
where:
F
j
is the probability distribution at time t ¼ j t,
Step-up case F
j
¼ 1
Cð t
0
þ jtÞ
Cð1Þ
Decay case F
j
¼
Cðt
0
þ j tÞ
Cð t
0
Þ
ð3:16Þ
N is the last measurement integrated using the trapezium method,
"
n
ðN;
d
) is the rest of the integral, evaluated using an exponent ial fit on the last
measurements (see below).
The number of measurements, N , could be large enough to ensure that the
sum of the terms for j > N are negligible, or, in other words, that C
N
is very
close to the steady-state value. In this case, the remaining parts, "
n
ðN;
d
),
are negligible. In practice, however, the measurement can be stopped before
reaching the steady state. In this case, the tail in the integral of the moments
is not measured, but is estimated. As shown in Figure 3.5, this tail is, in
most cases, exponential of the form:
Step-up: CðtÞ¼C
1
ð1 e
t=
c
Þ
Decay: CðtÞ¼Cð0Þe
t=
c
ð3:17Þ
Therefore, for time larger than t
N
¼ N t, it can be assumed that:
Fðt > t
N
Þ¼F
N
exp
t
N
t
d
ð3:18Þ
where
d
is a time constant determined by a fit on the last measurements in
the exponential part. In this case, the remaining part, "
n
ðN;
d
Þ, of the
moments are:
"
0
ðN;
d
Þ¼
ð
1
t
N
F
e
ðtÞdt ¼ F
N
ð
1
t
N
exp
t
N
t
d
dt ¼ F
N
d
"
1
ðN;
d
Þ¼
ð
1
t
N
tF
e
ðtÞdt ¼ F
N
d
ðt
N
þ
d
Þ
ð3:19Þ
The time required for reach ing an exponential decay depends not only on the
nominal time constant of the room, but also on the ventilation system. The
decay will be exponential from the beginning of the test where complete
mixing occ urs. In case of displacement ventilation, the decay should be very
sharp after a time equal to the age of air.
Error analysis
Using, mutatis mutandis Equation 2.22, the confidence interval of Fð) is:
bF
j
F
j
; F
j
þ F
j
c
46 Ventilation and Airflow in Buildings
Commentaires sur ces manuels