Design Considerations In Cold Climates.
VRF systems provide heating and cooling via refrigerant piped directly from air source or water source heat pumps to fan coils in each building zone. Using sophisticated control technologies, the variable speed of VRF systems allows the amount of refrigerant sent to each zone to be modulated independently and in tune with diverse and changing space loads. The key components of a VRF system are the outdoor unit, often called a condensing unit, the indoor unit, refrigerant, and the heat recovery unit where applicable (Figure 1). The main heating and cooling plant of a VRF system is usually an air source or water source heat pump.
The main benefit of VRF systems is energy savings that can range from $0.20/[ft.sup.2] to $0.40/[ft.sup.2] ($2.15/[m.sup.2] to $4.31/[m.sup.2]) of building area. VRF systems save energy in four primary ways:
1. Distribution of heating and cooling using refrigerant instead of air.
2. Variable speed compressors and fans.
3. Eliminating the reheating of air by providing only the needed heating and cooling.
4. Recovery of heat from cooling zones to heating zones in some systems.
An additional benefit of VRF systems is their small footprint, particularly for retrofit applications with limited space for ductwork. They require less area in new construction projects, freeing up leasable square footage. Their tight temperature control also has the potential to increase occupant comfort. The main barrier to increased VRF implementation is the general perception of a cost premium versus lower-efficiency systems.
Air Source VRF
To address the key outstanding design considerations, we developed a model* that predicts the energy consumption and corresponding utility cost of an air source VRF system under a variety of scenarios. The model primarily considers the following components of the system.
Outdoor units (heating or cooling). Building loads, coupled with heating and cooling efficiency of the outdoor unit, determine the outdoor unit's electricity consumption. The heating capacity and efficiency of outdoor units are highly dependent on the ambient air conditions. Figure 2 illustrates the heating efficiency as a function of ambient dry- and wet-bulb temperature of one manufacturer's VRF unit. The shape of this curve influences the results of this analysis. Heating performance is impacted by humidity--or wet-bulb temperature--due to defrost requirements between 0[degrees]F and 35[degrees]F (-17.8[degrees]C to 1.7[degrees]C). This curve shows the ratio of actual heating efficiency to the rated heating efficiency. For the majority of the heating season in a cold climate, the unit operates approximately 30% below its rated heating efficiency. However, at very cold temperatures (less than -10[degrees]F [-23.3 [degrees]C]), the ratio rapidly decreases, resulting in significant heating efficiency degradation. Capacity degrades in a similar fashion.
Supplemental heating. One approach to minimizing the effect of low ambient temperatures on VRF performance is to put the outdoor unit in a semi-enclosed, semi-heated mechanical room. A supplemental heating device maintains a temperature setpoint in the mechanical room, preventing the VRF unit performance from degrading significantly. A trade-off exists between the increased performance of the VRF unit and the additional energy consumption of the supplemental heater.
The mechanical room is semi-enclosed because it must be able to use the ambient air as a heat source/sink when temperatures are warmer or if it is in cooling mode. We analyzed a few strategies for controlling this enclosure:
* Fixed louvers. Louvers on the wall of the mechanical room are always open, regardless of the outdoor air dry-bulb temperature, allowing air to flow freely to the outdoor units.
* Operable louvers. Louvers are open and air flows freely to the outdoor units when the outdoor air dry-bulb temperature is higher than the mechanical room temperature setpoint. When the outdoor air dry-bulb temperature is lower than the mechanical room temperature setpoint, the louvers are closed.
* Fan assisted. A supplementary fan is used to deliver enough air to the outdoor unit when the outdoor air temperature is higher than the mechanical room temperature setpoint. When the outdoor air dry-bulb temperature is lower than the setpoint, the fan is off. This strategy has associated fan energy consumption.
We used the model to explore many different design decisions and their impacts on energy and cost.
Our base case for these design considerations was a 100,000 [ft.sup.2] (9290 [m.sup.2]) office building in Madison, Wis.
MECHANICAL ROOM CONFIGURATION
Our analysis indicates that the optimal mechanical room configuration is one with natural gas-fired supplemental heat and operable louvers. Table 1 shows the results for the three strategies described, as well as for a strategy with no supplemental heat. Note that these results use a mechanical room setpoint of -10[degrees]F (-23.3[degrees]C), which is just above the point at which this VRF equipment heating performance severely drops off.
The operable louver has the lowest annual utility cost and EUI of the three strategies described, followed closely by the fixed louver case. Also included in Table 1 is the associated supplemental heat capacity. The fixed louver case requires a higher supplemental heat capacity than the other cases to accommodate the high airflow through the louvers when the outdoor dry bulb is below the optimal mechanical room setpoint.
We also determined the impact of supplemental fuel type on system performance. Using gas-fired supplemental heat results in lower annual utility costs, due to the relative low cost of natural gas.
OPTIMAL MECHANICAL ROOM TEMPERATURE SETPOINT
Our analysis indicates that the mechanical room set-point for optimal energy performance occurs at a temperature close to but slightly higher than the temperature at which the VRF system's heating efficiency and capacity are significantly reduced.
Assuming some amount of supplemental heat is provided, an important consideration for the design engineer is the optimal mechanical room setpoint. This is the air temperature at which the supplemental device will activate, the louvers will close, and/or the fan assist will turn off. Figure 3 shows the impact of this control setpoint on operating cost. The optimum setpoint for this piece of VRF equipment falls between -15[degrees]F and 5[degrees]F (-26.1[degrees]C and -17.8 [degrees]C). The designer should select the system in this range that has the lowest first cost and potential maintenance costs considering the size of the unit heater and outdoor units.
The most significant factor influencing the mechanical room setpoint for optimal energy performance is the VRF unit's heating performance curve. The optimal setpoint would be higher or lower based on the specific heating efficiency curve for different VRF models and manufacturers.
Defrost has a significant impact on the operation and energy use of any air source heat pump, including air source VRF units. When these units are in heating mode and there is sufficient moisture in the air, it will condense and freeze on the outside of the unit coils.
VRF systems generally use a temporary reverse cycle for defrost that runs the unit In reverse to warm the outdoor coil and melt the frost.
There are two common methods for controlling reverse cycling. The simplest method is timed defrost that runs a defrost cycle for a set portion of each hour whenever the outdoor wet-bulb temperature is conducive to frost buildup. But timed defrost based only on wet bulb tends to overuse the reverse cycle, resulting in higher than needed energy consumption.
Demand defrost is a more sophisticated approach that uses a sensor to detect frost buildup and only operates the defrost cycle when necessary. This impacts the unit efficiency as shown in Figure 2. Our analysis showed that using a demand defrost approach as opposed to a timed approach is about 4% of the resulting VRF energy savings.
Our assumptions and analysis suggest it is optimal to design mechanical rooms to accommodate freezing temperatures. Therefore, the design engineer also must consider handling the melting frost at subfreezing temperatures to avoid ice throughout the mechanical room. A hard connected, insulated drain directly below the outdoor unit equipment is likely the best solution; in any case large amounts of heat trace should be avoided.
Another source of condensate may be condensing heaters to increase the efficiency of supplemental heat. For most climates, there are relatively few hours below the optimal mechanical room temperature setpoint, so increasing the efficiency of supplemental heaters does not have a large impact and may not be worth the additional cost and maintenance that comes with handling condensate.
NOTES ON SIZING OF VRF UNITS FOR HEATING
For this analysis, our primary focus was configuring and controlling outdoor units and supplemental equipment for air source VRF in a cold climate, with consideration for energy costs. But these design decisions also impact required equipment capacity, which affects first cost as well.
For our base case, which has a peak heating load of 2,400 MBtu/h (700 kW), a gas-fired supplemental device should be approximately 1,450 MBtu/h (430 kW). (Both numbers assume a 25% safety factor for heating.) This equates to 15 Btu/h x ft.sup.2] (47 W/[m.sup.2]) for our base building. Note that the electrical power to the compressors turns into heat, creating the difference between unit heater size and heating load.
For the same base case (peak heating load of 2,400 MBtu/h [700 kW]), the outdoor units must have a nominal capacity of 3,600 MBtu/h (1050 kW) to meet the load (due to derating at low ambient temperatures) if we maintain the ambient air to the compressor at 5[degrees]F (-15[degrees]C) using a supplemental heater. If no supplemental unit heater is installed and the unit must be able to operate at the -11[degrees]F ([dagger]) (-24[degrees]C) design condition in Madison, the nominal capacity must increase to 4,000 MBtu/h (1170 kW) if the designer chooses to only meet the design condition. This represents a 10% increase in equipment size. Sizing results are summarized in Table 2.
Water Source VRF
Water source VRF is an option to avoid exposing equipment to cold outdoor temperatures. In this system the condensing units are served by a water loop and can be located entirely inside the building, with a water-to-refrigerant coil on the water side. The heat source and sink for the water loop can be a boiler and fluid cooler, a ground heat exchanger, or any fluid loop in the building that is able to share heat. This approach eliminates some of the design challenges and maintenance issues associated with outdoor units in air source VRF systems in cold climates.
The water loop and heat source/sink add up-front costs, but they can lower maintenance and energy costs in cold climates. Adding a ground heat exchanger allows for additional energy savings by replacing the energy use of boiler and cooler equipment with a moderate increase in pumping energy.
To better understand potential outcomes of choosing the water source approach, we monitored and investigate two ground source VRF systems in the upper Midwest, (1) One building was a three-story, 70,000 [ft.sup.2] (6500 [m.sup.2]) multitenant office building in Madison, Wis,, with seven tenants, The building developer/owner had experience with air source VRF in this climate and chose water source for the project to lessen maintenance and control requirements and improve energy performance.
The other building, located in Minneapolis, was a 5,500 [ft.sup.2] (510 [m.sup.2]) portion of an office attached to a service facility, The building owner is in the geothermal business and wanted to test the new ground source VRF approach.
The ground heat exchangers in the two buildings provided moderate fluid temperatures to the condensing units as expected, The loop in the Madison building maintained a temperature of 40[degrees]F to 80[degrees]F (4[degrees]C to 27[degrees]C) throughout the year, while the loop in the Minneapolis office maintained a temperature of 50[degrees]F to 90[degrees]F (10[degrees]C to 32[degrees]C). Both temperature ranges are far above the low ambient temperatures experienced by air source systems in winter, In addition to measuring temperature, we measured fluid flow rate and power consumption in both buildings to determine operating EER in cooling and COP in heating, When conditions matched AHRI-rated entering fluid temperature and flow rate, with near-AHRI indoor temperature, the systems performed as shown in Table 3.
The Madison system was close to its rated performance at AHRI conditions, just slightly below in cooling and slightly above in heating, The Minneapolis system fell quite short in cooling, However, the primary energy efficiency benefit of VRF systems is not a function of peak rated performance, Rather it is from overall seasonal performance measured by Integrated Energy Efficiency Ratio (IEER), In this regard, both systems fell short of expectations to some extent, Figure 4 shows the operating COP in heating and cooling at various part loads for the Madison system, The lab-tested IEER rating for the system is shown for comparison, The expected increase in performance at part load does not appear to be fully realized for this system.
Our analysis above is mostly limited to heating-only and cooling-only operation of the system (based on unit status during measurement), VRF systems are capable of heat recovery during simultaneous heating and cooling that can increase efficiency, The Madison office exhibited approximately 0.1 kWh/[ft.sup.2] (1.1 kWh/[m.sup.2]) in additional energy savings potential from heat recovery, Direct heat recovery is only possible when heating and cooling zones are both connected to the same heat recovery unit in the system (Figure 1), Analysis of the design and coincident heating/cooling in the Madison building suggests that roughly half of this energy is directly recovered in a typical year due to limited connectivity in the system.
In addition to the documented EERs, which include compressor and indoor fan energy (about 73% and 7% of the HVAC energy, respectively), total HVAC energy includes the pump energy (19% of HVAC energy) required to move the water and a dedicated outdoor air unit to ventilate the space, At this level of total system energy, we compared water source VRF to other system types by using energy models. This approximation is necessary since direct comparison of two different HVAC systems in the same real building is impossible.
The VRF energy model is closely calibrated to the measured data from each site including VRF unit EER, component energy use, flow rates, and temperatures at various points. The HVAC system was then swapped in the model for other system options. The results of this comparison are shown in Table 4, with modeling assumptions for the different systems documented in column two. The system labeled "VAV, Common Deficiencies," however, reflects issues we typically observe in VAV systems, including higher minimum flows, imperfect temperature reset, and increased fan power. ([double dagger]) This allows a more representative comparison with the real-world VRF performance.
The ground source VRF system achieves a site EUI of about 35 kBtu/[ft.sup.2] x yr (0.40 GJ/[m.sup.2] x yr) --significantly better than typical buildings--and has the lowest operating cost of any system other than the ground source heat pump (GSHP). The GSHP saves $0.07/[ft.sup.2] ($0.75/[m.sup.2]) more than the ground source VRF, primarily because of the poorer part-load performance we observed in the VRF equipment, (1) though the GSHP system is not based on field measurement, and we cannot prove the two-stage GSHPs would achieve that rated performance. The ground source VRF system performs slightly better than the water source VRF and water source heat pump systems. This suggests two things: the addition of the geothermal borefield may have a longer payback than the switch to VRF alone; and a water source heat pump system performs similarly to VRF (assuming quality heat pumps that match their rated performance).
All of the Minneapolis office comparisons suggest that this particular ground source VRF system is not performing as would be expected. With the VRF model calibrated to the poor level of performance we measured, it is performing worse than even the typical VAV system performance, in terms of both cost and emissions.
In the Midwest where our field work took place, VRF generally replaces systems with gas-fired heating such as the VAV system we compared it to that has gas-fired hot water reheat. Therefore, the cost savings from VRF depends on gas and electricity prices. For example, the $0.18/[ft.sup.2] savings from VRF in the Madison example changes to $0.21/[ft.sup.2] ($2.30/[m.sup.2]) if electric rates are 30% lower, as they are in some areas of the country. On the other hand, if gas prices drop by 30%, the savings changes to $0.11/[ft.sup.2] ($1.20/[m.sup.2]).
The results suggest that water source VRF performance is dependent on many factors. It can outperform other systems under certain conditions and perform worse under other conditions. Our ability to make comparisons is limited to theoretical models or only two data points. Substantial additional data will need to be added to these two data points before an accurate judgement can be made on relative performance of water source VRF to other system types.
Other Lessons Learned
* Sizing condensing units for a space before tenants are known can be a challenge, Consider oversizing the mechanical room a bit, but wait to actually select and purchase the condensing units until tenants are known.
* Consider the non-energy benefits of VRF systems, Because of their small footprint, first costs for some construction projects can be substantially reduced by decreasing the floor-to-floor heights, providing flexibility, and increasing rentable square footage.
* Keep refrigerant piping runs as short as possible to prevent losses in capacity, Manufacturers will provide specific limits to the length and rise of refrigerant piping, Install condensing units in multiple locations to help meet these constraints.
* Have commissioning personnel investigate any potential for simultaneous heating and cooling between adjacent zones, or oscillation between heating and cooling within zones, This may require additional time at the site for observation and tuning of thermostat control loops.
Several areas of investigation remain for cold climate VRF, More field research is needed to validate how air source systems actually perform in cold climates, A project that measures energy use and other performance parameters from several air source systems would be best, Other studies have shown that actual field performance of such complex systems is often different than theory, More measurement is needed for water source systems, too.
The water source site findings published here are two of just three systems (2) studied and published to date, which cannot be considered definitive, Ideally, further study would also include life-cycle cost economics.
Another option to manage the mechanical room set-point is a dual temperature control scheme, In this scheme, whenever the outdoor temperature drops below the setpoint, a supplemental device would be used to maintain the temperature at the outdoor units well above the ambient temperature to provide improved efficiency. The trade-off between improved unit performance and mechanical room losses at high AT (to the ambient) should be addressed in future research.
Other sources of supplemental heat could also be investigated, One suggested approach has been a hybrid system that would allow the outdoor unit to use a water source heat exchanger below a certain temperature set-point instead of very cold ambient air.
The authors acknowledge Madison Gas and Electric for contributing funding, time, and information to make this project possible.
(1.) Hackel, S, 2015, "Performance of Water-Source Variable Refrigerant Flow: Measurement and Verification of Two Installed Systems," Seventhwave.
(2.) Piljae, I., X, Liu, 2014, "Case Study for ARRA-Funded Ground-Source Heat Pump (GSHP) Demonstration at Oakland University," Oak Ridge National Lab.
Swanson, G., C, Carlson, 2015, "Performance and Energy Savings of Variable Refrigerant Technology in Cold Weather Climates: Conservation Applied Research and Development Final Report," Prepared for Minnesota Department of Commerce, Division of Energy Resources.
Scott Sohuetter, P.E., is a senior engineer, and Scott Haokel, PE,, is a principal engineer at Seventhwave in Madison, Wis.
* For more detailed description of the model and results of this analysis, please see seventhwave.org/cold-climate-vrf.
([dagger]) -11[degrees]F (-24[degrees]C) is selected here, as it is the heating design condition in Madison, Wis. For sizing purposes, we felt this was an appropriate temperature to use. It was chosen completely separate/distinct from the previous analysis concerning mechanical room configuration and optimal mechanical room temperature setpoint.
([double dagger]) Our research shows that it would be inaccurate to assume that VAV systems operate according to code. We applied our typical field-measured findings for this comparison. More details on the methodology we used for this comparison are provided in Hackel. (1)
Caption: FIGURE 1 Variable refrigerant flow system, Courtesy of Cherie Williams, Seventhwave
Caption: FIGURE 2 Heating efficiency as a function of mechanical room dry-bulb temperature; the rated condition reflects AHRI testing conditions.
Caption: FIGURE 3 The energy cost of operating the system, by unit, as a function of mechanical room setpoint.
Caption: FIGURE 4 System performance, in COP, as a function of part load for the Madison, Wis.-based VRF system.
TABLE 1 Annual utility cost, EUI, and supplemental heat capacity for different mechanical room configurations. SUPPLEMENTAL HEAT ANNUAL EUI * SUPPLEMENTAL CONFIGURATION UTILITY KBTU/ HEAT COST, $/ [FT.sup.2] CAPACITY, [FT.sup.2] BTU/H x [FT.sup.2] No Supplemental $0.43 11.3 n/a Supplemental, Fixed $0.40 12.0 14.3 Supplemental, Operable $0.39 11.7 9.60 Supplemental, Fan Assist $0.44 13.6 9.50 * Within this article, energy use index is defined as site, as op- posed to source. TABLE 2 Nominal equipment sizes required to serve base case office building in Madison, Wis., for two different mechanical room temperatures. Units are shown in both MBtu/h and tons for convenience. MECHANICAL OUTDOOR UNIT SUPPLEMENTAL ROOM NOMINAL HEATING UNIT SIZE TEMPERATURE CAPACITY SETPOINT, MBtu/h [degrees]F MBtu/h Tons 5 3,600 300 1,450 -11 4,000 330 n/a TABLE 3 EER and COP for the Madison and Minneapolis offices at AHRI conditions. Not enough data was available in heating near- AHRI conditions (Minneapolis) to show a heating COP value. MADISON MINNEAPOLIS * HEATING COP COOLING EER COOLING EER BTU/WH BTU/WH Manufacturer 3.1 13.0 13.0 (AHRI) Rating Measured 3.3 11.9 8.5 Performance * Due to the nature of the office served by VRF in Minneaplis, there were not enough purely heating loads to get a representative measurement of the heating performance at rated conditions. TABLE 4 Energy performance of wafer source VRF in relation to other HVAC system choices, based on calibrated simulation models. Energy performance accounts for all end uses, including HVAC as well as lighting and other loads (all of which are kept constant, except for HVAC). MADISON ANNUAL ENERGY USE Model COST SITE SOURCE Assumptions $/ KBTU/ KBTU/ [[FT.sup [[FT.sup [[FT.sup .2]] .2]] .2]] VAV, Code Standard 1.06 56 109 Baseline 90.1-2007 Appendix G; Hot Water Reheat VAV, Standard 1.15 62 120 Common 90.1-2007 Deficiencies * Appendix G; Hot Water Reheat, but With More Typical Minimum Flows, Temperature Resets, and Fan Power ws vrf Identical 0.99 36 103 to GS VRF System, but served by Standard 90.1-2007 Boiler and Fluid Cooler gs vrf Calibrated to 0.97 35 100 Actual System Measured wshp Standard 0.98 38 100 90.1-2007 Compliant WSHP System gshp Standard 0.90 31 90 90.1-2007 Compliant GSHP System, but With Two-Stage Heat Pumps MINNEAPOLIS ANNUAL ENERGY USE Model COST SITE SOURCE Assumptions $/ KBTU/ KBTU/ [[FT.sup [[FT.sup [[FT.sup .2]] .2]] .2]] VAV, Code Standard 1.14 54 127 Baseline 90.1-2007 Appendix G; Hot Water Reheat VAV, Standard 1.22 60 136 Common 90.1-2007 Deficiencies * Appendix G; Hot Water Reheat, but With More Typical Minimum Flows, Temperature Resets, and Fan Power ws vrf Identical 1.28 50 136 to GS VRF System, but served by Standard 90.1-2007 Boiler and Fluid Cooler gs vrf Calibrated to 1.23 49 132 Actual System Measured wshp Standard 90.1-2007 Compliant WSHP System gshp Standard 90.1-2007 Compliant GSHP System, but With Two-Stage Heat Pumps # = Idealized, code-based models * The current local energy code for both of the sites we measured was based on this version of 90.1.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Variable Refrigerant Flow Systems|
|Author:||Schuetter, Scott; Hackel, Scott|
|Date:||Jan 1, 2017|
|Previous Article:||HVAC UV Germicidal Irradiation UV-C Fixtures.|
|Next Article:||Optimized Energy Recovery.|