engleski

Development of Fuzzy Neural Network Methodology for Improvement of Water Quality Analytical Processes

Water quality management processes and in particular water quality measurements might be considered routine but most certainly are economically and ecologically demanding and time consuming. Constant improvement paradigm yield multidimensional complex water quality process optimization that often results in insufficient improvements. Advanced fuzzy neural solutions can be implemented to overcome such results. The aim of this work is development of fuzzy neural network methodology for modelling and optimization of carbohydrate monitoring in water systems. The networks were optimized by means of training algorithm and, in addition, quantitative structure retention relationship models were developed, enabling prediction of retention parameters for other carbohydrate compounds in analytical system. The results were validated using external carbohydrate set. The obtained prediction ability showed a potential of the applied methodology for improving water quality analytical processes.

ion chromatography ; pulsed amperometric detection ; retention modelling ; fuzzy neural network ; quantitative structure-retention relationships

nije evidentirano