Reference Publication: Parker, D. S. and P. W. Fairey (2001). Preliminary Evaluation of Energy-Efficiency Improvements to Modular Classrooms. Report No. FSEC-CR-1272-01. Florida Solar Energy Center, Cocoa, FL 32922.
Disclaimer: The views and opinions expressed in this article are solely those of the authors and are not intended to represent the views and opinions of the Florida Solar Energy Center.
Evaluation of Energy-Efficiency
Improvements to Modular Classrooms
The objective of our investigation is to evaluate innovations that would enable modular classroom builders to improve the energy performance of their classrooms. We investigate improved insulation, better windows, daylighting, cross-ventilation, sensible and latent heat recovery of ventilation air and light colored surfaces and radiant barriers for cooling dominated climates. The tasks associated with this work are as follows:
Classroom Simulation Model
We created a building energy simulation to model the energy performance of modular classrooms. We utilized a modified version of EnergyGauge USA to accommodate classroom schedules and occupancies. The program uses DOE-2.1E as the simulation engine. The model and its validation are more completely described elsewhere (Parker et al., 1999; Fuerhlein, 2000).
Separate schedules were developed for
lighting and HVAC use based on a conventional school year (August
15th - May 15th) and one which includes
summer school. The schedule assumes the school day to extend
from 7 AM to 5 PM based on both occupancy and after-school activities.
The degree with which HVAC equipment and lighting systems are
positively switched off is based on previous monitoring and surveys
(Callahan et al., 1999). The performance of enthalpy recovery
ventilators (ERVs) and the application of ASHRAE Standard 62-1989
are based on a detailed monitoring evaluation of such equipment
(Shirey et al., 1997).
Our work plan had several tasks: collect data, prepare computer models, develop a baseline building, evaluate design possibilities, generate design alternatives and evaluate their performance. Later, we plan to analyze costs and select the final designs. The baseline modular classroom building (Figures 1 and 2) is 28' x 64' and has the following energy-related specifications: walls R-11, floors R-11, roof R-30, (2) aluminum frame slider type windows U = 1.17, (2) insulated steel doors, 2' x 4' recessed four-tube fluorescent light fixtures -50 Fc and a 3-ton air conditioner with 10 kW of strip resistance heating.
Figure 1. Baseline modular classroom building
Figure 2. Floor plan of the baseline modular classroom.
We plan to analyze the heating and cooling loads of the baseline modular classroom building with its long side facing east in seven different climates: Tampa, Florida; Sacramento, California; Rochester, New York; Raleigh, North Carolina; Cincinnati, Ohio; Dallas, Texas and Phoenix, Arizona using a specially produced DOE-2.1E software. Each climate presents a different challenge to mitigate the energy consumption of the classroom against the prevailing weather conditions. However, for this initial analysis we simulated the two heating and cooling extremes (Rochester, New York and Tampa, Florida) as well as a mixed climate (Raleigh, North Carolina). The analyses were used to redesign the modular classroom to integrate daylighting, windows and envelope design, and heat recovery for ventilation air as appropriate for each climate. Based on individual parametric analyses, we assemble successful groupings of measures (packages) for each climate.
While modular classrooms come in different sizes and configurations a double-classroom structure with a common wall and a low-slope roof appears to dominate the market. To characterize the modular classroom market, we evaluated submitted designs from the manufacturers: United Modular (Glenrose, TX), Roger Carter (Kingston, NC) and Walden Structures (Capistrano, CA). By far the most models are double classroom units with dimensions of twenty-four (24') to twenty-eight feet (28') in width and from fifty-six (56') to sixty-four (64') feet in length. The study baseline unit is one half of a 28' x 64' two classroom unit with a restroom (see Figures 1 and 2). This type of unit is comprised of two 14' wide by 64' long modular sections. This unit has a pitched roof with the ridge running in the 64' direction. A T-bar suspension ceiling is typical with the ventilation of the roof plenum as a construction variable. The unit is also equipped with two exterior door and two windows per classroom. The construction of this unit is consistent with The Uniform Building Code (UBC) type 5 non-rated construction, light frame wood. The occupancy classification is E-1, educational, with 25 occupants. The following is a brief summary highlighting some of the standard features of this baseline modular classroom.
Floor ConstructionThe floor construction is comprised of light wood framed joists with plywood floor sheathing, fiberglass batt insulation, and typically carpet for a finish.
Wall ConstructionThe exterior wall construction is comprised of light wood frame studs with a plywood combination siding and structural sheathing (i.e., T1-11), fiberglass batt insulation and an interior finish of a pre-decorated gypsum board.
Roof ConstructionThe roof/ceiling construction is made up of light wood framed trusses with plywood sheathing, fiberglass composition shingles, fiberglass batt insulation and a suspended acoustical tile ceiling system. Consistent with the submitted plans, the roof top structure is a low-slope metal roof.
Heating and Ventilating SystemThis unit is equipped with a 3-ton wall-hung package unitary heat pump with 10 kW back-up electric resistance heating. A 32 foot supply duct runs down the classroom centerline and is located in the plenum space (attic).
Electrical SystemThe main electrical service for this unit is a 200 amp 120/240 volt single-phase system controlled by a main distribution panel. Provisions are made for lighting, electrical convenience outlets, fire alarm systems, intercom and clock systems as well as serving power to the heating, ventilating and air conditioning system.
Energy Baseline for Various Climates
We plan to perform simulations for the following seven climates representing much of the market areas and climatic variation within the U.S.
For this draft preliminary report, an analysis was done on the two extreme climates (Tampa and Rochester and a mixed climate- Raleigh. The hourly simulation uses Typical Meteorological Year (TMY2) data for each location to perform the calculations. The energy baseline was simulated using the information gathered during the early phase of the project based on submitted plans. The simulations use a nine-month schooling period. The schedules for people, lights, and equipment are based on a study of classroom modules in Florida. The simulation was performed with the assumption that each classroom holds 25 students and 1 teacher (26 people total) that generate 240 Btu/hr/person of latent heat and 100 Btu/hr/person of sensible heat into the space. Impact of the large interactions of waste heat from the lighting system and heat loss and gains from the building envelope are intrinsically computed within the simulation. In order to accurately simulate the performance of the HVAC system, the performance data for the units was input for a Packaged Terminal Air Conditioner (PTAC) unit with heat-pump as the heat source. The data was obtained from the manufacturer (Bard, Wall-Mount heat pump, model WH361-A). Three different sets of performance data were input: the Cooling Capacity, the Heating Capacity, and the Heating Electric Input Ratio (EIR) and Cooling Electric Input Ratio (CEIR) of the system.
Comparison of Simulation Model to Measured Data
In order to verify the accuracy of our baseline simulation, we compared the simulation's fuel use predictions to energy use records collected for a Florida classroom building. Detailed monitoring has been performed within this project (Callahan, et. al., 1999). Daytona Beaches' climate data was used for simulating the module classroom in Port Orange, Florida due to its geographical proximity and climatic similarity.
Teachers in the modular buildings for the academic year 1998 were interviewed. The questionnaire included clarifications regarding the number of students occupying the classroom, the schedule for the equipment use, and the use of lights and hot water, and the schedule of the class for the whole academic year. No alteration was required from the established schedule. However, to approximately match the predicted performance of the baseline case with the actual energy use, the following rough changes were made to the baseline case.
Table 1. End Use Predicted and Simulated at Silver Sands Portable Classroom kWh per Day
|End Use||Measured kWh*||Predicted kWh||Error (%)|
Source: M.P. Callahan, D.S. Parker, J.R. Sherwin and M.T. Anello, 1999, Evaluation of Energy Efficiency Improvements to Portable Classroms in Florida, FSEC-CR-1133-99, November, 1999 (see Figure 3). The degree of correspondence is particularly good considering that the specific schedules and weather data were necessarily approximate to the corresponding conditions associated with the measured performance. Our prediction was within 5% of the actual energy use indicating that our simulation is fairly representative of the monitored modular classroom building. This also gives confidence in the simulation results.
Baseline Reference Building
We analyzed a series of efficiency measures in each climate location for the reference modular classroom. The base building in each location had R-11 walls and floor insulation, R-19 ceiling insulation (suspending between the trusses), single glazed windows with two insulated doors. The space conditioning system consisted of a 3-ton air conditioner (SEER 9) with 10 kW of electric strip heat. The supply duct was located in the attic space; the duct related leakage of the distribution system amounted to 168 cfm at a tested system pressure of 25 Pa (Qn = 0.15). This characterization is in agreement with an assessment done by Cummings (1998) on seven portable classrooms in Florida.
We note that many of the space conditioning systems in portable classrooms have significant leakage around the through the wall penetrations for the side-mount AC units. The lighting system consists of six four-tube flourescent fixtures with 40 W tubes and magnetic ballast (180 Watts each). Based on our experience with the lighting energy use in the two monitored portables in Port Orange, Florida (Callahan, et.al., 1999), we assume that lighting is accidently left on about 10% of the time over night and about 5% of the time over weekends and holidays. We also assume that the HVAC systems are accidently left on for similar periods. These assumed rates for human error allow evaluation of the potential savings of occupancy based controls.
Issues Associated with Ventilation and Infiltration
Based on our survey of manufacturers and installers, as well as from local units surveyed in Florida we learned that nearly all modular classrooms are shipped with outside air dampers closed. No installers with which we spoke indicated that these were opened when installed in the field although the mechanical engineers associated with the modular designs were aware of the potential issue. Although we analyzed a specific case with dampers closed as they are likely installed, we assumed that the dampers were placed in the first stop (150 cfm) for our reference case. Since we assumed average occupancy for the purposes of ASHRAE Standard 62-1989 (25 students and teachers in the classroom), this amounts to 6 cfm/student and corresponds well to previous studies (Callahan, et.al., 1999). It is noteworthy that this is not adequate to meet the requirements for ASHRAE Standard 62-1989 which requires 15 cfm/student. In recent years, much attention has been leveled at improving school ventilation as a means to improve indoor air quality (Construction Specifier, 1999; Energy Design Update, 1997). Thus, for our analysis of potential improvements we compare compliance with ASHRAE Standard 62-1989 with a case where only 6 cfm/student is being provided. We do this so that the predicted energy savings associated with the suggested changes to modular classrooms will not be altogether out of keeping with the actual achieved savings.
Another conservatism associated with the analysis is that we assume that the roof/plenum is not ventilated as is done with some models. When this is done with the dominated T-bar suspended ceiling, the attic vents become a site for major envelope air leakage. This has been observed in work done on building leakage and air change rates on seven portable classrooms in Central Florida (Cummings, 1998). For instance, those with T-bar ceilings with ventilated plenums were found to have total envelope leakages averaging 22 ACH @ 50 Pa pressure. On the other hand, with one building which had a rigid ceiling, the leakage was only 7 ACH at 50 Pa. Thus, within the analysis we assumed that although T-bar ceilings would be used, that the attic plenum would not be vented. This was the case for two of the three submitted plans which were reviewed.
Below we list and briefly describe the specific measures analyzed for each site.
As options we evaluate increases to roof insulation thickness from R-19 hr-ft2-F/Btu to R-30 or R-38. We also predict the impact of the addition of a perforated foil face to the roof insulation creating an attic radiant barrier. We also evaluate the impact of choosing a reflective white metal roof rather than an unfinished galvanized surface. Both of these measures are demonstrated to reduce cooling, although the reflective roofing will increase heating somewhat (Callahan et al., 1999).
We analyze how a thicker 2 x 6" wall with R-19 will improve performance. We also examine how adding 1 inch of isocyanurate sheathing to the exterior of the R-11 wall would alter heating and cooling. Finally, we also evaluate how a light pastel wall color would reduce cooling as compared to the base line medium tan color (solar absorptance = 0.6).
We evaluate double clear glass compared with the single glass normally assumed (U= 0.69 Btu/h-ft2-F). We also evaluate double glazed units with a low-e coating (U= 0.59). And finally, we evaluate the above with a solar control outer lite designed to reduce cooling needs (SHGC = 0.38). All glazing units were assumed to have aluminum frames.
We evaluate a sealed air distribution system where tested leakage to outdoors is only 54 cfm (Qn= 0.05). We also evaluate the impact of moving the supply duct to the interior of the conditioned classroom to avoid duct conduction losses.
We evaluate the substitution of slim-line T-8 lamps with electronic ballasts for the T-12 system with magnetic ballasts in the baseline system. This reduces the 4-tube fixture wattage from 180 watts to about 110 W. We also estimate the savings from using occupancy sensor control to positively shut off lights during unoccupied periods.
We evaluate a proposed daylighting package with a reduction to fixture lighting density. This involves adding three 22" tubular solar skylights spaced evenly down the center line of the classroom along with a strip of one foot high glazing located along the sides of the building up by the top of the door-line. These would include an interior light shelf to prevent glare and to reflect daylight to the ceiling. To achieve savings, the lighting fixtures would be altered from four tube to two-tube T-8 fixtures. This simple expedient would insure lighting systems, save on costs and result in a simple and robust system to harvest daylight.
Heating and Cooling Systems
The baseline air conditioner would be altered to a more efficient SEER 10 unit, or with a heat pump with supplemental strip heat. Finally, occupancy based HVAC controls would be evaluated to provide positive off during unoccupied periods.
Parametric evaluation considers no ventilation as many modular classrooms are shipped and operated, as well as compliance with Standard 62-1989 without heat recovery, but with a 4-ton heating and cooling system. Other evaluation would be made of use a 60% effective enthalpy recovery ventilator (ERV) in compliance with 62-1989 providing 375 cfm of outdoor air to the classroom. Based on performance data, the required fan power for such systems is approximately 0.9 W/cfm which was modeled within DOE-2. Such a system was modeled with standard strip heat and with a heat pump. Finally, we evaluate the impact of using operable windows in the daylighting package to provide cross ventilation to reduce cooling needs.
Based on the parametric evaluation of the specific measures, ones were chosen which showed the potential to produce promising systems. These are noted in the Tables A, B and C that follow and are marked with an asterisk (*). The combined performance of the packages is shown for the standard school schedule and for one including summer school.
The results of the simulation parametric analysis and the grouping of measure packages is given in Tables A, B and C (pages 10-12) for Rochester, NY, Tampa, FL and Raleigh, NC, respectively. Even though the annual energy consumption in the simulated model matched to within 5% of the actual energy consumption in our test evaluation where metered data was available, it would be desirable to obtain similar data from a cold climate location. Even thought the measures differed with different locations the predicted energy savings were approximately 45% in all locations. An implicit assumption within our analysis is that the attic/plenum space will be sealed. This is important given the findings by Cummings (1998) showing that T-bar ceilings in portable classrooms are inherently leaky. A general summary of the results is as follows:
In hot, cooling dominated climates such as Tampa, measures which reduce lighting and its internal heat generation show greatest promise to reduce building energy needs. Daylighting looks particularly attractive in this location although solar control glass looks important to reduce the space cooling liability. Similarly, light colored surfaces and solar control glazing looks more important than insulation. Heating system type is not as critical as cooling efficiency. Floor insulation is counterproductive.. In cold, heating dominated climates such as Rochester, New York, insulation measures and duct air leakage control measures look to be most important. The results would further suggest that ground source heat pumps (GSHPs) or natural gas heating may be attractive alternatives to air-source heat pumps. Daylighting, while producing savings in lighting energy, tends to increase heating budgets and is not as attractive as insulation and heating system measures. It appears important that a successful daylighting strategy in such location utilize highly insulated glazing (double glazed low-e was assumed) in order to be successful. Floor and wall insulation is important, solar control glass is counterproductive.. Mixed climates, such as Raleigh, North Carolina, evidence an amalgam of the preceding extremes. Insulation with a radiant barrier looks to be the best strategy for the roofing system as it helps control both heating and cooling needs. However, daylighting is quite attractive as it reduces lighting energy and substantially reduces space cooling. Insulation measures, duct leakage control and a more efficient heating system all look to be promising measures. Specification of a heat pump is important to controlling heating costs.
All packages show that ventilation heat recovery is vital to prevent energy savings from envelope and lighting measures from being swamped by the thermal impacts of complying with ASHRAE Standard 62-1989. Similarly, occupancy based control of lighting and HVAC systems is beneficial in all locations. A summary of results is given below in Figure 3.
|Figure 3.||Annual Energy Savings Potential for Typical Modular Classrooms for 3 Climate Types: 1) a heating dominated climate, 2) a cooling dominated climate and 3) a mixed climate.|
ASHRAE Standard 62-1989, Ventilation for Indoor Air Quality, American Society of Heating Refrigerating and Air Conditioning Engineers, Atlanta, GA.
Callahan, M.P., D. S. Parker, J.R. Sherwin and M.T. Anello, Evaluation of Energy Efficient Improvements to Portable Classrooms in Florida, FSEC-CR-1133-99, Florida Solar Energy Center, Cocoa, FL, November, 1999.
Callahan, M.P., D. S. Parker, W.L. Dutton and J.E.R. McIlvaine, Energy Efficiency for Florida Educational Facilities: The 1996 Energy Survey of Florida Schools, FSEC-CR-951-97, Florida Solar Energy Center, Cocoa, FL, July 1997.
Clark, W.H.,II, "Outside Air Units for Schools," Engineered Systems, April, 1997.
Construction Specifier, 1999. "Poor IAQ in Schools Raising Concern," October, 1999, p. 10.
Cummings, J.B., "Brevard County Schools: Portable Classrooms at Kennedy and McNair Middle Schools," Florida Solar Energy Center, Cocoa, FL, April, 1998.
Cutter information Corp, 1997, "Investigators Look to Improve Ventilation in Portable Classrooms," IEQ Strategies, June, 1997, p. 6.
Fuehrlein, B.S., S. Chandra, D. Beal, D.S. Parker, and R. K. Vieira , "Evaluation of EnergyGauge® USA, A Residential Energy Design Software, Against Monitored Data," Proceedings of ACEEE 2000 Summer Study, pp 2.115 - 2.126, American Council for an Energy Efficient Economy, Washington, DC, August 2000.
Parker, D.S., P.A. Broman, L.B. Grant, L. Gu, M.T. Anello, and R.K. Viera, "EnergyGauge® USA: A Residential Building Energy Simulation Design Tool", Proceedings of Building Simulation '99, Kyoto, Japan, 1999
Shirey, D.B., R.A. Raustad and M. Brassard, Evaluation of an Enthalpy Recovery Ventilator Installed in a Portable Classroom, FSEC-CR-1133-99, Florida Solar Energy Center, Cocoa, FL, 1999.
National Association of State Energy Officials
1414 Prince St., Suite 200
Alexandria, VA 22314