Fathom QiQuac M5, Руководство по эксплуатации

Страница: 14 / 33

14
“There is an art, it says, or rather, a knack to flying. The knack lies in learning
how to throw yourself at the ground and miss.” (Adams D. 1979)
Uncertainty Example
The following example is taken from a single injection using 1kg in ~1.3 m³/s
of flow and shows the inherent uncertainty in a flow measurement with a nat-
urally variable background. The plots in Figure 1 show three interpretations of
the same measurement in the Salt Portal and the variance in the Q and Uncer-
tainty. The QiQuac reported this measurement as a Grade B measurement and
recommended increasing the salt dosage. Note that the dosage needed to be
increased due to the variation in the BG ECT and not the stability of the sensor
signal
4
, which was actually quite low (0.02%).
Figure 7. a, b & c
Figure showing the same trace in the Salt Portal with 3 interpretations a) 1.32
m³/s ±4.0% assuming a sloping background. This is probably the most ac-
curate interpretation. Note the rise in EC.T is only 2.5 μS/cm. b) The default
interpretation taking the entire trace misses the natural dip in EC.T before the
pulse arrives and therefore underestimates the area. With “Assume Sloping BG
ECT” OFF, this results in a Q of 1.47 m³/s ±12.5%. c) moving the Pre- BG EC.T
sample just before the pulse and extending it out to the start of the rise results
in a Q of 1.22 m³/s ±5.3%. In c) we’ve zoomed in to show the natural BG EC.T
variation.
If we tabulate these results and compare them, only interpretation a) and c) are
significantly different at 95% confidence. Note that the min Q from b) doesn’t
include the max Q from c). Note that b) Uncertainty is quite high due to the
length of the pulse AND the inherent uncertainty between the Pre- and Post-
BG ECT because Assume Sloping BG ECT was OFF.
C:\gsentlin\FATHOM_SCI\PROJECTS\QiQuac\Manuals\01_Aug 2019 Revision\[Uncertainty Example.xlsx]Sheet1 2019-08-21 8:41
Case Q (m
3
/s) Unc. (95%) Max Q (m
3
/s) Min Q (m
3
/s) Sig Diff from a) Sig Diff from b) Sig Diff from c)
a 1.32 4.0% 1.37 1.27 FALSE FALSE
b 1.47 12.5% 1.65 1.29 FALSE TRUE
c 1.22 4.9% 1.28 1.16 FALSE TRUE
Notes
A] The Uncertainty and Significant Difference calculations are based on 95% confidence.
Ver 0.2
From this example, the user can see the value of uncertainty analysis, the influ-
ence of natural BG ECT variability, and the choices that the user must make in
post-processing.
4 The Stability is the standard deviation of the 10 pre and 10 post BG EC.T measurements.
a)
b)
c)
?
C H A P T E R 2 : T H E ORY