Feasibility of EMF-induced pipeline wall friction reduction by PLS2 intercalibration of acoustic chemometrics and reference laser velocimetry

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.