Berlin 2012 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

MM: Fachverband Metall- und Materialphysik

MM 56: Biomaterials

MM 56.4: Talk

Thursday, March 29, 2012, 16:30–16:45, H 0106

How to detect protein fouling and cleaning in heat exchangers using ultrasonic based measurements and classification methods — •Eva Wallhäußer, Mohamed Hussein, and Thomas Becker — TU München. Brau- und Getränketechnologie, (Bio-)Prozesstechnik und Prozessanalyse, Weihenstephaner Steig 21, 85354 Freising

Fouling and cleaning are severe issues in industries using heat exchangers. In particular food and pharmaceutical industry is concerned. Due to fouling unsupervised cleaning takes place which leads to high production costs. A setup was developed to produce and measure dairy protein fouling under realistic conditions. This setup was characterised theoretical using wave equations, port theory, and sensitivity analysis. Also, experimental validation was done. Dairy protein fouling was produced under flow conditions and it was shown that it is possible to detect its presence and absence using an ultrasonic measuring method. For this, an algorithm was developed which is based on a combination of acoustic and signal parameters which are fed into classification methods like artificial neural network (ANN) and support vector machine (SVM). Offline, detection accuracy of the ANN was 97.80 % and of the SVM 97.38 %. The code was then adapted and implemented for online analysis using an ANN as classification method. First online tests for cleaning succes were done.