Use of a spectroscopic sensor to monitor droplet size distribution in emulsions using neural networks.

Monitoring of emulsion properties is important in many applications, like in foods and pharmaceutical products, or in emulsion polymerisation processes, since aged and 'broken' emulsions perform worse and may affect product quality. This study reports the use of an 'in-line' turbidity sensor coupled with a neural network model to monitor droplet size distributions of metal working fluid emulsions (MWF), a case where emulsion aging affects product quality. The data from the sensor were used to fit the model for droplet size distribution estimation. The technique was applied to monitor the destabilisation of commercially available MWF with good accuracy. [ABSTRACT FROM AUTHOR]