427-8/2025
www.
inf
o
rmacjain
s
tal
.
c
o
m
.
pl
Z
Than
k
s
t
o
the re
s
earch, accurate inf
o
rmati
o
n
o
n
w
ater c
o
n
s
umpti
o
n
w
a
s
o
btained,
w
hich i
s
a valuable
s
o
urce
o
f
k
n
ow
ledge f
o
r
s
cienti
s
t
s
,
de
s
igner
s
and engineer
s
managing the
w
ater
s
uppl
y
net
wo
r
k.
Thi
s
re
s
earch indicated the
difference
s
bet
w
een the data
o
btained during
mea
s
urement
s
and the literature value
s
.
Similar re
s
earch
w
a
s
carried
o
ut in
wo
r
k
b
y
G
wo
ź
d
z
iej-
M
a
z
ur
[
5
]
,
w
hich e
x
amined the
w
ater meter
s
i
z
e
s
electi
o
n b
y
determining the
di
s
tributi
o
n
o
f characteri
s
tic fl
ow
s
in the f
o
rm
o
f
the minimum and ma
x
imum v
o
lume fl
ow
o
f
w
ater at the
w
ater
s
uppl
y
c
o
nnecti
o
n during
24-h
o
ur and an attempt t
o
determine the num-
ber
o
f apparent l
o
ss
e
s
.
A
n ultra
s
o
nic
w
ater
meter
w
ith the highe
s
t mea
s
urement accurac
y
o
f
R
800
w
a
s
u
s
ed f
o
r the te
s
t
s
.
D
uring the
wo
r
k
, the ma
x
imum, minimum and average
characteri
s
tic fl
ow
s
at the
w
ater
s
uppl
y
c
o
n-
necti
o
n at
s
pecific da
y
h
o
ur
s
w
ere calculated
f
o
r t
wo
s
elected multi-unit building
s
.
The anal
y
-
s
i
s
pr
o
ved that the
s
hare
o
f minimum fl
ow
s
in all
w
ater c
o
n
s
umpti
o
n i
s
50
%
.
Theref
o
re, it i
s
rea-
s
o
nable f
o
r the
ow
ner
s
o
f
w
ater meter
s
, i
.
e
.
w
ater
s
uppl
y
c
o
mpanie
s
, t
o
utili
z
e
w
ater me-
ter
s
w
ith
s
i
z
eable metr
o
l
o
gical accurac
y
due t
o
the p
o
tential apparent
w
ater l
o
ss
e
s
.
D
a Silva et al
.
carried
o
ut a c
o
mprehen-
s
ive anal
y
s
i
s
o
f
w
ater c
o
n
s
umpti
o
n
.
Thi
s
wo
r
k
e
x
amined
w
ater c
o
n
s
umpti
o
n rec
o
rded in
78 building
s
f
o
r 25 m
o
nth
s
[
1
].
A
t that time,
s
i
x
m
o
del
s
w
ere devel
o
ped de
s
cribing the
m
o
nthl
y
w
ater c
o
n
s
umpti
o
n in the building,
the be
s
t
o
ne
w
ith an accurac
y
o
f 91
%
.
Than
k
s
t
o
thi
s
wo
r
k
, parameter
s
affecting the am
o
unt
o
f
w
ater c
o
n
s
umpti
o
n in the building
w
ere
determined, the m
o
s
t imp
o
rtant
o
f
w
hich i
s
the age
o
f the building, number
o
f fl
oo
r
s
,
area per fl
oo
r, area per per
s
o
n, number
o
f
inhabitant
s
per r
oo
m and apartment, number
o
f bedr
oo
m
s
, and t
y
pe
o
f
s
y
s
tem rec
o
rding
c
o
n
s
umpti
o
n
.
Thi
s
s
tud
y
highlight
s
the c
o
m-
ple
x
it
y
o
f the pr
o
blem
o
f de
s
cribing and
f
o
reca
s
ting
w
ater u
s
e
.
Similar re
s
earch
o
n mete
o
r
o
l
o
gical fact
o
r
s
w
a
s
pre
s
ented in Nia
z
mardi et al
.
[
9
].
The
wo
r
k
anal
y
s
ed data fr
o
m 5
y
ear
s
o
btained
fr
o
m mea
s
urement
s
in 566 re
s
idential build-
ing
s
u
s
ing the
B
a
y
e
s
ian
Es
timat
o
r
o
f
A
brupt
Change, Sea
s
o
nal, and Trend (
BE
A
ST) alg
o
-
rithm
.
I
n additi
o
n t
o
s
imultane
o
u
s
l
y
e
x
amining
mete
o
r
o
l
o
gical fact
o
r
s
, the l
o
cati
o
n
o
f mea
s
-
urement p
o
int
s
(building
s
)
w
a
s
al
s
o
c
o
n
s
id-
ered
.
The re
s
ult
s
did n
o
t indicate an
y
s
patial
pattern
s
influencing
w
ater u
s
e
.
There
w
ere al
s
o
n
o
change
s
in the v
o
lume
o
f
w
ater c
o
n
s
ump-
ti
o
n, e
x
cluding the initial peri
o
d
o
f the C
O
V-
I
D
-19 pandemic
.
The fact
o
r
s
that influence
s
ea
s
o
nal
w
ater c
o
n
s
umpti
o
n are temperature,
average
w
ind
s
peed and air humidit
y.
A
n imp
o
rtant a
s
pect that i
s
c
o
n
s
tantl
y
devel
o
ping i
s
f
o
reca
s
ting
w
ater demand
.
I
n
2019, a c
o
mprehen
s
ive cr
o
ss
-
s
ecti
o
nal
wo
r
k
w
a
s
publi
s
hed, illu
s
trating the re
s
earch direc-
ti
o
n
s
related t
o
thi
s
i
ss
ue
[
8
].
The article brief-
l
y
di
s
cu
ss
e
s
re
s
earch
wo
r
k
s
u
s
ing artificial
intelligence (
A
I
) m
o
del
s
and related alg
o
-
rithm
s
,
s
uch a
s
vari
o
u
s
t
y
pe
s
o
f neural net-
wo
r
k
s
, regre
ss
i
o
n m
o
del
s
, genetic alg
o
-
rithm
s
, fu
zz
y
l
o
gic,
s
upp
o
rt vect
o
r machine
s
,
etc
.
The auth
o
r
s
p
o
int
o
ut that currentl
y
,
o
ne
o
f the alg
o
rithm
s
cann
o
t be
s
aid t
o
be m
o
re
accurate, but the
y
o
utperf
o
rm
s
tati
s
tical m
o
d-
el
s
in their adaptabilit
y.
Neverthele
ss
, re-
s
earch
wo
r
k
require
s
a br
o
ader de
s
cripti
o
n
o
f the m
o
del
s
u
s
ed, a
s
w
ell a
s
c
o
n
s
idering
a m
o
re
s
ignificant number
o
f fact
o
r
s
,
s
uch a
s
humidit
y
, pre
ss
ure
o
r temperature,
w
hich
w
ill
s
ignificantl
y
increa
s
e the c
o
mple
x
it
y
o
f the
pr
o
blem but
w
ill al
s
o
c
o
ntribute t
o
a greater
s
electi
o
n
o
f re
s
earch meth
o
d
s
.
O
ne e
x
ample
o
f the u
s
e
o
f
A
I
in detect-
ing
w
ater c
o
n
s
umpti
o
n pattern
s
i
s
the
wo
r
k
o
f
Jaquim Leitã
o
et al
.
[
6
]
,
w
hich e
x
amined the
p
o
ss
ibilitie
s
o
f anal
y
s
ing
w
ater demand b
y
clu
s
tering time
s
erie
s
.
The c
o
nducted re
s
earch
indicated that alth
o
ugh the accurac
y
o
f the
clu
s
tering perf
o
rmed depend
s
o
n the target
number
o
f clu
s
ter
s
and the ad
o
pted data
prepr
o
ce
ss
ing meth
o
d
s
, data clu
s
tering al-
l
ow
s
f
o
r a pict
o
rial pre
s
entati
o
n
o
f the nature
o
f dail
y
w
ater c
o
n
s
umpti
o
n than
k
s
t
o
the al-
g
o
rithm
s
u
s
ed,
w
hich all
ow
s
f
o
r fa
s
ter detec-
ti
o
n and de
s
cripti
o
n
o
f e
x
i
s
ting pattern
s
ba
s
ed
o
n
s
cientific meth
o
d
s
.
A
n e
x
ample
o
f re
s
earch dealing
w
ith the
i
ss
ue
o
f
s
tati
s
tical re
s
earch i
s
the
wo
r
k
o
f Wa
w
-
r
z
o
s
e
k
et al
.
[
12
].
The
wo
r
k
pr
o
p
o
s
ed a
s
tati
s
-
tical inference meth
o
d
o
l
o
g
y
u
s
ing
w
hich an
anal
y
s
i
s
o
f
w
ater c
o
n
s
umpti
o
n
w
a
s
carried
o
ut
f
o
r data
o
btained fr
o
m rand
o
m mea
s
urement
peri
o
d
s
,
w
here there
w
a
s
n
o
s
ignificant
am
o
unt
o
f accurate data, and the mea
s
ure-
ment
s
w
ere n
o
t carried
o
ut
s
imultane
o
u
s
l
y.
A
s
a re
s
ult
o
f the calculati
o
n
s
, it
w
a
s
f
o
und that
l
o
ng-term mea
s
urement
s
o
f the accumulated
value
s
o
f m
o
nthl
y
di
s
tributi
o
n
s
can pr
o
vide
a l
o
t
o
f inf
o
rmati
o
n if the
y
are
s
upplemented
w
ith a certain number
o
f
w
ee
k
l
y
and dail
y
mea
s
urement
s
,
w
ith even a much
s
h
o
rter
mea
s
urement peri
o
d
.
The ab
o
ve e
x
ample
s
c
o
nfirm that
w
ater partiti
o
n
s
and their nature
and anal
y
s
i
s
are c
o
mple
x
i
ss
ue
s
that
w
ater
c
o
mpanie
s
can u
s
e t
o
impr
o
ve the functi
o
ning
o
f the
w
ater
s
uppl
y
net
wo
r
k.
C
o
n
s
idering climate change and pr
o
-
gre
ss
ive urbani
z
ati
o
n,
w
hich negativel
y
af-
fect the
s
ecurit
y
and c
o
ntinuit
y
o
f
w
ater
s
up-
plie
s
[
7
]
, increa
s
ing
k
n
ow
ledge ab
o
ut the
w
ater
s
uppl
y
net
wo
r
k
and the recipient
s
w
h
o
u
s
e it ma
y
be nece
ss
ar
y
f
o
r the management
and
o
perati
o
n
o
f the net
wo
r
k.
Thi
s
nece
ss
it
y
re
s
ult
s
n
o
t
o
nl
y
fr
o
m the
s
cientific re
s
earch
cited but, in man
y
place
s
, al
s
o
fr
o
m applica-
ble regulati
o
n
s
.
The relativel
y
recentl
y
intr
o
-
duced
D
irective (
E
U) 2020/2184
o
f the
E
u-
r
o
pean Parliament and the C
o
uncil
o
n the
qualit
y
o
f
w
ater intended f
o
r human c
o
n-
s
umpti
o
n aim
s
t
o
limit
w
ater l
o
ss
e
s
cau
s
ed b
y
lea
k
s
[
2
].
A
m
o
re detailed under
s
tanding
o
f
the
w
ater
s
uppl
y
net
wo
r
k
b
y
e
x
amining the
w
ater c
o
n
s
umpti
o
n
o
f vari
o
u
s
entitie
s
w
ill
c
o
ntribute t
o
meeting thi
s
target
.
O
ne
o
f the m
o
s
t acce
ss
ible
s
o
luti
o
n
s
i
s
t
o
devel
o
p fundamental
w
ater di
s
tributi
o
n pat-
tern
s
f
o
r individual gr
o
up
s
o
f c
o
n
s
umer
s
, a
s
w
a
s
d
o
ne b
y
D
z
imiń
s
k
a et al
.[
4
].
F
o
r e
x
ample,
a
s
imilar anal
y
s
i
s
can be carried
o
ut f
o
r a
s
tu-
dent d
o
rmit
o
r
y
, c
o
mm
o
nl
y
k
n
ow
n a
s
an
“Ak
ademi
k”
in P
o
li
s
h
.
A
s
tudent d
o
rmit
o
r
y
ma
y
differ fr
o
m a t
y
pical re
s
idential facilit
y
a
s
s
tu-
dent
s
have a
s
lightl
y
different circadian rh
y
thm
[
11
]
,
w
hich i
s
reflected in
w
ater demand
.
M
a
t
e
ri
a
l
s
a
n
d
me
t
ho
d
s
a
the re
s
earch included five
s
tudent d
o
rmi-
t
o
rie
s
in Wr
o
cła
w
,
s
upplied
w
ith
w
ater
thr
o
ugh t
wo
metered
w
ater c
o
nnecti
o
n
s
.
The
s
tud
y
aimed t
o
devel
o
p a dail
y
w
ater di
s
tri-
buti
o
n pattern
w
ith a 1-h
o
ur interval f
o
r the
s
tudent h
o
u
s
e t
o
anal
y
s
e and de
s
cribe it
.
Water demand pattern
s
w
ere al
s
o
created
and c
o
mpared f
o
r a
wo
r
k
ing da
y
and a da
y
o
ff (
w
ee
k
end)
.
D
ata f
o
r the re
s
earch
w
ere
o
btained fr
o
m central
w
ater meter
s
bel
o
ng-
ing t
o
the
M
unicipal Water and Se
w
erage
C
o
mpan
y
(
M
PWiK S
.A.
) in Wr
o
cła
w.
The data anal
y
s
i
s
included
:
l
D
etermining the accurac
y
o
f the reading
s
–
verif
y
ing
w
hether there
w
ere an
y
inter-
rupti
o
n
s
in the mea
s
urement
s
.
l
Calculati
o
n
o
f h
o
url
y
w
ater c
o
n
s
umpti
o
n
f
o
r b
o
th c
o
nnecti
o
n
s
,
l
Summari
z
ing the
o
btained h
o
url
y
data
fr
o
m b
o
th c
o
nnecti
o
n
s
t
o
o
btain a
s
ingle
time
s
erie
s
f
o
r the te
s
ted
o
bject
s
,
l
C
o
nver
s
i
o
n
o
f h
o
url
y
data t
o
dail
y
data
f
o
r detecti
o
n and rem
o
val
o
f
o
utlier
o
b-
s
ervati
o
n
s
,
l
D
evel
o
pment
o
f a
w
ee
k
l
y
c
o
n
s
umpti
o
n
pattern
w
ith a 1-h
o
ur interval,
l
D
evel
o
pment
o
f a dail
y
di
s
tributi
o
n pat-
tern
w
ith a 1-h
o
ur interval f
o
r a
wo
r
k
ing
da
y
and a da
y
o
ff (
w
ee
k
end)
.
Calculati
o
n
s
, anal
y
s
i
s
and data vi
s
uali-
z
ati
o
n
w
ere perf
o
rmed in
Ex
cel
.
T
o
c
o
nduct
a
s
tati
s
tical te
s
t
o
f mean
s
, it
w
a
s
a
ss
umed that
the di
s
tributi
o
n
s
o
f b
o
th data
s
et
s
w
ere un-
k
n
ow
n
.
The number
o
f element
s
in the data-
s
et
s
w
a
s
s
ufficientl
y
large t
o
perf
o
rm the te
s
t
.
B
y
the
s
e a
ss
umpti
o
n
s
, n
o
rmal di
s
tributi
o
n ta-
ble
s
w
ere u
s
ed t
o
read the critical value u
and the f
o
ll
ow
ing f
o
rmula
[
3
]
w
a
s
empl
oy
ed
f
o
r the te
s
t
s
tati
s
tic
s
:
(2
.
1)
