Stratégie d'intensification des procédés

Abstract : Since a few years, intensified and microstructured reactors are an alternative to stirred tank reactors. Thanks to their high heat and mass-transfer rates, a better control of operating conditions is possible leading to a better product quality. However, it is difficult to predict quantitatively the real interest of intensified reactors and also to transpose a priori a batch process to a continuous process, without an extensive experimental work aiming at selecting the best reactor technology and optimal operating conditions. The acceleration of this exploratory step, costly in time and money, is therefore a challenging task. Hence, an iterative strategy based on the combination of experimental design and modeling has been developed to take advantage of new extended experimental windows offered by intensified reactors and minimize the number of experiments. This method considers a predictive phenomenological model which describes reactions, heat and mass-transfer processes in various reactors. The iterative method alternates experimental steps and prediction steps. During the prediction step, used for data acquisition, the model proposes, using the D-optimal criterion, the optimal experiments which give the maximal information on the model parameters. After each experiment, the strategy decides whether a new iteration is required. If the parameters accuracy is judged sufficient, then the strategy looks for optimal operating conditions in the available reactors, which maximize several process performance criteria (e.g. yield, cost, operation time, etc.), and proposes optimal operating conditions. If the new experimental results are not validated, a new global iterative step is performed to improve the accuracy of the model and propose better operating conditions. To demonstrate its performance, the strategy has been applied to a classical parallel-consecutive reaction scheme. A specific experimental set-up based on three different reactors (batch, semi-batch and plug-flow reactors) has been developed. This work highlights the interest of using several reactors for data acquisition, demonstrates the importance of data acquisition for process performance optimization and emphasizes the superiority of mechanistic models on black-box models for the choice of reactor technologies while minimizing experimentation step. Keywords: Intensification, Modelisation, Experimental strategy, Process performance
Document type :
Theses
File URL :
.
Complete list of metadatas

https://hal.univ-lorraine.fr/tel-01750510
Contributor : Thèses Ul <>
Submitted on : Thursday, March 29, 2018 - 12:47:58 PM
Last modification on : Friday, March 30, 2018 - 1:30:12 AM

Identifiers

  • HAL Id : tel-01750510, version 1

Collections

Citation

Florent Mathieu. Stratégie d'intensification des procédés. Autre. Université de Lorraine, 2013. Français. ⟨NNT : 2013LORR0210⟩. ⟨tel-01750510⟩

Share

Metrics

Record views

11