Acoustic chemometrics on the liquid flow in pipe

In-line measurements of liquid quality parameters by acoustic chemometrics.

Acoustic chemometrics is a new in-line measurement method, which can be used to monitor processes generating sound/vibrations (Esbensen et al. 1998). The principle of acoustic chemometrics is as follows; obtaining acoustic signals induced by the process, followed by signal transformation and chemometric data analysis.

The acoustic emissions from the liquid flow in the pipe can be obtained by introducing a pressure loss element to the flow, e.g. by installing a valve or an orifice plate. Figure 1 illustrates the experimental set-up for acoustic chemometrics on liquid flow using orifice plate. The pressure drop that develops over the constriction leads to vibration and sound. The frequency of the sound is related to the physical and chemical properties of the liquid. Example of the passive acoustic spectra is presented on Figure 2. The method is rapid, non-invasive and simple. Main advantage of this method compared with other in-line methods for process monitoring is that it uses so-called “clamp-on” accelerometers to register the vibration. These sensors can easily be installed almost anywhere on the pipe wall. In many down- or upstream processes it is preferred to use non-invasive sensors, because invasive sensors may cause disturbances like fouling or film layers inside the process equipment. From an industrial point of view it is also positive that the sensor can withstand high temperatures, dirty environments and is easy to maintain and clean. The fact that almost all processes produce some kind of acoustic emission opens up for many potential applications. The only requirement is that the acoustic emission must contain relevant information, which can be correlated with the parameters of interest.

Our research on the acoustic chemometrics led to a novel method for multivariate regression using frequency shift in the passive acoustic spectra. The new method led to lower prediction errors and more parsimonious models compared to the regression on the complete spectra. We also successfully related the acoustic data to the physical theory on vibration of pipe and bubble oscillation in the liquid flow.



Figure 1. Experimental set-up. From left to right: The measurement section of the equipment; magnified pipe with the mounted accelerometer; and the pipe where the accelerometer was mounted, (a,b) the pipe, (c) rubber ring, (d) orifice plate and (e) metric ruler.



Figure 2. Passive acoustic emission spectra of the ethanol-water mixtures flow.

Involved people
Tomas Isaksson
Elling-Olav Rukke
Reidar Barfod Schüller
Andriy Kupyna

Financial support
The project is funded by the Norwegian University of Life Sciences

Related publications1-3
1. Kupyna A, Rukke E-O, Schüller RB,Isaksson T. Passive acoustic measurements for monitoring quality characteristics of liquid foods. In NOVEM 2005: Noise and Vibration: Emerging Methods International Congress. Saint Raphaël, France, 2005
2. Kupyna A, Rukke E-O, Schüller RB, Helland H,Isaksson T. Partial least square regression on frequency shift applied to passive acoustic emission spectra. Journal of Chemometrics 2007; in press.
3. Kupyna A, Schüller RB, Rukke E-O,Isaksson T. Acoustic chemometrics on liquid flow: Shift in the frequency spectra and its relationship to the physical properties of the liquid and the pipe. Chemometrics and Intelligent Laboratory Systems 2007; submitted.

Back      Print page    Home