No abstract

Pure cultures of Pseudomonas putida (ATCC 17484) were grown in continuous culture on phenol at dilution rates of 0.074-0.085 h(-1) and subjected to step increases in phenol feed concentration. Three distinct patterns of dynamic response were obtained depending on the size of the step change used: low level, moderate level, or high level. During low level responses no accumulations of phenol or non-phenol, non-glucose-dissolved organic carbon, DOC(NGP), were observed. Moderate level responses were characterized by the transient accumulation of DOC(NGP) with a significant delay prior to phenol leakage. High level responses demonstrated a rapid onset of phenol leakage and no apparent accumulations of DOC(NGP). The addition of phenol to a continuous culture of the same organism on glucose did not result in transient DOC(NGP) accumulations, although transient phenol levels exceeded 90 mg l(-1). These results were consistent with intermediate metabolite production during phenol step tests coupled with substrate-inhibited phenol uptake and suggested that traditional kinetic models based on the Haldane equation may be inadequate for describing the dynamics of phenol degrading systems.

This paper outlines the use of a process model directly in a control algorithm. The process considered, a forced circulation single-stage evaporator, is a nonlinear interacting process. The control strategy employing a process model derived from fundamental mass and energy balances is shown to outperform single loop and predictive control strategies by a significant amount. The control structure is first presented in general form and then specifically applied to this process.Nous dCcrivons ici I'utilisation d'un modBle de procCdt directement dans un algorithme de rCgulation. Le procedi considCrC, un Cvaporateur ti une phase circulation forcCe, est un procCdC d'interaction non IinCaire. Nous montrons que la stratCgie de contr8le employCe qui utilise un modble de procede calcul6 h partir des bilans de masse et d'hergie, permet de surpasser largement la stratCgie de boucle unique et la stratCgie de contr8le prkdictif. La structure de contrBle est d'abord prksentte sous sa forme gCnCrale, puis specifiquement appliquCe au procCdC.

No abstract