No abstract

Whey is an abundant effluent in the production of cheese and casein. The biotechnological utilization of this economically important and nutritive source is limited mainly because of the presence of high percentages of lactose. This disaccharide has poor solubility, which can cause crystallization and insufficient sweetness in dairy food; additionally, part of the adult population suffers from associated lactose intolerance diseases. There are several methods to determine lactose such as spectrophotometry, polarimetry, infrared spectroscopy, titrimetry and chromatography. However these methods are tedious and time-consuming due to long sample preparation. These disadvantages stimulated the development of an enzymatic lactose biosensor. It employs two immobilized enzymes, beta-galactosidase and glucose oxidase and the quantitative analysis of lactose is based on determination of oxygen consumption in the enzymatic reaction. The influence of temperature on the biosensor signal was experimentally studied. It was observed that a nonlinear relationship exists between the electric response of the biosensor - provided by CAFCA (Computer Assisted Flow Control & Analysis - ANASYSCON, Hannover) - and lactose concentration. In this work, attempts were made to correlate these variables using a simple nonlinear model and multilayered neural networks, with the latter providing the best modeling of the experimental data

No abstract

Industrial archived process data represent a convenient source of information for data-driven models, such as artificial neural network (ANN), that can be used for safety and efficiency improvement like early or even predictive fault detection and diagnosis (FDD). Nonetheless, most of the data used for model generation are representative of the process nominal states and therefore are not enough for classification problems intended to determine abnormal process conditions. This work proposes the use of techniques to augment the original real data standards, dismissing the need for experiments that could jeopardize process safety. It uses the Monte Carlo technique to artificially increase the number of model inputs coupled to the nearest neighbor search (NNS) by geometric distances to consistently classify the generated patterns in normal or faulty statuses. Finally, a radial basis function neural network is trained with the augmented data. The methodology was validated by a study case in which 3381 pulp and paper industrial data points were expanded to monitor the formation of particles in a recovery boiler. Only 5.8% of the original process data were examples of faulty conditions, but the new expanded and balanced data collection leveraged the classification performance of the neural network, allowing its future use for monitoring purpose.

-The main objective of the present work is to analyze the influence of some important operational reaction parameters (agitation speed, polybutadiene -PB -content and initiator concentration) on the final properties of High Impact Polystyrene (HIPS) produced in bulk. Variable effects are analyzed both qualitatively and quantitatively with the help of a fractional factorial design. Physical, chemical and mechanical properties were evaluated through measurement of the weight-average molecular weight (M w ), polydispersity (PD), volume-average diameter of PB particles (D(4,3)) and impact strength (Izod). It was found that PD and D(4,3) depend strongly on the initiator concentration, rubber concentration and agitation speed; Mw depends on initiator and rubber concentrations; and Izod depends on the rubber concentration, PD and D(4,3) in the analyzed experimental range. As a consequence, it was shown that control of final polymer properties can be easily performed through proper manipulation of the analyzed operational variables.