Biorefinaria de soro de queijo: engenharia de bioprocessos e sistemas aplicada à transformação de um resíduo poluente em produtos com valor agregado
Abstract
The continuous advances in process computing have provided in recent years several sophisticated tools for process analysis and simulation. The effective use of these tools
demands the capability to interface with a pre-existent process system hierarchy. Within this perspective, an important issue is to provide meaningful and useful simulation and optimization applications for complex systems that require integration with data-intensive experimentation. This study proposes an integrated environment, using Internet as development platform, for simulation, monitoring, control and optimization of a cheese whey refinery, employing immobilized and stabilized enzymes as catalysts. The multipurpose process
described here, the cheese whey biorefinery, provides, besides lactose, whey protein concentrates and hydrolysates that can be applied in food and pharmaceutical formulae. In
these circumstances, it is possible to add value to this significant by-product of the dairy industry, avoiding its disposal in natura (what is mostly done by small cheese manufacturers). In parallel, relevant aspects of the enzymatic reactions within the cheese whey biorefinery were investigated. The lumping of substrate molecules in pseudo-components, using a hybrid phenomenological-neural approach for description of the enzymatic depolymerization
kinetics, is the suggestion to follow the complex reaction dynamics in biorefinery. During the estimative of model parameters, global search algorithms and different post-fitting statistical strategies were implemented and evaluated.
The work, thus, is concerned with two of the main dilemmas/challenges of the present millennium: reduction of environmental problems related to the disposal of agro-industrial residues (i.e., cheese whey) and development of processes for food production from alternative sources (whey refinery). The integrated web application here presented may allow small companies to access a remote engineering centre , with know-how on plant design, economic evaluation and advanced control/optimization techniques. The idea can also be extended to large dairy companies, providing the remote control of sites of production geographically sparse.
All implemented algorithms for remote process monitoring and control were validated with data from laboratory-scale assays. The validation of the hydrolytic kinetic models
followed the same procedure. According to the achieved results, is possible to conclude that the robustness of the integrated computational environment was demonstrated. The prediction capability of the approaches employed for description of the proteolytic enzymatic reactions was verified, as well.