Labview Sensors

19th European Symposium on Computer Aided Process Engineering – ESCAPE19 J. Je owski and J. Thullie (Editors) © 2009 Elsevier B.V. All rights reserved.

309

A LabVIEW-based intelligent system for monitoring of bioprocesses
Elmer Ccopa Rivera, Félix de Farias Junior, Daniel Ibraim Pires Atala, Rafael Ramos de Andrade, Aline Carvalho da Costa, Rubens Maciel Filho
School of ChemicalEngineering, State University of Campinas, P.O. Box 6066, 13081970, Campinas, SP, Brazil, elmer@feq.unicamp.br, maciel@feq.unicamp.br

Abstract
The application presented in this study illustrates the usefulness of an automated monitoring system carried out in LabVIEW environment. The results obtained have show that it is possible to infer into a real-time basis the key variables in bioethanolfermentation using pH, turbidity, CO2 flow rate and temperature on line measurements and a MLP-based Software Sensor. On-line monitoring system provided accurate online predictions of the concentrations during the fermentation process even when the secondary variables measurements were noisy. This study also will illustrate the usefulness of an automated monitoring system carried out in graphicalprogramming environment. Keywords: Modeling, bioreactor, software sensor, artificial intelligence.

1. Introduction
Artificial intelligence, such as Artificial Neural Network (ANN) has been used successfully for solving biotechnological complex problems related to the field of measurements and instrumentation. Probably, the most popular ANN used in engineering applications is the Multilayer PerceptronNeural Network (MLP) due to its easily understandable architecture and a simple mathematical form, which results in an easy tool for modeling and implementation. In addition, it is known from previous works that MLPs can be used to offer adaptive solutions, since the reestimation of their parameters is a straightforward procedure [1]. By exploiting the relationship among the process variables ofbioprocesses, MLPs could be used to implement advanced control techniques, such as software sensors, algorithms for on-line estimation of state variables and model parameters that are not measurable in real-time [2, 3]. This is a promising research area with significant impact on biotechnological industry, which requires an efficient monitoring with reliable sensors to control setting of the process.Thus, for a reliable performance prediction through modeling, the MLP-based software sensor should be implemented on a platform able to provide a powerful toolset for process identification and control with interface directly to instruments, sensor and actuators. A programme to fully automate the implementation of software sensors can be developed using a graphical programming environment, calledLabVIEW (Laboratory Virtual Instrument Engineering Workbench). The LabVIEW has extensive library of functions and subroutines for most programming task. It also contains an application specific library for data acquisition, serial instrument control, data processing, analysis presentation and storage. Applications created with LabVIEW are referred to as virtual instruments (VIs) [4, 5].

310E.C. Rivera et al.

In this study, it was used built-in LabVIEW function and library VIs to develop a software sensor based on MLPs for on-line monitoring of bioethanol production process. This program can be adapted to a wide range of instrumentation control and optimization applications. The proposed monitoring system is based on LabVIEW as driver software, the data acquisition system andthe sensors for the secondary measurements (pH, turbidity, CO2 flow rate and temperature). The system collects the variables of the fermentation process by means of sensors. The measured values are delivered to the computer program through the data acquisition system for data processing and prediction of state variables (concentrations of biomass, product and substrate) using the MLP-based...