Professional Documents
Culture Documents
Final Report
August 20, 2004
Prepared by:
Wrightsoft and Oklahoma State University (“Contractors”) have compiled this report and accompanying
1199-RP Source Code and Data CD (together the “Report”) with care. However, neither the Contractors
nor ASHRAE warrants that the information in the Report is complete or free of errors. ASHRAE does
not necessarily agree with any statement or opinion in the Report. The appearance of any technical data
or editorial material in the Report does not constitute endorsement, warranty, or guaranty by the
Contractors or ASHRAE of any product, service, process, procedure, design, or the like. ASHRAE has
not investigated and ASHRAE expressly disclaims any duty to investigate any product, service, process,
procedure, design, or the like which may be described herein.
THE ENTIRE RISK OF THE USE OF ANY INFORMATION IN THE REPORT IS ASSUMED BY
THE USER.
See also “1199RP_TermsOfUse.txt” on the 1199-RP Source Code and Data CD.
Executive Summary
ASHRAE Research Project 1199-RP, Updating the ASHRAE/ACCA Residential Heating and Cooling
Load Calculation Procedures and Data, developed two new residential heating and cooling loads
calculation procedures:
• Residential Heat Balance (RHB), a detailed heat balance method that requires computer
implementation; and
• Residential Load Factor (RLF), a simplified procedure suitable for hand or spreadsheet use.
RHB is based on first-principles heat balance models. A research-oriented FORTRAN implementation of
RHB, designated ResHB, was developed using ASHRAE Loads Toolkit as a starting point. ResHB
includes algorithms for calculating sensible cooling loads with temperature swing (temperature excursion
above the cooling set point) and to handle master / slave control (room cooling controlled by a thermostat
in another room). ResHB extends Toolkit capabilities to multiple rooms, zones, and systems, thus
supporting load calculations for real buildings. Also added were infiltration, internal gain, and duct loss
models appropriate for residential problems. Because RHB is a first-principles heat balance procedure, it
can be directly validated and refined using empirical data.
ResHB was tested against the general-purpose building model ESP-r. Testing procedures and results are
presented in the final report. Additional analytical, emprical, and inter-model validation of RHB is
underway and will be reported in the literature when complete
The RLF method was developed from RHB results. The form of RLF resembles prior methods.
However, the sensible cooling load procedure was derived using linear regression to find relationships
between design conditions, building characteristics, and peak cooling load predicted by RHB. This
eliminated the need for semi-empirical adjustments, such as averaging, that have been used in the
development of other methods. Results comparing RLF to RHB are presented. The RLF heating load
calculation is also described; it uses the traditional UA∆T formulation with some improvements to
procedures for slab and basement ground losses and infiltration air leakage.
Several conclusions are presented:
• RHB is the method of choice for residential load calculations. Due to its first-principles
approach, it handles virtually any residential configuration, including extreme cases, such as large
or concentrated fenestration area, that are problematic for prior methods. Currently available
computing power is now sufficient to support routine use of heat balance procedures.
• Inter-model validation is useful for testing models, identifying sensitivities, and verifying results.
• Linear regression is an effective and rational approach for development of simplified loads
calculation methods such as RLF.
• Simplified methods, such as prior residential procedures and RLF, are necessarily approximate.
They cannot capture all the effects modeled in RHB without addition of self-defeating complexity
and do not inherently provide indication of when they are inapplicable. For these reasons,
methods of this type should be superceded by rigorous heat balance approaches such as RHB.
Accompanying the report is the 1199-RP Source Code and Data CD containing ResHB source code and
other development material produced during the work.
This project was supported in part, by the Air Conditioning Contractors of America Educational Institute.
Start
1.000
Relative cooling load
0.900
0.800
0.700
0.600
0.500
0.400
0.00 0.83 1.67 2.50
Temperature swing (°C)
RHB definition
The Residential Heat Balance Method is a specialized application of the ASHRAE Heat Balance Method.
The following HB changes and extensions define RHB --
• Multi-room, multi-zone, and multi-system. The fundamental RHB modeling unit is the room.
Independent heat balances are performed for each room. Zones and systems are accounting
structures to which loads are accumulated to provide overall results.
• Specialized algorithms. Temperature swing and master / slave control can be modeled to produce
realistic sensible cooling load estimates.
• Residential models and assumptions. Component models and assumptions used for RHB are
appropriate for the residential application.
• Simple heating and latent cooling procedures. As discussed above, simple UA∆T model has
proven satisfactory for heating load calculations. Similarly, latent load can be estimated from
moisture gain from infiltration, ventilation, duct leakage, and occupants. These simple
approaches are retained in RHB.
It should be noted that RHB is not a fully elaborated cooling system design procedure. In particular,
RHB does not specify how temperature swing and master / slave control be considered during the design
process. RHB can model rooms with or without swing, allowing choice on the much-debated question as
to whether systems should be sized to allow swing at the thermostat on the design-day. Slave room
temperature results from a case-specific combination of limited capacity and control profile mismatch, so
its design implications are more complex. It may be that RHB master / slave capabilities should be used
for investigation of zoning options only after primary load calculations are done on an independent room-
by-room basis (with or without temperature swing).
The remaining sections of this paper provide details about the above aspects of RHB in its current form.
One major advantage of a heat balance formulation is that it can be tested and refined via direct
comparison to empirical data. It is expected that RHB will evolve as additional research results become
available.
Calculation Algorithms
The HB method is a design day procedure that requires iteration to find the steady-periodic solution at
which all heat flows correctly balance. RHB adds the additional requirement of finding loads under
floating temperature conditions, in order to handle temperature swing and master/slave control, as
described here.
repeat swing
repeat day
for hour = 1 to 24
for all rooms
repeat
for all surfaces
perform surface heat balance
end for surfaces
perform air heat balance
until room convergence for current hour
end for rooms
end for hours
until day convergence
determine room supply air flow rates for next swing iteration
until swing convergence
The convergence criteria are discussed below. The sequence was modified several times during
development and its logic is worth examining:
• The outer loop handles temperature swing (discussed below). Temperature swing occurs when
cooling capacity is less than required to hold a room at the set point. The swing search algorithm
adjusts each room supply air flow rate and repeats the entire calculation until the specified swing
is achieved.
• The hour loop is outside the room loop. This means that current hour conditions are available for
all rooms (either from the current day iteration or, at worst, from the prior day iteration), allowing
inter-room references.
One of the issues with a design day heat balance procedure is determining when the solution has
converged. A common technique is to continue iteration until calculated or temperatures change a very
small amount between iterations. The difficulty is to determine a “small amount” that truly represents
convergence. Unfortunately, there are cases that change very little iteration to iteration, but will continue
to change, resulting in significant drift in results. Various convergence criteria were attempted for ResHB
and the following are the best found to date:
• Hour. For each room, the current hour calculations are repeated until the sum of the absolute
change in surface temperatures plus air temperature is less than .0005 °K (.0009 °F), indicating
that a fully simultaneous solution has been closely achieved.
• Day. The day calculations are repeated until all rooms meet a) the fractional difference between
daily total inside and outside surface flux is less than .005 and b) the area-weighted total absolute
temperature change for all surfaces plus air is less than .0002 °K (.00036 °F). Note that the all-
room requirement means that some rooms will be iterated beyond this point.
• Swing. The swing search is continued until swings for all rooms are within .01 °K (.018 °F) of
specified. Each room has a specified swing. This allows different swings in master and slave
rooms, for example. Again, the all-room requirement means extra iteration for some rooms.
In addition to these basic criteria, there are various safety checks that detect oscillation or excessive
iteration and attempt to find an adequate result.
Temperature swing
ResHB uses a secant method search algorithm to search for the load when temperature swing is permitted.
Note that the calculations are based on varying system air volume flow rate with an assumed supply
temperature. This means the maximum extraction rate varies as the room temperature changes and sub-
cooling is self-limiting. If the modeling were done in terms of heat extraction, room air temperature
could be driven below the supply air temperature, which is impossible except under naturally floating
conditions.
The following calculation sequence is used:
• The required cooling air flow rate is found first for the 0 swing situation (that is, maximum
available air supply volume is unlimited and room temperature held at the set-point or floating
below it with no supply air flow).
• This maximum supply air flow rate is then reduced by 20% per °K (11% per °F) of target swing
and the room is calculated again. Generally, this reduction in supply air flow rate will produce a
significant temperature swing.
• The supply air flow rate is iteratively adjusted in proportion to the error in temperature swing, as
indicated by the secant method.
This algorithm is extremely efficient, because local linearity allows each subsequent estimate of supply
air flow rate to be much better than the prior one. Convergence to within .01 °K (.018 °F) of the target
swing usually occurs in less than 10 cycles. However, specific room characteristics can cause the search
to fail. As noted above, if room convergence is not essentially perfect, the secant method can produce
wildly unstable supply flow rate estimates. Several limits have been implemented to reduce the number
of cases that fail. Also attempted was an alternative algorithm based on a least-squares fit to the last
several points (as opposed to only the last two in the secant method); this approach did not work as well
as the secant method.
Master/slave control
The modeling of master/slave control is handled by the RHB temperature swing algorithm. By definition,
a slave room has the same air supply flow rate profile as its master, so the problem reduces to finding the
peak slave room supply air volume flow rate such that the maximum room temperature is the set point
(plus allowed swing if any). At each swing iteration, the peak flow rate is adjusted up or down using the
26 78.8
25 77
24 75.2
Room air temperature (°C)
22 71.6
21 69.8
20 68
19 66.2
18 64.4
17 62.6
0 4 8 12 16 20 24
Hour Master (south)
Slave (south)
25 77
24 75.2
23 73.4
22 71.6
21 69.8
20 68
19 66.2
18 64.4
17 62.6
0 4 8 12 16 20 24
Hour
Master (west)
Slave (east)
Models
RHB development involved review, refinement, and extension of Loads Toolkit models, as described in
the following sections.
Comparison studies were done for a range of conditions, with the expected result that loads depend
strongly on coefficient model. The Toolkit ASHRAE model was eliminated because it lacks sensitivity to
heat flow direction. The Fisher correlations are appropriate for ceiling diffuser configurations that are
typical in only a fraction of residential buildings.
The model selected for RHB is a variant of TARP simple. The “sys off” values are ASHRAE-based
natural convection values. To improve convergence stability, the transition between heat flow up and
down values is made linearly over 2°C (3.6°F), rather than abruptly. The “sys on” value is enhanced due
to air motion in the room. 5 W/m2-K (.88 Btu/h-ft2-F) was chosen as the “sys on” value based on analysis
of experimental data from ASHRAE research projects 529-RP and 664-RP for air change rates of
approximately 8 ACH (typical for residential systems). For each hour, the coefficient used in the heat
balance is the system-run-fraction-weighted combination of the “sys on” and “sys off” values.
1.000
0.950
h/h0
0.900
0.850
0.800
0.750
0.000 0.050 0.100 0.150 0.200 0.250 0.300
1 - P/P0 Forced
Natural
.24+.76 * P/P0
where
h = convective coefficient at pressure P (units consistent with h0)
h0 = convective coefficient at sea level pressure
P = atmosphere pressure at site elevation (units consistent with P0)
P0 = sea level atmospheric pressure
Buffer spaces
One of the many advantages of the heat balance approach is that buffer space temperatures can be
predicted by simply modeling an unconditioned room. These temperatures can be used as outside
boundary conditions for surfaces of adjacent conditioned spaces. Figure 6 shows typical ResHB results
for an attic with a dark asphalt roof. Changes in roof solar absorptance and inside surface long wave
emissivity (to represent radiant barriers) have the expected effects on predicted temperatures.
60 140
Temperature (°F)
Temperature (°C)
50 122
40 104
30 86
20 68
10 50
0 32
0 4 8 12 16 20 24
Hour Roof surface
Attic air
Infiltration
After review of available models, the AIM-2 model was selected for RHB (Walker and Wilson 1990,
Walker and Wilson 1998, and “enhanced model” in Chapter 26, ASHRAE 2001). As with other
simplified residential models, AIM-2 requires several input values that are difficult to determine,
including effective leakage area, leakage area distribution, and wind shelter parameters. ResHB provides
typical default values for these inputs. Leakage area can be specified based on pressurization test or
defaulted based on leakage classes defined by ANSI/ASHRAE Standard 119 (ASHRAE 1994).
AIM-2 is a single zone model. Infiltration leakage is determined for the entire building. In RHB, this
overall rate is allocated to rooms in proportion to volume – that is, the same air change rate is assumed to
apply to all rooms. Prior methods have variously allocated infiltration in proportion to exposed surface or
window area. Actual room leakage can be inward or outward and depends on room position relative to
the building neutral level and wind-induced pressure field. Thus there is no simple method for allocating
overall leakage other than using the average for all rooms.
Modeling of the interaction between mechanical ventilation and infiltration follows Palmiter and Bond
(1991) and Sherman (1992).
Distribution losses
ResHB duct losses are calculated using models specified in ANSI/ASHRAE 152-2004, Method of Test for
Determining the Design and Seasonal Efficiencies of Residential Thermal Distribution Systems and
Palmiter and Francisco 1997. These models are fully implemented in the RHB method, allowing room-
specific losses and gains to be included in estimates of air requirements and equipment capacity.
Framed constructions
The Loads Toolkit CTF-based conduction model assumes one-dimensional heat flow and thus requires
layer-by-layer construction descriptions as input. Framed constructions are common in residential
Modeling Assumptions
Practical application of the HB method requires that fixed or default values be established for as many
inputs as possible, both as a practical aid to the user and to achieve consistent results. The following
sections document the assumptions developed for RHB.
Internal gain
RHB internal gain assumptions are based on Building America 2003, which provides gain intensities and
schedules for significant residential end uses as a function of building floor area and number of
occupants. When estimating residential internal gains, care must be taken to distinguish between energy
consumption and space gain. For example, a clothes dryer uses significant energy, but most is exhausted
outside the space. In addition, RHB requires the radiant/convective/latent split for each gain source,
which Building America 2003 does not fully define. Estimates were developed from ASHRAE 2001 and
other sources as needed. These values are shown in Table 2 and have been incorporated into ResHB.
Note that both sensible and latent heat gain due to hot water use are neglected because they are not
covered in Building America 2003; these gains are probably small due to intermittent use and shower
exhaust fans but deserve further investigation.
Table 2. Fractional components of internal heat sources
Source Internal gain (to space)
Exhausted
Radiant Convective Latent
Refrigerator 0 1 0 0
Range .24 .16 .30 .30
Dishwasher .51 .34 .15 0
Clothes washer .40 .60 0 0
Clothes dryer .09 .06 .05 .80
Lighting .79 .21 0 0
Other appliances and plug loads .54 .36 .1 0
People (living) .33 .22 .45 0
People (sleeping) .30 .30 .40 0
1800 6143
1600 5461
1400 4778
1200 4095
Gain (Btu/h)
Gain (W)
1000 3413
800 2730
600 2048
400 1365
200 683
0 0
0 3 6 9 12 15 18 21 24
Hour Sensible
Latent
Figure 7. Hourly internal gain (all sources) for a 186 m2 (2000 ft2) house with 4 occupants
The gains specified in Building America 2003 are average values derived from energy consumption
measurements. It could be argued that higher values should be used for loads calculations, given that 600
W is less than the output of a single range burner. However, gains from normal residential activity are
intermittent and are absorbed by small space temperature swings. Higher assumptions are appropriate
only in cases where a significant gain is routinely expected. Occasional high gain situations, such as
social functions, should not be considered or handled with a parallel system designed for that condition.
Internal mass
The presence of internal mass (such as furniture) in a room has two effects on cooling load: the load is
increased due to enhanced surface area available for convective exchange and decreased due to the
storage (depending on its construction, space temperature swing, and presence of other mass such as a
slab floor). The ultimate impact of internal mass depends on which of these effects dominates in a
particular case. Note also that residential buildings typically have relatively small rooms and thus the
partition-to-floor area ratio can easily exceed 1 and often exceed 2. Partitions function as additional
internal mass.
ResHB parametric studies showed the impact of internal mass on sensible cooling load to be –2% to
+23%. Typical results are shown in Figure 8 and Figure 9 for a 162.6 m2 (1750 ft2) house in Washington,
DC with various amounts of 12 mm (.5 in) wood internal mass.
1.250
1.200
Relative cooling load
1.150
1.100
1.050
1.000
0.950
0.900
0 0.5 1 1.5 2 2.5
IM area / floor area Swing = 0
Swing = 0.83°C (1.5°F)
Swing = 1.67°C (3°F)
Swing = 2.5°C (4.5°F)
1.300
1.250
1.200
Relative cooling load
1.150
1.100
1.050
1.000
0.950
0.900
0 0.5 1 1.5 2 2.5
IM area / floor area Swing = 0
Swing = 0.83°C (1.5°F)
Swing = 1.67°C (3°F)
Swing = 2.5°C (4.5°F)
Other assumptions
Surface absorptance. Table 3 summarizes absorptance values recommended for typical load calculations.
Other values should be used as appropriate, particularly for roof outside absorptance if a reflective color
is specified.
Table 3. Default surface absorptances
Surface Short wave (solar) Long wave (thermal)
Outside Inside Outside Inside
Roof / ceiling (tilt < 60 °) .8 .45 .9 .9
Wall (180° < tilt ≤ 60°) .6 .45 .9 .9
Fenestration per FenClass per FenClass .84 .84
Floor (tilt = 180°) 0 .6 .9 .9
Material properties. ResHB includes default material properties gathered and reconciled from multiple
sources, as documented in Appendix A.
Toolkit Re-engineering
As stated above, the starting point for ResHB was the ASHRAE Loads Toolkit (Pedersen et. al. 2001).
Specifically, ResHB was created via incremental re-engineering of the Toolkit successive substitution
sample zone example. That module is a fully functional single zone model that uses the conduction
transfer function (CTF) calculations for building surfaces and the MRTBal radiant exchange model
(Liesen and Pedersen 1997; Walton 1980). Changes were introduced one at a time. Comparison of
output from each incremental change minimized the possibility of unintended changes.
The Toolkit is intended to allow exploration, understanding, and comparison of alternative loads models.
Its structure emphasizes clarity and modularity over efficiency. These objectives are secondary for
ResHB and ultimately an almost entirely new application resulted as the original code was adapted to the
residential loads application. Some of the significant changes include --
• Application framework. ResHB includes various utility capabilities such as error handling,
command line processing, and common report formatting functions.
• Streamlined input processing. The data dictionary file (IDD) is now simplified and embedded in
the application, making ResHB.exe a complete, standalone package. Unused input processing
features have been removed. An INCLUDE capability is now available, allowing common input
to be maintained in a single file. In addition, a rudimentary PARAMETER scheme has been
added.
• Unused models removed. As ResHB algorithms are determined, the alternative methods provided
in the Toolkit have been dropped. This makes the code smaller and eases development.
• Residential models added. Models not included in the Toolkit have been added for infiltration
and duct losses. Other models have been enhanced.
• RHB calculation algorithms. Temperature swing and master / slave control algorithms were
added.
Introduction
Background
An important aspect of 1199RP is demonstrating the validity of results, diagnosing deficiencies in the
calculation method or its computer implementation, and evaluating the system designs resulting from the
procedure.
A number of systematic validation and assessment methods have been developed for building
simulation/energy calculation programs in the last two decades, e.g. Judkoff, et al. (1983); Judkoff and
Neymark (1995), BRE (1988) and Bloomfield et al. (1988). Much less emphasis has been placed on
design load calculation procedures, perhaps since design load calculation methods have historically relied
less heavily on computer implementation than annual energy calculation (Rees and Spitler 1999)
Three types of test methods applied to whole building energy simulation programs can be identified
(Judkoff et al. 1983). These are experimental validation, inter-model comparison and analytical testing. In
any type of test, where results of a given code are compared with data from another source, three things
are implicitly tested, each of which may contribute to the overall ‘error’ in the results:
• The interpretation of the input data
• The model(s) or algorithm(s)
• The computer implementation of the algorithm(s)
Being most abstracted from the full complexities of real building simulation problems, analytical testing
has the advantage of offering the most certain form of reference or ‘truth’ model with which comparisons
can be made. The nature of analytical testing also makes it only applicable to limited cases for which
analytical solutions can be derived. Errors arising from the integrated performance of all the sub-models
and algorithms in a program are beyond the scope of an analytical test.
Experimental validation offers an approximate truth standard within the accuracy of the data acquisition
system and can be applied with any level of complexity. However, detailed, high quality, measurements
are usually expensive and time-consuming, even for one or a few cases.
Inter-model comparison cannot directly address the issue of a truth standard, but can be a very powerful
way of identifying errors by doing many comparisons quickly and inexpensively. Inter-model
comparisons of results from annual energy analysis codes have been attempted by a number of groups.
One notable attempt has also been made at devising a systematic diagnosis tool, BESTEST, based on
inter-model comparisons (Judkoff and Neymark 1995).
A further type of inter-model comparison can be identified that consists of a large number of test cases (of
the order one thousand) where certain parameters are systematically varied. Such a study using ‘design
day’ cooling load calculation methods was performed as part of ASHRAE 942-RP (Rees et al. 1998;
Spitler and Rees 1998). In order to check the residential heat balance (RHB) procedure, it is desirable
that this type of inter-model comparison be made in a highly-automated fashion, so that the RHB
procedure can be tested and evaluated throughout the development cycle.
Methodology
Types of comparison
Two types of inter-model comparison between the residential heat balance (RHB) load calculation
procedure and ESP-r are performed:
• Load Comparisons. In these comparisons, loads predicted by RHB are compared to those
calculated by ESP-r. These comparisons serve to verify RHB calculations and to reconcile RHB
and ESP-r input used for system design evaluation.
• System Design Evaluation. In these comparisons, the RHB procedure is used to generate design
system capacities. In ESP-r, the buildings are simulated with a residential cooling system, with
system capacity and room airflows taken from the RHB procedure. Room temperatures and
Parametric code
Both types of inter-model comparisons depend on a parametric code that identifies the specific case to be
run by both programs. A typical code is:
R2HsLCASDO000Q00m00AmmmmCwmhtmWwmmG1000mmmmPwwMwmFwm0m
Each sub-field of the code controls one aspect of input file generation. In this example, “Hs” indicates
that building “s” (shoebox) should be used, “LCASD” means that the location is San Diego, CA, and
O000 specifies building orientation of 0 (building front facing north). “Q00m00” means zero internal
gain, zero infiltration, medium cooling set point, zero room temperature swing, and zero set point setoff in
the controls. “Ammmm” means all interior surfaces have medium short wave and long wave
absorptivities. Other fields determine surface construction and configration, fenestration type and amount,
shadings, internal mass, etc. (e.g. “Cwmhtm” means the exterior ceiling is of wood frame construction
with medium insulation, and the roof above the hip attic has tile construction with medium short wave
absorptivity.) A complete description of the parametric code is found in the RHBGen documentation.
See Appendix D.
There are literally millions of combinations that can be generated using this scheme. In the results
presented below, cases have been chosen to capture typical residential building features and cover a
variety of applications.
Figure 10 illustrates the combined test process, which includes both types of comparisons. It starts with
the parametric run generator RHBGen (described below), which takes as input the basic parametric case
list, a description of the parameter variations and building prototypes. It then generates an expanded case
list and RHB input files, and runs RHB. Then, the ESP-r input generator (described below) takes in the
expanded case list and the system capacities and design airflow rates determined by RHB, and creates the
ESP-r input files. Perl scripts (described below) will then run the ESP-r simulations, both with an ideal
system (to determine the loads for a load comparison) and with a residential cooling system with RHB-
determined system capacities and design airflow rates (for the system design evaluation.) ESP-r results
are then processed by the ESP-r output processor and post processor (described below).
Prototypes
Expanded System
designs Loads
case list
ESP-r
Input files ESP-r output processor &
post processor
ESP-r
Design evaluation Loads comparison
ESP-r system
Figure 11. Predicted Percentage of Dissatisfied (PPD) as a function of Predicted Mean Vote (PMV).
Source: Ch. 8, Figure 13, 2001 ASHRAE Handbook of Fundamentals
This relationship is given by Equation 64 in Chapter 8 of the 2001 HOF:
[(
PPD = 100 − 95 exp − 0.03353PMV 4 + 0.2179 PMV 2 )]
Being able to determine PPD for any given room at any given time step is useful, but does not yet allow
us to develop a single figure of merit. Unfortunately, no single figure of merit was identified that
distinguished between a “good” system and a “bad” system. As described in the section “System design
evaluations”, a procedure has been developed which gives integrated PPD values that distinguish between
undersizing, inadequate distribution / control problems, and nighttime ambient overcooling.
1
Ground temperature at 1.2 m (4 ft) depth calculated from the procedure of Kusuda and Achenbach (1965) was
input to ESP-r. ESP-r calculates the ground surface temperature with a 6-node finite volume scheme, assuming the
1.2 m (4 ft) depth ground temperature as a constant.
2
The average temperature of the surrounding external buildings for a surface is estimated by using the temperature
of the zone surface, which has an azimuth difference of 180 Deg. This method is assumed to be applicable for
surfaces with elevations > -30 and < 30 deg. For all other surfaces (ie elevations < -30 and > 30 deg.) use a mean
weighted average of all external surfaces in all zones under simulation. (From ESP-r code)
9000
8000
15%
7000
RHB Loads (W)
6000
5000 -15%
4000
3000
2000
1000
0
0 1000 2000 3000 4000 5000 6000 7000 8000 9000
Figure 12. Peak cooling load comparison for the Great room
10000
9000
8000
15%
7000
6000
RHB Loads (W)
-15%
5000
4000
3000
2000
1000
0
0 1000 2000 3000 4000 5000 6000 7000 8000 9000
Figure 13. Peak cooling load comparison for the Living room
40
35
Temperature (C)
30
25
20
15
10
3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00
Hour
RHB ESP-r
Figure 14. Outside ground surface temperatures from RHB and ESP-r for Alamosa, CO
60
50
40
Temperature (C)
30
20
10
0
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00
Hour
Ambient dry bulb Great south surface Great north surface Great west surface
Figure 15. Comparison of the ambient dry bulb temperature and the outside surface temperatures
calculated from ESP-r (Great room)
2000
Cooling Load (W)
1500
1000
500
0
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00
Hour
RHB ESP-r
Figure 16. Comparison of cooling load calculated by RHB and ESP-r: with external long wave radiation
2500
2000
Cooling Load (W)
1500
1000
500
0
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00
Hour
RHB ESP-r
Figure 17. Comparison of cooling load calculated by RHB and ESP-r: without external long wave
radiation
In summary, ESP-r predicts higher cooling loads than RHB because of the different longwave radiation
source/sink temperature models. RHB uses very simple models, assuming that both other building
temperatures and the ground surface temperature are the same as the air temperature. ESP-r has
Figure 18 through Figure 21 show the result of the system design evaluations for 576 cases where the
system capacities and airflow rates were designed with the RHB program, using zero temperature swing.
Figure 18 shows the processed PPD values for the category of nighttime ambient cooling. For
convenience in identifying specific cases, they are plotted against case number. For most cases, the
processed PPD values fall in the range of 0-10%. The group of high values in this figure is for the light-
weight construction cases from Alamosa, Colorado, which has the cooling design condition of 27.8ºC
(82.0ºF) maximum dry bulb temperature and a 17.3ºC (31.1ºF) daily dry bulb range. As the outdoor air
temperature drops to 10.5ºC (50.9ºF) at night, and the sky temperature drops to –11.7ºC (-11.0ºF) at night,
it is not surprising that a lightweight house falls well below the setpoint of 24ºC (75.2ºF) at night.
Figure 19 shows the processed PPD values for the categories of system under-sizing and inadequate
distribution. For most cases, the processed PPD values of the category of system under-sizing fall in the
range of 0-2%; the processed PPD values for the category of inadequate distribution fall in the range of 1-
6%.
To understand the meaning of these values, Figure 20 and Figure 21 plot out the hourly PPD values and
hourly room air temperatures for two particular cases (case 17, case 20). In Case 17, the PPD attributed
to undersizing is zero; the PPD attributed to inadequate distribution is 6.3%, the worst case of all 576.
Looking at Figure 20, we can see that the actual hourly distribution of temperatures and PPD shows
problems in the slave zone (the living room).
On the other hand, Case 20, which has PPD attributed to undersizing equal to zero; and PPD attributed to
inadequate distribution of 2.7%, a more typical value. In Figure 21, we see some overcooling of the
living room in the morning, and a slight drift above the setpoint in the early evening.
Case 17 and 20 are both wood frame, medium insulation cases located in San Diego. For both cases, the
great room (master zone) has a north and an east window; the living room (slave zone) has a south and a
west window. However, Case 17 has single pane glazing with twice the area of the low-e double pane
glazing used in Case 20. This results in a significantly higher required system capacity – the Case 17
system capacity is over three times as high as the Case 20 system capacity. The combination of higher
transient heat gains, with different profiles in the two rooms, and larger system capacity causes Case 17 to
have significantly higher thermal discomfort.
30
25
20
PPD (%)
15
10
0
0 100 200 300 400 500 600
Case number
Figure 18. Processed PPD showing night time ambient cooling, system designed with zero temperature
swing
case 17
6
case 449
4
PPD (%)
2
case 20
0
0 100 200 300 400 500 600
Case number
Figure 19. Processed PPD showing system under-sizing and inadequate distribution, system designed
with zero temperature swing
40 24
Temperature (C)
30 18
PPD (%)
20 12
10 6
0 0
1 3 5 7 9 11 13 15 17 19 21 23
Hour
Figure 20. Hourly PPD vs. temperature for case 17, system designed with zero temperature swing
50 30
40 24
Temperature (C)
30 18
PPD (%)
20 12
10 6
0 0
1 3 5 7 9 11 13 15 17 19 21 23
Hour
Figure 21. Hourly PPD vs. temperature for Case 20; system designed with zero temperature swing
Figure 22 to Figure 24 show the results of the test set with RHB sizing the system capacities and room
airflow rates allowing a 1.67ºC (3.0ºF) temperature swing.
Figure 22 shows the processed PPD values for the category of nighttime ambient cooling. The processed
PPD values are slightly lower than those with zero design temperature swing. This is due to the fact that
the smaller system capacities cause the zone temperatures to swing upward during the day. Presumably,
the stored energy slightly reduces the downward temperature swing at night, resulting in slightly lower
values of PPD due to nighttime ambient cooling.
35
30
25
20
PPD (%)
15
10
0
0 100 200 300 400 500 600
Case number
Figure 22. Processed PPD showing night time ambient cooling, system designed with medium
temperature swing
case 355
7
case 149
6
5
PPD (%)
0
0 100 200 300 400 500 600
Case number
Figure 23. Processed PPD showing system under-sizing and inadequate distribution, system designed
with medium temperature swing
50 30
40 24
Temperature (C)
30 18
PPD (%)
20 12
10 6
0 0
1 3 5 7 9 11 13 15 17 19 21 23
Hour
Figure 24. Hourly PPD vs. temperature for case 355, system designed with medium temperature swing
where
qs = sensible cooling load, W (Btu/h)
ql = latent cooling load, W (Btu/h)
Ai = area of ith surface, m2 (ft2)
CFi = cooling factor of ith surface, W/m2 (Btu/h-ft2)
Opaque surfaces
The cooling load per unit area of opaque walls, ceilings, and non-slab floors is calculated as follows:
CFopq = U ⋅ (OFt ⋅ ∆T + OFb + OFr ⋅ DR ) (4)
where
CFopq = opaque surface cooling factor, W/m2 (Btu/h-ft2-F)
U = construction U-factor, W/m2-K (Btu/h-ft2-F)
∆T = design dry-bulb temperature difference (outdoor – indoor), °K (°F)
DR = daily range of outdoor dry-bulb temperature, °K (°F)
OFt, OFb, OFr = coefficients from Table 7
Table 7. Opaque surface coefficients
Surface type Construction OFt OFb OFr
Ceiling or knee wall Wood frame .62 14.1 ⋅ ∝roof – -0.23
adjacent to vented attic 4.3
Ceiling/roof assembly Wood frame 1 39 ⋅ ∝roof – 6.8 -.42
where
CFslab = slab cooling factor, W/m2 (Btu/h-ft2-F)
Fenestration
Fenestration cooling factors are calculated as follows:
CFfen = U ⋅ ( ∆T − .49 ⋅ DR ) + FFs ⋅ PXI ⋅ SHGC ⋅ IAC (6)
where
CFfen = fenestration cooling factor, W/m2 (Btu/h-ft2)
U = fenestration NFRC heating U-factor, W/m2-°K (Btu/h-ft2-F)
DR = daily range of outdoor dry bulb temperature, °K (°F)
∆T = cooling design temperature difference, °K (°F)
FFs = load factor (see Table 8)
PXI = peak exterior irradiance, including shading modifications (see below), W/m2 (Btu/h-ft2-F)
SHGC = fenestration rated or estimated NFRC Solar Heat Gain Coefficient
IAC = interior shading attenuation coefficient
Table 8. Fenestration coefficients
Exposure FFs
N .17
NE .09
E .17
SE .25
S .45
SW .54
W .48
NW .34
Horiz .66
Peak exterior irradiance (PXI) is the hourly maximum solar gain incident on the surface.
PXI = Tx Et (unshaded) (7)
PXI = Tx ( E d +(1 − Fshd ) ED ) (shaded) (8)
where
PXI = peak exterior irradiance for exposure, W/m2 (Btu/h-ft2-F)
Et, Ed, ED = peak total, diffuse, and direct irradiance for exposure, W/m2 (Btu/h-ft2-F)
Tx = Transmission of exterior attachment, see Table 10
Fshd = fraction of fenestration shaded by permanent overhangs, fins, or environmental obstacles
For horizontal or vertical surfaces, irradiance values can be obtained from Table 9 for primary exposures
or Algorithm 1 for any exposure. Skylights with slope less than 30° from horizontal should be treated as
horizontal. Steeper slopes, other than vertical, are not supported by the RLF method.
Horizontal surfaces
Et = 970 + 6.2 L - 0.16 L Ed = MIN(Et ,124 )
2
E D = Et − E d
Vertical surfaces
ψ
Ψ = (normalized exposure)
180
Et = 462.2 + 1625Ψ + 6183Ψ + 3869Ψ + 32.38ΨL
3 4
2
1.040L
+.3237ΨL − 12.56L − .8959L +
2 2
Ψ +1
121 4 L
Ed = MIN( Et , 392.1 - 138.6Ψ + 2.107Ψ L − )
Ψ +1
E D = Et − E d
where
Et, Ed, ED = peak hourly total, diffuse, and direct irradiance, W/m2 (multiply by
.317 to convert to Btu/h-ft2)
L = site latitude, °N or °S
ψ = exposure (surface azimuth), ° from south (-180 - +180)
The shaded fraction, Fshd, can be taken as 0 for fenestration in full sun and 1 for any fenestration that is
shaded by adjacent structures or other obstacles during peak hours. Fshd for simple overhang
configurations can be calculated as follows (more complex configurations should be analyzed with the
RHB method):
⎛ SLF ⋅ Doh − X oh ⎞
Fshd = MIN ⎜1, MAX (0, )⎟ (9)
⎝ H ⎠
where
SLF = Shade line factor from Table 11
Doh = Depth of overhang (from plane of fenestration), m (ft)
Xoh = Vertical distance from top of fenestration to overhang, m (ft)
H = Height of fenestration, m (ft)
The shade line factor (SLF) is the ratio of the distance a shadow falls beneath the edge of an overhang to
the width of the overhang (Table 11). Therefore the shade line equals the SLF times the overhang depth.
The tabulated values are the average of the shade line values for 5 h of maximum solar intensity on
August 1 on each wall exposure shown. North-, northeast-, and northwest-facing windows are not
effectively protected by roof overhangs; in most cases, they should not be considered shaded.
Table 11. Shade line factors (SLF)
Window Latitude, degrees N
exposure 24 32 36 40 44 48 52
East 0.8 0.8 0.8 0.8 0.8 0.8 0.8
SE 1.8 1.6 1.4 1.3 1.1 1.0 0.9
South 9.2 5.0 3.4 2.6 2.1 1.8 1.5
SW 1.8 1.6 1.4 1.3 1.1 1.0 0.9
West 0.8 0.8 0.8 0.8 0.8 0.8 0.8
Note: Shadow length below the overhang equals the shade line factor times the overhang depth.
where
Qbal = balanced ventilation air flow rate, L/s (cfm)
Qsup = total ventilation supply air flow rate, L/s (cfm)
Qexh = total ventilation exhaust air flow rate (including any combustion air requirements), L/s
(cfm)
Qunbal = unbalanced air flow rate, L/s (cfm)
Qvi = combined infiltration/ventilation flow rate (not including balanced component), L/s (cfm)
Note that unbalanced duct leakage can produce additional pressurization or depressurization. This effect
is included in distribution losses, discussed below.
The cooling (or heating) load due to ventilation and infiltration is calculated as follows, taking into
account the effects of heat / energy recovery ventilation (HRV / ERV) equipment:
qvi ,s = Cs ⋅ (Qvi + (1 − ε s ) ⋅ Qbal ,hr + Qbal ,oth ) ⋅ ∆T (14)
where
qvi,s = sensible ventilation/infiltration load, W (Btu/h)
Cs = air sensible heat factor, 1.23 W/(L/s)-K (1.1 Btu/h-cfm-F) at sea level
εs = HRV/ERV sensible effectiveness
Qbal,hr = balanced ventilation flow rate supplied via HRV/ERV equipment, L/s (cfm)
Qbal,oth = other balanced ventilation supply air flow rate, L/s (cfm)
∆T = indoor-outdoor temperature difference, °K (°F)
qvi,l = latent ventilation/infiltration load, W (Btu/h)
Cl = air latent heat factor, 3010 W/(L/s) (4840 Btu/h-cfm) at sea level
∆W = indoor-outdoor humidity ratio difference
qvi,t = total ventilation/infiltration load, W (Btu/h)
Ct = air total heat factor, 1.2 W/(L/s)-(kJ/kg) (4.5 Btu/h-cfm-(Btu/lb)) at sea level
εt = HRV/ERV total effectiveness
∆h = indoor-outdoor enthalpy difference, kJ/kg (Btu/lb)
Internal gain
The contributions of internal gains to peak sensible and latent loads are:
qig , s = G0, s + Gcf , s ⋅ Acf + Goc , s ⋅ N oc (18)
qig ,l = G0,l + Gcf ,l ⋅ Acf + Goc ,l ⋅ N oc (19)
where
Distribution losses
The allowance for distribution losses is calculated as follows:
qdl = Fdl ⋅ qs (20)
where
qdl = distribution loss, W (Btu/h)
Fdl = distribution loss factor, from Table 12
qs = building sensible load, W (Btu/h)
Development Overview
The RLF formulation is conceptually transparent and hand-tractable: the “loads” from each wall, window,
and other gain sources are calculated and summed to get the total load. Unfortunately, however, it is not
possible to find invariant models for each load component because of interactions among them. For
example, a major interaction occurs between opaque surfaces and fenestration – the load resulting from
solar gain is lagged and moderated by differing amounts depending on surface construction. Even simple
convective gains, such as infiltration and ventilation, present difficulty because they should be evaluated
at the building-dependent peak hour. Note that RHB completely avoids these difficulties: 24 hour
calculations allow gains to combine according to their case-specific profiles and the heat balance
procedure accurately represents component interactions.
Development of a load-component method such as RLF requires that the significant interactions be
identified and addressed (via configuration-specific load-component calculations), eliminated (by
restricting the configurations to which the method is applicable), or neglected if the effects are deemed
Room temperature swing is one reason adjustment is required. Assuming a fixed indoor temperature, as
is typically done in non-residential procedures, results in excessive loads for the residential case. Better
overall system performance and cost effectiveness results when equipment is sized to allow some
temperature variation at design conditions. Averaging of gains derived assuming fixed room temperature
mitigates their excessive peak.
The RLF development procedure avoided adjustments by relying on RHB cooling loads calculated with
temperature swing and deriving required factors using linear regression. Equation (2) was treated as a
model for which sub-models and coefficients were needed. Later sections of the paper present the
approaches used for each load component. The regression approach has two advantages. First,
significant independent variables and efficient model forms are naturally identified by the regression
process. If a model does not accurately predict load, that is revealed by poor statistical figures of merit.
Second, no averaging or other semi-empirical adjustments are required.
From a processing point of view, RLF was developed using three PC applications: ResHB as described
above (loads calculations), RHBGen (parametric case generator), and the R statistics package.
RHBGen was developed to support ResHB testing and RLF development (see Appendix D and the
accompanying 1199-RP Source Code and Data CD). RHBGen generates and runs ResHB input files
under the control of multi-character parametric codes. Fields within the code control various aspects of
the case to be run, such as base prototype (fundamental building geometry), location, orientation,
constructions, fenestration type and area, and so forth. RHBGen combinatorically varies code fields,
allowing sets of hundreds or thousands of ResHB runs to be constructed and executed. On typical
Pentium-based PCs, the RHBGen / ResHB system can complete several hundred cases per minute.
ResHB writes suitable results files for regression analysis and other post-processing.
The R package (R 2004) is an open-source system with extensive statistical and data visualization
capabilities. For RLF development, the linear regression and data plotting procedures were used. R is
particularly suitable for RLF development because it includes a script language for automation of
complex analysis sequences. The R scripts used in this work can be found on the 1199-RP Source Code
and Data CD that accompanies this report.
Modeling Assumptions
As discussed above, RLF cooling load procedures were developed by using linear regression analysis of
sets of ResHB results. To generate the input data for regression, specific combinations of inputs were
varied while others were maintained at typical values. For most variables, three levels were identified: L
= minimum, M = typical, and H = maximum. Even with only three levels and relatively few variables,
Prototype building
Cooling loads for many variations of a single prototype building were the basis for the regression
analysis; Table 13 summarizes the prototype characteristics and Table 14 shows construction details. The
floor plan of the prototype building was square with 4 rooms, one in each corner. Note that ResHB
surfaces need not be geometrically consistent, allowing exterior wall area to be based on an assumed
typical-width rectangular plan and that area to be distributed equally on all facades. The 4 room plan was
chosen so there was a reasonable ratio of interior partition to floor area and to limit radiant transfer among
exterior walls. The sensible cooling load used as the regression independent variable was the maximum
value of the combined 24 hour profile derived by summing the room loads for each hour.
Table 13. Prototype building characteristics
Item Value Notes
Conditioned floor area 168 m2 (1808 ft2) Typical size
Height 2.5 m (8.2 ft) Single story
Exterior wall area 142.4 m2 (1533 ft2) Average width assumed to be 8.5 m (28 ft),
yielding perimeter = 57 m (187 ft)
Interior partition area 140 m2 (459 ft2) 83% of conditioned floor area
Nominal fenestration 27.2 m2 (89.2 ft2) windows window area = 16% of floor area
2 2
1.68 m (18 ft ) skylight skylight area = 1% of floor area
clear double glazed (U = 2.73
2 2
W/m K (.48 Btu/h-ft -F), SHGC
= 0.76)
Fenestration variation All cases run with 200%
nominal area. IAC values
varied L = 0, M = 0.5, H = 1.
Internal mass 168 m2 (1808 ft2) of 12 mm 1199-RP default
(0.5 in) wood
Indoor design 24 °C (75.2 °F)
temperature
Indoor temperature 1.67 °K (3 °F)
swing
Infiltration Leakage class E (normalized Reasonably tight contemporary construction
leakage = .34) (ASHRAE Standard 119)
Internal gain Default Based on Building America 2003, see below.
Surface exterior solar walls: 0.6
absorptance roof: 0 – 1 (varied)
Surface interior Beam solar gain: floor: 0.6,
absortance internal mass: 0.3, other: 0
Diffuse solar gain: all surfaces:
0.6
Orientation 0 ° and 45 ° All 8 primary orientations considered.
The ResHB application uses these inputs to generate 24-hour design sequences that drive the heat balance
simulation. Hourly incident solar radiation was calculated using the ASHRAE clear sky model
(ASHRAE 2001) with updated coefficients (Machler and Iqbal 1985).
The combination of 8 design conditions and 3 latitudes resulted in 24 runs for each prototype variant.
Regression Strategy
It was not practical to perform one regression analysis to identify all RLF coefficients because the
overwhelming number of case combinations that would have been required. Instead, an iterative series of
linked regressions was performed. Equation (2) was applied to ResHB loads results and rearranged to
isolate the envelope load component:
qenv = qs ,rhb − qig ,rhb − qvi ,rhb (21)
where
qenv = Envelope cooling load component = ∑ Ai ⋅ CFi in Equation (2), W (Btu/h)
qs,rhb = ResHB sensible cooling load, W (Btu/h)
qig,rhb = ResHB sensible internal gain at peak hour (simultaneous with qs,rhb), W (Btu/h)
qvi,rhb = ResHB sensible ventilation/infiltration at peak hour (simultaneous with qs,rhb), W (Btu/h)
The envelope cooling load is the sum of the load components from the various envelope elements:
qenv = q fen + qceil + qwall + q floor (22)
Each component term of Equation (22) was estimated using a separate data set described in Table 17.
Each data set contains ResHB loads based on varying inputs relating to the term under consideration
while fixing other inputs at M or nominal values. The component regressions were performed in the
sequence shown in Equations (23) - (26) and the results of each were applied to the next step. (The
details of each component model are discussed below.) Initial (iteration 0) estimates were set by hand
using suitable prior results. It was determined by trial and error that 5 iterations achieved essentially
complete convergence.
+1
qˆ ifen = qenv − qˆceil
i
− qˆwall
i
− qˆ ifloor (23)
i+1
qˆ
ceil = qenv − qˆ i+1
fen − qˆ i
wall − qˆ i
floor (24)
i +1 i +1 i +1
qˆ
wall = qenv − qˆ fen − qˆ ceil − qˆ i
floor (25)
i +1 i +1 i +1 i +1
qˆ floor = qenv − qˆ fen − qˆ ceil − qˆ wall (26)
where
qˆ xi = ith iteration estimated load component for fenestration, ceiling, wall, or floor, W (Btu/h)
Component Models
Ventilation and infiltration
As discussed above, typical infiltration was included in the ResHB runs used to generate regression data,
but the cooling load induced by this air leakage was subtracted from the load used in the envelope
regressions. Thus the loads predicted by the regression models implicitly assume 0 air leakage.
Equation (10) was developed to provide a simple method for estimating infiltration leakage for RLF.
ResHB calculates infiltration using the AIM-2 model (Walker and Wilson 1990, Walker and Wilson
1998), which is too complex for practical hand application. The AIM-2 model was exercised over a range
of temperature differences and building heights. Other assumptions included shelter class 4, flue shelter
class 2, and wind speed multiplier values from Table 10, Chapter 26, ASHRAE 2001. Leakage
distribution was assumed to be walls=.5, ceiling=.25, floor=.25 (R=1, X=0), all proportionately reduced if
flue is present. The maximum flue leakage fraction considered was 0.5. Regression was used to find the
form of Equation (10) and the Ix coefficients. The underlying functional form of the AIM-2 model is not
linear but the simple form of Equation (10) was maintained for ease of application. The regression model
yielded an adjusted R2 of .94. Figure 25 compares results from the regression to those from AIM-2.
Because of minimal air density dependence, Equation (10) is valid at any elevation.
The procedure for combining mechanical ventilation with infiltration air flows, shown in Equations (11) -
(13), follows Palmiter and Bond 1991 and Sherman 1992.
150
RLF infiltration (L/s)
100
50
0
0 50 100 150 200
Figure 25. Predicted infiltration leakage rates, AL = 1000 cm2 (155 in2) and representative range of
stack height, temperature difference, and flue fraction. RLF values from Equation (10), see text. (1 L/s =
2.12 cfm)
Internal gains
RHB internal gains are based on Building America 2003, which specifies gain intensities and schedules
for residential appliances, lighting and occupants. Experiments with these gains and schedules in ResHB
revealed that the sensible cooling load attributable to internal gains is generally approximated by the total
sensible internal gain during the peak cooling hour. This is not necessarily expected, since a significant
fraction of the gain is radiant and has a delayed load impact. The removal of load due to internal gain in
Equation (21) is based on this approximation.
For RLF, Equations (18) and (19) are the aggregated Building America gains using 4 PM schedule
values, that time being a common peak cooling hour for typical residential construction. Consideration
was given to developing a model that predicts the peak cooling hour so a more accurate internal gains
formulation could be included. However, such an addition to RLF was deemed excessively complex.
Opaque surfaces
The model forms for opaque surface CFs were found by experimentation. Prior methods, both residential
and non-residential, have used an Equivalent Temperature Difference (ETD) or Cooling Load
Temperature Difference (CLTD) form, where ETD or CLTD = A + ∆T - DR/2 (where A is a constant, ∆T
is the outdoor-indoor temperature difference, and DR is the daily range). This was taken as a starting
point for RLF. A coefficient was added for ∆T and multipliers other than 0.5 were allowed for DR. In
some cases, the ∆T coefficient was found to be a value very close to 1, in which case it was dropped from
the regression and forced to be 1. In other cases, coefficients were found to be not significant and
Fenestration
A goal for the fenestration model was to separation of latitude-dependent exterior effects from building-
dependent effects. It was determined by experimentation that this is achieved by factoring out peak hour
irradiance incident on the fenestration exterior, leading the PXI formulation shown in Equation (6). As
with the opaque surface models, this form is similar to prior residential and non-residential methods.
Many combinations of effective window aperture were included in the regression data set used to find the
FFs coefficients. The final adjusted R2 value was over .995.
An attempt was made to eliminate exposure-specific FFs values, leaving PXI as the only exposure-
dependent input. This produced significantly worse regression results. The FFs coefficients (Table 8)
show a physically reasonable relationship with exposure. East surfaces produce less cooling load per unit
irradiance than do west surfaces, as is expected. Prior methods that relied on averaging show a less
plausible E/W and SE/SW symmetry.
Distribution losses
Duct losses can be calculated using models specified in ANSI/ASHRAE 152-2004, Method of Test for
Determining the Design and Seasonal Efficiencies of Residential Thermal Distribution Systems and
Palmiter and Francisco 1997. These models are fully implemented in the ResHB. Using typical input
values, ResHB was exercised to produce Table 12 suitable for use with RLF hand calculations
Heating Loads
The RLF procedure for heating loads calculation is identical, in most respects, to previously published
ASHRAE 2001 residential heating load calculation procedures. The heating load calculation is based on
a steady-state UA∆T calculation, with no solar radiation and no internal heat gains. Infiltration leakage
rate is based on Equation (10). The calculation procedure for heat losses from surfaces in contact with the
ground has been revised as described in the following sections.
Heat flow
lines
Radial isotherms
(from intersection
of grade and Radial isotherms
basement wall) (from intersection
of basement wall
and floor)
Figure 26. Heat flow from basements (2001 ASHRAE Handbook of Fundamentals)
In the RLF procedure, the original Latta and Boileau work was revisited and reformulated in a more
flexible manner. The revised procedure allows for variation of the soil thermal conductivity and, if
desired, partial wall insulation with any thermal resistance. Furthermore, an analytical expression for the
average U-factor has been developed, along with new tables. This may be summarized as follows.
In cases where the basement wall is partially insulated, it will be desirable to calculate the heat loss
separately for portions of the wall with differing amounts of insulation. Consider the region between
depth z1 and z2 in Figure 27. (Here z1 and z2 can be any region of the wall, including the entire wall.
z1
zf z2
For the region of interest, in steady-state heat transfer, there are several thermal resistances of interest –
the soil, the concrete wall, the insulation (if any) and the inside surface resistance. If all thermal
resistances besides the soil are lumped into a single value, Rother, the average U-factor between the
basement air and the ground temperature is:
2k soil ⎡ ⎛ 2k soil Rother ⎞ ⎛ 2ksoil Rother ⎞ ⎤
U avg ,bw = ⎢ ln ⎜ z2 + ⎟ − ln ⎜ z1 + ⎟⎥ (27)
π ( z2 − z1 ) ⎣ ⎝ π ⎠ ⎝ π ⎠⎦
where
Uavg,bw = average U-factor between basement air and ground temperature over region of interest
shown in Figure 27, W/m2-K (Btu/h-ft2-F)
ksoil = soil thermal conductivity, W/m-K (Btu/h-ft-F)
z1 = depth of upper bound of region of interest (see Figure 27), m (ft)
z2 = depth of lower bound of region of interest (see Figure 27), m (ft)
Rother = combined resistance of wall, insulation and surface conductance, m2-K/W (ft2-h-F/Btu)
While values of soil conductivity vary widely with soil type and moisture content, a typical value of 1.4
W/m-K (0.8 Btu/h-ft-F) was used in past editions of the Handbook of Fundamentals to tabulate U-factors.
Rother is the sum of the resistance of the concrete wall, insulation (if any) and the inside surface resistance.
In past editions of the Handbook of Fundamentals, Rother was approximately 0.25 m2-K/W (1.47 ft2-h-
F/Btu) for uninsulated concrete walls. For these parameters, Uavg,bw is tabulated for a range of depths and
insulation levels in Table 18.
Table 18a. Average U-factor for basement walls with uniform insulation (SI units)
Table 18b. Average U-factor for basement walls with uniform insulation (I-P units)
Uavg,bw from grade to depth, Btu/h-ft2-Fa
Depth (ft) Uninsulated R-5 R-10 R-15
1 0.432 0.135 0.080 0.057
2 0.331 0.121 0.075 0.054
3 0.273 0.110 0.070 0.052
4 0.235 0.101 0.066 0.050
5 0.208 0.094 0.063 0.048
6 0.187 0.088 0.060 0.046
7 0.170 0.083 0.057 0.044
8 0.157 0.078 0.055 0.043
a
Soil conductivity is 0.8 Btu/h-ft-F; insulation is over entire depth.
For other soil conductivities and partial insulation, use
equations.
a
Soil conductivity is 1.4 W/m-K; floor is uninsulated.
Table 20b. Heat loss coefficient F2 of slab floor construction (I-P units)
Construction Insulation F2 (Btu/h-ft-
F)
8 in. block wall, Uninsulated 0.68
brick facing R-5.4 from
0.50
edge to footer
4 in. block wall, Uninsulated 0.84
brick facing R-5.4 from
0.49
edge to footer
Metal stud wall, Uninsulated 1.20
stucco R-5.4 from
0.53
edge to footer
Poured concrete wall Uninsulated 2.12
with duct near R-5.4 from
a 0.72
perimeter edge to footer
a
Weighted average temperature of the heating duct was assumed at 110ºF during the
heating season (outdoor air temperature less than 65ºF).
Verification of Results
The RLF method was added to the ResHB application, allowing RLF vs RHB cooling load calculations to
be performed on test cases. Figure 28 shows typical results of such a comparison for a building not
involved in the regression process and using design weather data for 20 diverse U. S. locations. As can be
seen, there is generally good agreement but a trend remains in that RLF predicts too high for low loads
and too low for high loads. This is being investigated and may lead to model refinement.
0 10 20 30 40 50 60 70
70
20
60
15
50
RLF (kBtu/h)
40
RLF (kW)
10
30
20
5
10
MAE: 3.3% = 548 W (1870 Btu/h)
RMSE: 4.1% = 679 W (2318 Btu/h)
MBE: 0.51% = 84.8 W (289 Btu/h)
Cor: 0.966
0
0
0 5 10 15 20
RHB (kW)
Figure 28. RLF vs. RHB sensible cooling load comparison. Test building calculated for representative
range of climate and construction conditions (1280 cases)
Fundamental Research
• Interior convection correlations. 1199-RP has shown significant sensitivity to interior convection
coefficients for some cases. Yet, the ASHRAE Handbook still uses single values determined for
natural convection conditions over 60 years ago. More recent research has been useful to some
degree, but is not well aligned to common building practice. The suggestion is to develop
convection correlations that are harmonized to standard diffuser design practices. There are five
outlet categories in Chapter 32 of the 2001 HOF. Convection correlations ought to be developed
that match these categories for realistic ranges of ADPI.
• Exterior convection correlations. There has been some recent work applicable to both residential
scale buildings (Yazdanian and Klems 1994) and high rise buildings (Loveday and Taki 1998).
However, in both cases, there are still gaps in applicability – neither covers roofs and the latter
does not give any guidance for still air conditions.
• Longwave radiation – building to sky. The Berdahl and Martin model seems to be well-accepted
for predicting clear sky temperature seen by horizontal surfaces. However, sky temperature seen
by vertical surfaces, which “see” the sky through longer path lengths through the atmosphere, has
only been handled heuristically (Walton 1983, McQuiston et. al. 2000).
• Longwave radiation – building to surroundings. As shown by 1199-RP, peak cooling loads for
some buildings are surprisingly sensitive to longwave radiation between the building and
surroundings. Two sets of seemingly reasonable assumptions give differences on the order of
10% of the peak cooling load.
Applied Research
• Sky models – solar radiation. The current ASHRAE ABC model has some questionable features,
e.g. how, precisely, to treat the clearness number and it is only applicable for the continental U.S.
More detailed models are available but usually require parameters that are unknown to design
engineers.
• Ground heat transfer – A whole range of slab-on-grade and basement heat transfer models have
been developed. However, relatively few are suitable for design load calculation procedures.
The best prospects are two correlation-based methods (Beausoleil-Morrison and Mitalas 1997,
Krarti and Choi 1996). However, both methods have large numbers of coefficients and users of
either method would face some difficulties.
Fenestration
The Loads Toolkit fenestration model requires input of angular SHGC and absorption values. This is not
practical for a routine use.
ResHB implements an idea designated “fenestration class.” A fenestration class can be thought of as a
family of fenestration types that exhibit common behavior. The fenestration class embodies the ratio of
transmission to absorption and the angular characteristics of the fenestration system. An actual
fenestration is specified by its NFRC-rated U-factor and SHGC plus its fenestration class. The required
angular characteristics are taken from the fenestration class and are scaled by the ratio of rated SHGC to
nominal (fenestration class) SHGC.
As shown in Figure 29, 6 classes are sufficient represent the overwhelming majority of residential
glazings (although there is some error for reflective glazing) --
• Clear single
• Heat absorbing single (for heat absorbing and reflective single glazings)
• Clear double (for clear double plus low-E double)
• Heat absorbing double (for heat absorbing and reflective double)
• Clear triple (for clear and low-E triple)
• Heat absorbing triple (for heat absorbing and reflective triple)
1H
1R
2C
Glass Type
2EH
2EL
2H
2R
3C
3H
1.300
Relative cooling load
1.250
1.200
Swing = 0
1.150
Swing = .83 C
1.100
Swing = 1.67 C
1.050
Swing = 2.5 C
1.000
0.950
0.900
0 0.5 1 1.5 2 2.5
IM area / floor area
1.300
Relative cooling load
1.250
1.200
Swing = 0
1.150
Swing = .83 C
1.100
Swing = 1.67 C
1.050
1.000 Swing = 2.5 C
0.950
0.900
0 0.5 1 1.5 2 2.5
IM area / floor area
1.300
Relative cooling load
1.250
1.200
Swing = 0
1.150
Swing = .83 C
1.100
Swing = 1.67 C
1.050
1.000 Swing = 2.5 C
0.950
0.900
0 0.5 1 1.5 2 2.5
IM area / floor area
1.300
Relative cooling load
1.250
1.200
Swing = 0
1.150
Swing = .83 C
1.100
Swing = 1.67 C
1.050
1.000 Swing = 2.5 C
0.950
0.900
0 0.5 1 1.5 2 2.5
IM area / floor area
Figure 33. Effect of internal mass, single exposure room, slab floor
For several of the walls, a comparison has been made for a design day where the exterior surface
temperature varied cyclically between 30 C and 15 C and the interior surface temperature was held at 20
C. Interior heat fluxes calculated with the 1145-RP CTFs are compared to interior heat fluxes calculated
with the equivalent homogeneous layer method. (After using the method to determine the properties of
the equivalent homogeneous layer, CTFs are calculated using procedures from the Loads Toolkit. These
CTFs are used to calculate the interior heat fluxes.) Sample results for walls 1 and 18 are shown in Figure
34 and Figure 35. As can be seen, the approximate method gives approximately correct results. It is not
perfect, but may be the best that can be done at present. (Also note that wall 18 has some error in the
CTF coefficients – it may be that some of the discrepancy is due to this problem.)
Wall Assembly -- 1
5
4.5
4
3.5
3
Flux (W/m2)
2.5
2
1.5 Flux – 1145RP CTF
1 Flux-Homogenous
0.5 Approx.
0
-0.5
-1
-1.5
-2
-2.5
0 2.5 5 7.5 10 12.5 15 17.5 20 22.5 25
Time
Figure 34. Wall 1 of interior heat flux calculated with 1145-RP CTF and CTF generated with equivalent
homogeneous layer method
1.5
1 Flux – 1145RP CTF
0.5 Flux-Homogenous
Approx.
0
-0.5
-1
-1.5
-2
0 2.5 5 7.5 10 12.5 15 17.5 20 22.5 25
Time
Figure 35. Wall 18 interior heat flux calculated with 1145-RP CTF and CTF generated with equivalent
homogeneous layer method
Material properties
In spite of the fact that the building simulation field is approaching 40 years old, it remains difficult to
find a coherent and well-documented list of the properties of common building materials suitable for use
in detailed computer models.
In order to assemble material property assumptions for this project, a number of sources have been
reviewed and best-estimate values selected and built into ResHB. Sources included Andersen et. al. 1999,
CEN 1999, Forest Products Laboratory 1999, ASHRAE 2001 (Table 4, Chapter 25 and Table 22, Chapter
29), Berkeley Solar Group 1986, NIST 2000, Pedersen et. al. 1998, and ESRU 2000.
runID is the base name for input and output files for the run (see below)
Options can be any combination of the following --
Option Effect
-C<filepath> Case file. If present, requests that a 1-line, comma-separated result summary be appended to the
specified file. This file is useful for import into Excel.
-G<filepath> Regression export file. If present, requests that a 1-line comma-separated result summary be appended to
the specified file. This results summary is similar to the case file summary but includes additional building
characteristics data.
-I<dirList> Include directories. Specifies a ;-delimited list of directories that ResHB searches for include files (the
current directory is searched first). Example: -Iinc1;c:\inc2
-M<xxx> Flags selecting alternative models, consisting of one or more upper case plus lower case pairs. The upper
case letter identifies the model, the lower case selects the specific alternative, as described below. For
example –MSvIt selects the visibility solar model and the Threlkeld incident solar model. The choice with *
is the default.
Model Descripton
C Inside surface convective coefficient
x* Runtime weighted. System on: 5 W/m2-K on all surfaces. System off: ASHRAE
still air values with flow-direction enhancement; linear transition of value from ∆T =
-1 °K - +1 °K.
s Same as ‘x’ except no linear transition.
y Same as ‘x’ except system on = 4.5 W/m2-K
a Fixed coefficients: 1.25 ceiling, 4.67945 wall, 4.37 floor (all W/m2-K)
d TARP detailed model (∆T^.333 and tilt dependent) (from Toolkit)
v Fisher ACH model (from Toolkit)
E Elevation adjustment of inside and outside convective coefficients
1* Enabled
0 Disabled
I Sloped surface incident diffuse solar (i.e. short wave sky model)
t* ASHRAE Threlkeld-Jordan “Y” model
i Isotropic sky model: view factor = 0.5 · (1 + cos( tilt))
p Perez
K Sky temperature
b* Berdahl/Martin
6 tAir – 6 °K
S Clear sky solar
a ASHRAE HOF
r* ASHRAE with Machler/Iqbal revised coeffiients
v Machler/Iqbal visibility
0 No incident solar (for testing)
X Outside surface convective coefficient
c* Constant 12.5 W/m2-K
m MoWiTT (ignoring wind direction)
flag effect
+ Append report to pre-existing runID.rep (if it exists). Default action is to overwrite prior file.
A building geometry (surface areas)
B room energy balance
C construction CTFs (for all constructions found in input file)
c construction CTFs (for constructions used in current run)
D generate additional report runID.dbg containing detailed debugging information; the
contents of this file changes as needed
E input echo with all built-in and include file text
e input echo, primary input file only
G Room internal gains with GAINITEM breakdown
g Room internal gains, totals only
S hourly surface info
s surface input summary
U building UA
W weather data summary
-X<xxx> Results file options. Flag characters which specify what is include in the results file. The results file,
runID.res, is a comma-separated file intended for import into Excel and other applications. By default, the
results file includes only a loads summary.
flag effect
B room energy balance
D room details (numerous 24 hour values for plotting)
S Solar model(s) output (re inter-model comparison)
W weather data in format suitable for plotting
Files
ResHB reads or writes the following files.
File name Use Description
<runID>.rhb Input, must exist Input file.
ResHB.log Output, append (created if does not Log file. Information is appended to this file
exist) documenting each run, showing date/time of run,
general information about the run, and any error
messages.
ResHB.err Output, append (created if does not Error file. Only written if errors are detected. Can serve
exist) as an error flag – delete before a group of runs, if exists
at end, there has been at least one error.
-C<filepath> Output, append (created if does not Case results file. One line of data appended to this file
exist). Written only if valid file per run in comma separated format. Useful for import
specified and if no errors into Excel for analysis and plotting of run sets (see
below).
-G<filepath> Output, append (created if does not Regression export file. One line of data appended to
exist). Written only if valid file this file per run in comma separated format. Useful for
specified and if no errors import into regression applications for post processing
(see below)
<runID>.rep Output, overwritten. Report file. Loads results plus optional info selected via
the –R command line switch.
<runID>.res Output, deleted, then created if run Results file. Room, zone, and system level results in a
successful format for import into post-processing programs (format
described below). Additional info can specified via the
–X command line switch
If –C<filename> is specified on the command line, each ResHB run adds one line <filename> (creating
the file if needed). The line consists of comma-separated items to support simple import into Excel and
other applications, as follows:
where:
• runID runID from command line (root file name of input, report, and results files)
• timeStamp date and time of execution
• loadHtg calculated heating load (W) (peak for all systems)
• loadClgSens peak sensible cooling load (W) = MAX( 24 hour loadClgSen)
• loadClgLat peak latent load
• 24 hour loadClgSen 24 hour load profile (W) (sum of all systems)
Note that the case file loads are building-wide totals derived by summing all systems. These results
meaningful for single-system buildings but must be used with care in multi-system situations.
The results file, runID.res, always contains a loads summary, documented below. In addition, other
information can be added via the –X command line switch. The format of these additional summaries is
self-documenting.
The entire results file is comma-separated for convenient import into Excel.
runID
! timeStamp
B, , <# systems>, <QLH>, <Ta>, <Ts>, <QLCSen>, <QLCLat>, <AVFClgDes>
S, <systemID>, <# zones>, <QLH>, <Ta>, <Ts>, <QLCSen>, <QLCLat>, <AVFClgDes>
Z, <zoneID>, <# rooms>, <QLH>, <Ta>, <Ts>, <QLCSen>, <QLCLat>, <AVFClgDes>
R, <roomID>, 0, <QLH>, <Ta>, <Ts>, <QLCSen>, <QLCLat>, <AVFClgDes>
R, <roomID>, 0, <QLH>, <Ta>, <Ts>, <QLCSen>, <QLCLat>, <AVFClgDes>
R [repeats for each room in zone]
Z, <zoneID>, <# zones>, <QLH>, <Ta>, <Ts>, <QLCSen>, <QLCLat>, <AVFClgDes>
Z [repeats for each zone in system]
S [repeats for each system in building
Where
AVFClgDes Design air volumetric flow (l/s) (generally = peak of hourly AVF)
QLCLat Latent cooling load (W)
QLCSen Sensible cooling load (W)
QLH Heating load (W)
Ta Indoor air temp (C)
Ts Supply air temperature (C). For rooms = at register. Average values for systems and zones
Format
An ResHB input file is a free-format text file, prepared in NotePad or suitable alternative text editor.
On each line, all text beginning with an exclamation point is treated as a comment and is dropped.
Commands consist of a list of comma-separated items and are terminated by a semi-colon.
Items are either numeric (N) or alphanumeric (A). Items may be enclosed in quotes (allowing inclusion
of commas and/or semi-colons). The maximum length of A items varies but is generally at least 40 chars.
Some items may be either N or A. ResHB examines the input text and treats it as N if it can be
successfully decoded as a number or A if not. This feature is used, for example, to allow 24 hourly values
to be input as single value to be used for all 24 hours or a schedule reference.
In some cases, items in a command command are optional. That is, they can be omitted (blank(s)
between commas) or, if all remaining items are optional, the command can be terminated with a semi-
colon after the last non-optional item. Optional items take default values as indicated in the detailed
documentation.
Command and ID matching are done case-insensitive. “SCHEDULE” and “Schedule” are both
recognized as a SCHEDULE command. A schedule with ID = “MySched” can be referenced as
“mysched”.
Error Checking
ResHB performs some amount of input validation and checking. Warning or error messages are issued
when anomalies are detected and the loads calculation is suppressed.
Caveat user: ResHB is a research program. Input data is not checked to the same level of detail in all
areas of the application. Results should be checked carefully. In addition, certain kinds of errors are not
detected – for example, file access problems are not handled.
Command Order
With the exception of ROOMs, SURFACEs, and GAINs (see below), ResHB commands can be in any
order. The entire input file is read into memory at startup. Input processing is done from the in-memory
image, allowing any command to be located when needed.
Exception: SURFACEs and GAINs belong to ROOMs and this relationship is determined from input file
order. All SURFACEs and GAINs found in the file belong to the ROOM that they follow.
Note also that INCLUDE commands, PARAMETER commands, and parameter resolution are handled as
the file is initially read, making order significant for these.
Include Files
Input files may include other input files via the INCLUDE command. INCLUDE is a powerful feature
that allows single-copy maintenance of standard constructions, materials, and other common items.
INCLUDE, “fileName”;
Note that the -I command-line option can be used to simplify access to include files.
Built-in Commands
ResHB has a number of so-called “built-in” items, such as MATERIALS and FENCLASSs. These are
discussed in the detailed command description. This capability is implemented as an implicit (invisible)
INCLUDE file that is read prior to all other input. Thus, all input rules apply with respect to built-in
items – for example, a built-in MATERIAL cannot be redefined.
Parameters
The PARAMETER command allows simple string substitution at the item level, as shown in this
example:
PARAMETER, area, 20;
(other input)
SURFACE, Ceiling, Wood, @area, ...
The PARAMETER command associates a string with a name (in this case “20” with “area”). When an
item is encountered that begins with “@”, the string associated with the name is substituted.
Rules:
• Parameters are defined and used in the order they are encountered in the file. Parameter values
can be changed by subsequent PARAMETER commands referring to the same name.
• Parameters are resolved at the whole item level. That is, parameter references are identified by
the initial @, so they cannot be embedded within items.
Example
The following example calculates loads for a residence in Atlanta, GA. For simplicity, the house is
treated as two rooms corresponding to living and sleeping areas. The main file references two include
files, containing location-specific data (LocGAAT.rhi) and standard definitions (RHBLib.rhi). The
contents of those files is shown after the listing of main file Example.rhb.
File Example.rhb
! Example.rhb
BUILDING, 0, 3;
INFIL, Std, AIM2_LCLS, G, 2.4, 1;
SYSTEM, Sys01, *ALL*;
ZONE, Zone01, 20., 24.,, *ALL*;
! end of file
File LocGAAT.rhi
! lat long tzn elv
Location, "Atlanta, GA", 33.65, -84.42, -5.0, 315;
! dst db dr wb tgnd
DesCondCooling, JUL21, 7, 21, D, 32.6, 9.6, 23.4, 0;
DesCondHeating, -7.9, 6.7;
File RHBLib.rhi
!=============================================================================
! RHBLib.rhi: common definitions for ResHB
!=============================================================================
!=============================================================================
! SCHEDULES
!-----------------------------------------------------------------------------
! California Energy Commission internal gain schedule (fraction of daily total)
SCHEDULE, CECGain, 0.024, 0.022, 0.021, 0.021, 0.021, 0.026, 0.038, 0.059,
0.056, 0.060, 0.059, 0.046, 0.045, 0.030, 0.028, 0.031,
0.057, 0.064, 0.064, 0.052, 0.050, 0.055, 0.044, 0.027;
! Building America lighting scheds: consolidated, living, sleeping (frac of day total)
SCHEDULE, LightAll, 0.0085, 0.0085, 0.0084, 0.0084, 0.0237, 0.0500, 0.0558, 0.0499,
0.0219, 0.0151, 0.0151, 0.0151, 0.0151, 0.0151, 0.0151, 0.0261,
0.0557, 0.0783, 0.1052, 0.1252, 0.1270, 0.0878, 0.0487, 0.0203;
SCHEDULE, LightLiv, 0.0081, 0.0081, 0.0081, 0.0080, 0.0161, 0.0322, 0.0374, 0.0361,
0.0186, 0.0186, 0.0186, 0.0186, 0.0186, 0.0186, 0.0186, 0.0348,
0.0736, 0.1027, 0.1268, 0.1341, 0.1208, 0.0695, 0.0373, 0.0161;
SCHEDULE, LightSlp, 0.0091, 0.0091, 0.0091, 0.0091, 0.0365, 0.0801, 0.0871, 0.0733,
0.0275, 0.0092, 0.0092, 0.0092, 0.0092, 0.0092, 0.0092, 0.0115,
0.0252, 0.0366, 0.0686, 0.1098, 0.1374, 0.1191, 0.0682, 0.0275;
! Living / sleeping
GAINZONE, GZLiv,
Kitchen, AppPlug, Laundry, AppPlug, Plug, AppPlug,
Lighting, .63 LightLiv,
PrsnLiv, OccLiv;
GAINZONE, GZSlp,
Plug, AppPlug,
Lighting, .37 LightSlp, PrsnSlp, OccSlp;
! consolidated
GAINZONE, GZAvg,
Kitchen, AppPlug, Laundry, AppPlug, Plug, AppPlug,
Lighting, LightAll, PrsnAvg, OccAll;
!=============================================================================
! FENESTRATIONS
!-----------------------------------------------------------------------------
! wood operable frames, default SHGC and U-factor
FENESTRATION, FEN-WOC1, OpWood, Clear, 1;
FENESTRATION, FEN-WOC2, OpWood, Clear, 2;
FENESTRATION, FEN-WOE2, OpWood, LoEHS, 2;
FENESTRATION, FEN-WOX2, OpWood, LoELS, 2;
FENESTRATION, FEN-WOC3, OpWood, Clear, 3;
!==============================================================================
! FENCLASS
!------------------------------------------------------------------------------
!=============================================================================
! CONSTRUCTIONS
!-----------------------------------------------------------------------------
! wood frame sub-constructions (10% and 25% framing factor)
! thickness adjusted when ref'd
Construction, WF10Empty, PP, $F 90 .10 FrameWood Empty;
Construction, WF10FbrGl, PP, $F 90 .10 FrameWood FbrGlsBatt;
Construction, WF25Empty, PP, $F 90 .25 FrameWood Empty;
Construction, WF25FbrGl, PP, $F 90 .25 FrameWood FbrGlsBatt;
! Internal mass
Construction, IMASS-W, PP, FrameWood 12;
! Floors
Construction, FLOR-WL, PP, WF10Empty 190, Plywood 19, OakFloor 8;
Construction, FLOR-WM, PP, WF10FbrGl 190, Plywood 19, OakFloor 8;
Construction, FLOR-WH, PP, FoamBoard 50, WF10FbrGl 190, Plywood 19, OakFloor 8;
! end RHBLib.rhi
SIMCONTROL
DOCUMENTATION
Provides run documentation passed to output files to aid parametric studies. Optional.
DOCUMENTATION, runDoc, caseTag;
LOCATION
SCHEDULE
Specifies a named sequence of 24 values. Schedules are widely used in ResHB to input data by hour.
SCHEDULE, ID, 1 or 24 value(s);
Note: The time base of schedules is local time (standard or daylight, per DESCONDCOOLING). This is
generally appropriate for modifying occupancy-related values such as internal gain. However, care must
be exercised when applying a schedule to items that involve solar time, such as surface shading or
outdoor temperature profiles.
Defines a type of internal gain. GAINITEMs have no effect until they are included in a room via
reference in GAIN commands. See example with GAIN (below)
GAINITEM, ID, gain, gainA, gainCFA, gainBR, gainBR, schedMult, fConv, fSWRad, fLWRad,
fLat, fExh;
The schedMult value consists of any number of space-separated numeric constants or schedule IDs that
are multiplied together to yield a 24-hour profile for the gain. A numeric constant is treated as a 24 hour
schedule with the constant value for all 24 hours.
Examples
schedMult Result
.3 all hours = .3
.3 .4 all hours = .12
.3 AllDay each hour = 0.3 · hour value of schedule AllDay
AllDay OccProf each hour = (AllDay hour value) · (OccProf hour
value)
GAINZONE
Provides default values for several SURFACE arguments, optional command. Note that input is not
provided for all SURFACE arguments, since only some of them commonly vary. Any value can be
specifically set on a surface-by-surface basis (see SURFACE).
ROOMDEFAULTS
Specifies a geometric shading configuration. Geometric shading can be applied to any vertical surface,
whether opaque or fenestration.
Left and right in OHFIN are defined as looking from the outside of the surface.
Note that srfWidth and srfHeight are used for geometric calculations only. The area implied by these
dimensions need not equal the area(s) of referencing surface(s).
OHFIN, ID, srfWidth, srfHeight, srfReveal, ohDepth, ohSrfDist, ohLeftExt, ohRightExt,
ohFlapHeight, lfDepth, lfSrfDist, lfTopExt, lfBotShort, rfDepth, rfSrfDist,
rfTopExt, rfBotShort;
CONSTRUCTION
Layer descriptions are multipart specifications of layer thermal properties. Several options provide
flexibility in how layers are described. A layer description consists of 1, 2, or 5 values, as follows (note
that these values are delimited by whitespace, not commas):
rmc [thk [fFrm matFrm matCav]]
Default U-factor: If the cooling U-factor (UorCons) is not specified, a default value is derived from
glzTy and frameTy using values from Table 4, Chapter 30 (Fenestration), 2001 HOF, using values shown
below.
Default SHGC: If SHGC is not specified, a default value is derived from glzTy and frameTy. The SHGC
is calculated using frame fraction, frame SHGC, and glazing only SHGC, using the values given on p.
30.39 (Fenestration), HOF 2001, as follows:
Frame Frame
FrameTy codes
fraction SHGC
Clear 1 1a U 5.91 7.24 6.12 5.14 5.05 4.61 6.42 6.07 5.55 5.55 5.35
SHGC 0.86 0.75 0.75 0.64 0.64 0.64 0.78 0.78 0.75 0.75 0.75
2 5a U 2.73 4.62 3.42 3.00 2.87 2.53 3.61 3.22 2.86 2.84 2.72
SHGC 0.76 0.67 0.67 0.57 0.57 0.57 0.69 0.69 0.67 0.67 0.67
3 29a U 1.76 3.80 2.60 2.25 2.19 1.91 2.76 2.39 2.05 2.01 1.93
SHGC 0.68 0.60 0.60 0.51 0.51 0.51 0.62 0.62 0.60 0.60 0.60
LoELS 2 25a U 1.70 3.83 2.68 2.33 2.21 1.89 2.75 2.36 2.03 2.01 1.90
SHGC 0.41 0.37 0.37 0.31 0.31 0.31 0.38 0.38 0.36 0.36 0.36
3 40c U 1.02 3.22 2.07 1.76 1.71 1.45 2.13 1.76 1.44 1.40 1.33
SHGC 0.27 0.25 0.25 0.21 0.21 0.21 0.25 0.25 0.24 0.24 0.24
LoEHS 2 17c U 1.99 4.05 2.89 2.52 2.39 2.07 2.99 2.60 2.26 2.24 2.13
SHGC 0.70 0.62 0.62 0.52 0.52 0.52 0.64 0.64 0.61 0.61 0.61
3 32c U 1.42 3.54 2.36 2.02 1.97 1.70 2.47 2.10 1.77 1.73 1.66
SHGC 0.62 0.55 0.55 0.46 0.46 0.46 0.56 0.56 0.54 0.54 0.54
HtAbs 1 1c U 5.91 7.24 6.12 5.14 5.05 4.61 6.42 6.07 5.55 5.55 5.35
SHGC 0.73 0.64 0.64 0.54 0.54 0.54 0.66 0.66 0.64 0.64 0.64
2 5c U 2.73 4.62 3.42 3.00 2.87 2.53 3.61 3.22 2.86 2.84 2.72
SHGC 0.62 0.55 0.55 0.46 0.46 0.46 0.56 0.56 0.54 0.54 0.54
3 29c U 1.76 3.80 2.60 2.25 2.19 1.91 2.76 2.39 2.05 2.01 1.93
SHGC 0.34 0.31 0.31 0.26 0.26 0.26 0.31 0.31 0.30 0.30 0.30
Rflct 1 1l U 5.91 7.24 6.12 5.14 5.05 4.61 6.42 6.07 5.55 5.55 5.35
SHGC 0.31 0.28 0.28 0.24 0.24 0.24 0.29 0.29 0.27 0.27 0.27
2 5p U 2.73 4.62 3.42 3.00 2.87 2.53 3.61 3.22 2.86 2.84 2.72
SHGC 0.29 0.27 0.27 0.22 0.22 0.22 0.27 0.27 0.26 0.26 0.26
3 29c U 1.76 3.80 2.60 2.25 2.19 1.91 2.76 2.39 2.05 2.01 1.93
SHGC 0.34 0.31 0.31 0.26 0.26 0.26 0.31 0.31 0.30 0.30 0.30
FENCLASS
Defines a fenestration class (a collection of data that defines the angular heat gain characteristics of a
fenestration system).
ResHB has a number of built-in FENCLASSs that support default characteristics (see table below).
Additional FENCLASSs can be input to handle special cases.
FENCLASS layers are defined outside to inside. That is, layer 1 is the outside (or only) layer.
The number of values required for a FENCLASS depends on the number of glazing layers and the
number of angles for which data is available. See example under General Information (above).
ResHB includes a number of built-in FENCLASSs that are used to provide defaults for typical
fenestration configurations; they can also be referenced as needed for custom situations. The associated
data is taken from Table 13, Chapter 30 (Fenestration), ASHRAE HOF (2001), refer to that table for
complete information.
ID NGlz SHGC (glazing Description
only)
FCA-1a 1 .86 clear
FCA-1c 1 .73 bronze heat absorbing
FCA-1l 1 .31 reflective
FCA-5a 2 .76 clear / clear
FCA-5c 2 .62 bronze heat absorbing / clear
FCA-5p 2 .29 reflective / clear
FCA-17c 2 .70 clear / low-E (“high solar”)
FCA-25a 2 .41 low-e / clear (“low solar”)
FCA-29a 3 .68 clear / clear / clear
FCA-29c 3 .34 green heat absorbing / clear / clear
FCA-32c 3 .62 clear / clear / low-e (“high solar”)
FCA-40c 3 .27 low-e / low-e / clear (“low solar”)
FENCLASS Example:
FENCLASS, FCA-25a, 2, 7,', & ! double LE CLR (low-e low solar)
0, 40, 50, 60, 70, 80, 90,
.41, .40, .38, .34, .27, .14, 0, .36, ! SHGC
.37, .35, .33, .29, .22, .11, 0, .31, ! TRANS
.24, .26, .26, .27, .28, .23, 0, .26, ! ABS1 (outer)
.04, .04, .04, .04, .03, .03, 0, .04; ! ABS2
Specifies infiltration parameters. ResHB includes a full implementation of the AIM-2 infiltration model
as well as a simpler air change model.
INFIL, ID, meth, L, Hstack, Heave, FCrawl, ShelClass, PRef, N, R, X, Y, ShelClassFlue;
4 Hstack N (m) (req) Average stack height = height difference between floor (or grade if
lowermost floor is below grade) and ceiling averaged over
conditioned footprint of building
5 Heave N (m) Hstack + Eave height above grade, determines wind modifier.
1
6 FCrawl N 0 Fraction floor area above crawl space; 1 – FCrawl is assumed
above slab or basement.
7 ShelClass N (1-5) 4 Building shelter class (see below)
8 Pref N (Pa) 4 Reference pressure for leakage areas. Unused if meth = AIM2_C
9 N N .67 Pressure exponent
10 R N .5 Leakage distribution – (floor + ceiling) / total
11 X N 0 Leakage distribution – (floor – ceiling) / total
12 Y N 0 Leakage distribution – flue / total
13 ShelClassFlue N 2 Flue shelter class (see below)
Specifies a duct system. Note: as of version 1.00, the duct model is not integrated with the loads model.
These inputs are experimental.
DUCTSYS, ID, config, supF, supLoc, supLF, supR, supMat, retF, retLoc, retLF, retR,
retMat;
BUILDING
SYSTEM
Specifies a system (collection of zones). Not that ResHB is does not perform any system modeling. A
system is simply an accounting structure to which loads are accumulated for reporting purposes.
SYSTEM, ID, zoneList;
Specifies a building zone (a group of rooms). The rooms in a zone share HVAC controls.
ZONE, ID, tspHeat, tspCool, rhClg, roomList;
ROOM
Specifies direct room ventilation. Vent air flow adds to infiltration and any other outdoor air assumed in
the model.
0 or 1 VENT statements are allowed in each room.
VENT, vFlow [, vACH[, vMult]];
Vent air volume flow rate is calculated for each hour as follows:
Total vent flow (l/s) = (vFlow + vACH • RoomVol • 1000 / 3600) • mult
Note that air volumes are stated in terms of indoor conditions. ResHB uses constant (i.e. approximate) air
properties evaluated at indoor conditions.
GAIN
Specifies an internal gain to a room. References, scales, and schedules a GAINITEM, GAINITEMSET,
or GAINZONE (see above). An example showing the use of GAIN is below.
GAIN, gainItemID, schedMult;
Any number of GAINs can be included in a ROOM. All gains are combined. Negative gains are
permitted.
If the referenced gainItemID is a GAINITEM, then the total possible gain = (gain + gainA · RoomCFA +
gainCFA · bldgCFA + gainBR · bldgNBR + gainOcc · bldgNOcc). The fractions fConv, fSWRad,
fLWRad, fLat, and fExh are then applied to the total. The fExh fraction is discarded and the other
fractions are applied as indicated in the heat balance model.
If the referenced gainItemID is a GAINITEMSET, each GAINITEM in the GAINITEMSET is treated as
above, with the application of the additional GAINITEMSET schedMult.
If the referenced gainItemID is a GAINZONE, then the treatment is identical to GAINITEMSET except
Example –
SCHEDULE, OccSched, 1, 1, 1, 1, 1, 1, 1, 1,.2,.2,.2,.4,
.2,.2,.2,.5,.5, 1, 1, 1, 1, 1, 1, 1;
SCHEDULE, LightSched, 0, 0, 0, 0, 0, 0, 0,.3, 0, 0, 0, 0,
0, 0, 0, 0, 0,.2,.5,.8, 1,.7,.4,.2;
Note that only the first 5 arguments are required (ID – tilt). All remaining arguments take default values
as specified in their descriptions.
# Item Format Default Description
1 ID A (ID) (req) Identifies surface for error messages. Must be unique within room.
2 cons N (W/m2-K) (req) Surface U-factor (W/m2-K). If cons is a valid numeric value, the
surface is treated as having massless opaque construction. U-factor
value should include surface conductances (34 W/m2-K outside and
8.3 W/m2-K inside)
A Ref to defined Construction (surface is opaque)
(Construction
ID)
‘$REV’ A Ref to defined Construction (surface is opaque). Construction is
(Construction reversed, allowing the same construction to be referenced in adjacent
ID) rooms (such as the ceiling of conditioned space and floor of attic
above). Without this feature, the ceiling construction would have to be
defined twice with opposite layer order.
A Ref to defined Fenestration (surface is fenestration)
(Fenestration
ID)
3 area N (m2) (req) Net (actual) heat transfer area (ResHB makes no adjustment for sub-
surfaces such as fenestration).
• Opaque: generally based on outside-to-outside dimensions
(measured on building exterior, less sub-surface area)
• Fenestration: NFRC projected area (approximately equal to
rough opening)
4 azm N (deg) (req) Surface azimuth (0 = north, 90 = east, 180 = south, 270 = west). Note
actual azimuth depends on BUILDINGNORTH.
5 tilt N (deg) (req) Surface tilt (0 = horiz facing up (roof), 90 = vertical, 180 = horiz facing
down (floor)).
6 bcCode A (code) AMB Boundary condition code (see below)
7 bcAdj N or sched ID (blank) Additional data re boundary conditions (see below).
8 shade ‘none’ none Shaded fraction = 0
‘shaded’ Shaded fraction = 1
N (0 – 1) Shaded fraction for surface (beam solar multiplied by this value)
A (sched ID) Schedule specifying shaded fraction by hour
A (Ohfin ID) Ref to defined Ohfin specifying geometric shading
9 grRflct N (0 – 1) Note 1 Ground reflectivity adjacent to surface.
10 Opq: N (0 – 1) Note 2 Outside short wave absorptance.
SWAbsO
Fen: EAC N (0 – 1) 1 Exterior attenuation coefficient = incident solar reduction factor re
exterior shade screens
11 Opq: N (0 – 1) Note 2 Inside short wave absorptance. Ignored for fenestration.
SWAbsI
Fen: IAC N (0 – 1) 1 Interior attenuation coefficient (re interior shading)
12 LWAbsO N (0 – 1) Note 2 Outside long wave absorptance.
Generates an unconditioned room of type ‘attic’, including all required surfaces and internal mass.
Surface defaults are used for all values for which there is no input. If finer control is required, the ROOM
and SURFACE commands should be used.
ATTIC, ID, config, l, w, ht, azm, ACH, TRBelow, ceilCons, rfCons, rfSWAbsO, rfLWAbsI;
Generates an unconditioned room of type ‘crawlbase’ (crawlspace or basement) including all required
surfaces. Surface defaults are used for all values for which there is no input. If finer control is required,
the ROOM and SURFACE commands should be used.
CRAWLBASE, ID, config, l, w, ht, azm, ACH, TRAbove, HsFlrCons, WallCons, CBFlrCons;
The ceiling surface (that is, the house floor) generated by CRAWLBASE has ID ‘HsFlr’. Thus, to couple
to an crawlspace / basement, the boundary conditions for the floor(s) of adjacent room(s) should be set to
TSRS=<cbID>:HsFlr.
runID is the base name for input and output files for the run (see below).
Options can be any combination of the following --
Effect
Option
-I<dirList> Include directories. Specifies a ;-delimited list of directories that RHBGen searches for include
files (the current directory is searched first).
Example: -Iinc1;c:\inc2
Exit code
RHBGen returns an exit code indicating the outcome of processing --
Exit code Meaning
0 Success, no errors: all cases generated and run without error
1 Internal program error
2 – 19 Input or file generation error(s), review RHBGen.log.
20 All cases successfully generated, but ResHB detected errors and at least one case was not
successfully run. Review RHBGen.log and/or ResHB.log.
Files
RHBGen reads or writes the following files.
File name Use Description
<runID>.rhg Input, must exist Input file.
RHBGen.log Output, append (created if does not exist) Log file. Information is appended to this file
documenting each run, showing date/time of run,
general information about the run, and any error
messages.
RHBGen.err Output, append (created if does not exist) Error file. Only written if errors are detected.
Can serve as an error flag – delete before a
group of runs, if exists at end, there has been at
least one error.
<outDir>\<runID>.scl Output, replace (overwritten if exists) Successful case list. Lists parametric codes for
all cases successfully generated and run (or just
successfully generated if no RUNCASE)
File “Example.rhg” –
! Example.rhg
RUNSET, 0, TestSet,,,
L=DCWJ|AZPH|FLKW|CAPR|TNME|CASD,
O=000|090|180|270;
!================================================================
! CaseSets
CASESET, TestSet,
R2HSLCOALO000QmEm00AmmmmCwmhamWwmmGx110llllPwxMwmFwmax,
R2HSLCOALO000QmEm00AmmmmCwmhtmWwmmGx110mmmmPwxMwmFwmax,
R2HSLCOALO000QmEm00AmmmmCwmhamWwmmGx11mhhhhPwxMwmFwmax;
! Prototypes
INCLUDE, "Prototypes.rhg";
! Example.rhg end
File “Prototypes.rhg”
! Prototypes.rhg -- RHBGen prototype geometry definitions
!---------------------
! Prototypes.rhg end
GENDIR
RUNCASE
Specifies application for analysis of each generated case. Optional, if omitted, no program is run.
If present, RHBGen executes the following command for each successfully generated file:
<runCmd> <runID> <runArgs> <caseArgs>
RUNCASE, runCmd, runArgs;
CASESET
Specifies cases to be generated and run. Any number of RUNSETs can be included.
RHBGen is driven by a list of parametric codes that are optionally modified via variants.
RUNSET, control, caseSetFile, caseTag, caseResPath, caseArgs [, caseVar(s)];
The base set of parametric codes used for a RUNSET are defined by the referenced CASESET or read
from a case file. Parametric codes are described in a later section of this document.
Case sets or case files contain lists of parametric codes to be run. These files are external to the main
input file to allow sets of cases to be generated using automated schemes. The format is one parametric
code per line, with the following rules –
• Blank lines are ignored
• Any characters after ‘!’ are ignored (for comments)
• A line consisting of “*STOP*” terminates processing. This is useful for testing.
Variants allow generation of multiple cases from a single base parametric code.
A variant is a parametric code prefix followed by a list of alternative values for that prefix field. For
example, O=000|090|180|270 means “generate/run each case at the 4 orientations 0, 90, 180, 270.” If
multiple caseVars are specified, all combinations are generated and run (assuming control = 0). Rules for
variants –
• Any position containing ‘*’ retains the original parametric code character from the case file.
• Any position containing ‘~’ retains the current parametric code character. This feature allows
nested combinations within the same field. For example, O=1|2, O=~10|~20 would generate/run
cases with orientation 110, 120, 210, and 220.
• The variant code may be shorter than the associated parametric code field (but not longer). Short
variants are implicitly padded with ‘*’s to the required length.
• Orientation variants have a special case: 0=999 causes an orientation value 0 – 359 to be
randomly selected.
Example 1 --
RUNSET, 0, myCases.gfl, , , , O=000|090|180|270, W=*l|*m|*h;
This command generates and runs 100 cases with random orientation (1-360) and random wall insulation.
PROTOTYPE
Along with PRROOM and PRSURF, provides basic geometry for prototypes used to generate ResHB
input files. The prototype used is selected via the Hx field in the parametric code, where ‘x’ is the single
letter designation of the prototype
Any number of prototypes can be defined.
PROTOTYPE, name, comment, hInfil, NBR, NOcc;
PRROOM
Specifies the geometry of a room that belongs to the current prototype. These values are passed to
ResHB, see ResHB documentation for exact definitions.
PRROOM, name, cfa, height, vol, gainzone;
Notes
• SW absorptance codes: 0: 0, l: .3, m: .6, h: .8, d: .85, b: .95, 1: 1
• All codes case sensitive, however upper and lower case variants are not used in same code position
because this would result in non-unique generated file names (file name are not case sensitive)
• Floor exterior SW abs set to same value as wall (required only for exposed floor)
• Possible additional variable = wall shading (handle with absorp?)
Ceiling constructions
Config Boundary Cons code Type Insul codes ‘i’ Constructions
code ‘f’ conditions ‘c’
x AMB 0 Massless l, m, h MASSLESS-L, -M, -H
w Wood frame l, m, h CEIL-WL, -WM, -WH
c Concrete l, m, h CONC-L, -M, -H
f AMB 0 Massless l, m, h MASSLESS-L, -M, -H
w Wood frame l, m, h CLRF-WL, -WM, -WH
c Concrete l, m, h CONC-L, -M, -H
h Attic 0 Massless l, m, h MASSLESS-L, -M, -H
w Wood frame l, m, h CEIL-WL, -WM, -WH
c Concrete l, m, h CONC-L, -M, -H
Wall constructions
Cons code ‘c’ Type Insul codes ‘i’ Constructions
0 Massless l, m, h MASSLESS-L, -M, -H
w Wood frame l, m, h WALL-WL, -WM, -WH
c Concrete l, m, h CONC-L, -M, -H
Fenestration
Type code ‘t’ Glazing Frame ASHRAE NFRC U-factor SHGC
ID W/m2-K
1 Single clear none 1a 5.91 .86
2 Double clear none 5a 2.73 .76
e Double low-e high none 17c 1.99 .70
solar
x Double low-e low none 25a 1.70 .41
solar
3 Triple clear none 29a 1.76 .68
!=============================================================================
! CONSTRUCTIONS
!-----------------------------------------------------------------------------
! wood frame sub-constructions (10% and 25% framing factor)
! thickness adjusted when ref'd
Construction, WF10Empty, PP, $F 90 .10 FrameWood Empty;
Construction, WF10FbrGl, PP, $F 90 .10 FrameWood FbrGlsBatt;
Construction, WF25Empty, PP, $F 90 .25 FrameWood Empty;
Construction, WF25FbrGl, PP, $F 90 .25 FrameWood FbrGlsBatt;
! Internal mass
! Floor
Construction, FLOR-WL, PP, WF10Empty 190, Plywood 19, OakFloor 8;
Construction, FLOR-WM, PP, WF10FbrGl 190, Plywood 19, OakFloor 8;
Construction, FLOR-WH, PP, FoamBoard 50, WF10FbrGl 190, Plywood 19, OakFloor 8;
!=============================================================================
! MATERIALS
!-----------------------------------------------------------------------------
! k rho Cp R dflt L
! W/m-K kg/m3 kJ/kg-K m2-K/W mm
Material, Empty, 0, 0, 0, .18;
Material, R1000, 0, 0, 0, 1000;
Material, FbrGlsBatt, .039, 19, .96, 0;
Material, FoamBoard, .029, 30, 1.2, 0;
Material, FrameWood, .12, 500, 1.6, 0;
Material, Plywood, .14, 550, 1.5, 0;
Material, OakFloor, .17, 720, 1.65, 0, 8;
Material, GypBoard, .16, 800, 1.09, 0, 13;
Material, ConcHvy, 2.2, 2240, .9, 0;'
Material, Brick, 0.9, 2000, .84, 0, 90;
Material, CMUMed, 1.3, 1840, .84, 0;
Material, Soil, 1, 1900, 1, 0;
Material, AsphaltShingle, .12, 1100, 1.26, 0;
Material, Carpet, .06, 300, 1.40, 0, 13;