Energy and Buildings 39 (2007) 792–801

www.elsevier.com/locate/enbuild

Environmental performance of a naturally ventilated city centre library

Birgit Krausse *, Malcolm Cook, Kevin Lomas
Institute of Energy and Sustainable Development, De Montfort University, Queens Building, The Gateway, Leicester LE1 9BH, UK

Abstract
To tackle climate change it is essential to reduce carbon dioxide emissions. To this end, it is important to reduce the energy demands of non-
domestic buildings. Naturally ventilated buildings can have low energy demands but the strategy is difficult to implement in deep plan, urban
locations. The Frederick Lanchester Library at Coventry University, UK, incorporates natural ventilation, daylighting and passive cooling
strategies. By using lightwells and perimeter stacks to supply and exhaust air, it can be ventilated by natural means despite its deep plan form and
sealed façade. This paper describes the building and presents the energy consumption and the internal temperatures and CO2 levels recorded in
2004/2005. The building’s performance is compared to the original design criteria and good practice guidelines. Recommendations for the design
of such buildings are made and the likely performance in other UK cities is assessed. It is concluded that the building uses under half the energy of a
standard air-conditioned building and yet, in summer, can keep the interior comfortable and up to 5 8C below ambient. The design would perform
equally well in the typical weather conditions experienced at 13 other UK cities, but not in London. It is concluded that deep-plan, naturally
ventilated buildings with sealed facades, if well designed, could maintain thermal comfort in all but a very few UK locations, whilst consuming
much less energy than even good practice standards.
# 2007 Elsevier B.V. All rights reserved.

Keywords: Building design; Natural ventilation; Energy efficiency; Monitoring; Temperatures; CO2

1. Introduction periods of occupancy are all seen as barriers to natural

ventilation. It has been shown that these perceived barriers can
Global warming is perhaps the most significant challenge be overcome by designing buildings which use the centre-in,
facing mankind. The emission of CO2 enhances the greenhouse edge-out, buoyancy-driven stack ventilation approach [3,4].
effect and is therefore seen as a significant contributor to global One example of such a building is the Lanchester Library at
warming. Reducing the emission of CO2 from buildings, by Coventry University (Fig. 1).
reducing their energy consumption, is an important plank of the The design of the library has been described elsewhere by
UK’s carbon reduction strategy [1]. Increasing energy costs are members of the design team [3–8] and by others [9–11]. This
causing building owners to take a greater interest in the design paper therefore only briefly describes the building but focuses
and management of their buildings. on its in-use energy and environmental performance. Adven-
Large non-domestic buildings which are naturally ventilated titious use is made of data recorded by the Building Energy
are likely to consume less energy than those which are air Management System (BEMS) to provide an insight into the
conditioned, partly because they tend to make more effective internal temperatures, CO2 concentrations and energy con-
use of natural light and partly because the electrical energy sumption. The measured temperatures are compared with
consumed by fans, chillers and pumps is avoided [2]. However, current design criteria for naturally ventilated buildings and
the desire to maximise the use of urban sites, through the use of with performance predictions made at the design stage.
deep plan built forms, the imperative of sealed facades to Similarly, the energy use is compared with UK benchmarks
reduce the ingress of urban noise and to ensure security, and the for office buildings.
increases of internal heat gains due to computers and longer If such innovative building forms are to gain wider
acceptance amongst building designers, information is needed
about how such buildings will perform in different UK
* Corresponding author. Tel.: +44 116 257 7957; fax: +44 116 257 7981.
locations and under different weather conditions. This paper
E-mail address: bkrausse@dmu.ac.uk (B. Krausse). compares the weather recorded at Coventry, which is in the
URL: http://www.iesd.dmu.ac.uk English midlands, in 2004/2005 with the typical weather
0378-7788/$ – see front matter # 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.enbuild.2007.02.010
 B. Krausse et al. / Energy and Buildings 39 (2007) 792–801 793

2. Description of the building

The Frederick Lanchester Library at Coventry University

has a gross floor area of 9103 m2 and consists of four floors and
a basement, within a 50 m2 footprint. The brief called for a
highly energy efficient building, but the site, and the desire for a
simple, legible library layout, demanded a deep plan form.
Additional constraints were the close proximity of the site to
main roads, resulting in high noise levels and pollutant
concentrations. Despite these constraints a design was
developed which relied purely on natural ventilation for the
main four library floors (Figs. 2 and 3). A basement houses a
book archive and computer suite with 24 h access which, due to
the high and prolonged heat gains, requires air conditioning.
Fig. 1. View of Lanchester Library from the West. Fresh air is introduced into the building via a plenum
between the first floor and the basement which serves four
lightwells, one in each quadrant of the building. Heat gains
from building occupants and computers warm the internal air
and create the buoyancy forces that cause the air to rise and
accumulate in a layer below the 3.9 m high ceilings. The ‘stack
effect’ generated by the 20, 1.8 m2, perimeter stacks and the
tapering central lightwell, draws the warm stale air out of the
building.
In winter the incoming air is warmed by pre-heating coils,
which lie horizontally across the base of the 6 m2 supply
lightwells, and trench heating at the point where air from the
lightwells enters onto each floor. Cooling in the warm summer
months is provided by passive methods. Night-time venting is
used to cool the exposed thermal mass of the building so that it
can absorb heat during warm periods of the following day.
Ventilation of the top floor is ensured by four separate
ventilation stacks, which were added to solve the problem of
Fig. 2. Typical floor layout. backflow of exhaust air from the central lightwell which was
identified by airflow simulations during the design phase [5].
conditions in 14 other cities in England, Ireland, Scotland and The positioning of the lightwells is intended to provide good
Wales and the conditions these cities experience in only 1 year fresh air distribution and daylight provision across the deep
in 10. Inferences about the performance of buildings like the plan floors. Solar gains are minimized by moveable translucent
Coventry library when located in other UK cities can then be horizontal blinds at the head of the supply lightwells, careful
made. window placement and the use of overhangs and metal shading

Fig. 3. Sections through the building showing (a) the central exhaust lightwell and stacks and (b) the supply lightwells.
794 B. Krausse et al. / Energy and Buildings 39 (2007) 792–801

fins (see Fig. 1). This helps to reduce the risk of overheating and for night-time cooling is based on the BEMS’s self-learning
improves the effectiveness of the natural ventilation system. algorithm to ‘predict’ the likely (passive) cooling requirement
Since the library is used for a variety of purposes, the basic for the next day. Over-cooling is prevented by monitoring slab
square open plan floor layout has been adapted to provide a temperature.
range of suitable spaces. All floors remain predominantly open The building has been a marked success, with students and
plan, with a significant area taken up by book stacks (apart from library staff reportedly enjoying learning and working in the
the ground floor, which serves as an unofficial meeting place space. In addition to the 2500 entries per day, which it was
and provides access to the issue desks and main staff offices). initially designed for, the building attracts a large number of
Internal partitions create additional spaces. On each floor a ‘visitors’ who use it as a stop-over between lectures, increasing
number of zones with different occupancy and usage the daily throughput to 5000.
characteristics can therefore be identified, such as open plan
area with book shelves, silent study rooms, group study rooms, 3. Predicted performance
study desks with and without PCs (open plan), print and
photocopy rooms and offices (see for example Fig. 4). The main challenge in the design of naturally ventilated
The building is controlled by a Building Energy Manage- buildings in the UK is to provide comfortable indoor
ment System (BEMS) which operates dampers and windows temperatures during periods with high external temperatures.
depending on indoor and outdoor temperatures, wind speed and The dynamic thermal simulations carried out during the design
direction, and CO2 concentrations inside the library. Ventilation phase, using ESP-r [12] and the Kew67 weather data [13],

Fig. 4. Floor plan of the second floor showing the location of BEMS temperature and air quality sensors (each circle represents a sensor pair: one temperature and one
CO2 sensor). Data from sensor pairs A–D are currently logged while data from the other sensors (S) are currently not logged but may be included in further monitoring
phases.
 B. Krausse et al. / Energy and Buildings 39 (2007) 792–801 795

indicated that the passive cooling and ventilation methods This paper makes opportunistic use of these data. However,
should be sufficient to provide comfortable conditions inside since the sensors from which data have been logged were not
the library even during the hottest periods. In their description chosen with performance analysis in mind, the data available do
of the operating concept Cook et al. state that ‘‘dry-resultant not allow a thorough statistical analysis. This paper will thus
temperatures would always be below 28 8C and that 27 8C use time series plots and summary statistics to evaluate the
would be exceeded for only 11 h of the year’’ [5]. Refined building’s thermal and ventilation performance. It is worth
BEMS controls (not simulated) were expected to ‘‘be capable noting, however, that additional BEMS sensors have recently
of reducing internal temperatures even further’’. been selected for data logging to aid future performance
These performance indicators suggest that the building analyses and to provide further explanations of the initial results
should satisfy the recently published overheating criterion presented in this paper.
published by the Chartered Institution of Building Services
Engineers (CIBSE), that: 4.2. Temperature sensors

 dry resultant temperature should not exceed 28 8C for more The temperature sensors are designed to monitor air
than 1% of the occupied hours (CIBSE Guide A [14] and temperature. However, due to the close proximity of the
TM36 [15]); sensors to the wall surface, the logged temperature values may
not be equivalent to the true air temperature in the space and can
and may well also satisfy the criterion that: thus only be used as indicators for the conditions experienced
by the occupants. In order to quantify the relationship between
 dry resultant temperature should not exceed 25 8C for more the sensor readings and parameters typically used to assess
than 5% of the occupied year (CIBSE Guide J [16]). thermal comfort, short-term monitoring studies are currently
being carried out. These include the measurement of operative
During the design phase, alternative design propositions temperature, PPD, PMV, air velocity and humidity. Once
were considered from a variety of view points, including cost, available, these data can be used to determine whether, and how
efficiency of space use, legibility of floor layouts, etc. The accurately, the BEMS temperature readings relate to the air
annual space heating and lighting energy use for each design temperature levels experienced by the occupants.
was also estimated, using the LT method [17]. This method is It is further worth noting that the CIBSE overheating
intended for comparison of design alternatives rather than for criteria, used here for performance assessment, are based on
reliably estimating actual energy demands. dry resultant temperature (DRT) as their target parameter. It is
It is interesting to see whether the measured internal tempe- probable that, during the summer days, the DRT is lower than
ratures concur with the predicted values and to compare the actual the air temperature because the thermal mass will have been
energy use with that which is typical of other building types, most cooled during the night. Therefore, it is likely that the
notably air-conditioned buildings—air conditioning is the frequency of occurrence of high temperatures as measured by
‘standard’ approach to conditioning a sealed, deep plan building. the BEMS sensors is an overestimate of the actual frequency
of occurrence of high dry resultant temperatures, i.e. the
4. Monitoring of temperature and CO2 levels building may actually perform better than the results
presented in this paper indicate. Further work is underway
4.1. BEMS sensors to clarify this matter.

In order to ensure comfortable internal conditions for all 4.3. CO2 sensors
occupants, the BEMS must be able to control ventilation and
temperatures in the different zones, i.e. open plan areas as well The BEMS air quality sensors log CO2 concentrations at
as enclosed study room and offices. Therefore, a large number hourly intervals. Inspection of the logged data showed that
of sensors are distributed throughout the space (Fig. 4). some CO2 sensors read consistently higher than others.
Based on readings from these sensors, individually However, comparison of the time series traces showed that
controllable dampers are adjusted to provide ventilation for the overall shapes of the profiles of all the sensors matched well
thermal comfort and air quality in each zone. In addition to their and that the values for individual sensors decreased to a
function as part of the building control system, the sensors can consistent minimum value over night. Considering that internal
contribute to the assessment of the building’s performance if CO2 concentrations during unoccupied periods (e.g. over night)
their readings are continuously logged. For a comprehensive are likely to be equivalent to the ambient CO2 concentration, it
analysis, long term monitoring data are required for a variety of was assumed that minimum values which exceeded the ambient
zones, ideally from all sensor locations. However, logged data concentration were in error due to calibration drift. The data
is currently only available for a limited number of BEMS from these sensors were therefore adjusted by an offset value so
sensors. Temperature and CO2 data, for example, are available that the baseline values of all sensors matched that of the sensor
for two locations on the ground and third floors, and four with the lowest baseline readings, i.e. night-time values of
locations on the second floor. Logged data for the first floor approx. 360–370 ppm, which is equivalent to typical back-
includes temperature only for three locations. ground CO2 concentrations [18].
796 B. Krausse et al. / Energy and Buildings 39 (2007) 792–801

Fig. 5. Internal and external temperatures during the monitoring period (June
2004–June 2005).

5. Measured thermal performance

Fig. 7. Average temperatures on each floor and the ambient temperature during
a ‘hot spell’.
The following is based on temperature data logged during
the period June 2004–June 2005 (although 3 weeks are missing
in autumn 2004 and 1 week in January 2005). first three working days and are similar from Wednesday to
The results presented are derived from time series data Friday. On Saturday the peak temperature and the duration of
logged at hourly intervals by eight BEMS sensors, two for each warmer temperatures is less, presumably due to the shorter
floor, typically located on two different walls at a height of period of occupancy.
about 1.5 m. During the warmer periods of the year the internal
The data show that the average temperature in the building temperatures are strongly influenced by those outside
remains relatively stable throughout the year (Fig. 5). During (Fig. 5). However, because of the thermal mass and night
the heating seasons the daytime indoor temperatures are venting strategy, individual hot days do not significantly raise
dominated by the heating schedule, heating set points and the the internal temperatures, see for example the days around 01
internal heat gains. The temperatures remained below 24 8C July, 22 July and 06 September in 2004. Even during the two
during the daytime and decreased to approximately 21 8C periods of prolonged high ambient temperatures, reaching up to
during the night, which is the minimum mid-week temperature 35 8C (in August 2004 and June 2005), the internal
set by the facilities managers and the temperature set for the air temperatures only occasionally exceeded 25 8C.
supplied by the lightwells. Decreases in temperature below During the hot spell in August 2004 (Fig. 7), the ambient
21 8C can be observed at weekends and more obviously during temperatures rose to over 30 8C. However, during this period,
the Christmas and Easter breaks – when the building was not the night-time ambient temperatures remained below 18 8C and
occupied and not heated. the diurnal temperature swing was in excess of 9 8C, with, on
There is a regular pattern to the temperatures during each the hottest day, a swing of 15 8C. There was therefore a
week of the heating season (Fig. 6). The building cools to its reasonable passive night-time cooling potential.
lowest temperature on Sunday nights, i.e. following the longest During the first 4 days of the hot spell (Fig. 7), the internal
period without occupancy of the week, but only to marginally temperatures remained relatively low, initially with morning
below 20 8C. The peak daily temperatures gradually rise for the temperatures in the library of around 21–22.5 8C and peak
temperatures under 25.5 8C. The diurnal variations in internal
temperature were between 2 and 3 8C. There was a gradual
increase in the peak internal temperature on successive warm
days, and this continued as the external daytime temperatures
became higher (04–08 August 2004). However, the night
ventilation cooling, together with the exposed thermal mass,
prevented the internal temperatures exceeding 26 8C, which, on
the hottest day, represented a temperature depression of over
5 8C below ambient. During the entire 2-year monitoring
period, the maximum internal temperature recorded was
26.4 8C, which occurred on the third floor on 19 June 2005
when the ambient temperature was 35.4 8C – a temperature
depression of 9 8C (Fig. 5).
During the last 5 days of the hot spell (Fig. 7), the diurnal
Fig. 6. Extract of time series data (March–April 2005) showing daily and swing in internal temperature remained between 2 and 3 8C on
weekly variability in indoor temperatures. all floors, with the ground floor, which benefits from the
 B. Krausse et al. / Energy and Buildings 39 (2007) 792–801 797

third floor and the ground floor, whilst on the first floor
temperatures remained below 25 8C throughout the entire
monitoring period. However, even on the third floor the CIBSE
2002 overheating criterion (less than 5% of occupied hours over
25 8C) was met, with temperatures greater than 25 8C only
occurring during 3.8% of the hours of use. The internal
temperatures never exceeded 27 8C, i.e. less than the 11 h
predicted at the design stage; which confirms the expectations
stated by Cook et al. [5] – that with intelligent BEMS control a
better building performance can be achieved than the
simulation results suggested.
Clearly the building meets the current CIBSE 2005/2006
criterion [14,15] that less than 1% of occupied hours should
exceed 28 8C.

Fig. 8. Example time series for all floors and ambient during moderate summer 7. Spatial variability of temperature and CO2 levels
weather.

greatest height of ventilation stack, and thus the greatest In order to investigate whether uniform conditions are
buoyancy driving forces, having the greatest night-time achieved across the deep plan library floors, time series data
temperature reductions. Regarding the other three floors, the from different locations on the second floor are compared.
third floor tended to be warmer than the second, which was Data for temperature and CO2 levels is available for a 6-
warmer than the first. Considering the relative stack heights on week period in summer 2005 (10 May–22 June 2005) from four
each of these floors, and yet their similar occupancy sensor pairs at locations A–D, as shown in Fig. 4. Sensors at A,
characteristics, this is perhaps to be expected. It reinforces C and D are located above study desks (without PCs) around the
the notion that it is the top floors of naturally ventilated side walls of the building, aligned with perimeter ventilation
buildings that are the most susceptible to overheating. stacks. The sensors at location B are located on an internal wall
Towards the end of the warm spell (Fig. 8), when ambient near the lift adjacent to the entrance door. All sensors are
temperatures generally stayed below 25 8C, the pattern of installed at a height of 1.5 m.
internal temperatures returned to that described above: daytime Time series plots, such as the 2-week plot in Fig. 9, clearly
peaks of around 24 8C and night-time minima of 21 8C. It is illustrate the different variability of the temperatures and CO2
interesting to note, however, that during this period, it is the concentrations. Although both parameters are influenced by
temperature profiles for the first and second floor that are very ventilation settings and occupancy levels, the CO2 curves show
similar. The third floor and ground floor are warmer, by 1 8C, distinct peaks, usually in the early afternoon, while the
perhaps because, on the third floor, the night ventilation cooling temperature curves rise throughout the day before decreasing
is less effective due to reduced stack height. The ground floor sharply at night due to the night-time venting strategy.
result is unexpected but could be due to a number of factors: The plots showed that there were slight temperature
observations from site visits and anecdotal evidence (from library differences across the floor during the period investigated.
staff) indicate that the ground floor is more densely occupied than Temperatures seem to be typically higher at locations A and B
anticipated, which, together with office partitions which act as a than at locations C and D, the average difference between A and
barrier to airflow, will lead to higher temperatures. It is also D being 0.52 8C. This could be due to movement of people
possible that modifications of the BEMS sensors and controls are through the doors leading to the main staircase causing an
required. Future monitoring studies will investigate this issue. exchange of air into the adjacent zones where sensors A and B
are located. It is interesting to note however that this pattern
6. Comparison of measured thermal performance with sometimes changes, with temperatures at location D rising to
guidelines and predictions similar levels as A while temperatures at the adjacent location C
remain lower. A likely cause is that there is a rather localised
The overheating statistics (Table 1) show that the effect of desk users on temperature levels in the close proximity
temperature levels most frequently exceeded 25 8C on the of sensor D. The presence of people at the desks below the

Table 1
Number of hours during which various temperature thresholds were exceeded between 26 June 2004 and 24 June 2005 during the occupied period
Guideline temperature (8C) Number of hours over stated temperature (h)/percentage of occupied hours over stated temperature (%)
Ambient Ground floor First floor Second floor Third floor
25 149 h/4.1% 78 h/1.95% 0 h/0% 32 h/0.8% 152 h/3.8%
27 73 h/2.0% 0/0 0/0 0/0 0/0
28 48 h/1.3% 0/0 0/0 0/0 0/0
798 B. Krausse et al. / Energy and Buildings 39 (2007) 792–801

Fig. 9. Temperature and CO2 levels measured at locations A–D on the second floor during the period 30 May 2005–12 June 2005.

sensor will warm the air and cause it to rise up towards the comparison for the whole building is possible, by presenting the
ventilation openings of the perimeter stack, i.e. vertically past library and benchmark data normalised by both floor area and
the sensor location, while sensor C on the adjacent stack (7 m period of occupancy (Table 2).
away) is likely to register little change in temperatures if the With an annual consumption of 0.049 kWh/(m2 h), the
desks directly below it are unoccupied. building performs significantly better than recommended in the
A similar observation can be made for the CO2 concentra- ECON19 good practice guidelines for office buildings [20]
tions. While the profiles of the data from the three sensors seem (Fig. 10). The building uses 51% less energy than the typical
to match in general, concentrations at one location occasionally air-conditioned building and 35% less than the typical naturally
increase significantly above the levels measured at the other ventilated open plan building. In fact, the Lanchester Library
locations. This often coincides with elevated air temperature at also performs better than an office building built to the good
the same location. Again, increased occupancy density at the practice standard for naturally ventilated open plan offices.
desks near the sensors is thought to be a likely reason for these
peaks. 9. Likely performance in other UK cities
The CO2 concentrations for occupied periods during the 6
weeks investigated were typically between 400 and 500 ppm, By comparing the weather data recorded in Coventry with
with occasional peaks of up to around 700 ppm. The maximum that for other UK location it is possible to infer how the the
CO2 concentration recorded was 720 ppm, which is below the library would perform at these other locations, and in particular
limit of 1000 ppm recommended for school buildings [19]. whether acceptable internal comfort conditions would be
These preliminary investigations suggest that the natural achieved. The UK CIBSE has recently produced weather data
ventilation strategy is working well, ensuring a sufficient
supply of fresh air and providing relatively uniform conditions
Table 2
across the second floor. More comprehensive monitoring
Energy consumption of library in 2004
campaigns, using additional sensors located throughout the
zones and covering a wider range of operating conditions and End use
comfort parameters, will reveal whether comfortable indoor Heating Electricity Cooling
conditions are consistently achieved. Total annual consumption (MWh) 1117 1012 205
Consumption per m2 (kWh/m2) 95 86 17
8. Energy consumption Consumption per m2 and per occupied 0.024 0.021 0.004
hour (kWh/(m2 h))
The library’s measured annual energy consumption of
electricity and gas for 2004, as determined from meter readings,
was 198 kW/m2. This includes the heating, lighting and power
consumption of the basement and the four levels of the library;
the two cannot be disaggregated. The basement is a computer
suite, accessible 24 h each day, with high power and lighting
loads and so is mechanically cooled. The library itself is
accessible for approximately 4000 h each year.
With this data it is difficult to make comparisons between the
energy consumption of the library and benchmark figures for
purely naturally ventilated buildings; or with energy use Fig. 10. Comparison of the library’s annual energy consumption during 2004
predictions for the library made at the design stage. However, a with ECON19 benchmark values for typical and good practice offices [20].
 B. Krausse et al. / Energy and Buildings 39 (2007) 792–801 799

Fig. 11. Comparison of recorded exceedance hours from Coventry with Fig. 12. Comparison of recorded exceedance hours from Coventry with
CIBSE’s Test Reference Year data (TRY) from 14 UK cities. CIBSE’s Design Summer Year data (DSY) from 14 UK cities.

for 14 UK cities with locations ranging in latitude from Plymouth although in the middle of the city, where there is a substantial
and Southampton in the South to Edinburgh and Glasgow in urban heat island affect, comfort may well have been
Scotland and Belfast in Northern Ireland. Close to Coventry, data compromised [24]. The results indicate that the building would
is available for Birmingham and Nottingham. The other cities are also have maintained comfort during hot years (i.e. those that
Cardiff (Wales), Leeds, Manchester and Newcastle in the North are only exceed for 1 year in 10) at all locations except perhaps
of England, Norwich in the East, and Swindon and London in the Birmingham, Leeds and London. It is worth noting however,
South East. For each city there is a Test Reference Year (TRY), that the library performed well within the overheating criteria
which typifies conditions experienced at the site, and a Design limits and so it may well tolerate the higher temperatures
Summer Year (DSY) that is intended for use in analyses to assess experienced in Birmingham and Leeds in 1 year in 10.
the risk of summertime overheating in naturally ventilated
buildings. The TRYis composed of 12 individual months chained 10. Recommendation for designers
together, where each month is the most typical of that
experienced during a 20-year period, and the DSY is the third Whilst the library has performed well there are a number of
hottest year in the 20-year period, i.e. there is only 1 year in 10 recommendations that flow from the study that might assist the
that is likely to be hotter. The derivation of the years and some of designers of future buildings of this type. These are, in addition
their characteristics have been presented by others elsewhere to those commonly adopted in low energy building design (e.g.
[16,21–23]. exposed thermal mass and night-time ventilation, solar shading,
The total number of hours for which the recorded ambient high levels of insulation, good glazing specification, maximis-
temperature in Coventry exceeds various temperatures1 is ing daylighting):
compared with the corresponding values for the 14 TRYs and
DSYs in Figs. 11 and 12, respectively, with the number of hours
Table 3
over 25, 27 and 28 8C being listed in Tables 3 and 4. It is evident Number of hours for which various ambient dry-bulb temperature thresholds
that the Coventry temperatures show a similar trend to the TRY were exceeded at Coventry and in each CIBSE Test Reference Year
data but that that the number of hours with temperatures over
City Temperature threshold (8C)
30 8C is greater than in any of the other 14 cities. In the 26–
29 8C range, the Coventry temperatures exceed those for all 25 27 28
a
locations except London (Fig. 11). Comparing the Coventry Coventry 162 79 52
data with the DSYs, it can be seen (Fig. 12) that Coventry was Belfast 0 0 0
warmer in the range 26–29 8C than all but three of the other Birmingham 71 30 14
Cardiff 13 1 0
cities (Birmingham, Leeds and London) but only Leeds and Edinburgh 6 0 0
London had a higher occurrence of temperatures above 28 8C. Glasgow 0 0 0
Given that in Coventry the library satisfied all three of the Leeds 51 5 2
overheating criteria considered (Table 1), we may conclude that London 106 42 28
the building would have maintained comfort in 13 of the 14 UK Manchester 56 22 8
Newcastle 18 3 1
cities in a typical year. It may also have remained comfortable, Norwich 54 24 8
as defined by hours over 28 8C in the London environs, Nottingham 50 3 0
Plymouth 3 0 0
Southampton 46 11 7
1 Swindon 49 13 4
The Coventry hours have been normalised to account for the 679 missing
a
hours in September/October 2004 and January 2005. Year from 26/6/04 to 24/6/05.
800 B. Krausse et al. / Energy and Buildings 39 (2007) 792–801

Table 4 11. Conclusions

Number of hours for which various ambient dry-bulb temperature thresholds
were exceeded at Coventry and in each CIBSE Design Summer Year
The Lanchester Library at Coventry University has a deep
City Temperature threshold (8C) plan and a sealed façade and yet is naturally ventilated and
25 27 28 daylight using internal lightwells and perimeter stacks to
Coventry a
162 79 52
ventilate the four floors of library and study areas.
Belfast 8 0 0 The building benefits from the exposed thermal mass and the
Birmingham 109 45 27 night ventilation strategy. Isolated warm days caused minimal
Cardiff 18 3 0 rise in internal temperatures. Even during prolonged hot spells,
Edinburgh 10 3 0 which, in the period June 2004–June 2005, included outside air
Glasgow 11 3 0
Leeds 178 93 58
temperatures as high as 35.4 8C, the internal temperature did
London 267 107 63 not exceed 26.4 8C: thus internal temperatures were up to 9 8C
Manchester 52 20 14 below peak ambient temperatures. All floors of the building
Newcastle 6 0 0 therefore comfortably met the prevailing CIBSE Guide A [14]
Norwich 35 4 1 thermal comfort criterion: that there should be no more than 1%
Nottingham 25 10 3
Plymouth 36 13 9
of occupied hours with a dry-resultant temperature above
Southampton 26 2 0 28 8C. In fact, the building also met the tougher criterion,
Swindon 66 14 4 mentioned in CIBSE Guide J [16], that there should be no more
a
Year from 26/6/04 to 24/6/05.
than 5% of occupied hours over 25 8C.
The peak summertime temperatures on the third floor of the
library are higher than those on floors 1 and 2; the top floor has
 Areas which are to be occupied for longer periods of time, the smallest stack height by which to drive a flow. This
especially into the evenings and possibly at night, can be co- indicates that, in stack ventilated buildings, it is the top floor
located (in an accessible but access controlled) part of the which is likely to be critical when trying to meet overheating
building. This enables the majority of the building to be criteria.
operated with night-time ventilation without affecting the Temperature and CO2 data from four sensor locations on the
comfort of night-time occupants. second floor indicate that the natural ventilation strategy is
 The summertime performance can be further improved if air working well. It provides relatively uniform conditions across
inlet dampers are closed down when the internal dry-resultant the floor and ensures a sufficient supply of fresh air; CO2
temperatures are below ambient. They can be controlled in concentrations rarely exceeded 700 ppm, which is below the
response to internal CO2 levels under these circumstances. recommended maximum value of 1000 ppm [19].
 The performance of buildings is dramatically improved if With a total annual consumption of 0.049 kWh/(m2 h), the
there is a tenacious building manager at work, with an interest building performs significantly better than recommended in the
in monitoring the building to meet the demands of the good practice guidelines for offices [20]. The library uses 51%
occupants whilst satisfying the requirement to use energy less energy than a typical air-conditioned office and 35% less
efficiently. than a typical naturally ventilated open plan office. However,
 It may be necessary to provide seasonal control variations these ‘energy savings’ are conservative as the values include the
whereby the limiting extent of ventilation openings is supply to the 24-h computer suite which is not part of the
reduced in winter relative to summer. This has the benefit of natural ventilation strategy.
reducing the risk of over-ventilation in winter, thereby The monitoring covered a particularly warm period that had
reducing energy consumption and the risk of draughts. a greater occurrence of summertime ambient temperatures over
 More generally, there is a difficulty in advanced naturally 26 8C than is typically experienced in any of the 14 UK cities
ventilated buildings in translating the control strategy studied, except for London. The temperatures were higher than
envisaged at the design stage into an operational program those experienced in 1 year in ten at all but three of the cities.
in a BEMS. This is largely because there can be a chain of Overall, therefore, given the geographical spread of the cities,
professionals involved, the environmental designer, a we may conclude that deep-plan, naturally ventilated buildings
building services engineer and then a controls specialist, with sealed facades, if well designed, could maintain
and partly because each of these is contracted at a different acceptable thermal comfort standards in all but a very few
time and for a different specific task. It is hard therefore for UK locations.
early design concepts to be retained and for information to The building is a marked success based on occupant
pass back and forth along the chain of individuals involved. feedback and the number of students attracted to use the library.
Further, whilst concepts will be relayed down to the BEMS Nevertheless some observations about how to improve the
programmers, it is rare, in the authors’ experience, that a building’s operation and the performance of advanced naturally
programmed control strategy will be passed back for approval ventilated buildings in general are made. It is hoped that the
by an engineer or environmental designer. This control paper will give designers added confidence with which to
strategy/implementation interface is an area which is embark on the design and construction of naturally ventilated
probably worthy further study. buildings—even in tough urban environments.
 B. Krausse et al. / Energy and Buildings 39 (2007) 792–801 801

12. Further work [8] M.J. Cook, C.A. Short, Natural ventilation and low energy cooling of
large, non-domestic buildings—four case studies, The International Jour-
nal of Ventilation 3 (4) (2005) 283–294.
This brief analysis has focused on long term temperature and [9] A. McDonald, Celebrating outstanding new library buildings. Society of
CO2 data. However, in order to assess the building’s College, National and University Libraries, 2002 [Accessed August 2006].
performance more thoroughly, and to diagnose the reasons http://www.sconul.ac.uk/pubs_stats/newsletter/27/ARTICL27.RTF.
for the performance features observed in this paper, detailed [10] J. Field, Breeze blocks, Building Services Journal, December 2000, pp.
medium-term measurements of indoor thermal comfort and 18–22.
[11] S. Pidwell, Lanchester Library by short and associates, Architecture Today
occupant satisfaction are planned. Hopefully, these will include 115 (2001) 38–49.
questionnaires to capture the users’ perception regarding these [12] Data Model Summary ESP-r version 9 series, report no. TR96/2, Energy
issues. Systems Research Unit, 1996.
[13] M.J. Holmes, E.R. Hitchin, An example year for the calculation of energy
demands in buildings, Building Services Engineering 45 (1978) 186–190.
Acknowledgements
[14] CIBSE Guide A: Environmental Design, The Chartered Institution of
Building Services Engineers, London, 2006.
The building was designed by Short and Associates [15] TM 36: Climate change and the internal environment, a guide for
Architects with staff in the Institute of Energy and Sustainable designers. Technical Memorandum, The Chartered Institution of Building
Development as the Environmental Design Consultants. We Services Engineers, London, 2005.
gratefully acknowledge the continuing support from Caroline [16] CIBSE Guide J: Weather, solar and illuminance data, The Chartered
Institution of Building Services Engineers, London, 2002.
Rock (head librarian) and her colleagues in the library, and [17] N. Baker, K. Steemers, The LT Method Version 2.0: An Energy Design
from the Estates Department at Coventry University, particu- Tool for Non-Domestic Buildings, Cambridge Architectural Research
larly Jim Skelhon. Ltd., 1994.
[18] Global Monitoring Division, Trends in atmospheric carbon dioxide,
References National Oceanic and Atmospheric Administration [Accessed August
2006]. http://www.cmdl.noaa.gov/ccgg/trends/.
[19] Building Bulletin 101: Ventilation in School Buildings, Department for
[1] Our energy future—Creating a low carbon economy, Energy White paper, Education and Skills, 2006.
Department of Trade and Industry, HMSO, London, 2003. [20] Energy use in offices, Energy Consumption Guide 19, Building Research
[2] B. Bordass, R. Cohen, M. Standeven, A. Leaman, Assessing building
Energy Conservation Support Unit, Energy Efficiency Best Practice
performance in use: energy performance of the PROBE buildings, Build- Programme, 2000 [Accessed August 2006]. http://www.cibse.org/pdf/
ing Research and Information 29 (2) (2001) 114–128. egg019.pdf.
[3] K.J. Lomas, M.J. Cook, Sustainable buildings for a warmer world, in: [21] D.H.C. Chow, G. Levermore, P. Jones, D. Lister, P.J. Laycock, J. Page,
Proceedings of the World Renewable Energy Congress, Aberdeen, May
Extreme and near-extreme climate change data in relation to building and
22–27, 2005. plant design, CIBSE Building Services Engineering Research and Tech-
[4] K.J. Lomas, Architectural design of an advanced naturally ventilated nology 23 (4) (2002) 233–242.
building form, Energy and Buildings 39 (2007) 166–181.
[22] M. Ren, N. Doylend, G.J. Levermore, The impact of new CIBSE weather
[5] M.J. Cook, K.J. Lomas, H. Eppel, Design and operating concept for an data on natural ventilation design, CIBSE Building Services Engineering
innovative naturally ventilated library, in: Proceedings of the CIBSE Research and Technology 24 (2) (2003) 83–922.
National Conference, Harrogate, UK, October, 1999. [23] CIBSE, CIBSE/Met Office TRY/DSY Hourly Weather Data Set (CD-
[6] M.J. Cook, K.J. Lomas, H. Eppel, Use of computer simulation in the
ROM) - 14 sites, Available form the Chartered Institution of Building
design of a naturally ventilated library, in: Proceedings of the PLEA99 Services Engineers, 2003.
Conference, Brisbane, Australia, 1999. [24] R. Watkins, J. Palmer, M. Kolokotroni, P. Littlefair, The London Heat
[7] C.A. Short, K.J. Lomas, A. Woods, Design strategy for low-energy Island: results from summertime monitoring, CIBSE Building Services
ventilation and cooling within an urban heat island, Building Research
Engineering Research and Technology 23 (2) (2002) 97–106.
and Information 32 (3) (2004) 187–206.