1、 James D. Lutz is a consultant in Oakland, CA. Getting Heat Pump Water Heaters Into California James D. Lutz, PE Member ASHRAE ABSTRACT Californias single-family residential building stock is dominated by gas-fired storage water heaters. This is a result of the building energy efficiency code. For d
2、ecades the water heater energy consumption calculated in the budget compliance tools has made it hard to justify using an electric resistance water heater. As a result very few electric water heaters have been installed. The compliance tools were written in a way that poorly calculates the hot water
3、 load and the calculated time of water heater energy use does not account for the buffering effect of a storage tank. An important part of the budget calculation uses a time dependent valuation of electrical generation to capture the societal costs of using electricity for every hour of a typical ye
4、ar. Meaning the actual effects of the time difference between the hot water use and the energy consumption of electric storage water heaters are not being evaluated properly. The combined effect of these oversights has inadvertently blocked adoption of heat pump water heaters in new construction in
5、California, a major obstacle for reaching the states net-zero energy and greenhouse gas emission targets. This paper describes the way the building energy efficiency code currently calculates the water heater energy budget. Problems in the calculation procedure are explained. Our knowledge about res
6、idential hot water systems has increased greatly in recent years, significantly improving our ability to characterize these systems. Revisions to the building code calculations are suggested based this increased knowledge. An enhanced hot water load calculator has recently been adopted by RESNET. De
7、tailed field studies over the past several years of residential hot water draw patterns provide a source for more realistic draw schedules to use in the calculations. An open source water heat simulation model developed for utility incentive programs in the Northwest could be adapted to calculate th
8、e amount and timing of energy use. The role of demand response controls to reduce the impact of electric heat pump water heaters on the grid are also discussed. INTRODUCTION The California building energy efficiency code, colloquially known as Title 24 after the section number in the California Code
9、 of Regulations, sets minimum energy efficiency requirements for new and retrofit buildings in California. It is an asset rating calculation which compares the energy budget of a proposed design against the energy budget of an equivalent basecase building. The energy used to heat water used in the b
10、uilding is included in that budget. The building energy simulation models used to develop the efficiency code and to certify compliance with the code is described in the Alternative Calculation Method (ACM 2013). A clearly defined method is necessary to assure consistency across all buildings. The w
11、ay the people who will eventually live in the house use energy and the future weather obviously cannot be known in advance, therefore standardized assumptions are made about typical use and weather. “Overall hot water system performance is dependent upon the complex interactions between the househol
12、d occupants behavior, plumbing configuration, climate, and water heater characteristics” (Backman and Hoeschele 2013). This makes modeling and calculating the energy used to heat that water difficult. The current water heating module in the ACM is based on calculations initially developed in the ear
13、ly 1990s (ACM 1992). At that time water heater annual consumption calculations were done as a side calculation separate from the rest of the building energy simulation. The hot water load scales with conditioned floor area up to 2500 square feet (232.2 m2). Above that the daily hot water load is con
14、stant. The impact of plumbing configurations other than the baseline assumption of trunk and branch, such as home run and recirculation systems are approximated by distribution system multipliers, as is pipe insulation. In the early 2000s the calculations were converted to hourly using average daily
15、 hot water draw patterns from ASHRAE (Thrasher et al. 1990). There have been several subsequent modifications, but the basic calculation methodology for single-family homes remains essentially the same. In the time since the original water heating energy algorithms were adopted there have been a ple
16、thora of research projects on hot water use, hot water distribution systems (HWDS) and water heaters. Many new tools and simulation models have been developed. This paper suggests ways to combine the results of those research projects to better evaluate hot water system energy performance in the ACM
17、. An important part of the budget calculation is a Time Dependent Valuation (TDV) of electrical generation to capture the societal costs of using electricity for every hour of a typical year. An hourly energy calculation is necessary to appropriately apply the TDVs that are used to evaluate whether
18、the proposed building is in compliance with the regulations (Price et al. 2011). The existing water heating energy calculations assume the energy used to heat water is consumed the same hour the hot water is drawn. While this is true for tankless water heaters, the energy to heat the cold water ente
19、ring in the tank of storage-type water heater is used only after hot water has been drawn. Depending on the storage volume and controls, it could be several hours after the hot water is drawn before energy is consumed. An example of this is a large volume, electric heat pump water heater (HPWH). Cur
20、rently 88% of Californias single-family homes heat water with natural gas (RASS 2009). As more electricity is generated from renewable sources, meeting the states goals of reducing greenhouse gas emissions may require water heating be electrified. Correctly calculating the electricity use of this ty
21、pe of water heater will be essential to the success of this effort. The method proposed here to calculate hot water energy use for rating a house consists of several parts. The first part is to determine the number of people who would be living in the house. Then the incoming cold water mains temper
22、ature is determined from the weather at the location of the house. The next part is to combine the characteristics and location of the house along with information about the people living in the house to calculate an average daily hot water load. The daily hot water draw patterns are determined once
23、 the daily hot water load and the number of people in the house is known. These detailed daily hot water draw patterns are then used by the water heater simulation model. The final part is the calculation of hourly water heater energy use by the water heating simulation model. A schematic flowchart
24、of the hot water energy calculation is shown in figure 1. A detailed description and suggested algorithms for each part are included in the discussions below. Figure 1. Schematic flowchart of hot water energy calculations. It would be good to have these calculations as separate algorithms. This will
25、 make the energy calculation scheme easier to understand and would allow those algorithms to be updated as conditions change or as better data becomes available. People Hot water is used by the people in the house. Therefore it is important to have a consistent algorithm to select the number of peop
26、le in the house. Hot water use correlates much better with the number of people by age group than with the total number of people in the house. Anecdotal stories of teenagers taking long showers are supported by field data (Fairey et al. 2015). However for an asset rating method, such as this, the s
27、pecific people in a house are not known. A sample of California houses in surveys such as RECS could be used to develop regression equations to correlate the number and age of people in newly built California homes with features such as conditioned floor area, number of bedrooms, number of bathrooms
28、, etc (RECS 2009). Currently the appliances, miscellaneous energy use and internal gains section of the ACM assumes the number of people is 1.75 + 0.4 * NumberOfBedrooms (ACM 2013, 136). The RESNET daily hot water load calculator uses 1.09 + 0.54 * NumberOfBedrooms (RESNET 2014). Mains Temperature T
29、he existing water mains temperature calculation uses the daily average air temperature from the typical meteorological years used for the house. Other calculators are available. The RESNET amendment uses a mains water temperature calculator from a paper by Burch and Christensen (Burch and Christense
30、n 2007). The NEEA HPWH Model Validation Study uses a rolling average air temperature of several prior weeks (Ecotope 2015). The accuracy of these mains water temperature calculator should be checked against available data from field monitoring projects in California (Kosar et al. 2012). If either of
31、 these calculators are more accurate than the current algorithm, they should be used instead. Daily Hot Water Load Daily hot water use is highly variable both between residences and from day to day within the same residence. For asset rating calculations, a consistent daily hot water load must be us
32、ed. For this reason the average daily hot water load should be calculated. Average daily hot water load can be calculated based on the people, the plumbing configuration and fixtures in the house and the incoming cold mains water temperature. Because the mains temperature varies so much by time of y
33、ear, the average daily hot water load should be calculated for each month of the year. The best documented daily hot water load calculator is a recent amendment to the RESNET standard for calculating the HERS index (RESNET 2014). This algorithm sums the hot water loads due to people, dishwashers and
34、 clothes washers. The structural and behavioral waste due to the plumbing configuration is also calculated. The RESNET amendment uses the method developed in Fairey et al. to normalize the occupancy by mains water temperature (Fairey et al. 2015). This normalized occupancy number is then used to cal
35、culate the daily fixture water use and separately the hot water waste due to distribution system losses. Daily Hot Water Draw Patterns A hot water draw pattern is the record of the timing, duration, and flow rate of hot water draws at the water heater over a day. As with the daily hot water loads, t
36、here is significant variation in hot water draw patterns among households and from day-to-day (Lutz and Melody 2012). Hot water use within a single home is composed of clusters of end use draws grouped together with many hours of no hot water use. The hot water is delivered to the fittings and fixtu
37、res for those draws through a hot water distribution system. The configuration of the pipes can have a significant impact on the duration of the draws, especially waiting for hot water at showers. Several computer models capable of simulating HWDSs have recently been developed (Lutz et al. 2013). Ho
38、wever at this time it is probably not appropriate to attempt detailed modeling of HWDSs in Title 24. The configuration of plumbing systems is not typically specified in the design documents for single family homes. It would be better to directly develop the daily hot water draw patterns at the water
39、 heater based on the number of occupants, the mains water temperature at the house and general information about the HWDS in the house. Given the same daily total amount of hot water use, the exact timing and volume of individual draws can dramatically change equipment efficiency. This is particular
40、ly true for HPWHs. For a water heater simulation model to produce reasonable performance estimates, it is important to develop draw profiles that reflect typical hot water draw patterns. Several typical hot water draw patterns and draw pattern generators are available that could be adopted. The Ecot
41、ope HPWH model validation study created typical weekly draw profiles with one minute time-steps rooted in field measurements collected in the Northwest (Ecotope 2015). The patterns were derived using an algorithm that accounts for the number of small, medium and large draws clustered into windows of
42、 activity spread across a day. The houses were analyzed by groups based on number of occupants. These typical weekly draw patterns were derived from field data collected in the Northwest and are not directly applicable to California. However they could be adjusted to account for varying mains temper
43、atures and monthly average daily hot water loads. Another means of generating hot water draw patterns is to use the Building America DHW Event Schedule Generator (Hendron et al. 2010). That generator uses a Monte Carlo approach to generate specific end use draws. The timing, duration and flow rates
44、of the different events are randomly chosen according to probabilities derived from other studies. The DHW event schedule generator was used to generate standard annual draw patterns in six second time-steps for houses by number of bedrooms. These standard annual draw patterns could be adjusted base
45、d on user inputs for climate, water heater tank temperature, and mixed use water temperature. The draw patterns in the new Department of Energy Uniform Energy Factor test procedure could also be used as generic patterns. An advantage of this is that the draw patterns in the test procedure are the on
46、es used for rating the efficiency of water heaters. (DOE 2014). Because mains water temperature varies seasonally, the average daily hot water load also varies seasonally. To match the seasonally varying daily hot water load with the same daily hot water draw pattern, either the flow rate or the dur
47、ation of the individual draws must be changed. The optimal number of daily hot water draw patterns still must be determined. Because the calculations are used for an asset rating perhaps only one draw pattern is sufficient. A substantial difference in the timing of draws has consistently been observ
48、ed between weekday and weekends. Perhaps limiting draw patterns to one weekday and one weekend pattern is sufficient for single-family homes. When more than one dwelling unit is served by the same water heater the diversity of draw patterns becomes more important. If every dwelling unit had the same
49、 draw pattern then as more are served by the same water heater, individual draws in the pattern would be artificially amplified. From field studies it is clear that draw patterns are not consistent from unit to unit. A consistent way of incrementally time shifting each additional draw pattern may be the best way of modeling this diversity. Water Heater Simulation To accurately predict hourly energy consumption by the water heater given a draw pattern and environmental conditions, a simulation model is necessary. The simulation model needs to be accurate, while simple enough to use for com
