A. I. Cabrera Llanos, M. G. Ramírez Sotelo
Unidad Profesional Interdisciplinaria de Biotecnología del IPN, Barrio La Laguna Ticoman s/n, GAM, C. P. 07340 México D. F.
Recibido: 18 de marzo de 2007.
Aceptado: 9 de mayo de 2007.
]]> Resumen
Se describe el desarrollo de un sensor virtual o suave aplicado a la estimación de las variables de un proceso de fermentación en lote alimentado para el cultivo de Sacharomyces cereviseae. Se desarrolla el algoritmo del estimador basado en redes neuronales diferenciales también llamado neuro-observador, se detallan los elementos que lo conforman y los resultados obtenidos al aplicarlo a una base de datos de dos experimentos de fermentación, con duración de 24 horas cada uno, finalmente se comparan la evolución temporal de las variables estimadas como la concentración de biomasa y la concentración de trehalosa en contra de las variables en una base de datos para determinar el desempeño del neuro-observador.
Palabras clave: Sensor suave; Redes neuronales diferenciales; Procesos de Fermentación en lote alimentado; Sacharomyces cereviseae.
Abstract
In this paper we describe the virtual sensor or soft sensor applied to the fed batch fermentation process estimation with Sacharomyces cerevisiae yeast culture. The algorithm based on the differential neural networks is developed and this algorithm is called neuro-observer. The details for this structure are described and the application on the data for two fermentation process during a period of 24 hours of evolution of each one. Finally the estimate variables (biomass concentration and trehalose concentration) are compared versus the measured variables evolution to show the neuro-observer performance.
Keywords: Soft sensor; Differential neural networks; Fed batch fermentation process; Sacharomyces cereviseae.
DESCARGAR ARTÍCULO EN FORMATO PDF
]]> Referencias
[1] J. de A Adilson, R. M. F., Soft sensors development for on-line bioreactor state estimation, Computers and Chemical Engineering, 24, 1O99 (2OOO). [ Links ]
[2] L. Fortuna, S. Grazini, Soft sensors for monitoring and control of industrial processes, 1st ed. (New York, Springer, 2006). [ Links ]
[3] A. Cabrera, A. Poznyak, T. Poznyak, J. Aranda, Identification of a fed batch fermentation process: comparison of computational and laboratory experiments, Bioprocess and Biosystems Engineering, 24, 319 (2002). [ Links ]
[4] O. Sotomayor, S. W. P., Software sensor for on-line estimation of the microbial activity in activated sludge systems, The instrumentation, systems, and automation society, 41, 127 (2001). [ Links ]
[5] L. Hudson, Neural Network and artificial intelligence for biomedical engineering, (Wiley-IEEE Press, 2000). [ Links ]
[6] S. Haykin, Neural Networks: A comprehensive Foundation. New York, (New York, Cambridge University Press, 1994). [ Links ]
[7] A. V. W. Ph. Bogaerts, Parameter identification for state estimation - application to bioprocess software sensor, Chemical Engineering Science, 59, 2465 (2004). [ Links ]
[8] B. McNeil,. L. M. Harvey, Fermentation a Practical Approach. (New York, IRL Press, 1990). [ Links ]
[9] N. G. McDuffie, Biorreactor Design. (Oxford,Butterworth-Heinemann., 1991). [ Links ]
[10] A. S. Poznyak, E.N. Sanchez and W. Yu, Dynamic Neural Networks for Nonlinear Control: Identification, State Estimation and Trajectory Tracking. 1 st ed. (New Jersey, World Scientific, 2001). [ Links ]
[11] J. S. Aranda, E. Salgado and P. Taillandier, Biochem. Eng. J., 17, 129 (2004) [ Links ] ]]>