You are on page 1of 6

EI I[| G't

AND
UILDING-
ELSEVIER Energy and Buildings 23 (1996) 251-256

LT Method 3 . 0 - a strategic energy-design tool for Southern Europe


Nick Baker, Koen Steemers
The Martin Centrefor ArchitectJtral and Urban Studies, Departmentof Architecture Universityof Cambridge, 6 ChaucerRoad, Cambridge CB2 2EB, UK

Abstract

The LT Method is an energy-design tool which responds to parameters available early in the design development. It provides an output of
annual primary energy for lighting, heating cooling and ventilation. This paper introduces LT 3.0 i a version for Southern Europe which
includes a procedure to evaluate the affect of shading devices on cooling loads and on lighting.

Keywords: Energy-design tools; Strategic; Architecture; Southern Europe

1. Introduction In considering the factors which influence energy use in a


building, we can classify them into three broad categories:
Energy modelling has developed to a high standard in the (i) building; (ii) services, and (iii) occupants. Studies of
past decade. The growing power of computers together with Baker have shown that all three categories independently
rigorous procedures for validation, relying upon high-quality have about equal impact on energy use. Within building fac-
data from test cells, has greatly increased the accuracy with tors we can identify two sub-categories building-design par-
which thermal models can predict temperatures and energy ameters and engineering parameters. For example, the plan
use. However, this is often at the expense of complication - depth of a building is a fundamental building-design param-
in order to develop the potential accuracy, more detailed eter, interacting with many other parameters and having an
building data is required than is likely to be available until impact on the form and performance of the building in terms
the design is fully developed. This means that if the model is of daylight, heating and cooling, and natural ventilation. In
to be used as a design tool, the iterative loop of design pro- contrast, the U-value of the wall, whilst it will have significant
posal: testing/design modification/re-testing, etc., is a wide effect on the thermal performance, can take on values inde-
one. The conceptual momentum of a design team, and prac- pendently from other parameters.
tical considerations, make this iterative process unlikely to More to the point is that at the strategic stage of design the
occur. architect is really not very interested in U-values or the lumi-
A second problem with these powerful models is that few nous efficacy of the lamps; he or she is free to choose them
integrate other energy uses into the analysis; the emphasis at a later stage. However, the plan, section, facade design,
has been on thermal models. But energy use for electric light- orientation, position on the site in relation to other buildings,
ing, mechanical ventilation and refrigeration, often consumes organisation of functions within the building are all of great
a much greater proportion of the total energy. This is espe- interest during the phase when the 6B-pencil sketches are
cially true for non-domestic buildings in milder climates. being created.
The LT 2 Method has atlempted to address these two prob- In developing the LT Method we have concentrated upon
lems. This has been at the. expense of modelling precision, these design parameters. The engineering parameters do
but there is clearly no virtue in being able to take account of influence energy performance, but these have been given
a parameter, if information of that parameter is not yet avail- 'sensible' default values, or reduced to simple categories such
able. as 'high internal gains' or 'low internal gains'.
It so happens, that to assist in the dissemination of the LT
t The full data and documentation is published by the Energy Research Method, it has been developed predominantly in a manual
Group, Department of Architecture, University College Dublin, Dublin 14,
form. That is to say that the procedure is carried out using
originally as supporting material for the EC sponsored 'Zephyr' Architec-
tural Competition. scale rule, pocket calculator and pencil. This may seem anach-
2 LT: Lighting and Thermal energy. ronistic in a time when the computer prevails, but in fact the

0378-7788/96/$15.00 © 1996 Elsevier Science S.A. All fights reserved


SSD10378-7788(95)00950-3
252 N. Baker, K. Steemers / Energy and Buildings 23 (1996) 251-256

constraint has usefully forced us to simplify data requirc- MWh/m 2


0.30
ments and output options. It also means that the method has
to rely upon pre-computed data presented in the form of lotal ~
0.25 ......... cool 1.0
graphs and tables. ...... heat
\. .............
Although we are now developing electronic versions, the
0.20 .7
presentation ofpre-computed data in this way is still adopted• s.r"
This maintains the transparency of the method and confronts
the user with his or her design proposal set against other 0.15 •, , " "" .,,,0.35

performance data. It shifts the emphasis from the number to


the trend; from evaluation to comparison. 0.10 " "" "*' "'"*

0.05 ~
"'•. .... fan p o v ~

2. The L T Method .......... :_...-...:.-...:.....; ......


0.00
1"o 2o ~o 4"o 5o ~ 7o ~o go
SOUTH % glazing rabo
The LT Method relies upon the concept of the passive zone
as shown in Fig. 1. This is an area in the building within a
Fig. 2. LT Curve for south-facingofficein Southern Europe.
maximum distance from a perimeter wall, or an area under a
roof. These passive zones can receive the benefit of daylight, (iii) multiply the areas by the appropriate specific energy
natural ventilation and useful solar gains in winter, but also consumptions and sum for all zones and uses.
the disbenefit of heat loss through the envelope and solar The results can either be presented by the different uses:
gains in summer. Non-passive zones are away from the enve- heating, lighting, cooling, etc., or a total, or on a per m 2 basis•
lope, and thus require mechanical ventilation and artificial The latter is most useful when comparing the performance of
lighting, but do not suffer from unwanted solar gains or fabric the building with a target.
heat loss• In the LT Method the passive zone depth from a
side wall is defined as twice the floor to ceiling height, or 2.1. The LT Model
6 m as a default. (This is quite a useful way to assess a
building even if it is not used as part of the LT procedure.)
The LT Curves are derived by a mathematical model. Fig. 4
Annual Primary Energy consumption per m z is presented
shows the energy flows which are modelled. First, the model
for the zones as a function of glazing ratio and orientation. A
evaluates the heat conduction through the external envelope,
typical LT Curve is shown in Fig. 2.
and ventilation heat loss (or gain). Using monthly mean
In using the LT Method there are three main steps:
temperatures and a thermal reference point with a correction
(i) define passive and non-passive zones and their orien-
factor to allow for intermittent heating, a monthly gross heat-
tation; measure the areas from a sketch plan and enter on the
ing load is calculated. The model then evaluates the solar gain
LT Worksheet (Fig. 3);
and applies a utilisation factor to this.
(ii) read off the Annual Primary Energy consumption per
At the same time the monthly hours of available daylight
m 2 from the LT Curves for lighting, heating and cooling
are calculated from the average hourly sky illuminance on
applying any correction factors due to the presence of shading
the facade, the daylight factor, and an internal lighting datum
devices, and
value. This gives a monthly electrical consumption for arti-
~, ~J' t / ficial lighting, and a monthly heat gain.
~'!' - -
The lighting heat gains and useful solar gains, together
with a fixed casual gain from occupants and equipment, are
then subtracted from the gross heating load to establish the
net heating load. When the gains are greater than the gross
load, there is a cooling load. The cooling loads contain an
allowance for fan and pump energy.
As soon as a cooling load exists, the model eliminates a
percentage of the solar gain as defined by the shading device•
However, the ventilation remains at the low rate for calculat-
ing the cooling load, assuming there are no openable windows
and no 'free cooling'. This is to establish what the cooling
load would be, if it were mechanically cooled.
Note also that the model assumes 'sensible light-switch-
ing', i.e. that lights are only on when the daylighting value
Fig. !. Passive zones (white) and non-passive zones (hatched) on sketch drops below the datum value. In practice this would almost
plan. certainly require automatic, light-sensing switching. Thus to
N. Baker, K. Steeraers / Energy and Buildings 23 (1996) 251-256 253

~ t,a~ t'mmmU~

I ~"~'" "'8' I .o/_~ I . I .o~? I 0 6 o I .o~rs I

A to~ M~t y 0

vent

¢omnte~ts 0 SUMMARY
n ~ annual ol~naw net annual
e¢~rgy consumptior CO2 ernisr,~
MWtl kWWm' t k~tm 2 %
~,~ . / v o . , - /f"""a~s~;~..~ ~ , ~ .

TOTALS ~"7~' /07 / 2$'-~' 100

Fig. 3. LT worksheet.

a certain extent the model already assumes a considerable 3. Shading


degree of good design.
The monthly energy consumption is calculated for a 'cell' Shading is recognised as a vital measure to prevent over-
and then reduced to the value of energy consumption per m 2. heating, in particular in warmer climates. Yet it is clear that
This is then totalled for the year and plotted as a function of most shading devices reduce the daylight, as well as solar
glazing ratio. A 'cell' corresponds to a room surrounded by gain, within the building. If reduced below a threshold, this
other rooms. This implies a zero conductive heat loss through then causes an increased lighting energy demand. How can
all surfaces except the external (window) wail or in the case this be taken account of in an integrated energy tool?
of rooflighting, the ceiling. Appropriate efficiency factors are Again the LT Method has to make simplifications. Two
applied to reduce all energy to primary energy. kinds of shading system are identified: (i) those which only
The reason that primary energy is used is that it allows the reduce unwanted solar gain (type A), and (ii) those which
different 'fuel' inputs for lighting, heating and cooling to be reduce the daylight by the same proportion that they reduce
reduced to one common unit. Primary energy is the energy solar gain (type B). These two extremes may rarely exist in
value of the fuel at source. In the case of fuels such as gas or practice, but due to the way LT works, it is more meaningful
oil when it is to be used for heating, there is an energy over- than it may seem at first.
head required for extracting, refining and distribution, and Take the case of movable shading devices such as fabric
then the loss of heat due to combustion losses, at the point of roller blinds or venetian blinds (Fig. 5). These need only be
use. In the case of electricity, used for lighting and mechanical deployed when solar radiation is failing onto the window, i.e.
when there is a surplus of illuminance. Even if the fabric or
power for cooling and ventilation, an energy overhead occurs
louvre system has a diffuse transmission of only 10%, applied
at the power station due to the thermodynamic efficiency of
to direct solar radiation with an illuminance of around 80 000
the conversion of heat to raechanical power. This is a large
lux, this still leaves 8000 lux transmitted through the window
factor: 1 unit of delivered electricity is equivalent to 3.7 units
plane. Converting this to the equivalent horizontal illumi-
of primary energy. nance in order to apply a conventional Daylight Factor (DF),
For fossil fuel, Primary Energy relates well to CO2 and a 2% DF value would give a horizontal illuminance of about
other pollution production, and to cost. Problems arise when 500 lux. Furthermore, in principle the louvre system can be
a large proportion of electricity is generated by renewable adjusted to give sufficient diffuse illuminance transmission,
sources such as hydro or nuclear sources. On a national basis, without admitting direct sunlight. Thus although the system
there are wide variations of the fuel used for electricity gen- usefully reduces unwanted solar gain, it permits sufficient
eration. However, bearing m mind that electricity is exported daylight to exceed the threshold value for artificial lighting.
and imported on the European grid, a single European factor Now consider a light shelf (Fig. 6), a fixed system. High-
is appropriate. angle direct sun is completely prevented from reaching the
254 N. Baker, K. Steemers / Energy and Buildings 23 (1996) 251-256

lighting power

heaU~

coc~ir~ ~gh~ng

~I~, conduc~i(xlthroughglass

~ ~ )"~c(mducfiontlvough opaque

Fig. 4. Energy flows in the LT Model.

= ,i = / Q

.........................
300
0 L _ ~ shading 0
• transmission M m
500 lux 700 lux 10% 600 lux SO00lux

oll
lux

ref poi~ . . . . . . . . "


1" i :°21
300]
J

ref Point
,*,'

0 ~ shading 0 __ . shading
transmission ~ ~ = ~ I ~ M transmission
300 lux 600 lux 100% 800 lux 1000 lux 35%
Fig. 5. Moveable shading is only deployed when there is an excess of
illuminance. Type-Al shading.

workplane, most being reflected by the external part of the


shelf whereas diffuse radiation is reflected to the back of the
room. Even in diffuse sky conditions, although the illumi-
300
nance is reduced near to the window, due to intereflection, it
0
is not reduced at the back of the room. Thus the transmission
of the system is geometrically selective. 300 lux 450 lux
For the evaluation of artificial lighting energy, LT is only Fig. 6. A light shelf reduces excess illuminance close to the window but not
interested in critical conditions. In both these cases LT at the critical reference point. Type-A2 shading.
assumes that the shading device causes no reduction in day-
light at the critical reference point in the room (4.5 m from of the shading device, or because it is not deployed at times
the window wall), and at the critical time (i.e. as the internal of heating demand. So the heating curve is read off with no
daylight level approaches the switch-on threshold value). We modification.
will call these two shading types A1 and A2, respectively. In contrast, type B shading devices are really rather dumb,
When reading off from the LT Curves, if the shading device and cannot be recommended. These devices reduce daylight
has no effect on illuminance at the critical switch-on time, it by as much as they reduce unwanted solar gain (Fig. 7). This
will not effect lighting energy. Thus for a given glazing ratio, is the same effect as opaque wall - in other words it is just as
no modification to lighting energy has to be made. For cooling good to reduce the glazing area. In fact this is the better
energy, however, the LT Model has been used to produce a solution, because shaded glass still has a high U-value,
family of curves at different shading transmission factor 0.7 whereas opaque wall can be well insulated, and at lower cost.
and 0.35. For heating energy, it is assumed that low-angle If shading devices such as these are to be evaluated using LT,
useful winter sun is uninterrupted, either due to the geometry the reduced light transmission must be accounted for. This is
N. Baker, K. Steemers / Energy and Buildings 23 (1996) 251-256 255

m
0.30 ~ I l
lux o.25 ~ I /

:t I re,poi~.......... . o'~ 0.20 ~. . . . . . . ~ - ~ _~. ~.,


s" '" p'~

iiim
60,3 lux
. . . .

300 lux
'" d htt"~-
0.10 ~M~-~ ,

AC 0.05 ' '-q. i

600 t . , - ....
300 .... x ........... ~..,....e ..... 0.00

0 10 60 70 80 90
0 ~ < ~ ~ ~ transmission nST ..~.~/.~,~ ~
105 lux 210 lux 35%

Fig. 7. Fixed screens and grids reduce daylight and solar gain by the same
amount. Type-B shading. Fig. 8. The effect of fixed shading on lighting energy.

done by multiplying the actual glazing ratio, by the shading development, are overheating prediction curves. These, how-
transmission coefficient, when reading off the lighting energy ever, are based upon the cumulative frequency of tempera-
curve, as shown in Fig. 8. tures above a datum value, and thus are driven by the
It is interesting to note that several advanced daylight sys- departure of conditions from the norm, rather than the average
tems, combining reflective louvres or prismatic glass with conditions. Until this feature is available, we believe that
shelves, claim to improve the distribution of the daylight. potential cooling loads are of interest, since from a compar-
This could be described as having a negative shading factor ative viewpoint they can help to optimise design.
for daylight, actually increasing the daylight at the critical
position, (from the LT viewpoint this would be equivalent to
increasing the depth of the lrassive zone) and a positive shad- 5. LT and education
ing factor for solar gain. This discussion suggests the urgent
need for shading devices to be classified in this way - that is The messages given by the use of LT are quite simple.
a time-integrated ratio of solar shading to critical daylight After having used the procedure a number of times, it would
transmission. probably be possible to sum up the conclusions in a few lines
of good advice. Why then go to the trouble of producing such
a complex procedure?
4. Cooling loads and passive buildings We believe that, quite rightly, designers are sceptical about
prescriptive advice particularly when it requires many qual-
A most important question that designers ask is 'can LT ifications to make it sufficiently robust. Humans have basi-
tell me if my building will be OK without air conditioning?' cally quantitative minds, and advice which is backed by
The short answer is no. LT can produce an annual cooling quantities is much more convincing. Furthermore it is more
load which responds to solar gains, internal gains, and ambi- useful because it enables us to rank measures in order of
ent temperature, but it is quite difficult to translate this into priority. For example, it may be important to know by how
an overheating probability. much the application of shading devices reduces cooling load
This is because overheating in subtropical regions is a compared with, say, reducing the internal gains. Even simple
result of non-uniformity in conditions, rather than the result global messages such as 'avoid deep plans' seem to be more
of steady or average conditions. This is in contrast to the convincing when backed up by numbers.
situation for heating. In most of Europe, the average winter
temperature is at least 7 °C below comfort temperature and 5.1. Validation
in many cases 15 °C. This contrasts with the most severe
summer condition, e.g. the average temperature in Athens in So, although strategic energy tools like LT are limited in
August of 29 °C which is only 2 °C above an accepted upper the number of parameters they can address, they have a role
comfort limit. Most of Southern Europe's average summer in creating their own advice. This raises the question: is it
temperature lies well within the comfort zone. good advice?
Thus it is difficult to translate annual cooling load direct In assessing energy models, validation is regarded as a pre-
to overheating frequency. A new LT feature, currently under requisite for their use. This usually consists of comparing
256 N. Baker. K. Steemers/ Energy and Buildings 23 (1996) 251-256

measured data from simplified buildings such as test cells, Bibliography


with results from blind simulations using the model. This is
N. Baker, Low energy strategies for non-domestic build-
by no means as easy as it sounds, but model validation meth-
ings, in S. Roaf and M. Hancock (eds.), Energy Efficient
ods have become sophisticated enough to cope with most of
Building, Biackwell, Oxford, 1992.
the difficulties.
Other versions of the LT Method are as follows:
With simplified energy-design tools, the same approach is
(i) J. Goulding, J.O. Lewis and T. Steemers, LT 1.2,
hardly appropriate. So much of the data has to be given
(Office buildings in Europe), Energy in Architecture, Bats-
assumed values, that comparing measured and predicted ford for the CEC, 1993.
energy-consumption figures would really become a test for (ii) LT 2.0, (Office buildings in UK), Cambridge Archi-
the selection of default values. Far more useful is the criterion tectural Research Ltd. for Building Research Energy Conser-
'are the general messages given by this tool in accord with vation Unit, BRECSU.
the expert view?' This softer criterion does not appear to lend (iii) LT 4.0, (Multi-residential Buildings in Eastern
itself to rigorous application. Rather, the continued use of Europe), Documentation for architectural competition Liv-
design tools such as LT in education and professional training ing in the City, Energy Research Group, Department of
contributes to a slowly developing confidence. Architecture, University College Dublin, Dublin 14.

You might also like