Arcadio Sotto

Nanofiltration removal of pharmaceutically active compounds

Abstract Capability of nanofiltration membranes (NF) to remove pharmaceutical active compounds from wastewater streams was investigated. Sulfamethoxazole, diclofenac sodium, hydrochlorothiazide, 4-acetamidoantipyrine, nicotine and ranitinide hydrochloride were selected as model compounds since they are widely produced as pharmaceutical agents. Two commercially available polyamide nanofiltration membranes (NF-90 and NF-270 from Dow FilmTec) were tested. Solute retention by NF-90 membrane was very high in all cases (over 95%), whereas NF-270 retention systematically appeared lower ranging from 75% (nicotine) to 95% (ranitidine hydrochloride). Temporal evolution of flux decline was also investigated. The influence of physicochemical properties of both membrane and solutes on membrane performance was analyzed to explore the main solute-membrane interactions that determine the solute transfer across the membrane. The influence of operation pressure on NF-90 and NF-270 rejection was also studied.

Prediction of the Amount of Wood Using Neural Networks

Coniferous trees such as eucalyptus used to be preferred for papermaking because the cellulose fiber in the pulp of these species are longer, therefore making for stronger paper. In this study, the proposed neural network method solves in an efficient way, how to build prediction models in engineering. The system has been applied to predict amount of wood for production of paper, in which the coefficients can explain the variable with more influence over the variable to forecast. Obtaining a good prediction and as simple as possible, i.e. with the least number of forecast variables.

Data Mining with Enhanced Neural Networks-CMMSE

This paper presents a new method to extract knowledge from existing data sets, that is, to extract symbolic rules using the weights of an Artificial Neural Network. The method has been applied to a neural network with special architecture named Enhanced Neural Network (ENN). This architecture improves the results that have been obtained with multilayer perceptron (MLP). The relationship among the knowledge stored in the weights, the performance of the network and the new implemented algorithm to acquire rules from the weights is explained. The method itself gives a model to follow in the knowledge acquisition with ENN.