Performance Modeling Comparison of a Solar Combisystem and Solar Water Heater

Solar combisystems utilize solar thermal collectors for two residential thermal load applications: active solar thermal space heating and DHW. One key advantage of solar combisystems as compared to SWH systems is that combisystems increase the solar collector's utilization independent of occupant hot water use because the space heating is supplemented by heat collected by the solar collectors. In terms of disadvantages as compared to SWH systems, combisystems require a relatively large incremental capital investment and the design of combisystems can be intricate.

Much of the previous research on solar combisystems has focused on the optimal sizing and design of systems. Since solar thermal systems exhibit transient behavior, the most commonly used research tool to evaluate the impact of design on system performance is trnsys, a modular component-based tool originally developed by Klein et al. [1]. Specifically, trnsys was utilized to simulate the annual system performance of solar combisystems and examine the system performance as a function of the collector area and storage capacity [2]. The research of Duffie and Mitchell [2] led to the development of F-Chart, a solar thermal system analysis and design program based on correlation coefficients generated by trnsys simulations, that has the ability to quickly estimate the performance of generic solar heating systems.

Over the years, researchers have evaluated the performance of unique combisystem designs and have also evaluated the impact of sizing, collector efficiency, and draw profiles on performance of combisystems. In order to yield large energy savings from the system, the research showed the importance of small auxiliary volumes, low auxiliary set points, and good thermal stratification [3]. Stratification can be enhanced in tanks by adding heat to the top of the tank and removing heat from the bottom.

Another method for improving stratification, which has been popularized in Europe, has been to introduce stratifying tubes within the tank. A stratifying tube is an immersed tube with several outlets where the incoming water is directed into the tank at the level where the temperature is the same as the incoming water. Andersen and Furbo [4] demonstrated the impact of stratifying tubes on the thermal performance of systems. Researchers found that the thermal performance of combisystems increases by 7–14% by using stratifiers for the solar collector loop and space heating loop rather than immersed heat exchangers. Additionally, the study found that because loads vary so dramatically throughout the year with combisystems as opposed to SWHs, stratifiers are much better choice as compared to internal heat exchangers and direct inlets because stratifiers are less sensitive to varying operating temperatures.

Studies have also examined the impact of loads on the performance of combisystems. Lund [5] investigated the sizing of solar thermal combisystems with different heating loads. The study found that oversizing a solar thermal system proved to be more advantageous for less efficient buildings as compared to more efficient buildings. Jordan and Vajen [6] studied the impact of realistic load profiles on the performance modeling of combisystems, since DHW draws can have a severe impact on the temperature stratification in the tank. The study found that the fractional energy savings between models with simplified profiles and models with realistic discrete draws can differ by up to about 3%. Additionally, Bales and Persson [7] concluded that modeling systems with a realistic draw profile has a significant impact on the predicted energy savings of modeled systems. These studies concluded that an optimized combisystem design using a realistic load profile can differ significantly from a combisystem optimized using a simplified load profile.

The focus of this paper is on the impact of system size, location, and loads on the performance of solar combisystems relative to SWHs. It is expected that smaller combisystem, consisting of 6 m² of collector area, will offset most DHW loads year round (except during the winter) and will only contribute minimally to the space heating in the spring and fall when there is still heating loads and the solar collectors are operating at higher efficiencies as compared to the winter months. As collector area increases, the amount of solar energy utilized is expected to increase more rapidly in a solar combisystem configuration as compared to a SWH. However, the amount that the system will offset auxiliary space heating loads will depend on several factors such as the loads and vacation periods.

To determine how combisystems compare to SWHs in terms of energy performance, the annual performance of these systems was simulated using a trnsys model of a typical system. Figure 1 represents the combisystem that was modeled using trnsys [8]. The modeled combisystem has a single tank system with a solar-side lower heat exchanger and a space heating load-side upper heat exchanger. The DHW ports are directly connected to the tank. Several standard components from the trnsys library, such as the data reader, radiation processor, space thermal loads, and radiant floor systems, were used to develop the combisystem model [8]. Moreover, the combisystem model uses trnsys components developed for multinode storage tanks with immersed heat exchangers [9], and solar collectors with capacitance effects [10]. These two models are discussed in more detail later in this paper.

The combisystem model was validated using data from a residential combisystem installed in Carbondale, CO, which was monitored for more than 2 yrs as part of a Building America research project [11]. The combisystem used to validate the model consists of flat-plate collectors and a single tank with two immersed heat exchangers as shown in Fig. 1. The solar collector loop utilizes a glycol/water mixture, and it transfers heat to the tank through the lower immersed heat exchanger. The space heating utilizes the upper heat exchanger to transfer heat to and from the tank. The tank is pressurized and the DHW is directly heated by the tank. As part of the monitoring protocol, four water flow meters and ten thermocouples were installed to facilitate measurement of the solar and auxiliary energy to the DHW and space heating loads [11].

Figure 2 summarizes the comparative analysis between the model predictions and measurements for the heat transfer rates both at the lower heat exchanger (connect to the solar collector) and the upper exchanger (connected to the space heating load) as noted in Fig. 1. Specifically, Fig. 2 shows that most of the predicted values are within the experimental uncertainty error bars. More discussion of the experimental measurements and the validation analysis is provided by Sustar [12].

For the parametric analysis to assess the performance of combisystems, the model of Fig. 1 is utilized without any significant modifications. In particular, the storage design of the model is not modified, while possible, in order to study the optimal combisystem design. The model was based solely on the Carbondale system, which used a single tank to serve as both the solar storage and the auxiliary storage. Moreover, since radiation floor heating systems require lower supply temperatures as opposed to baseboard convectors or forced-air heating coils, the model utilized a radiation floor heating system to give the combisystem as much of an advantage as possible. Additionally, it was decided to substantially oversize the auxiliary heater capacity in the system, so that there would be minimal issues with the system not being able to meet both the space heating and DHW loads. Lastly, the auxiliary heating element was placed within the upper portion of the tank between the upper and lower inlets of the upper load-side heat exchanger. This placement allows the upper heat exchanger to utilize both solar storage (below the auxiliary heater) and auxiliary storage (above the auxiliary heater) to meet the space heating load. The DHW ports are directly connected to the tank.

For the analysis, a two-story home is modeled with a 12.8 m × 9.2 m (118 m²) footprint with a total floor area of 232 m² as shown in Fig. 3. In the cold climates, which included Denver, Boston, and Chicago, the buildings were modeled with an unconditioned basement. In the warmer climates of Atlanta, San Francisco, and Phoenix, the model simulations assume a slab-on-grade construction.

For the benchmark case, it is assumed that the building construction meets 2009 IECC code [13] and that the daily DHW usage is 227 l. The space heating setpoint is set year round at 20 °C. Other base case model values for relevant house model and solar system model parameters are given in Table 1. The cold water temperature is assumed to be the same as that of the water mains.

A couple of important modeling assumptions were made when developing the model. First, it was assumed that the heat losses from the tank do not interact with the home's thermal zone. This modeling assumption implies that the heat losses do not decrease the space heating load in the winter or increase the space cooling load in the summer. For an 820 l tank, which is the selected storage size for a 9 m², the tank losses were on the order of 2 GJ during the heating season. For a Denver benchmark house, the tank losses during the heating season are equivalent to 3% of the space heating load.

The SWH modeled utilizes the same collector and tank model parameters as the combisystem model, however, rather than having an upper heat exchanger, the space heating load is met with a separate auxiliary heat source that is connected to the radiant floor space heating loop. The control scheme for the solar loop and the auxiliary heating element are the same as the control scheme for the combisystem solar loop and auxiliary heating element. The capacity of the auxiliary heater in the SWH is set to 4.5 kW.

The study compares the auxiliary energy use of both the solar combisystem and SWH to a conventional system without solar, which consists of a 136 l auxiliary hot water tank and an auxiliary-heated radiant floor space heating system. In the conventional DHW tank, two 4.5 kW electric resistance elements heat the tank.

A flat plate solar collector with capacitance effects was used to model the flat plate collector [10]. The solar thermal collector model considers the effects of the collector mass, which includes the collector fluid, tubes, and absorber plate, on the performance of the system. Capacitance is important to model in thermal collectors since the capacitance impacts the instantaneous efficiency of the system. The model utilizes system parameters such as the transmittance–absorbance efficiency (a₀), the loss coefficient (a₁), the second order loss coefficient (a₂), and the incidence angle modifier (b₀), taken directly from the quoted collector's SRCC rating values as noted in Table 2 [14].

The total absorbed radiation (S) is a function of the incidence angle modifiers for the beam and diffuse radiation terms. The collector efficiency factor (F′) represents the ratio of the actual useful energy gain to the useful energy gain if the collector absorber was the temperature of the fluid.

For the purposes of this study, the mounting of the solar collectors in the model was held constant year round. For all the climates and systems, the mounted collector azimuth angle was set to 0°. The tilt angle of the SWH and the combisystem were set based on design guidelines presented by Ramlow [15]. For the SWH, the tilt angle is set at the site's latitude, and for the combisystem, the mounting tilt of the collectors is set at the latitude + 15 deg. The increased mounting tilt for the combisystem is implemented in order to maximize incident radiation in the winter months.

The stratified tank is modeled using the vertically cylindrical fluid-filled, constant volume storage tank with immersed heat exchangers model. In this model, the tank is divided into 18 nodes and energy balances are calculated for each section of the tank and temperatures of the each node are calculated for each time step [9]. The nodes interact thermally with nodes above and below through fluid conduction and fluid movement. The fluid in the storage tank also interacts thermally with the fluid in the immersed heat exchangers, the surrounding ambient temperature, and directly through inlet and outlet ports.

The heat exchangers are modeled as coiled tube immersed heat exchangers. The bottom heat exchanger is connected to the solar collector loop to transfer heat from the solar collector outlet to the storage tank. The top heat exchanger transfers heat to the space heating load. An inlet port at bottom node and outlet port at the top node of the tank are used for the DHW flows. Both solar side heat exchanger and load/auxiliary side heat exchanger validation studies were performed using test data in order to calibrate the heat transfer coefficients for the heat exchangers. Table 3 shows the key storage tank and heat exchanger model parameters.

Since the variability in occupant use patterns can greatly impact water heater performance, discrete DHW load profiles were generated using the Building America DHW Event Generator [16]. The Building America Event Generator is a spreadsheet-based tool that generates random DHW profiles based on probability functions that describe the overall daily average usage, event duration, event flow rate, and clustering of draws. The day-to-day standard deviation of the daily hot water draws over the course of the year is 50% of the average daily draw. The tool also accounts for vacation periods, weekend versus weekday usage variation, and the impact of seasonality on tempered loads (sinks, shower, and bath). The tool was used to develop a hot water profile for a 76 l per day, 227 l per day, and a 379 l per day household. In each of the three discrete DHW profiles generated, there is a total of 2 weeks of vacation during the year (7 days in May, 3 days in August, and 4 days in December). Figure 4 shows the annual DHW loads for the selected locations and the three DHW load profiles.

As noted earlier, all the buildings analyzed in this study are two-story, 232 m² finished floor homes. The buildings were modeled using the trnsys multizone building component. In an attempt to model several building performance types for each climate zone, the buildings are modeled according to a low-performance building type which is modeled as a 1960s retrofit house [17], a Building America benchmark performance building type which is modeled as a 2009 IECC code house [18], and high-performance building type which is modeled as a 50% source energy savings house relative to the Building America benchmark house [19]. The heating loads of the trnsys buildings were compared to equivalent building energy optimization (BEopt) simulations to ensure the trnsys heating loads are reasonable [12]. Figure 5 shows the annual space heating loads for all locations and building types.

Using the trnsys model, parametric studies were performed to study the model's sensitivity and the impact of variations in system size and system loads on the performance of the systems (see Table 4 for a summary of the parametric study). The selected cities represent a wide range of U.S. climates ranging from warm climates (Phoenix and Atlanta), mild climates (San Francisco), to cold climates (Boston, Denver, and Chicago).

where $δQload$ is a penalty value used to normalize the auxiliary energy usage, which is equal to the amount of load that the system was not able to meet over the course of the year. This penalty factor was applied to the saved energy equation because there are periods during the year where the heater capacity was not large enough to keep up with the demand of the system's water heating and space heating load. In Eq. (2), the first parenthesis term is the auxiliary energy usage for a conventional system without solar and the second parenthesis term is the auxiliary energy usage for the system with solar collectors.

The system efficiency differs from the collector efficiency in that collector efficiency evaluates the ratio of the collector's total energy input to the useful energy output, whereas the system efficiency examines the ratio of the system's saved energy as compared to a conventional system to the collector's energy input from radiation. Therefore, system efficiency takes into account the energy losses from the tank and the pipes in the solar collector loop. As compared to the collector efficiency, the system efficiency will typically be about 5–10% points lower than the solar collector's efficiency due to the fact that system efficiency takes into account the extra energy losses in the solar thermal system.

Figure 6 shows the monthly total load and the collected useful solar energy for a combisystem in Denver benchmark house with 96 ft² collector area and a 60 gal per day DHW draw. The plot shown in Fig. 6 is divided into three zones, which are labeled by numbers in the plot. Zone 1, which is the zone that covers the largest portion of the plot, is the total load in the building that exceeds the amount of useful solar energy. In a combisystem, where the system is serving both the space heating and DHW load, the time period when the load exceeds the solar resource will be during the winter months, when the load is large. During this time, the auxiliary heater is required to meet the building load. Zone 2, which is portion of the plot where the useful solar energy section overlaps the total load section, represents is the building's load that can be met by the solar energy. Zone 3 is the solar energy that exceeds the load. In a combisystem, solar energy will likely only exceed the load in the summer months, when the load is at its minimum.

Figure 7 shows the monthly total load and the collected useful solar energy for a SWH in Denver with 96 ft² collector area and a 60 gal per day DHW draw. In contrast to Fig. 6, which shows that zone 1 occupies the majority of the plot, the SWH plot shows that zone 2 occupies the majority of the plot. Zone 2, which represents is the building's load that can be met by the solar energy, means that the 96 ft² collector area SWH in Denver will yield high solar fractions. Additionally, the large SWH also means that in the majority of the months, the useful solar energy exceeds the DHW load.

Figure 8 shows the saved energy and the system efficiency for both a SWH system and combisystem for a 227 l/day DHW load in a benchmark house in Denver, CO. The performance metrics for the combisystem are depicted with a solid line, and the performance metrics for the SWH systems are depicted with the dashed line.

Figure 8 shows several key performance trends with regards to the saved energy between the combisystem and the SWH. First, as collector area increases, the saved energy in both systems also increases, although the rate of increased savings diminishes as collector area increases. Second, as collector area increases, the saved energy difference between the combisystem and the SWH system grows. This is due to the SWH energy savings being limited by a smaller load. Combisystems on the other hand serve a larger load which allows for greater energy savings as collector area increases. The third trend is that with systems with 3 m² of collector area, the combisystem offers a little to no savings as compared to the SWH.

Figure 8 also shows that the system efficiency of both the combisystem and the SWH system decreases as collector area increases, with the SWH system efficiency decreasing more quickly than the combisystem system efficiency as a function of collector area. There are a couple of important factors that go into determining the system efficiency. First, the collector efficiency plays a large role in the system's efficiency. Since collector efficiency decreases with increasing inlet fluid temperature, it is expected that the collector efficiency will decrease as collector area increases due to the fact that the higher temperature inlet fluid has a lower capacity to the heated. Smaller collector area systems have lower inlet fluid temperatures than larger system because the tanks in small systems run cooler than the tanks in larger systems. Additionally, larger systems will have increased tank losses due to high tank temperatures.

In terms of comparing the combisystem system efficiency to the SWH system efficiency, combisystem tank temperatures are lower which translates into lower tank losses and lower inlet fluid temperatures. Tank temperatures are kept lower in combisystems because they have higher loads to dump heat into as compared to SWH systems.

Figure 9 shows the impact the loads on the annual energy savings for the combisystem as compared to the SWH for a 9 m² system in Denver. The annual energy savings for the combisystems compared to the SWH systems with the same size are referred to as incremental savings. Figure 9 shows that for systems with low DHW loads, the difference between the saved energy for combisystems and SWHs will be larger than with systems with high DHW loads. Figure 9 also shows that the combisystem incremental savings are more significant in the 1960 retrofit house than in the Building America 50% house. Based on these results, a combisystem serving a house with high space heating loads and low DHW loads will yield the largest incremental savings. In contrast, a combisystem serving a high performance house and high DHW loads will yield the lowest incremental savings.

From the Denver results, it is clear that combisystems will provide incremental energy and cost savings in comparison to a standard SWH, however, the magnitude of these incremental savings is highly dependent on the system size and the loads. The parametric simulations were also performed using Typical Meteorological Year 3 (TMY3) data sets from five additional U.S. cities, which were Chicago, Boston, San Francisco, Atlanta, and Phoenix. Figure 10 shows the incremental energy savings for the three collector area sizes for homes with benchmark loads in all the locations studied. As with Denver, the incremental savings in these locations are minimal with 1 collector, but become significant with 2 and 3 collectors.

Based on Fig. 10, it is clear that the locations of Denver and San Francisco yield the highest incremental savings, largely due to their relatively significant space heating loads and relatively high incident solar radiation during the space heating months. The monthly incremental savings provided by the combisystem relative to the SWH for Denver and San Francisco are shown in Fig. 11. In particular, Fig. 11 shows that San Francisco has much larger incremental savings in April through August as compared to Denver. This is due to the higher space heating loads and the higher incident radiation in San Francisco during these months. Additionally, the incremental savings are more constant throughout the year in San Francisco as compared to Denver. A system in Denver yields significantly larger incremental savings in the spring and fall as compared to the winter.

As determined for the Denver study, the largest incremental savings from a combisystem relative to a SWH will occur when DHW loads are small and space heating loads are high and the smallest incremental savings from a combisystem relative to a SWH will occur when DHW loads are high and space heating loads are low. Based on these two bounds, the incremental site energy and cost savings from combisystems can be evaluated across all the locations.

In evaluating the economics between SWH and combisystems, both the incremental energy savings and the cost of energy will play a significant role in the incremental annual cost savings between the combisystem and the SWH. The economics of combisystems are evaluated for all-electric systems in Fig. 12 assuming that the cost of electricity is $0.10 per kWh. The costs for the combisystems are based on reported IEA data [20] while for SWH systems are based on data collected as part of the California Solar Initiative residential SWH installations [21]. The error bars represent the savings for a benchmark house and 227 l daily DHW draw and the error bars represent the lower and upper bounds of the incremental cost savings. For Denver and San Francisco, the annual incremental savings m² of collector area will be about $27 for $34 for 6 m² and 9 m² systems, respectively, for the upper bound of incremental savings. For the lower bound of incremental savings, the annual incremental savings per unit area will be less than $3 for all climates and system sizes.

A solar combisystem and SWH model was developed and used to study the incremental improved performance of a solar combisystem as a function of loads, size, and location. In Denver, the incremental savings supplied by the combisystem as compared to the SWH were 0.5%, 7%, and 18.6% for the 3, 6, and 9 m² collector are, respectively. The storage/collector ratio for all the systems analyzed was held constant at 92 l/m².

The incremental savings were highly sensitive to both the DHW loads and the space heating loads. The largest incremental savings occur when DHW loads are low and space heating loads are high, while the smallest incremental savings occur when DHW loads are high and space heating loads are low. To illustrate the impact that loads have on incremental savings, for a benchmark house in Denver with a 9 m² system, increasing the daily DHW draws from 227 l to 379 l will decrease the incremental savings by 59%, while decreasing the daily DHW draws from 227 l to 76 l will increase the incremental savings by 140%.

The combisystem performance was studied for six different climates that included Phoenix, Atlanta, San Francisco, Denver, Boston, and Chicago. The combisystem performed the best in climates where there was a significant solar resource during the heating load months. San Francisco and Denver yield 27% and 19% increased savings, respectively, as compared to a SWH with a 9 m² system that serves benchmark space heating and DHW loads.

As noted in the analysis presented in this paper, the combisystem design was not optimized. In particular, the storage size may not be adequate for the collector area. Recommended future work includes: (1) examining how the storage volume to collector area ratio can impact the annual performance of the modeled system in U.S. climates, (2) examining how replacing the solar-side heat exchanger with an inlet stratifier impacts the performance of the combisystem, (3) examining the performance between a combisystem paired with a radiant floor heating system and a combisystem paired with a forced-air heating system, and (4) investigating whether evacuated-tube collectors provide any increased savings in climates such as Boston and Chicago as compared to flat-plate collectors.

The authors acknowledge the financial support of the National Renewable Energy Laboratory (NREL) and the technical support of Greg Barker and Bob Hendron who were extremely helpful in providing critical information on the experimental data for the monitored combisystem.

A =

area (m²)

a₀ =

intercept efficiency

a₁ =

first order efficiency coefficient

a₂ =

second order efficiency coefficient

b₀ =

first order incidence angle modifier

C =

collector capacitance (kJ/ °C)

C_p =

specific heat (kJ/kg °C)

F′ =

collector efficiency factor

I_T =

tilt radiation (kJ)

ṁ =

mass flow rate (kg/hr)

Q =

energy (kJ)

T =

temperature (°C)

S =

total absorbed radiation (kJ/hr)

U_L =

loss coefficient per unit area (kJ/hr m² °C)

Subscripts

amb =

ambient

ann =

annual

aux =

auxiliary

coll =

collector

in =

inlet to collector

j =

isothermal node

no sol =

without solar

sav =

saved energy

sol =

with solar