Solar Energy 137 (2016) 516–530

Contents lists available at ScienceDirect

Solar Energy
journal homepage: www.elsevier.com/locate/solener

Hourly consumption profiles of domestic hot water for different

occupant groups in dwellings
Kaiser Ahmed a,⇑, Petri Pylsy a,c, Jarek Kurnitski a,b
a
Aalto University, Department of Civil Engineering, Finland
b
Tallinn University of Technology, Faculty of Civil Engineering, Estonia
c
The Finnish Real Estate Federation, Finland

a r t i c l e i n f o a b s t r a c t

Article history: Hourly consumption data of domestic hot water (DHW) is essential to compute the energy demand, and
Received 2 May 2016 for system sizing. Few on-site measured data and simulation based studies are available to forecast DHW
Received in revised form 1 August 2016 consumption focusing on a daily average, hourly average, appliance consumption, and occupant number.
Accepted 21 August 2016
This study derived the hourly DHW profiles for 5 different groups of 1 person, 3 people, 10 people, 31
Available online 7 September 2016
people and more than 50 people as a function of the number of occupants. Weekday (WD), Weekend
(WE) and Total day (TD) consumption variation were analyzed. The study accomplished with on-site
Keywords:
hourly DHW consumption measurement from 86 apartments with 191 occupants during one year and
DHW consumption
Hourly profile
findings was also validated against a larger sample from previous study. A specific selection procedure
Weekday was developed to find out the most representative DHW profile among measured candidates of each
Weekend group fulfilling the selection criteria. Selected profiles had similar daily consumption (L/per./day) to an
Residential building average of all profiles and also followed similar consumption pattern during a day. Two sharp peaks with
large consumption variance were found in each day and smaller groups had higher consumption during
peak hours. Result also found higher evening peak compared to the morning peak and the average con-
sumption of peak hours was 2–4 times higher than non-peaks hours. Morning peaks of WE shifted 2–3 h
later from WD’s and kept similar position during the evening. Profiles of 5 groups were necessary to nor-
malize with scaling factor to maintain the daily average value. Derived hourly values could be used with
monthly factors to deliver hourly profiles of all months and the format of hourly and monthly factors
used allows to define DHW consumption in relevant simulation and sizing software.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction
Residential building sector has accounted 16–50% of total
energy in most of the countries (Saidur et al., 2007) where domes-
tic hot water (DHW) is reported a good percentage of it. In Finnish
residential building it requires the second largest of energy
demand next to space heating (Hakala, 2014). Dutch residential
sector accounted for 72% of total DHW consumption, whereas
12% and 16% had reported for small and large scale of office users
(Geudens, 2008). The information of DHW in residential building is
very significant because of having a large contribution on overall
energy demand.
DHW consumption profile is complicated and strongly fluctu-
ated over time. Graphical location, weather condition, occupant
number, occupant behavior toward DHW usages, lifestyle, social
⇑ Corresponding author at: Rakentajanaukio 4 A, FI-02150 Espoo, Finland.
E-mail address: kaiser.ahmed@aalto.fi (K. Ahmed).
and economic condition, etc. were found most significant variables
in literature body (Aydinalp et al., 2004; Meyer and Tshimankinda,
1998a; Papakostas et al., 1995; Wolf et al., 1980). Perlman and
Mills (Perlman and Mills, 1985) had identified demographics of
the location, occupant number, occupant attitude toward usages,
ownership, etc. as the most significant factors, whereas Tso
(Tso and Yau, 2003) addressed occupant income, ownership of
household, housing type, seasonal variation as additional factors.
Reference study (Vine et al., 1987) monitored DHW consumption
in four apartment buildings in San Francisco for 4–6 months and
interviewed the usage pattern of DHW. It reported education, cul-
tural and social norms as significant variables. Abrams and Shedd
(Abrams and Shedd, 1996) found a variation of DHW usages from
day-to-day and seasonal variation, other studies also marked sea-
sonal variation as a significant factor (Becker and Stogsdill, 1990;
Perlman and Mills, 1985). Among all factors Parker (2003) noticed
an occupant number as the weightiest factor and occupancy pat-
tern also had a significant impact on domestic energy demand

http://dx.doi.org/10.1016/j.solener.2016.08.033
0038-092X/Ó 2016 Elsevier Ltd. All rights reserved.
 K. Ahmed et al. / Solar Energy 137 (2016) 516–530 517

Nomenclature

v t;a average consumption of apartment ‘a’ at hour ‘t’ (Liter) f t;G hourly consumption factor at ‘t’ hour of selected profile
n days in a month; n = 1, 2, 3, . . . , 31 for given group (L/person)
N total days in a month (NTotal = 30 days, NWD = 20 days, Rt consumption ratio at specific hour
N WE = 10 days for November 2014) DHW t hot water consumption at specific hour (L/per./h)
v t;a;o average consumption of each occupant of apartment ‘a’ DTW t total water consumption at specific hour (Hot + Cold) (L/
at hour ‘t’ per./h)
Oa occupant number at apartment ‘a’ vm average hourly consumption at ‘m’ month (L/per./h)
vt average consumption at hour ‘t’ (L/person) v t;m;G hourly consumption at ‘t’ hour of ‘m’ month of year for
v t;o occupant ‘O’ consumption at hour ‘t’ (Liter) given group (L/per./h)
Ni eliminated total occupant number f t;G hourly consumption factor
V a;m hourly average consumption (L/per./h)
V annual daily average consumption (L/per./day) Indices
fm monthly consumption factor t hour in a day; t = 1, 2, 3, . . . , 24
VG daily consumption of given group i.e., sum up consump- a apartment or dataset; a = 1, 2, 3, . . . , 86
tion of 24 h (L/per./day) i elinated occupant number (for November iTD = 164,
V G;a hourly average consumption of given group (L/per./h) iWD = 164, iWE = 163)
Sm;G scaling factor at ‘m’ month for given group m month of year
v t;G consumption at ‘t’ hour of selected profile for given G group
group (L/person)

(Stokes et al., 2004). The study found a higher consumption rate of
3 or less people compare to 4 or higher people occupied Canadian
households.
Daily average consumption and hourly profile of DHW are not
similar in all countries. The annual average of DHW for Finnish
and Swedish people were 43.0 and 33.0 L/per./day (Ahmed et al.,
2015b; BBR, 2012). Schipper (1982) found that people of certain
developed European countries consumed 7 times less DHW than
Americans (Schipper, 1982). In addition, the hourly consumption
during a day are also varied along nationwide. German have higher
consumption during the morning and low consumption during the
evening, whereas Finnish people have a higher consumption dur-
ing the evening (IEA, 2007). Hourly consumptions were also found
diverse along the week (Hidalgo et al., 2012; Papakostas et al.,
1995). Papakostas et al. (1995) monitored DHW usages in four
apartment buildings in a solar village 3 in Greece and presented
the average hourly, daily, yearly DHW usage profiles (Papakostas
et al., 1995). The hourly consumption pattern during WD were
nearly similar with morning and evening peaks, whereas usage
pattern were more uniform during WE. In this study monthly con-
sumption factor in Table 1 are recommended to be used for deliv-
ering the hourly profile.
Meyer and Tshimankinda had investigated the DHW consump-
tion in South African traditional houses (Meyer and Tshimankinda,
1996), townhouses (Meyer and Tshimankinda, 1998b) and apart-
ment buildings (Meyer and Tshimankinda, 1998a) over a period
of one year. The study reported the consumption in L/per./day as
a function of summer WD, summer WE, winter WD and winter
WE. Low density traditional houses (Meyer and Tshimankinda,
1996) and townhouses (Meyer and Tshimankinda, 1998b) con-
sumed 30% and 44% more DHW than high density houses and
townhouses. Also, 70% of consumption increased from summer
to winter for both cases. On the other hand, low density occupied
apartments used more than four times hot water than high density
apartment buildings and consumption increased by 80% from sum-
mer to winter (Meyer and Tshimankinda, 1998a). The daily con-
sumption also varied during WD and WE with two peaks, i.e.
morning and evening peaks. In apartments and townhouses, the
morning peaks during WD occurred at 6:00 and 8:00 h for high
and low density respectively. For evening peaks, it found at
19:00 and 21:00 h, respectively (Meyer and Tshimankinda,
1998a,b). In traditional houses, morning and evening peaks were
noticed at 7:00 and 20:00 h in both cases. Though two peaks were
available in WD and WE, the WE peaks were shifted 1–2 h later
than the WD peaks (Meyer and Tshimankinda, 1998a).
Many researchers also had drawn the DHW consumption pat-
tern based on on-site field measurement since 1970 (Becker and
Stogsdill, 1990; Burch and Salasovich, 2002; Perlman and Mills,
1985). The data sets were collected three to four decades ago and
might not reflect the current usage pattern. A Danish study found
the mean DHW consumption reduction over 200 L/day to about
100 L/day for Danish families during 1989–1996. The reasons were
water price, water saving campaign, shorten the family sizes
(Knudsen, 2002), energy efficient appliance (Bansal et al., 2011),
new regulation and market forces (Koomey et al., 1999) etc.
More review of forecasting method of DHW consumption are
well explained in literature body (Aydinalp et al., 2004; Bagge
and Johansson, 2011; Blokker et al., 2010; Good et al., 2015;
Paatero and Lund, 2006). Blokker et al. (2010) proposed an end
use water demand model which could predict the water demand
pattern with 1 s resolution. The model based on statistical data
of end uses in the household as well as information about occupant
number, ages, usage frequency, flow intensity, event duration, vol-
ume flow of each usages event, number of occurrences. In this
approach end uses and user information required initially. After-
ward, information regarding frequency, intensity, and duration of

Table 1
DHW monthly consumption factor for Finnish apartment buildings (Ahmed et al., 2015a,b).

Annual average specific Jan. Feb. Mar. Apr. May Jun. Jul. Aug. Sep. Oct. Nov. Dec.
consumption, L/person/day
Arithmetic mean (Total) 43 1.105 1.040 1.033 1.005 0.977 0.895 0.880 0.964 0.969 0.980 1.111 1.041
Arithmetic mean (WD) 43 1.119 1.049 1.020 0.999 0.941 0.912 0.892 0.964 0.982 0.987 1.100 1.030
Arithmetic mean (WE) 43 1.067 1.017 1.062 1.023 1.062 0.860 0.845 0.964 0.940 0.961 1.136 1.067
518 K. Ahmed et al. / Solar Energy 137 (2016) 516–530
usage collected from local statistics and survey report, which fur-
ther correlated with the diurnal pattern of local individuals. Once
this information was put into the simulation model it could gener-
ate water demand pattern of multiple days for single or multiple
households. The approach was required a large number of statisti-
cal information about appliance, occupant, and diurnal pattern
which were not available in the current study. Another study
(Paatero and Lund, 2006) had applied slightly different approach
to generate realistic hourly consumption of domestic electricity
for a small number to thousands of households. The first part of
bottom up model defined the appliance stocks in each household
and diurnal consumption level by determining the social random
factors in daily basis. The probability density function, and stan-
dard deviation of social random factors were calculated from
domestic consumption data of two Finnish datasets. The following
part simulated the electricity consumption profile of each appli-
ance separately based on consumption cycle. The consumption
cycle of each appliance also considered the seasonal probability
factor with hourly probability factor for WD and WE, social proba-
bility factor, active consumption, standby consumption, starting
frequency. The mentioned factor was obtained from public reports,
statistics, and appliance information. This model generated hourly
load consumption during WD and WE for individual appliance or
appliance group in each household.
Utilization of solar energy depends on availability of the solar
resource, the efficiency of solar energy system, storage, user profile
of energy consumption, etc. In Nordic countries the solar resource
is available very short duration in a day during mid-autumn to mid
spring. For an instance, the availability of solar energy in Finland is
1170 kW h/m2 which is nearly same as in central Europe but most
of the radiation generates in the southern part of Finland during
May to August (Pihlakivi, 2015). Many researchers had already
shouted about the importance of accurate DHW profile for perfor-
mance evaluation of solar energy technologies (Jordan and Vajen,
2000; Spur et al., 2006). Though many studies used a repeating
daily profile while designing the solar domestic hot water and solar
Combi-systems (Enteria et al., 2014; Hobbi and Siddiqui, 2009;
Lima et al., 2006), the consumption volume and pattern has large
variability along days of the week and months of the year. Jordan
and Vajen (2000) studied the importance of the draw-off duration
and flow rate of DHW for sizing the solar Combi-system. In addi-
tion, DHW draw-off had a severe influence on stratification of
the storage tank. In reference study (Knudsen, 2002) used two dif-
ferent DHW load profiles and found higher utilization of solar
energy by a system with unrealistic profile compared to the more
realistic profile which could lead a wrong sizing of the system. In
the same context, other studies (Buckles and Klein, 1980; Spur
et al., 2006) found the influence of daily DHW draw-off profiles
on the storage system and identified the needs of realistic DHW
hourly profile for testing and simulation of storage system. Open
literature shows a large number of applications of the DHW profile
while dimensioning the solar thermal plant, geothermal heat
pump, DHW storage tank (Chow et al., 2006; Kalogirou, 2000;
Sterling and Collins, 2012; Wolf et al., 1980). Hidalgo et al.
(2012) used dynamic DHW consumption profile of 202 Spanish
inhabitants while dimensioning the storage system. The study con-
sidered the average hourly consumption of 202 inhabitants which
could dampen the actual inhabitant hourly consumption pattern.
Hence, an under estimation of solar thermal storage could place.
In similar context, Edwards et al. (2015) identified the draw-off
timing as most significant variable of solar thermal system sizing.
Thus, realistic DHW profiles are evidently important for correct
dimensioning of solar thermal system. Application of DHW profiles
in previous studies is shown in Table 2.
Our work demonstrates a quite detailed and realistic end use
DHW consumption data of 86 Finnish apartments during a year.
The hourly consumption of DHW with monthly correction factor
can apply to forecast the demand from a single person up to a large
number of people. The proposed method did not required an
extensive information about consumer and their appliances.
This study was validated against previous study where daily
DHW consumption of four Finnish apartment buildings in Tampere
with 182 apartments and 379 occupants during two consecutive
years was analyzed (Ahmed et al., 2015a,b) and daily consumption
(L/per./day) values with monthly correction factors were developed.
This study accomplished of hourly consumption of 86 rental apart-
ments with 191 occupants during one year and found similar daily
consumption over successive year. It was found 2–6% of consump-
tion variation compared to previous study. In this study, authors
considered an hourly basis consumption of each occupant to draw
an hourly consumption pattern during winter, summer and spring.
Hourly consumption of 1 person, 3 people, 10 people, 31 people, over
50 people groups during WD and WE were generated. It portrayed
real peaks which might absence if considered an hourly average of
all occupants. In addition, authors reported the procedure to gener-
ate profiles of all 12 months in a year with consideration of monthly
consumption variation. The proposed profiles can be used in rele-
vant simulation work and system sizing software which can account
hourly DHW take-off pattern, individual consumption, and avail-
ability of solar energy during a day being all very important input
data for solar thermal and storage systems sizing.
2. Method
2.1. Description of target building and apartments
The study was accomplished based on hourly consumption of
domestic hot water (DHW) and domestic cold water (DCW) data
from 86 Finnish apartments. All apartments with 191 occupants
belong to one large limited liability housing company (owner occu-
pied apartments) which is located in Helsinki, Finland. The exact
information about occupant age, gender, and income data were
not collected and not taken into account due to concern on occu-
pant privacy. Apartment number, type and family size are listed
in Table 3.
Most of the Finnish population are living in Helsinki, Tampere
and Turku which are covered by climate zone 1 and 2. The annual
average outdoor temperature of climate zone 1 and 2 are 5.3 and
4.6 °C respectively. The study data represents well the majority of
the Finnish population, however, some Northern part of Finland hav-
ing extreme winter may apply different profiles of DHW. The
obtained profile with a monthly correction factor from reference
(Ahmed et al., 2015b) could consider as an hourly user profile of
DHW for Finnish apartment building. According to Finnish building
code the minimum and maximum value of supply water tempera-
tures are 55 and 65 °C respectively, which are usually in the range
between 55 and 60 °C. In this study, the supply water temperature
has been at least 55 °C which fulfills minimum requirements of the
Finnish building code (D1, 2007). The family size of the apartments
is shown in Table 3 which further matched with an exact occupant
number of each apartment that was collected from the register book
of a building. Each apartment had two individual meters, which took
the readings of DHW and DCW separately. Occupants were using
DHW for bathing, shower, washing hands and kitchen facilities. In
Finland dishwasher and washing machine use only DCW. District
heating system was available for space and DHW heating.
2.2. Data measurements
Each apartment had an individual metering system to measure
DHW and DCW separately, which further used for billing purpose.
 K. Ahmed et al. / Solar Energy 137 (2016) 516–530 519

Table 2
Application of DHW profiles in previous studies.

Reference Research problem Methods Outcomes

Hampel Studied the influence of the DHW draw-off Simulation based study that evaluated the Collector outputs decreased almost 5% during
et al. timing of a multi-family house collector outputs while comparing the DHW evening draw-off of DHW instead of morning
(1999) draw-off during the morning and evening for a draw-off for a SDHW system
SDHW system
Jordan and Studied the influence of the draw-off duration Simulation based study that made a comparison More realistic profiles saved 2% of energy. DHW
Vajen and flow rate of DHW on the solar Combi – of energy savings between DHW load profile draw-off duration and flow rate had significant
(2000) system (prEN-12977, 1997) and a more realistic profile influence on the temperature stratification in the
with 1-min time scale storage tank
Knudsen Studied the effect of a constant consumption A theoretical investigation that considered 12 A higher utilization of solar energy of a system
(2002) pattern during a day and a realistic DHW load different SDHW systems with different tank with unrealistic profile compared to the more
profile for a Danish family on SDHW system volumes and collector sizes. Constant and realistic profile. Also, no influence of different
sizing realistic profiles were used and made a thermal DHW profiles was found for the highest thermal
performance comparison between systems performed system
Bales and Studied 7 external DHW units with different A short survey, laboratory test and simulation Maximum possible primary flow, low
Persson control methods. The study determined the most work were proceed to identify the most temperature returned water to tank and hot
(2003) effective method and influential factors of influential design factors water load profile were distinguished as the
energy savings most important factors
Furbo et al. Studied the thermal performance of SDHW A simulation model was developed which A SDHW system with smart solar tank had 5–
(2005) system with smart solar tanks considered the tank dimension, collector 35% higher thermal performance compared to
parameters, four different daily hot water the tradition SDHW system depending on the
consumption volume, 5 different draw off consumption volume and pattern.
patterns
Spur et al. Studied the influence of daily DHW draw-off A TRNSYS simulation model was used to show The study concluded to use realistic daily profile
(2006) profiles on the storage system the effect of realistic daily profile and five other for performance analysis of thermal storage
existing profiles on the DHW storage system system
Hobbi and Studied the optimum design solution of a forced Simulation based study to optimize the solar By utilizing the solar energy, the proposed
Siddiqui circulation solar water heating system for a fraction of the entire system. The annual average system could provide 83–97% and 30–68% of
(2009) single-family residential unit daily DHW consumption of the household was DHW during May to September and October to
246 L/day and it was distributed according to February respectively
Rand profile, where the morning and evening
peaks were available from 9:00–10:00 and
19:00–20:00
Hidalgo Studied the dimensioning criteria of DHW solar A transient simulation program had developed. By considering the transient effect of input value
et al. plant under transient process of solar irradiance, The environmental data of 10 min interval and and delivered water temperature, the study
(2012) consumption, and thermal accumulation hourly consumption profile of 215 inhabitants found the storage tank size had an influence on
were used instead of constant values solar plant performance
Lauterbach Studied the integration of solar thermal system A TRNSYS model was developed and findings The collector parameter, load characteristic such
et al. into the hot water supply system of a brewery. were validated with the measured heat as mass flow and temperature were identified as
(2014) Also, evaluated the system performance with quantities the most influential factor for the system
different faults which occurred during the performance
system operation
Dongellini Studied the hourly energy production of the solar Simulink and F-chart method were evaluated the Solar coverage factor was strongly dependent on
et al. thermal panels, and the temperature of the hot solar coverage factor by using static and dynamic the hourly consumption profile of DHW
(2015) water produced by the system with static and DHW profile
dynamic hourly consumption profile
Edwards Studied the thermal performance of SDHW and Realistic profile was developed based on on-site Realistic profile had an effect of thermal
et al. solar Combi-system with repeating daily measured data from 73 houses in Quebec. performance of a solar DHW system. Timing of
(2015) consumption profile and more realistic Simulation of a solar DHW system was DHW draw-off had a significant impact on
consumption profile conducted with these two profiles thermal performance of a system

Table 3 The main objective of the study was to find out the hourly con-
Apartment type, apartment number and family size in a building.
sumption variation along 24 h in a day. Consumption at different
Apartment type No. Occupancy in apartment No. months in winter period followed a similar pattern, however, the
One room apartment 2 One person occupied 24 hourly consumption volume was varied. Authors considered the
Two room apartment 20 Two people occupied 31 month of November and August as the representative months of
Three room apartment 27 Three people occupied 15 summer and winter DHW consumption. Because of long winter
Four room apartment 30 Four people occupied 15
duration, authors also considered the consumption of the month
Five room apartment 7 Five people occupied 1
of January to show the consumption variation. Consumption pat-
tern during November and January was similar, except the hourly
consumption volume. The monthly consumption factor of DHW for
each month was obtained in previous study, where found a higher
The accuracy level of measurement devices was good i.e., ±2%. The consumption level during November (Ahmed et al., 2015a,b). It
measured data were stored in a database with a resolution of 1.0 L. could also possible to choose consumption pattern in October, Jan-
The individual smart water metering system provided measure- uary and February. December could not use as representative
ment for each apartment as follows month because of Christmas holidays. In addition, most of the Fin-
nish people spend summer holidays in June and July which may

Cumulative meter readings of hourly consumption of DHW for not represent the actual consumption profile. The same explana-
each apartment in liter. tion was also applicable for the summer period. Thus, authors con-

Cumulative meter readings of hourly consumption of DCW for sidered the month of August as representative month of summer.
each apartment in liter. The data were collected from 17.05.2014 to 05.02.2015 with the
520 K. Ahmed et al. / Solar Energy 137 (2016) 516–530
1 h interval period. Some data cleaning was also accomplished dur-
ing data processing work. Data were not taken into account if the
apartment was vacant for a whole month or daily average con-
sumption of individual was less than 20 L/per./day. There was only
one 5 person occupied apartment which also ignored in this study.
Some justification has been done due to the quality of measured
data. The occupant number beyond 2 in apartment are considered
as children and they have similar consumption volume of adult
people.
2.3. Data analysis
Hourly basis data of 86 apartments were separated according to
the month of November, August and January. Cumulative meter
readings of hourly consumption of DHW and DCW for each apart-
ment in liter for the month of November were processed. It found
86 datasets for 86 apartments respectively, and each dataset con-
tained each single hour consumption in a day and 30 days of a
month. To obtain the hourly profile of an apartment in a day, each
hour average consumption of the apartment was obtained from the
following equation
Xn
v t;a;n
v t;a ¼ 1
ð1Þ
N
With t, hour in a day, (h), a, apartment or data set, v t;a , average
consumption of apartment ‘a’ at hour ‘t’, (Liter), n, days in a month,
N, total days in a month (N Total = 30 days, N WD = 20 days,
N WE = 10 days for November 2014).
The processed data represented the hourly consumption of
each apartment which further simplified to hourly consump-
tion of individual in that apartment by using the following
equation
v t;a
v t;a;o ¼ ð2Þ
Oa
With v t;a;o , average consumption of each occupant of apartment
‘a’ at hour t (L/person), Oa , occupant number at apartment ‘a’.
The simplified form was given an hourly average consumption
profile for all occupants in respective apartment. It is assumed that
each individual in the respective apartment had a same DHW con-
sumption profile. The total number of occupants was found 191 in
a building and hourly profiles were generated respectively.
Authors made an assumption and not to consider those occupants
profiles whose average daily consumption for any month was less
than 20 L/per./day. The number of occupants eliminated to 164
and ensured minimum daily consumption of 20 L/per./day for
November total days. After the elimination November WD and
WE occupant number were 164 and 163 respectively. The average
of each hour (1.0–24 h) of 164 profiles were obtained which could
present the hourly profile of daily consumption for the month of
November.
Pi
v t;o
vt ¼ 1
ð3Þ
Ni
With v t , average consumption at hour ‘t’, (L/person), v t;o , occu-
pant ‘O’ consumption at hour ‘t’, (Liter), i, eliminated occupant
number (For November iTD = 164, iWD = 164, iWE = 163), N i , elimi-
nated total occupant number.
This section had considered a large number of occupants which
ignored the actual peak of DHW consumption, but it portrayed the
idea of consumption pattern. Moreover, it was found large varia-
tion of consumption during different hours in a day and authors
had figured out the frequency of consumption for each hour.
Though the study was accomplished based on one building with
86 apartments, the daily consumption also justified by previous
study (Ahmed et al., 2015b) where daily DHW consumption of four
buildings with 182 apartments in Tampere was analyzed. In addi-
tion, frequency of daily consumption of present study compared to
previous study.
Many studies (Parker, 2003; Stokes et al., 2004) gave the occu-
pant number as the weightiest factor of DHW consumption. In this
study profiles are drawn as a function of occupant number. In order
to draw DHW hourly consumption profiles from single to large
number of occupants, 5 different group sizes, i.e. 1 person, 3 peo-
ple, 10 people, 31 people, 164 people were considered to explain
properly the hourly variations. The group sizes were selected while
considering that it could draw an hourly DHW profile for any num-
ber of occupants. Consideration a large number of samples always
preferable compare to a small number of datasets and proposed
profiles are most representative one with real peaks. Due to the
large variety of consumption, it was hard to select one profile out
of 164. While selection the representative profiles authors consid-
ered the following circumstances

Selected hourly profiles had nearly similar daily consumption to
average daily consumption of 164 profiles,

Selected hourly profiles followed the similar hourly consump-
tion pattern of 164 occupant average profiles,

Smaller group expected large consumption in peak hour com-
pared to large group,

Selected profiles are most representative hourly DHW profiles
of the respective groups.
For one person hourly consumption profile authors considered
164 hourly profile or datasets whereas the study had 24 datasets
of one person occupied apartment (Table 3). Authors considered
a large number of datasets instead of the small number of datasets.
At initial phase 15 profiles closest to the average daily consump-
tion of 164 profiles were selected. Among 15 profiles 4 good candi-
dates named as P1, P2, P3 and P4 were selected based on similarity
with the shape of the average consumption profile of 164 occu-
pants. These 4 good candidates could represent the hourly con-
sumption profile for 1 person group. The study had 164
individual consumption profile and Excel RAND function was used
to pick 3 random profile out of 164 and formed a group of 3 people.
The total number of combination was 54 and each combination
had an own hourly profile that was accomplished from the hourly
average of 3 profiles, i.e. a set of 24 data corresponding 24 h. The
random selection procedure was repeated 3 times, i.e. 162 combi-
nations in order to sure that the selection of combination would
not affect the results. These 162 combinations produced 162 pro-
files and among of them 15 were selected which had nearly similar
daily consumption to the average daily consumption of all 164 pro-
files. Again 4 different candidates named as P1, P2, P3 and P4 were
chosen which could represent an hourly consumption profile for 3
people group. The same procedure was followed to draw the con-
sumption profile of DHW for 10 people group and 31people group.
The number of combinations were formed by 48 and 15 for 10 peo-
ple and 31 people groups respectively. In some cases there were
many similar candidates available and selection of the most repre-
sentative profile seemed very difficult.
To stress the importance of peak hours, consumption hourly
values between 7.00–9.00 and 20.00–22.00 were multiplied with
weighting factors. Weighting factors of 3, 5 and 10 were applied
for these hours, i.e. to improve the resolution of the selection pro-
cedure the values of these peak hours were increased (for 3 people
and 10 people group peak hour consumption of all 164 profiles
were multiplied by factor 3 and the rest of other hours were mul-
tiplied by factor 1). This resulted in more visible peaks. The authors
applied this approach, because some cases, profiles of 3 people and
10 people had a close result to the average daily consumption of
 K. Ahmed et al. / Solar Energy 137 (2016) 516–530 521

164 profiles, but hourly consumption pattern did not follow the
hourly consumption pattern of 164 occupants average profile.
Some candidates with factor 10 for 3 people group is shown in
Fig. 1.
The new daily average of 164 profiles with factor 3 was
obtained. Excel RAND function picked randomly 3 among 164 pro-
files and formed 54 combination. The same selection procedure
repeated for 3 times and extended the combination to162 times.
The hourly profile of each combination was accomplished from
the hourly average of 3 profiles, i.e. a set of 24 data corresponding
24 h. 10 profiles were selected which had nearly same daily aver-
age consumption to new daily average of 164 profiles with factor
3. Afterward, consumption ratio (Dx) between evening peaks to
morning peaks also considered if needed while selection of pro-
files. Among of them profiles of similar pattern to the average
one were selected. It produced some good candidates of 3 people
and 10 people groups. The similar procedure is applied of factor
5 and 10 to get better candidates. The profiles with different factor
for 3 people and 10 people groups are shown in Fig. 2.
The peak hours of all profiles (Fig. 2) were divided by respective
factor. Minimum 4 good candidates were available for each group
while consideration of factor or without factor. The selection of one
good candidate out of 4 had chosen by assuming that small group
had a higher consumption during peak hours compare to large
group. Selected one candidate for each group could use as a repre-
sentative profile for respective group. Weighting factors 3, 5, and
10 led to selection of the same candidates as without weighting.
Therefore the weighting just confirmed the result in this study
and provided additional confidence in the case of many similar
candidates.
The main objectives of the study were to find out the shape of
hourly DHW consumption for all months. The average daily con-
sumption in L/per./day for all months was taken from (Ahmed
et al., 2015b) because of larger sample and nearly similar daily con-
sumption compared to present study. Furthermore, selected pro-
files of respective group may be used to generate hourly profiles
of other months in the winter season if it is assumed that other
months of winter season kept similar shape of hourly consumption
as November. The daily consumption of DHW for each month was
obtained by multiplying the monthly consumption factor with
annual daily average consumption from (Ahmed et al., 2015b).
Afterward, normalizing scaling factor was assigned to each group
in order to maintain the daily average value. The following equa-
tions were used to form scaling factor
V  fm
V a;m ¼ ð4Þ
24

Fig. 2. Similar candidates named as P1, P2, P3 and P4 of November for group (a) 3
people with factor 5, (b) 10 people with factor 5, (c) 3 people with factor 10, and (d)
10 people with factor 10.

VG
V G;a ¼ ð5Þ
24

V a;m
Sm;G ¼ ð6Þ
V G;a

f t;G ¼ Sm;G  v t;G ð7Þ

Fig. 1. Similar candidate with factor 10 for 3 people group.
522 K. Ahmed et al. / Solar Energy 137 (2016) 516–530
With V a;m , hourly average consumption, (L/per./h), V, annual
daily average consumption, (L/per./day), f m , monthly consumption
factor, V G , daily consumption of given group, i.e. sum up consump-
tion of 24 h, (L/per./day), V G;a , hourly average consumption of given
group, (L/per./h), Sm;G , scaling factor at ‘m’ month for given group,
v t;G , consumption at ‘t’ hour of selected profile for given group,
(L/person), f t;G , hourly consumption factor at ‘t’ hour of selected
profile for given group, (L/person).
Furthermore, domestic cold water hourly profile for the month
of November has drawn. The hourly ratio of DHW to DCW (total) is
calculated by using the following equation
DHW t
Rt ¼
DTW t
With Rt , consumption ratio at specific hour, DHW t , hot water
consumption at a specific hour (L/per./h), DTW t , total water con-
sumption at a specific hour (L/per./h).
The same procedure was applied for WD, and WE hourly con-
sumption for month of November. It also applicable to obtain
hourly consumption of WD, WE and Total day of January and
August.
Fig. 3. Frequency of consumption at peak hours of 8 h, 9 h, 20 h, 21 h and non-peak hours of 14 h, 15 h during November weekday.
3. Results and discussion
3.1. Average daily consumption and variation at occupant level
The result was drawn based on the hourly DHW and DCW con-
sumption of 191 occupants from 86 apartments belongs to one Fin-
nish building in Helsinki region. The obtained daily DHW
consumption in L/per./day was shown similar results to previous
study where conducted the datasets of 4 Finnish buildings of 182
apartments with 379 occupants in Tampere, Finland (Ahmed
et al., 2015b). The daily average consumption of total occupants
were 38.5, 47.0, 45.2 L/per./day for the month of August, November
and January respectively, and found a maximum consumption
variation of 6% from the previous study. Another finding in the cur-
ð8Þ rent study was higher variation of consumption in L/per./day than
in previous study (Ahmed et al., 2015b). While data processing of
individual data set of 191 occupants the DHW consumption varied
from 0 to 300 L/per./day. Moreover, authors considered a dataset of
hourly consumption during one year, while the illustration of
hourly profiles and quality of proposed profiles might be improved
if authors could manage couple of consecutive year data. In Table 1
consumption variance of DHW during WD and WE were not
 K. Ahmed et al. / Solar Energy 137 (2016) 516–530 523

seemed very significant which also true for the present study. between 20.00 and 22.00 h similar to WD peak. The average con-
However, consumption patterns were varied during WD and WE. sumption at peak hours during WD and WE in November were
4.1 and 3.6 L/per./h respectively which also varied from 0 to
113 L/per./h For non-peaks hours average consumption were 1.1
3.2. Consumption frequency at the hourly level
and 1.4 L/per./h for WD and WE respectively. The consumption fre-
quency at peak and non-peak hours during November WE are
The proper identification of daily peak consumption was one of
shown in Fig. 4.
the objectives of the study and it is also important information for
Authors also mentioned the average consumption were 3.6 and
system sizing. The data analyzed to quantify the hourly variation of
1.5 L/per./h during peak and non-peak hours, respectively, while
DHW and found two peaks, i.e. morning and evening during a day.
considering the similar usage pattern for total days (TD) in Novem-
Morning peaks associated in between 7.00 and 9.00 whereas eve-
ber (Ahmed et al., 2016). Almost 90% of consumption for WD, WE
ning peaks were available in between 20.00 and 22.00 during
and total days (Ahmed et al., 2016) were found in between the
WD. In peak hours the consumption varied from 0 to 107 L/per./h
range of 0–20 L/per./h during the measured period. In some cases
whereas other parts of the days it seemed nearly uniform con-
the deviations of consumption were found 10 times higher than
sumption. During 11.00–16.00 h the consumption rate was low
the average one, but the probability of occurrence were very low
and almost no consumption was found from 1:00 to 5:00 h for
which could ignore. Information of consumption frequency of each
all representative months. The frequency of consumption at peak
hour is significant while sizing the renewable energy system. For
and non-peak hours during the weekdays of November are shown
instance, higher contribution of solar fraction was found during
in Fig. 3.
afternoon compared to morning (Buckles and Klein, 1980; Evarts
Morning peaks shifted 2–3 h later during WE compared to WD.
and Swan, 2013) and production variation of ground source heat
Expected peak consumption during the morning were available in
pump also changed along a day (Sterling and Collins, 2012). Fre-
between 10.00 and 13.00 h. However, evening peaks occurred in

Fig. 4. Frequency of consumption at peak hours of 11 h, 12 h, 21 h, 22 h and non-peak hours of 14 h, 15 h during November weekends.
 524
Table 4
Frequency of DHW consumption for month of November during weekday.

Consumption, 1:00 2:00 3:00 4:00 5:00 6:00 7:00 8:00 9:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00 23:00 0:00
L/per./h
0.0 0.81 0.92 0.95 0.95 0.98 0.93 0.72 0.42 0.41 0.66 0.74 0.75 0.78 0.78 0.78 0.75 0.63 0.50 0.48 0.44 0.40 0.43 0.42 0.59
1.0 0.11 0.05 0.04 0.04 0.02 0.04 0.12 0.14 0.19 0.14 0.12 0.10 0.10 0.11 0.11 0.14 0.18 0.23 0.22 0.20 0.21 0.22 0.21 0.19
2.0 0.04 0.01 0.00 0.00 0.00 0.01 0.06 0.09 0.10 0.05 0.04 0.05 0.03 0.03 0.04 0.03 0.08 0.10 0.09 0.07 0.09 0.06 0.09 0.06
3.0 0.01 0.00 0.00 0.00 0.00 0.00 0.01 0.08 0.05 0.04 0.02 0.03 0.03 0.02 0.02 0.02 0.04 0.05 0.06 0.06 0.06 0.05 0.06 0.05
4.0 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.05 0.05 0.02 0.02 0.01 0.01 0.01 0.01 0.02 0.02 0.04 0.04 0.05 0.04 0.04 0.03 0.02
5.0 0.01 0.00 0.00 0.00 0.00 0.00 0.01 0.03 0.03 0.02 0.01 0.00 0.01 0.01 0.01 0.01 0.02 0.02 0.03 0.03 0.03 0.03 0.03 0.01
6.0 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.03 0.03 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.02 0.02 0.03 0.02 0.02 0.01
7.0 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.02 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 0.01 0.02 0.01 0.02 0.03 0.01
8.0 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.02 0.00 0.00 0.01 0.01 0.01 0.00 0.01 0.01 0.01 0.01 0.02 0.03 0.02 0.02 0.01
9.0 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.02 0.01 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.01 0.01 0.01 0.01 0.01 0.01 0.00
10.0 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.02 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.02 0.01 0.02 0.01 0.00
11.0 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.01 0.01 0.01 0.01 0.01
12.0 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.02 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.02 0.01 0.01 0.00
13.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.02 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 0.00 0.00
14.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.02 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 0.01 0.00
15.0 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.00 0.00 0.00
16.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.02 0.01 0.01 0.00

K. Ahmed et al. / Solar Energy 137 (2016) 516–530

17.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 0.00
18.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 0.00
19.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.01 0.00
20.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.00

Table 5
Frequency of DHW consumption for month of November during weekend.

Consumption, 1:00 2:00 3:00 4:00 5:00 6:00 7:00 8:00 9:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00 23:00 0:00
L/per./h
0.0 0.77 0.83 0.86 0.86 0.96 0.95 0.92 0.83 0.67 0.50 0.44 0.42 0.44 0.48 0.56 0.57 0.51 0.48 0.49 0.46 0.47 0.43 0.48 0.62
1.0 0.11 0.10 0.10 0.10 0.02 0.03 0.05 0.11 0.16 0.19 0.19 0.18 0.21 0.18 0.18 0.19 0.21 0.21 0.20 0.19 0.16 0.24 0.21 0.18
2.0 0.04 0.02 0.02 0.02 0.01 0.01 0.02 0.03 0.06 0.08 0.09 0.09 0.10 0.08 0.09 0.08 0.06 0.09 0.08 0.08 0.08 0.09 0.09 0.07
3.0 0.03 0.01 0.01 0.01 0.00 0.00 0.01 0.01 0.04 0.05 0.06 0.07 0.04 0.07 0.04 0.06 0.06 0.05 0.06 0.07 0.04 0.05 0.04 0.04
4.0 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.01 0.02 0.05 0.04 0.04 0.05 0.05 0.03 0.02 0.03 0.02 0.05 0.02 0.04 0.03 0.03 0.02
5.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.02 0.03 0.03 0.02 0.03 0.01 0.01 0.02 0.03 0.03 0.03 0.03 0.03 0.03 0.02
6.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.02 0.03 0.02 0.02 0.02 0.01 0.02 0.03 0.03 0.03 0.02 0.01 0.02 0.02 0.02
7.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.02 0.02 0.02 0.02 0.02 0.01 0.01 0.01 0.01 0.01 0.02 0.03 0.01 0.02 0.01
8.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.02 0.03 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.02 0.02 0.02 0.01 0.01
9.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.02 0.01 0.01 0.00 0.00 0.01 0.02 0.01 0.02 0.01 0.02 0.01 0.00
10.0 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.00 0.01 0.01 0.00 0.01 0.01 0.01 0.00 0.01 0.02 0.01 0.01 0.00
11.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 0.01 0.01 0.00 0.00 0.01 0.01 0.01 0.01 0.01 0.01 0.00 0.00
12.0 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.02 0.01 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.01 0.00 0.01 0.00 0.00
13.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 0.00
14.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 0.00
15.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.00 0.01 0.00
16.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 0.01 0.01 0.01 0.00 0.01 0.00 0.00 0.01 0.00 0.00 0.00 0.00
17.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.00
18.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.00 0.00
19.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.01 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.00
20.0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.00
 K. Ahmed et al. / Solar Energy 137 (2016) 516–530 525

quency of DHW consumption for the month of November during are usually more dampened and smoother. Identification of real
WD and WE are shown in Tables 4 and 5 respectively. peaks also very significant while system sizing. Occupant number
Tables 4 and 5 have shown the frequency of consumption vari- was found as the key parameter of DHW consumption, which
ation as well as shifting of peaks. The findings also had shown a
good agreement with previous research works (Meyer and
Tshimankinda, 1998a). It also pointed out the importance of indi-
vidual profile of DHW for WD and WE. Hourly average consump-
tion of all occupants, those had a minimum 20 L/per./day for the
month of August, November and January are shown in Fig. 5.
Fig. 5 shows a clear movement of the peaks and significance of
WD and WK profile over total day profile. It found a good agree-
ment with previous studies (Table 1) (Ahmed et al., 2015a,b) where
maximum consumption was noticed during November. This study
reported 2–7% of consumption variation from previous study
(Table 1) during WD and 3–7% of variation for WE. The pattern
of average hourly consumption of all 164 profiles was nearly sim-
ilar for November, August and January.

3.3. Profiles for 5 different groups

Authors omitted average daily consumption of all 164 profiles

as representative profile because average profiles for a large group

Fig. 5. Hourly average profile of DHW consumption for month of (a) August, (b) Fig. 6. Similar candidates of November for group (a) 31 people, (b) 10 people, (c) 3
November and (c) January. people, and (d) 1 person.
526 K. Ahmed et al. / Solar Energy 137 (2016) 516–530
had also shown an influence along a day. Larger group had nearly
similar consumption pattern, whereas it found more turbulence for
smaller groups. As mentioned in the method section that each
group had a couple of good candidates and each candidate could
use as a representative profile for respective group. The similar
candidates for different groups of November are shown in Fig. 6.
Authors considered a large number of datasets to propose most
representative one with real peaks, but not the exact profile of the
respective group. Most precise profile out of good candidate pro-
files (Fig. 6) for respective group were selected while consideration
that a smaller group would have a large consumption in peak hour
compared to large group. Moreover, the total number of occupants
in a building is required while selecting a single profile for building
rather gets summed up multiple profiles. The variation of hourly
consumption for 5 different groups are shown in Fig. 7.
The reason of drawing separate profile for different groups is
shown in Fig. 7. However, there were some common findings such
as two peaks were available for all groups, peak consumption
hours were shifted 2–3 h later in WE from WD, more turbulence
Fig. 7. Proposed profiles of 5 groups for month of November (a) Weekday, (b)
Weekend, (c) Total day.
consumption during evening for all cases, more uniform consump-
tion of non-peak hours during WD compared to WE. Similar find-
ings also found for the month of August which is shown in Fig. 8.
The multi number of profiles for different group could able to
estimate more preciously the consumption of any number of occu-
pants. For instance, if the designer would like to draw a profile of 5
storied building with 35 occupants, he or she could easily choose a
profile of 31 people group. Similarly, profiles of 3 people group
could more suitably for 4 people occupied apartment. Moreover,
if anyone plans to design for 100 people, profile of 164 people
can be used. It also found very small difference between the con-
sumption profile of 164 people, 82 people, and 54 people which
is shown in Fig. 9.
The selection procedure of each profile was explained briefly in
the method section. Selected candidates of each group followed the
similar usage pattern to the average of all profiles. The most key
component of describing the shape of the profiles were morning
and evening peak ratios, consumption peaks and duration. Most
representative profiles of 5 groups were complemented by scaling
Fig. 8. Proposed profiles of 5 groups for month of August (a) Weekday, (b)
Weekend, (c) Total day.
 K. Ahmed et al. / Solar Energy 137 (2016) 516–530 527
factor and implied with monthly factor to deliver hourly profile of
0:00

1.14
1.08
0.94
0.74
0.37
0.31
0.81

0.26

0.98
1.20
0.59
0.47
0.32
0.80

0.05

all months in respective season. For November, daily consumption

of selected profiles for different groups was scaled with daily aver-
age consumption of November which was obtained from a previ-
23:00

1.98

1.47
1.67
1.38

1.62
1.82
1.85
1.78
1.89
2.01

2.20
1.60
0.57

0.87

0.55

ous study (Table 1). Afterward, factor for each hour of all groups
had assigned by using following formula (Ahmed et al., 2016).
22:00
2.41
2.78

1.32

1.55
1.56

1.72

2.54
1.23
3.04

0.96

2.02

0.46
2.09

2.20
0.96

The scaled hourly consumption factors of November and August

for 5 individual groups are shown in Tables 6 and 7.
21:00

V  fm
2.67
3.39
2.87
3.53
1.77
2.26
1.97
2.62
3.48

2.51

1.53
0.37

2.30
3.02
3.60

vm ¼ ð9Þ
24
20:00
1.99
1.82

4.58
3.25

1.58
2.49

1.97
1.87
3.46
1.92
1.90
2.09

0.14

3.06
0.90

v t;m;G ¼ v m  f t;G ð10Þ

With v m , average hourly consumption at ‘m’ month, (L/per./h),
19:00

1.26

1.29
1.13
1.35
1.06
0.87
0.43
0.99
2.09
1.04
0.91

0.79
0.14
1.06
0.94

v t;m;G , hourly consumption at ‘t’ hour of ‘m’ month of year for given
group, (L/per./h), f t;G , hourly consumption factor (Tables 6 and 7).
18:00

1.35
0.74
0.56
0.42
0.83
0.63
1.01
0.98
0.97
0.41

0.85
0.94
0.84
0.66
0.64

The hourly consumption factor of other months in winter fol-

lowed the November hourly consumption factor. Hourly average
17:00

profiles of October, December, January, February, and March could

1.17

1.53
0.57
0.57
0.39
0.66
0.33

0.86
1.30
0.28

0.82
0.95
0.73
0.47
0.53
be drawn with multiplying the average daily consumption of the
respective month (Table 1). In similar way August hourly factor
16:00

1.34

1.27
1.35
0.39
0.38
0.25

1.04
0.97

0.76
0.97
0.56
0.25
0.98
0.03
0.04

could use to generate hourly average profiles for different group

of April, May, June, July, and September.
15:00

1.22

1.62
0.38
0.57
0.26
0.12
0.26

1.03
1.03
2.03

0.71
0.78
0.53
0.41
0.77
3.4. Domestic hot water ratio
14:00

1.44
1.41
1.92
1.37
1.81
0.54
0.77
0.64
0.14

0.88
0.89
0.62
0.93
0.47
0.04

Required energy for DHW depends on supply hot water temper-

ature, incoming cold water temperature, volume flow, etc. As the
13:00

usage pattern of DHW during WD and WE were not similar, it

1.89
2.17
1.44

2.41
1.13

1.51
0.67
0.93
0.57
0.31

1.10
0.79

0.57
0.04

2.00
would expect also a different profile of domestic cold water. The
hourly consumption of domestic cold water and ratio of hot water
12:00

1.81

1.76
3.56

1.12

1.46
0.71
0.78

0.25
0.15

2.02

4.09

0.92
0.63
1.20
0.40
to cold (total) water in November is shown in Fig. 10.
The average consumption ratio varied from 0.30 to 0.33 for
November. The ratio seemed higher during peak hours of both
11:00

1.99
1.95
2.76
1.53

1.33
0.75

0.47
0.31
0.11

3.90
1.20
0.94

1.03
1.50
0.70
10:00

1.24

1.16
1.15

1.72

1.31
1.07
0.96

0.32
0.58

0.94
0.68

1.09
0.81

0.65
0.83
9:00
2.38
1.82
3.47
3.62

1.65
1.77
2.22
0.79
0.49
0.36
0.14
0.32
0.44

0.96
0.92
8:00

2.85
3.66
4.77
8.91

1.95
1.82
2.37

5.57
0.28

0.58

3.10
3.00

0.30
0.07
0.08
7:00

1.33

2.53

1.58
0.76
0.79

0.14

0.55
0.34
0.44
0.08
0.05
0.01
0.04
0.09
0.50
6:00
0.14

0.29

0.61

0.19

0.12

0.38
0.09

0.01

0.03
0.03
0.03
0.01

0.10

0.04
0.03
Hourly consumption factors of 5 groups for month of November.

5:00
0.01
0.02
0.01
0.01

0.05
0.04
0.05

0.03
0.04
0.01
0.01
0.00

0.00
0.00

0.00
4:00

0.11
0.13
0.05
0.02
0.02

0.09
0.03
0.07
0.07
0.06
0.03
0.01
0.00
0.00

0.00
3:00

0.11
0.13
0.05
0.02
0.02

0.09
0.03
0.07
0.07
0.06
0.03
0.01
0.00
0.00

0.00
2:00
0.14

0.23
0.18
0.33
0.21

0.17
0.14

0.28
0.07
0.03
0.03

0.06
0.00

0.00

0.00
1:00
0.43
0.33
0.37

0.55
0.83
0.47
0.54

0.47
0.45
0.21
0.83
0.05
0.04

0.02
0.00
P50 Per.

P50 Per.

P50 Per.
31 Per.

31 Per.

31 Per.
10 Per.

10 Per.

10 Per.
3 Per.
1 Per.

3 Per.
1 Per.

3 Per.
1 Per.
Weekend
Weekday

Total
Table 6
Fig. 9. Profiles of different groups for month of November (a) Weekday, (b)
Weekend.
 528 K. Ahmed et al. / Solar Energy 137 (2016) 516–530

cases. It was found very low value during 1.00–6.00 h, which might
0:00

reason of user behavior. As it seemed minimum consumption at

1.27
1.47
1.18

1.25
1.39

1.34
1.41
0.24

0.96
0.11

0.59
0.46
0.52
0.70

0.09
that period, occupant might use DHW for a very short duration. Ini-
tial delivered water temperature might low which could adjust
23:00
1.77
1.91
2.84
1.11

1.35

1.47

1.69
1.91
1.73
2.16
0.99

1.20

0.45
0.31

1.20
within few second. The supply water temperature of cold water
was found very significant in energy demand for DHW.
22:00
2.39

3.59
2.53
3.12
1.93
1.72
1.79
1.96
1.42
2.24

3.47
2.54
2.60

2.10
2.60
21:00

3.5. Application of proposed profiles

2.49
2.61

4.14
4.56
2.31

3.21

5.38
2.46
2.45
3.88

4.54
2.80

2.20

3.20

4.09
To show the effect of DHW profiles on solar thermal system per-
20:00
1.63
1.42
1.31
2.73
1.15
1.63
1.39
1.39
1.26
3.33

2.44
1.83
1.59
1.60
1.60

formance a simulation example was run with IDA Indoor Climate

and Energy 4.7 (IDA-ICE) building simulation tool (D1, 2007;
19:00

Sahlin et al., 2004). This tools is suitable for modelling of multi-

1.37

1.33

1.22

1.35
1.30
0.79
1.09

1.05

0.67

1.30
1.50
0.28
1.01
0.70

0.03

zones building, energy balance, HVAC system, building integrated

renewable system with consideration of the dynamic variance of
18:00

1.23
0.92
0.84
0.62
0.89
0.58
0.81
0.83

0.59
0.16
0.85
0.81
0.77
0.81
0.43

weather parameters, i.e. outdoor air temperature, relative humid-

ity, wind direction and speed, diffuse and direct normal solar radi-
17:00

ation etc. It also suitable to estimate the onsite energy production

1.18

1.32
0.58
0.67
0.62
0.49

0.92
0.79

0.76
0.76
0.58
0.66
0.06

0.06
0.00

from solar panel, solar thermal panel, wind turbine, ground source
borehole loop etc.
16:00
0.46
0.46
0.27
0.13

0.92
0.65
1.08
0.13
0.58
0.51
0.36
0.43
0.12
0.80
0.00

IDA Early Stage Building Optimization (ESBO) plant with a solar

thermal panel, hot water tank and top up heating system was used.
The annual average of DHW consumption with monthly consump-
15:00

1.31

1.57
0.52
0.39
0.42

1.10

0.77

0.79
0.71
0.47
0.42
0.88
0.25
0.04
0.00

tion factor for Finnish apartment building (Ahmed et al., 2015a,b)

and the hourly consumption factor for 5 different groups were
14:00

1.26

1.92
0.69
0.51
0.64
0.23

1.01
0.94
0.67

0.86
0.89
0.54
0.52
0.64
0.00

used as input parameters of the model. The model was developed

for max value of number of persons studied (50 people) in order to
keep the systems sizing parameters constant per person and then
13:00

1.46
1.33

1.59

1.25

1.15
0.86
0.92
0.93

1.09

0.66
1.04

0.76

0.61
0.60
0.00

studied profiles were applied as the only difference. System sizing

of 1.5 m2 of solar thermal panel and 0.03 m3 volume of hot water
12:00

tank per person were used. The incoming cold water and delivered
1.12

1.94

1.85

1.21
1.26
1.22
1.66
0.83

0.63

2.40

4.08
3.40
0.80

0.06
0.00

water temperature was assumed as 5.9 °C and 55 °C respectively

(Simonson, 2005). The simulation was run for five consecutive
11:00

1.76

2.55
2.56
5.91
1.28
1.21
1.05
0.92
0.97

2.06

1.03
0.63
0.55
0.30
0.00

working days selected from April, May, June, July, August and
10:00
1.33
1.43
1.17

1.81
2.12
2.37
1.62

1.43
1.24
1.50

0.35
1.50

1.03
0.34
0.00
9:00
2.25
2.37
2.54
3.42
2.14

1.65
1.81
1.32

1.66
0.53
0.37
0.68
0.67

2.01
0.03
8:00
2.39
2.35
1.98

1.63
1.63

1.31
7.23
3.30
9.70
0.18

0.19

2.02
0.07
0.04

0.03
7:00
0.52
0.64
0.11
0.36

0.12
0.24

0.22
0.61
0.13
0.04

0.06

0.40

0.03
0.00
0.00
6:00

0.13
0.04
0.04

0.20

0.08

0.03

0.05
0.04
0.01

0.03
0.00

0.00

0.00
0.00

0.00
5:00
0.02
0.03
0.01

0.03
0.05
0.03

0.03
0.03

0.02
0.06
0.00
0.00

0.00
0.00

0.00
Hourly consumption factors of 5 groups for month of August.

4:00
0.02
0.01

0.04
0.05
0.07
0.05
0.01

0.03
0.03

0.03
0.00
0.00

0.00

0.00
0.00
3:00

0.15
0.37
0.13
0.05
0.03
0.02

0.06

0.04

0.08
0.03
0.05

0.20
0.00

0.00

0.00
2:00
0.13

0.18
0.29
0.22
0.15
0.19

0.19
0.12

0.20
0.05

0.09
0.02
0.00
0.00

0.00
1:00

0.24

0.64
0.63
0.53
0.18

0.49

0.23
0.46
0.40

0.05
0.05
0.04

0.70

0.09
0.00
P50 Per.

P50 Per.

P50 Per.
31 Per.

31 Per.

31 Per.
10 Per.

10 Per.

10 Per.
3 Per.
1 Per.

3 Per.
1 Per.

3 Per.
1 Per.
Weekend
Weekday

Total
Table 7

Fig. 10. (a) Hourly average consumption profile of cold water (b) Ratio of DHW to
cold water (total) for month of November.
 K. Ahmed et al. / Solar Energy 137 (2016) 516–530 529

September so that both sunny and cloudy days were included, with
the following profiles:

Uniform distribution of DHW along 24 h,

1 person profile,

3 people profile,

10 people profile,

31 people profile,

50 or more people profile.
The shape factor of the hot water tank was 5. The stratified
tank model had eight vertical layers and the insulation was
0.25 W/m2 °C. The model also considered the circuit losses and
tank losses. Weather data of Helsinki (TRY2012) test reference year
was considered (Jylhä et al., 2011). The yearly average outdoor
temperature was 5.6 °C and the global horizontal solar radiation
was 975 kW h/m2/a.
Simulated solar thermal system was able to produce the most of
DHW in July while significantly more top-up heating was required
during other months, Fig. 11a. DHW heating energy use was in
between 2181 and 2442 Wh/day/person during summer period
(from April to September). The required top up heating power
showed remarkable dependency on applied profiles, Fig. 11b.
The simulated result also accounted all losses from the DHW
system. The simulated performance applies for Nordic countries
with long winter and higher consumption demand of DHW during
winter as well as short duration of sunlight. The required top up
heating during June and July was nearly similar for all profiles,
however, it has shown a visible variation of other months in sum-
mer. The hourly profile of 50 people considered the hourly average
consumption of large number individuals that could dampen and
smoother the actual inhabitant hourly consumption pattern. The
uniform profile had distributed equally the daily consumption
along 24 h, which could not show the actual required top up heat-
ing. In contrast, the hourly profile of 1 person accounted actual
peaks that represents a small apartment. The peak hour consump-
tion during the morning and evening was multiple times higher
compared to an average one (equal distribution of daily consump-
tion along 24 h) as well as non-peak hour consumption. The impact
of peak hour consumption on heating power is shown in Fig. 11b.
The required heating power was varied from 1.4 to 5.2 times
higher compare to uniform profile depending on the proposed pro-
files. Though the top up heating for uniform and other three pro-
files such as 50 people profile, 31 people profile and 10 people
profile were shown nearly similar result, the required heating
power were also varied from 1.4 to 2.5 times higher compare to
uniform profile depending on the proposed profiles. Furthermore,
the proposed profiles had a significant impact on storage system
which is shown in the following Table 8.
To keep the same peak power as in the case of the uniform pro-
file, the tank volume was varied from 1.1 to 8.8 times higher com-
pare to tank volume for uniform profile depending on the proposed
profiles.

4. Conclusion

This study derived hourly DHW consumption profiles for 5 dif-

ferent occupant groups. It also provided individual profiles of WD
and WE during August and November. The following conclusion
can be drawn based on the measured data from 86 Finnish apart-
ments with 191 occupants and with individual hot and cold water
metering system.

Daily average DHW consumption (L/per./day) in August,

November and January were nearly similar to previous study
with 6% of maximum variation. The variations were noticed of
2–7% and 3–7% for WD and WE respectively compared to previ-
ous study.

The large consumption variance during a day (0–300 L/per./day)
had shown a good agreement with previous study. Almost 90%
of hourly consumption were found in between 0 and 20 L/per./h
and following the lognormal distribution.

Two sharp peak consumption were available in WD. Morning
peak consumption associated in between 7.00 and 9.00 whereas
evening peak consumption were placed in between 20.00 and
22.00. The average consumption were 4.1 and 1.1 L/per./h for
Fig. 11. The effect of applied DHW profiles on required (a) Top up heating, (b) Top
peak and non-peak hours respectively.
up heating power.

Table 8
Storage volumes (m3/person) providing the same peak power as in the case of the uniform profile and 0.03 m3/person.

April May June July August September

Uniform profile 0.03 0.03 0.03 0.03 0.03 0.03
1 person profile 0.23 0.26 0.12 0.11 0.12 0.22
3 people profile 0.24 0.26 0.05 0.04 0.05 0.24
10 people profile 0.24 0.26 0.05 0.04 0.05 0.25
31 people profile 0.22 0.26 0.05 0.04 0.06 0.23
650 people profile 0.13 0.13 0.04 0.03 0.05 0.14
530 K. Ahmed et al. / Solar Energy 137 (2016) 516–530

In WE morning peak consumption were shifted 2–3 h later from
WD’s and kept similar position of evening peak consumption.
The average of peak hour consumption was 3.6 L/per./h and dis-
tributed into more consecutive hours. Non peak hours con-
sumptions seemed less uniform than WD’s and average was
found 1.4 L/per./h.

This study proposed hourly profile of DHW consumption for 5
occupant groups as a function of occupant number. This allows
to estimate more preciously the DHW consumption of a build-
ing with a small or large number of occupants. In the case of
a small number of occupants the hourly peak values were by
up to factor 4 higher compared to peak values of a large occu-
pant group revealing that the building size has important
effects on the DHW profile shape.

The hourly profile for WD, WE, TD of all months could be gen-
erated from obtaining profiles with monthly consumption fac-
tors (Table 1) and hourly consumption factors (Tables 6 and 7).

The required top up heating power showed remarkable depen-
dency on applied profiles, and it varied from 1.4 to 5.2 times
higher compare to uniform profile depending on the proposed
profiles (Fig. 11).

The tank required 1.1–8.8 times higher volume (depending on
the proposed profiles) compare to the tank volume for uniform
profile to keep the same peak power demand as in the case of
the uniform profile (Table 8).

An average hourly DHW ratio were 0.33, 0.30, and 0.33 for WD,
WE and TD respectively. It also noticed significant variation
(0.10–0.39) during a day.
Acknowledgement
K.V. Lindholms Stiftelse foundation, Säätiö L.V.Y. Foundation
and Vercon Oy are greatly acknowledged for grant support and
cooperation in data collection.
References
Abrams, D., Shedd, A., 1996. Effect of seasonal changes in use patterns and cold inlet
water temperature on water-heating loads. ASHRAE Trans. 102, 1038–1041.
Ahmed, K., Pylsy, P., Kurnitski, J., 2015a. Domestic hot water profiles for energy
calculation in Finnish residential buildings. Linz, Austria, Young researcher
International Conference on Energy Efficiency.
Ahmed, K., Pylsy, P., Kurnitski, J., 2015b. Monthly domestic hot water profiles for
energy calculation in Finnish apartment buildings. Energy Build. 97, 77–85.
Ahmed, K., Pylsy, P., Kurnitski, J., 2016. Hourly consumption profiles of domestic hot
water for Finnish apartment buildings. Aalborg, Denmark, CLIMA 2016 -
Proceedings of the 12th REHVA World Congress.
Aydinalp, M., Ugursal, I.V., Fung, S.A., 2004. Modeling of the space and domestic hot-
water heating energy-consumption in the residential sector using neural
networks. Appl. Energy 79, 159–178.
Bagge, H., Johansson, D., 2011. Measurements of household electricity and domestic
hot water use in dwellings and the effect of different monitoring time
resolution. Energy 36, 2943–2951.
Bales, C., Persson, T., 2003. External DHW units for solar combisystems. Sol. Energy
74, 193–204.
Bansal, P., Vineyard, E., Abdelaziz, O., 2011. Advances in household appliances: a
review. Appl. Therm. Eng. 31, 3748–3760.
BBR, 2012. Regelsamling för Byggande BBR-2012, Sweden: s.n.
Becker, B., Stogsdill, K., 1990. Development of a hot water use data base. ASHRAE
Trans. 96, 422–427.
Blokker, J.M.E., Vreeburg, H.G.J., Dijk, C.V.J., 2010. Simulating residential water
demand with a stochastic end-use model. J. Water Resour. Plan. Manage. 136
(1), 19–26.
Buckles, W.E., Klein, S.A., 1980. Analysis of solar domestic hot water heaters. Sol.
Energy 25, 417–424.
Burch, J., Salasovich, J.R., 2002. Flow Rates and Draw Variability in Solar Domestic
Hot Water Usage. Nevada, USA, s.n.
Chow, T.T., Fong, F.K., Lin, Z.C., 2006. Potential application of a centralized solar
water heating system for a high rise residential building in Hong Kong. Appl.
Energy 83, 42–54.
D1, 2007. Suomen Rakentamismääräyskokoelma D1, Helsinki: Kiinteistöjen vesi-ja
viemärilaitteistot Määräykset ja ohjeet.
Dongellini, M., Falcioni, S., Morini, L.G., 2015. Dynamic simulation of solar thermal
collectors for domestic hot water production. Energy Proced. 82, 630–636.
Edwards, S., Beausoleil, M.L., Laperriere, A., 2015. Representative hot water draw
profiles at high temporal resolution for simulating the performance of solar
thermal systems. Sol. Energy 111, 43–52.
Enteria, N. et al., 2014. Case analysis of utilizing alternative energy sources and
technologies for the single family detached house. Sol. Energy 105, 243–263.
Evarts, C.J., Swan, G.L., 2013. Domestic hot water consumption estimates for solar
thermal system sizing. Energy Build. 58, 58–65.
Furbo, S. et al., 2005. Smart solar tanks for small solar domestic hot water systems.
Sol. Energy 78, 269–279.
Geudens, P.J.J.G., 2008. Water Supply Statistics 2007. Vewin - Association of Dutch
Water Companies, Rijswijk, Netherlands.
Good, N., Zhang, L., Espinosa, N.A., Mancarella, P., 2015. High resolution modelling
of multi-energy domestic demand profiles. Appl. Energy 137, 193–210.
Hakala, J., 2014. Energy Consumption in Households 2014. Edita Publishing Oy,
Helsinki.
Hampel, M., et al., 1999. Einfluß des verbrauchsprofils auf den solaren Ertrag
ausgewahlter großerer Solaranlagen. Staffelstein, Germany. In: Proceedings of
9. Symposium Thermische Solarenergie.
Hidalgo, C.R.M. et al., 2012. Domestic hot water consumption vs. solar thermal
energy storage: the optimum size of the storage tank. Appl. Energy 97, 897–906.
Hobbi, A., Siddiqui, K., 2009. Optimal design of a forced circulation solar water
heating system for a residential unit in cold climate using TRNSYS. Sol. Energy
83 (5), 700–714.
IEA, 2007. Energy Conservation in Buildings and Community Systems Programme. s.
l.: IEA International Energy Agency, Annex 42.
Jordan, U., Vajen, K., 2000. Influence of the DHW load profile on the fractional
energy savings: a case study of a solar combi-system with TRNSYS simulations.
Sol. Energy 69 (1–6), 197–208.
Jylhä, K., et al., 2011. Test Reference Year 2012 for Building Energy Demand and
Impacts of Climate Change. s.l.: Finnish Meteorological Institute, Sitra Studies
53.
Kalogirou, S.A., 2000. Long-term performance prediction of forced circulation solar
domestic water heating system using artificial neural networks. Appl. Energy
66, 63–74.
Knudsen, S., 2002. Consumers’ influence on the thermal performance of small
SDHW systems—theoretical investigations. Sol. Energy 73 (1), 33–42.
Koomey, J., Mahler, S., Webber, C., McMahon, J., 1999. Projected regional impacts of
appliance efficiency standards for the US residential sector. Energy 24, 69–84.
Lauterbach, C., Schmitt, B., Vajen, K., 2014. System analysis of a low-temperature
solar process heat system. Sol. Energy 101, 117–130.
Lima, J.B.A., Prado, R.T., Taborianski, V.M., 2006. Optimization of tank and flat-plate
collector of solar water heating system for single-family households to assure
economic efficiency through the TRNSYS program. Renewable Energy 31 (10),
1581–1595.
Meyer, P.J., Tshimankinda, M., 1996. Domestic hot water consumption by
developing communities in South African traditional houses. Energy 21 (12),
1101–1106.
Meyer, P.J., Tshimankinda, M., 1998a. Domestic hot water consumption in South
African apartments. Energy 23 (1), 61–66.
Meyer, P.J., Tshimankinda, M., 1998b. Domestic hot water consumption in South
African townhouses. Energy 39 (7), 679–684.
Paatero, V.J., Lund, D.P., 2006. A model for generating household electricity load
profiles. Int. J. Energy Res. 30, 273–290.
Papakostas, T.K., Papageorgiou, E.N., Sotiropoulo, A.B., 1995. Residential hot water
use patterns in Greece. Sol. Energy 54 (6), 369–374.
Parker, D.S., 2003. Research highlights from a large scale residential monitoring
study in a hot climate. Energy Build. 35 (9), 863–876.
Perlman, M., Mills, B., 1985. Development of residential hot water use patterns.
ASHRAE Trans. 91, 657–679.
Pihlakivi, E., 2015. Potential of solar energy in Finland. Turku: Bachelor Thesis.
prEN-12977, 1997. Preliminary European Standard. Thermal solar systems and
components: Custom built systems. Test methods., s.l.: s.n.
Sahlin, P. et al., 2004. Whole-building simulation with symbolic DAE equations and
general purpose solvers. Build. Environ., 949–958
Saidur, R., Masjuki, H.H., Jamaluddin, M.Y., 2007. An application of energy and
exergy analysis in residential sector of Malaysia. Energy Policy 35 (2), 1050–
1063.
Schipper, L., 1982. Progress in more Efficient Energy Use: The Importance of
Keeping Score. Washington: Annual meeting of the AAAS.
Simonson, C., 2005. Energy consumption and ventilation performance of a naturally
ventilated ecological house in a cold climate. Energy Build., 23–35
Spur, R., Fiala, D., Nevrala, D., 2006. Influence of the domestic hot-water daily draw-
off profile on the performance of a hot-water store. Appl. Energy 83, 749–773.
Sterling, J.S., Collins, R.M., 2012. Feasibility analysis of an indirect heat pump
assisted solar domestic hot water system. Appl. Energy 93, 11–17.
Stokes, M., Rylatt, M., Lomas, K., 2004. A simple model of domestic lighting demand.
Energy Build. 36 (2), 103–116.
Tso, K.F.G., Yau, K.W.K., 2003. A study of domestic energy usage patterns in Hong
Kong. Energy 28, 1671–1682.
Vine, E., Diamond, R., Szydlowski, R., 1987. Domestic hot water consumption in four
low-income apartment buildings. Energy 12 (6), 459–467.
Wolf, D., Tamir, A., Kudish, A.I., 1980. A central solar domestic hot water system
performance and economic analysis. Energy 5, 191–205.