Kh. Esbensen et al., Feasibility of EMF-induced pipeline wall friction reduction by PLS2 intercalibration of acoustic chemometrics and reference laser velocimetry, J CHEMOMETR, 15(4), 2001, pp. 241
Various applied physics lines of research on the interactions between elect
romagnetics and flowing water have led to the suggestion that it might be p
ossible to reduce pipeline wall friction by applying a transverse electroma
gnetic field (EMF), We here conduct a feasibility study of the shape and ma
gnitude of internal fluid Row velocity profiles of hydropower turbine inlet
pipelines, with and without an EMF. The Row velocity profile itself can on
ly be measured in transparent pipeline sections with the aid of refined las
er velocimetric equipment, which can never be employed in realistic hydropo
wer settings however. The objective of the present studies is therefore to
use a much simpler alternative, i.e. an acoustic chemometric approach, rely
ing on 'clamp-on' acoustic sensors applied directly to the outer wall of th
e pipeline. Our results show that it is fully possible to verify the 'on/of
f' status of the applied EMF using either of the alternative modalities, la
ser velocimetry and acoustic chemometrics, i.e. to verify an increased wate
r flux when the EMF is 'on'. The interactions between this feature and real
istic changing temperatures (15-27 degreesC) as well as varying overall wat
er velocities (1-4-m s(-1)) are explored. It is concluded that the temperat
ure sensitivities of both modalities can be compensated for and, furthermor
e, that the temperature-compensated modelling is velocity-invariant. This o
pens up an interesting avenue for field on-line EMF-induced wall friction r
eduction monitoring and control. Our final feasibility demonstration concer
ns direct PLS2 intercalibration of the acoustic chemometric monitoring data
(X) with the laser velocimetric reference profiles (Y), This translates in
to a new acoustic chemometric possibility for 'seeing' the effects of the E
MF-modified velocity profiles directly from the suitably calibrated acousti
c signals alone. We find that a 42-Y-variable PLS2 model can be validated a
s X, Y: 30%, 70% respectively, fully satisfactory for prediction purposes o
f the velocity in an outer (near-wall) 5-6 mm annular Row regimen. With thi
s we have positive proof that acoustic chemometric monitoring can be used f
or predicting the effects of the EMF, Upscaling to field-scale pipelines re
mains. Copyright (C) 2001 John Wiley & Sons, Ltd.