, as water temperatures rise above 45°F, the circulation loops are in the upper portion of the lake. This pattern continues throughout the summer. The upper portion of the lake remains relatively warm, with evaporation cooling the lake and solar radiation warming it. The lower portion (hypolimnion) of the lake remains cold because most radiation is absorbed in the upper zone. Circulation loops do not penetrate to the lower zone, and conduction to the ground is quite small. The result is that, in deeper lakes with small or medium inflows, the upper zone is 70 to 90°F, the lower zone is 40 to 55°F, and the intermediate zone (thermocline) has a sharp change in temperature in a small change in depth. This condition is referred to as summer stagnation.

As fall begins, the water surface begins to cool by radiation and evaporation. With the approach of winter, the upper portion begins to cool toward the freezing point and the lower levels approach the maximum density temperature of 39.2°F. An ideal temperature-versus-depth chart is shown in Figure 34 for each of the four seasons (Peirce 1964).

Many lakes do exhibit near-ideal temperature profiles. However, (1) high inflow/outflow rates, (2) insufficient depth for stratification, (3) level fluctuation, (4) wind, and (5) lack of enough cold weather to establish sufficient amounts of cold water necessary for summer stratification can disrupt the profile. Therefore, a thermal survey of the lake should be conducted or existing surveys of similar lakes in similar geographic locations should be consulted (Hattemer and Kavanaugh 2005). When interpreting survey data, be aware that there are annual variations in temperature profiles that will not be reflected in measurements for a single year. Shallow ponds and lakes often destratify completely. This does not preclude their use for SWHP systems, but does reduce performance in cooling mode and may require larger heat exchangers compared to lakes that remain stratified.

The thermal and environmental effects of heat rejection and absorption on larger lakes and streams have been studied (Hattemer et al. 2006; Spitler 2014). However, the effect of SWHPs on thermal stratification profiles is not well characterized, and there is a lack of experimental data. The relative importance of heat transfer modes is not well known, especially during heating mode. There are no publicly available design tools that consider the many heat and mass flow modes of lakes, streams, and oceans. The model of unstratified ponds developed by Chiasson et al. (2000b) has been implemented in publicly available energy calculation programs, but its lack of accounting for stratification, freezing on the coil, or freezing at the surface limits its usefulness for analysis of deeper lakes and operation in heating mode. The model developed by Spitler et al. (2012) accounts for stratification and for freezing both on the coil and at the surface, but makes several approximations such as neglecting inflows, outflows, and water level variations. Furthermore, validation of all such models is necessarily limited to a small number of lakes for which experimental data are available. Therefore, some caution in using such tools is warranted.

Figure 34. Idealized Diagram of Annual Cycle of Thermal Stratification in Lakes

The closed-loop SWHEs shown in Figure 35 have several advantages over the open loop:

Disadvantages of closed-loop systems include the following:

Figure 35. SWHEs: (A) HDPE Coil Type and (B) Plate Type

High-density polyethylene (HDPE 3408) is recommended for in-lake piping. All connections must be either thermally socket-fused or butt-fused. These plastic pipes should also have protection from UV radiation, especially when near the surface. Polyvinyl chloride (PVC) pipe and plastic pipe with band-clamped joints are not recommended.

Plate heat exchangers are also available, and manufacturers typically provide estimated capacities for specified conditions. However, use care in applying metal heat exchangers in cold climates. For an equal temperature difference between the reservoir and fluid inside the coil, the surface temperature of metal heat exchangers is closer to the freezing point of water. The higher thermal resistance of HDPE results in a larger required surface area and a much lower heat transfer per unit of surface area. Thus, the surface temperature of an adequately designed HDPE coil tends to be higher than that of a metal tube or plate and less likely to develop ice on the coil exterior. In certain circumstances, ice build-up may occur on the closed-loop SWHE nonetheless. Use proper methods for anchoring the SWHE to resist the upward buoyant force caused by ice build-up.

The piping networks of closed-loop systems resemble those used in ground-coupled heat pump systems. Both a large-diameter header between the heat pump and lake coil and several parallel loops of piping in the lake are required. Loops are spread out to limit thermal interference, hot spots, and cold pockets. Although this layout is preferred in terms of performance, installation is more time consuming. Many contractors simply unbind plastic pipe coils and submerge them in a loose bundle. Some compensation for thermal interference is obtained by making bundled coils longer than the spread coils. A diagram of this type of installation is shown in Figure 35A.

Copper coils have also been used successfully. Copper tubes have a very high thermal conductivity, so coils only one-fourth to one-third the length of plastic coils are required. However, copper pipe does not have the durability of HDPE 3408, and if fouling is possible, coils must be significantly longer.

Open-loop surface water heat pump systems use heat pumps or chillers to provide heating and/or cooling, with surface water circulating through a heat exchanger to provide the heat source and/or sink. The heat pumps may be unitary or custom built.

Direct surface water cooling (DSWC) systems use cold lake or sea water to provide cooling without heat pumps. Because the total (horizontal + vertical) distance between the building(s) being cooled and the actual location of the cold water is often significant, the scale of the system that is economically feasible tends to be quite large.

In cases where the lake or sea water may not be cool enough to meet all demands, hybrid systems that can provide cooling with or without heat pumps or chillers have also been used. Published descriptions of all three types of systems have been reviewed by Mitchell and Spitler (2013).

As noted previously, open-loop systems with unitary heat pumps are not suitable for lake temperatures below 40°F because of the risk of freezing in the evaporator. However, larger installations that use custom-designed heat pumps are successfully operated in Scandinavia under even colder conditions. Their capacities can be as large as 100 million Btu/h. One such system takes water from the Baltic Sea at 37°F and returns it at 33°F using falling film evaporators (i.e., plate evaporators with the sea water sprayed on to the outside of the evaporator) (Mitchell and Spitler 2013).

Open-loop systems are generally designed with either a submersible pump in a wet sump pit connected to the water body or with a conventional centrifugal pump in a dry sump pit. It is also possible in some applications to place the pump just above the water level. Small (e.g., residential) systems are sometimes installed with a submersible pump in the water body.

Design guidelines for open-loop systems may be summarized as follows:

Closed-loop horizontal and surface water heat exchanger systems often require antifreeze in the circulating water in locations with significant heating seasons. Antifreeze may not be needed in a comparable vertical borehole heat exchanger, because the deep ground temperature is essentially constant. At a depth of 6 ft, a typical value for horizontal heat exchangers, ground temperature varies by approximately ±10°F. Even if the mean ground temperature is 60°F in late winter, ground temperature at a 6 ft depth drops to 50°F. The heat extraction process lowers the temperature even further around the heat exchanger pipes, probably by an additional 10°F or more. Even with good heat transfer to the circulating water, the entering water temperature (leaving the ground heat exchanger) is around 40°F. Lakes that freeze at the surface in the winter approach 39°F at the bottom, yielding nearly the same margin of safety against freezing of the circulating fluid. An additional 10°F temperature difference is usually needed in the heat pump’s refrigerant-to-water heat exchanger to transfer heat to the refrigerant. Having a refrigerant-to-water coil surface temperature below the freezing point of water risks growing a layer of ice on the water side of the heat exchanger. In the best case, coil icing restricts and may eventually block the flow of water and cause a shutdown. In the worst case, ice could burst the tubing in the coil and require a major service expense.

Several factors must be considered when selecting an antifreeze for a ground-loop heat exchanger; the most important are (1) effect on system life-cycle cost, (2) corrosivity, (3) leakage, (4) health risks, (5) fire risks, (6) environmental risks from spills or disposal, and (7) risk of future use (acceptability of the antifreeze over the life of the system). A study by Heinonen and Tapscott (1996) evaluating six antifreezes against these seven criteria is summarized in Table 22. No single material satisfies all criteria. Methanol and ethanol have good viscosity characteristics at low temperatures, yielding lower-than-average pumping power requirements. However, in concentrated forms they both pose a significant fire hazard. Methanol is also toxic, eliminating it from consideration in areas that require nontoxic antifreeze to be used. Propylene glycol had no major concerns, with only leakage and pumping-power requirements prompting minor concerns. Potassium acetate, calcium magnesium acetate (CMA), and urea have favorable environmental and safety performance, but are all subject to significant leakage problems, which has limited their use.