Portal de Periódicos da CAPES

Neural network modelling for on-line state estimation in fed-batch culture of l-lysine production

1996; Elsevier BV; Volume: 61; Issue: 1 Linguagem: Inglês

10.1016/0923-0467(96)03015-1

1873-3220

I Yet-Pole, Wu Wen-Tengu, Liu Yung-Chuan,

Advanced Control Systems Optimization

A moving window neural network was applied for dynamic modelling and on-line estimation of consumed sugar, cell mass, and product concentration in l-lysine fed-batch culture. The organisms were auxotrophic mutants for l-homoserine and were resistant to S-(2-aminoethyl)-l-cysteine. The reducing sugar concentration was calculated on line by using the estimated consumed sugar concentration and was maintained at a given value by a simple compensatory feeding strategy. With the estimator, the fed-batch culture could be satisfactorily carried out.