Last update:
10/28/2004
 |
Neural Networks for Membrane Process Analysis |
Deterministic modeling of membrane desalination is often complicated when membrane fouling occurs due to organics, colloids, microorganisms and mineral salts. The development of membrane foulant layer and the impact on membrane performance are processes that at present cannot be adequately described by deterministic models. Recent, advances in neural network system design have made it possible to map complex process behavior, given sufficient operational data, and to be able to make generalizations regarding the process and even assess potential process upsets. Although, there have been a number of scoping studies on the application of neural networks to membrane process modeling, our group is in a unique position having developed, in collaboration with the University of Tarragona (Spain), a novel superior neural network approach for modeling complex systems.
Our group is developing a comprehensive cognitive neural networks model, based on pilot plant data, to analyze and optimize membrane system performance. It is expected that the analysis tool developed using the present approach will have a positive impact on the rapidly expanding area of membrane desalination technology.
