This article was originally published in the July/August 1996 issue of Home Energy Magazine. Some formatting inconsistencies may be evident in older archive content.


| Back to Contents Page | Home Energy Index | About Home Energy |
| Home Energy Home Page | Back Issues of Home Energy | EREN Home Page |


Home Energy Magazine Online July/August 1996


Try These On for Size: 
New Guidelines for Multifamily Water Heating

by Fredric S. Goldner

Fredric S. Goldner, C.E.M., is principal of Energy Management and Research Associates in Brooklyn, New York. He is the author of the new 1995 ASHRAE guidelines discussed in this article.


ASHRAE has published new sizing guidelines for hot water systems, based on recent studies of water-heating use in apartment buildings. If adopted in practice, the new sizing method should prevent the costly oversizing that is now common.

The 60-unit building in the photo above houses middle-income families. As part of a study conducted in New York City, researchers monitored the building's hot water consumption, which fell into the medium usage category under the 1995 ASHRAE guidelines.

Energy professionals have long been frustrated by the lack of reliable data for sizing domestic hot water (DHW) equipment in multifamily buildings. To be on the safe side, many designers oversize the equipment, resulting in systems with higher equipment costs, operating costs, and energy use. Now ASHRAE has incorporated data from recent studies into new guidelines for properly sizing DHW systems.

Using previous (pre-1995) ASHRAE guidelines resulted in serious undersizing (see Figure 1). In practice, however, DHW systems and combined heating/DHW boilers are often oversized by 30%-200%, according to the New York City Department of Housing Preservation and Development, Energy Conservation Division. Discussions with designers in other parts of the country revealed similar oversizing.

What happens is that the individual responsible for installing a boiler will often size it with a what was there before, looks like ..., or other rule-of-thumb method. Even when they do try to calculate the loads, designers use enormous safety factors because they know the DHW demands estimated with the old methods tend to undersize (see Evolution of an Oversizing Rule). The safety factors cause considerable oversizing even when the space heating portion is calculated properly, which is rarely the case. I've seen factors that double the size of the boiler relative to the space heating load (a rule of thumb that is particularly inaccurate for the New York climate).

Evolution of an Oversizing Rule

A review of manufacturers' literature uncovered at least half a dozen different methods for sizing both combined heat/DHW units and stand-alone systems. 

Many of these methods were initially based on the pre-1995 ASHRAE approach. I once sat down with the VP of marketing and one of the design engineers for a prominent manufacturer and asked them how the data sheets in their catalog determine system size. They replied the ASHRAE Handbook method. After running some calculations, we found that in fact their results were somewhere between two to three times greater than the results obtained using the Handbook method. 

What probably happened was that the engineer who had written the sizing sheets (many years ago) started with the Handbook values as a base. But from his experience, he recognized that the numbers were not sufficient to meet a building's demand, so he added a safety factor based on that experience. Subsequently, as the catalog has been revised, each engineer given the responsibility to update the sheets has said to himself or herself, Well, I'm not going to be responsible for there not being enough hot water in a building and has added another safety factor on top of the previous one. And then the chief engineer in charge of the revision says, I'm not going to be responsible for there not being enough hot water . . . and adds yet another safety factor. Thus over time these values have in some cases become grossly inflated.

To aggravate this already bad situation, the contractor on the job may look at the data sheets and say, Well I'm not . . . and add another level of so-called safety factor. The job then gets sized out and a call is made to the warehouse, whose staff, feeling like all the other parties, applies the next-size-up approach before sending the heater/boiler out to the job site.

Figure 1. Comparison of monitored data to hot-water usage calculated with values from the 1991 ASHRAE HVAC Applications Handbook Table 7. For these New York City buildings, using the 1991 ASHRAE guidelines would have resulted in severely undersized equipment. 
The 1995 ASHRAE Guidelines The new guidelines update the previous ASHRAE hot-water demand values. In part, the new values reflect changes in numbers of water-consuming devices, personal habits, lifestyles, and sanitation needs since the late 1960s, when the previous values were determined. In addition, sophisticated computers and monitoring equipment have enabled us to gather more extensive data on which to base sizing criteria.

The 1995 guidelines also take a new approach. Rather than a single value for volume of water used per apartment, they offer a range of values for different types of users. The residents or likely residents of a building are separated by their demographic characteristics into three usage categories: low, medium, or high (LMH). Additionally, the usage factors are provided per capita rather than per apartment. This reflects the fact that people, not apartments or square footage, use water.

To help in the design process, the new ASHRAE tables give more detailed levels of consumption for the peak 5 minutes and the peak 15 minutes (the old tables had only 60-minute peak values). These values more closely represent the instantaneous demand peak that a building will experience.

Using the New Method The first step in calculating DHW demand is to determine the demographic profile of the project and building occupants. Different types of building occupants consume hot water in fairly predictable patterns. Users can be lumped into one of the three typical LMH categories of water consumers.

Table 1 lists a variety of occupant classifications, one or a combination of which should describe the occupants of any particular multifamily building. For example, a luxury condominium in an area inhabited predominantly by young couples will tend to fall into the all occupants work category of low anticipated water consumption. By contrast, a low-income housing project will generally fall somewhere between the low income and no occupants work categories of high-volume water consumption. An abundance of hot-water-consuming appliances, such as clothes washers or dishwashers, will tend to increase hot-water consumption. If the condominium building example above intended, or allowed, the future installation of a clothes washer in each unit, the demographic category should be augmented from low to medium. It is up to the system's designer to determine this category.

Once this LMH factor has been determined, values for hot-water consumption can be selected from Table 2. Values are indicated per capita in peak or maximum flows of 5 minutes, 15 minutes, one hour, two hours, three hours and one day, as well as average daily flow. From these values, anticipated demand can be determined for the estimated maximum building population.

Table 1. Demographic Characteristics Correlation to DHW Consumption
Demographic characteristics Usage factor
No occupants work
Public assistance and low income (mix)
Family and single-parent households (mix)
High percentage of children
Low income
Public assistance 
Single-parent households
Higher population density
Middle income
One person works, one stays home
All occupants work

Table 2. National DHW Sizing Guidelines (Low-Medium-High)
Hot Water Demands and Use for Multifamily Buildings
  Maximum hour Peak 15 minutes Maximum day Average day
Low 2.8 gal (10.5 l)/person 1 gal (4 l)/person 20 gal (76 l)/person 14 gal (53 l)/person
Med 4.8 gal (18 l)/person 1.7 gal (6.4 l)/person 49 gal (185 l)/person 30 gal (114 l)/person
High 8.5 gal (32.5 l)/person 3 gal (11.5 l)/person 90 gal (340 l)/person 54 gal (205 l)/person
  Peak 5 minutes Peak 30 minutes Maximum 2 hours  Maximum 3 hours
Low 0.4 gal (1.5 l)/person 1.7 gal (6.5 l)/person 4.5 gal (17 l)/person 6.1 gal (23 l)/person  
Med 0.7 gal (2.6 l)/person 2.9 gal (11 l)/person 8 gal (31 l)/person 11 gal (41 l)/person  
High 1.2 gal (4.5 l)/person 5.1 gal (19.5 l)/person 14.5 gal (55 l)/person 19 gal (72 l)/person
Note: These volumes are for DHW delivered to the tap at 120oF.
Sources: Data from Chapter 45: Service Hot Water, In 1995 ASHRAE Handbook: HVAC Applications, Atlanta: ASHRAE, 1995, and Goldner, F.S., and D.C. Price. Domestic Hot Water Loads, System Sizing and Selection for Multifamily Buildings. In 1994 ACEEE Summer Study on Energy Efficiency in Buildings Proceedings, 2.105-2.116. Berkeley: American Council for an Energy-Efficient Economy, 1994.

Author Fredric Goldner discusses meter equipment with building superintendent John Perkins. The meter he is pointing to monitors hot water recirculation, and above it is a domestic hot water consumption meter. 

The number of occupants per apartment should be estimated based on local standards or regulations. For example, in a given city, studios may accommodate two persons; one-bedroom apartments, three persons; two-bedroom apartments, three to five persons; and so on.

In buildings where corrective maintenance cannot be done, a safety factor of 20%-30% may be employed to compensate for poorly maintained fixtures and distribution piping. However, this should be done only in extreme cases.

The figures presented in Table 2 are for centrally fired systems; individual apartment water heater systems are likely to have lower levels of consumption because the resident usually pays for fuel directly, which encourages conservation. There isn't a set of values for individual systems in ASHRAE, but a suggested rule of thumb for sizing these would be to use a low-end estimate for a single-family home load.

ASHRAE based its 1995 guidelines (published in the 1995 HVAC Applications Handbook) on new research conducted in New York City (see Collecting Usage Data in New York City) as well as data from studies in seven other areas of the United States and Canada. Both research and practical experience in different areas of North America indicate that there are variances in DHW use among geographical locations. There is, however, no distinctive pattern that can be identified with the available data.

The joint ASHRAE/ASPE (American Society of Plumbing Engineers) Domestic Hot Water Design Manual, to be published this fall, will go into greater depth than the ASHRAE standards, including the patterns of consumption and demand derived from the New York study. Becoming familiar with these patterns can help designers choose the best equipment and help auditors troubleshoot related system problems.

Collecting Usage Data in New York City

In 1990-91, Energy Management & Research Associates (EMRA) gathered 14 months of real-time monitoring data in 30 New York City multifamily buildings. The New York State Energy Research and Development Authority sponsored the study. 

The data were collected by computerized heating controllers, which monitored burner on-off times and the following temperatures: apartment air, outdoor air, boiler water (aquastat), and DHW. Eight buildings had additional monitoring equipment installed to record stack temperature, boiler makeup water flow, DHW flow in 15-minute increments, oil flow, and DHW temperature before and after the mixing valve and on the return line.

In 1993, we equipped a subset of three of the sites to record DHW flow in 5-minute increments and to record recirculation flows. This was done to get a more precise picture of short-term/instantaneous demand peaks and to collect the missing information necessary to create an accurate simulation of real-time operations. We collected data in these three buildings for 100 days.

EMRA also collected building operation and tenant information from superintendents and property managers via questionnaires and interviews, and building and apartment occupancy records. We conducted energy audits to determine the type and condition of equipment and buildings.

Within the New York research, we tried to include a variety of building sizes, income levels, ethnic backgrounds, and locales. The study buildings are characteristic of the older and predominant stock of the over 120,000 New York City multifamily buildings. The buildings range in size from 17 to 103 apartments in either five or six above-ground stories. These buildings were built before 1902 or between 1902 and 1928. All have combination steam-space-heating and DHW-generating steel tube boilers, which use primarily #4 or #6 oil in air-atomizing burners. DHW is generated by a tankless coil just under the surface of the boiler water.

Energy Use Analysis

An evaluation of the energy used to produce DHW was conducted for the summer period, when the systems are used strictly for DHW purposes. This analysis revealed that an average of 150 gallons (ranging from 100 to 200 gallons) of DHW was produced and used at the tap for each gallon of #6 oil (or equivalent) consumed by the burner. Included in these figures are various levels of combustion efficiency, standby losses, pipe insulation, and other real-time factors that affect the operation of systems in occupied buildings. These numbers can be used as a check against results of energy savings predictions from audit calculations related to hot water conservation measures (such as low-flow showerheads).

For further details, a copy of Report No. 94-19, Energy Use and Domestic Hot Water Consumption: Phase 1, is available from NYSERDA. Tel:(518)465-6251, Ext. 250.

Figure 2. Seasonal variations in weekend consumption, gallons per person (composite of data from New York City apartment buildings).
The Variations Behind the Values   Seasonal and Daily Variations The multifamily buildings we studied show distinct seasonal variations of DHW consumption levels (see Figure 2). The average daily consumption rises 10% in the fall (from summer consumption), and rises 13% more in the winter. Consumption then drops slightly in the spring and drops significantly (19%) in the summer.

There is generally a slightly higher daily consumption on weekends than on weekdays. This holds true in all seasons. The average weekend daily consumption is 7.5% greater than the average weekday daily consumption.

Weekday and weekend hot-water consumption patterns have distinct differences (see Figure 3). Weekdays have little overnight usage; a morning peak; lower afternoon demand; and an evening or nighttime peak. Weekends have just one major peak, which begins later in the morning and continues until around 1 pm to 2 pm. The usage then tapers off fairly evenly through the rest of the day. The weekend peak is greater than any of the weekday peaks.

The highest peaking level occurs during winter weekends. Thus, the best tactic for an engineer who has the time and money to custom-design a retrofit system is to monitor current consumption for two or three winter weekends to determine a building's actual peak usage, rather than estimating it with Table 2. A system designed to meet these draws should satisfy all other year-round requirements.

Figure 3. Weekday versus weekend consumption (composite of data from New York City apartment buildings). Research in New York City found that apartment residents use the most water between 10 AM and 12 noon during winter weekends.

Two morning peaks occur on the weekdays, the first between 6 am and 8 am and the second between 9:30 am and noon. Individual buildings tend to exhibit one of these two peaks. Generally, the buildings with large numbers of working tenants and middle-income populations experience the early morning peak, while buildings with many children exhibit the later morning peak (especially during the summer period).

This knowledge of flow patterns can come in particularly handy when troubleshooting hot water complaints. For instance, a large fluctuation in water temperature at a time when the usage was extremely low recently helped me to identify a problem with a faulty hot water coil. If the fluctuation was observed only during a high usage period, the cause-perhaps an undersized coil or a problem with the mixing valve-would have been harder to determine.

Recirculation Systems DHW systems in multifamily buildings generally employ one of three types of return or recirculation system. The first option is to have no recirculation piping at all. This is most often found in the smallest end of the multifamily sector, where there are short runs between the supply source (boiler or heater) and the farthest tap. The second option is a gravity return system (thermosiphon circulation). The monitoring data indicate that these systems have a very small flow, ranging from 0 to 0.5 gpm. The third option is a forced recirculation system. These systems employ a small pump to keep water flowing, thus avoiding stagnation and the need for residents to run the tap for a long period (particularly on upper floors) to receive sufficiently hot water. The pumps either are run continuously or may be cycled on and off by an aquastat.

Although recirculation pumps should be sized to meet each individual building's requirements, common practice is one size fits all. Thus we found the same pump size at all sites. (A methodology for proper pump size selection can be found on page 45.5 of the 1995 ASHRAE HVAC Applications Handbook.)

Our monitoring showed that water consumption has an inverse relationship with recirculation flow. In the overnight period, when there is little or no consumption, the pump reaches its maximum capacity rate. Designers should consider this and the flow curves in Figure 3 when choosing between recirculation control strategies (see The Best Boiler and Water Heating Retrofits, HE Sept/Oct '95, p. 27, and Controlling Recirculation Loop Heat Losses, HE Jan/Feb '93, p. 9). A new study investigating three very low-cost approaches to reduce recirculation system losses while maintaining resident comfort and satisfaction should be completed in early 1997.

Peak Demands and Average Consumption In the New York City buildings, the average hourly consumption is only 42% of the consumption in the peak hour. Instead of sizing a system to be able to provide the peak demand, it's possible to generate and store hot water during the periods of average and below-average demand to meet the peak. This could be accomplished by installing a system with a heater designed to generate the average hourly load, running essentially continuously, and providing enough storage tank capacity to store unneeded hot water during the night and furnish it during periods of peak demand (such as morning shower time).

Figure 4. Parts of three-hour peak and 60-minute peak consumption.
Concurrence of Peaks Beyond the general usage patterns of a building, peaking times and flows are used to more closely identify demands on the boiler. Figure 4 shows how all of the peak volumes contribute to the one-hour and three-hour peak demands on the DHW generation and/or storage system. These relationships can be used to model various configurations of hot water supply systems (see A Sizing Example).

The 5-, 15-, 60-, 120-, and 180-minute peak demand times coincide with each other. These volumes should therefore be addressed as different (time length) measurements within the same peak DHW draw, so the system can be designed to satisfy this load. An instantaneous system designed to meet the peak 5-minute draw will have no problem meeting the rest of the load. Generation and storage systems should be designed both to provide hot water for the average load and to meet the short, sharp peaks.

A Sizing Example

Let's take a 58-unit apartment building whose occupants are a mix of families, middle income couples, and some singles. Most adults work outside the home. There is a public laundry in the basement with a few washers, and the leases prohibit both clothes washers and dishwashers in the apartments (although conversations with the building superintendent have confirmed that a number of people have such appliances.)

Step 1. Compute the maximum potential occupancy, based on local standards and expectations, and conversations with the building owner or manager.

                                     Maximum           Total         Apt size       Apts         people/apt         people         3-bedroom       4       x       5       =       20         2-bedroom       14      x       4       =       56         1-bedroom       25      x       3.5     =       87.5         Studios         15      x       2.25    =       33.75 Building total 198 

Step 2. Determine the Low, Medium, or High (LMH) usage factor of the building's occupants from Table 1, based on knowledge of the building, conversations with the building owner or manager, and observations. Consider the effect of either currently installed or potential future additions of appliances that might move a building up to a higher usage category.

 Based on the information above, the Medium usage factor was selected.

Instantaneous Systems

For either an instantaneous DHW-only system or a tankless coil in a combination heat/DHW boiler, first find the system load (gallons per hour) based on the peak 5-minute demand. Next, convert this to a Btu/h rating. This rating can then be used to select equipment.

Step 3a. Compute the system load using the 5-minute peak demand values in Table 2.

              Number       Peak 5-min                           Peak LMH factor   of people       demand            Periods/h    system load Medium          198    x  0.7 gal/person   x      12    =    1,663 gal/h Step 4a. Convert the system load to a Btu/h rating. (In New York City, the average year-round temperature rise required is about 90oF.)
                                                                  1/Boiler System load             Conversion              Temp rise       combustion efficiency           DHW load 1,663 gal/hr    x       8.33 lb/gal     x       90oF    x       1/0.8 (80% CE)  =       1,558,439 Btu/h Instantaneous DHW-Only Heater. The 1,558,439 Btu/h should be the size of the DHW heater. (Note that a higher combustion efficiency should actually be used for sizing an instantaneous heater; use 85% or the efficiency specified in the equipment documentation.)

Combination Heat/DHW Boiler. When sizing a tankless coil in a combination heat/DHW system, the 1,663 gallons per hour is the coil size to be ordered. The 1,558,439 Btu/h is the additional load capacity for DHW to be added to the space-heating load to size the boiler. (In an existing steam heating distribution system, the space-heating load should be computed by the EDR-equivalent direct radiation-methodology.)

Generation and Storage System

For a system with a mix of generation and storage, calculate the generator size based on the peak 30-minute demand to get a Btu/h rating. Calculate the storage tank volume based on the maximum three-hour demand. 

Step 3b. Compute the system load using the peak 30-minute and maximum three-hour hot water values in Table 2.

                Number          Peak 30-min LMH factor      of people       demand category         Periods/h       System load Medium          198     x       2.9 gal/person  x       2       =       1,148 gal/h Number Maximum 3-hours LMH factor of people demand category Storage volume Medium 198 x 11 gal/person = 2,178 gal

Step 4b. Next, convert the load into equipment ratings.

                                                                1/Boiler System load             Conversion              Temp rise       combustion efficiency   DHW load 1,148 gal/h     x       8.33 lb/gal     x       90oF    x       1/0.85 (85% CE) =       1,012,536 Btu/h The 1,012,536 Btu/h is the size of the hot-water heater. This heater should then be used to supply 2,100 gallons of unfired storage tanks.

Estimating Consumption in Existing Buildings

To estimate how much hot water is used in a building for energy consumption or savings calculations, use the LMH factor and the average day hot-water value in Table 2. In this calculation, replace the maximum potential occupancy from Step 1 with the actual current (or best-guess recent) occupancy level.

Step 3c. Calculate system load using the average day values in Table 2.

                        Current number  Average day         LMH factor      of people       demand                  System load         Medium          153     x       30 gal/person   =       4,590 gal/day
Straighten Up and Size Right There seem to be as many different types of DHW heating systems as there are people who design them. What they all attempt to accomplish is to provide the correct mix of generation capacity and storage to satisfy both the peaks and the average load. One major concern during the development of the LMH approach was the acceptance and use of the new system. Because it results in higher load estimates than the old guidelines, it is important that the new method be used correctly.

If the current practice of defensive oversizing is applied to the new guidelines, this will only exaggerate the capital and energy inefficiencies experienced in the past. It is therefore important for the designer to recognize the inherent safety nets in the new approach. The most significant of these is that the method uses the building's maximum potential occupancy, which may never actually occur. Also, using the new guidelines, an engineer designs a system to satisfy the higher-volume but short-duration peaks (not delineated in the old guidelines), which occur only a few times during the year. Even if the system were not able to satisfy that load, the problems would probably be minor-for instance, the occupants might experience slightly lower temperature hot water at their taps a few times per year.

The main question concerning acceptance and use of the new guidelines is whether the designers and energy professionals are comfortable with their reliability and professional backing. ASHRAE's Technical Committee 6.6 (Service Hot Water) was the main force in the call for a new sizing tool based on the vast quantity of real-time data that has been collected. The new joint ASHRAE/ASPE Domestic Hot Water Design Manual, scheduled for publication this fall, should also provide substantial support for those who wish to size systems properly. It includes a how-to sizing guide for 17 different building types-from residential buildings to commercial, industrial, and recreational facilities.

Further Reading Chapter 45: Service Hot Water, In 1995 ASHRAE Handbook: HVAC Applications, Atlanta: ASHRAE, 1995.

Goldner, F.S. DHW System Sizing Criteria for Multifamily Buildings. ASHRAE Transactions 100, No.1 (January 1994): 147-65.

Goldner, F.S. Energy Use and DHW Consumption Research Project, Report No. 94-19. Final Report: Phase 1. Prepared by Energy Management and Research Associates for New York State Energy Research and Development Authority, November 1994.

Goldner, F.S., and D.C. Price. Domestic Hot Water Loads, System Sizing and Selection for Multifamily Buildings. In 1994 ACEEE Summer Study on Energy Efficiency in Buildings Proceedings, 2.105-2.116. Berkeley, CA: American Council for an Energy-Efficient Economy,1994.

 | Back to Contents Page | Home Energy Index | About Home Energy |
| Home Energy Home Page | Back Issues of Home Energy | EREN Home Page |


Home Energy can be reached at:
Home Energy magazine -- Please read our Copyright Notice


  • 1
  • NEXT
  • LAST
SPONSORED CONTENT What is Home Performance? Learn about the largest association dedicated to home performance and weatherization contractors. Learn more! Watch Video