Efficient monitoring and automatic control systems for biological wastewater treatment processes, especially those dealing with bioinhibitory pollutants, such as phenol, are urgently required in order to meet increasingly stringent environmental regulations. Practical on-line sensors of variables that describe water quality, such as BOD or individual toxic pollutants such as phenol, are not commercially available; e.g. phenol is generally monitored off-line by spectrofluorometry. Inference software sensors could be an attractive alternative for on-line monitoring of these variables. As a first step towards the development of inferential sensors for biological wastewater treatment processes, we consider in this study, a simplistic version of such a process which consists of a continuous culture of Pseudomonas putida Q5 degrading phenol. In this work, we propose a neural network based inferential sensor for phenol monitoring using on-line biomass concentration measurements by spectrophotometry. The network was built with wavelets as the basis functions and the adaptive algorithm for the weights was based on a Lyapunov stability analysis. Predicted phenol output of the network showed good agreement with experimental data, over fairly broad ranges of inlet phenol concentration and dilution rate step changes. Simulations were conducted to find convergence conditions and to investigate possible sources for errors in phenol estimates.
Les systemes de surveillance efficace et de controle automatique des procedes de traitement des eaux usees biologiques, en particulier ceux ayant affaire a des polluants bio-inhibiteurs tels les phenols, deviennent absolument necessaires en vue de satisfaire les reglements environnementaux rigoureux. Des capteurs de variables montes en ligne pratiques decrivant la qualite de l'eau, comme la DBO ou differents polluants toxiques comme le phenol, ne sont pas disponibles dans le commerce, le phenol etant generalement surveille par des systemes non montes en ligne par la spectrofluorometrie. Les capteurs logiciels a inference pourraient etre une solution attrayante pour la surveillance en ligne de ces variables. Comme premiere etape vers la mise au point de capteurs inferentiels pour des procedes de traitement des eaux usees biologiques, nous considerons dans cette etude une version simplifiee d'un tel procede qui consiste en une culture continue de Pseudomonas putida Q5 degradant le phenol. Dans ce travail, nous proposons un capteur inferentiel base sur un reseau neuronal pour la surveillance du phenol utilisant des mesures de concentration de biomasse en ligne par spectrophotometrie. Le reseau a ete bâti avec des ondelettes comme fonctions de base, et l'algorithme adaptatif pour les ponderations a ete base sur une analyse de stabilite de Lyapunov. La reponse en phenol predit par le reseau montre un bon accord avec les donnees experimentales, dans des gammes assez larges de concentration de phenol en entree et de modifications en echelon de la vitesse de dilution. Des simulations ont ete effectuees pour trouver des conditions de convergence et etudier les sources possibles d'erreurs dans les estimations de phenol.
[1]
A. F. Gaudy,et al.
Analysis of growth data with inhibitory carbon sources.
,
1984,
Biotechnology and bioengineering.
[2]
P. Kumaran,et al.
Kinetics of phenol biotransformation
,
1997
.
[3]
Boris Tartakovsky,et al.
Application of multi-wavelength fluorometry for monitoring wastewater treatment process dynamics
,
1996
.
[4]
Peter A. Vanrolleghem,et al.
The potential of off-gas analyses for monitoring wastewater treatment plants
,
1996
.
[5]
G. R. Sullivan,et al.
Dynamics of phenol degradation by Pseudomonas putida
,
1993,
Biotechnology and bioengineering.
[6]
T. Fukuda,et al.
Presenting the Special issue on Control and robotics
,
1995
.
[7]
W. E. Inniss,et al.
Psychrotrophic bacteria in sediments from the great lakes
,
1978
.
[8]
Jean-Jacques E. Slotine,et al.
Space-frequency localized basis function networks for nonlinear system estimation and control
,
1995,
Neurocomputing.
[9]
Michael Häck,et al.
Estimation of wastewater process parameters using neural networks
,
1996
.
[10]
Bhavik R. Bakshi,et al.
Wave‐net: a multiresolution, hierarchical neural network with localized learning
,
1993
.
[11]
A. F. Gaudy,et al.
Selection of growth rate model for activated sludges treating phenol
,
1985
.
[12]
A. Lallai,et al.
Steady-state behaviour of an aerobic mixed culture growing on phenol in a continuous stirred reactor
,
1994
.
[13]
Robert M. Sanner,et al.
Gaussian Networks for Direct Adaptive Control
,
1991,
1991 American Control Conference.
[14]
Lyle H. Ungar,et al.
A hybrid neural network‐first principles approach to process modeling
,
1992
.