The study on modeling and control of double carbonylation reaction process

The reaction discussed in this paper is a complicated, multi-phase-catalyzed batch reaction to produce phenyl-pyretic acid by double carbonylation of benzyl chloride. In order to industrialize the production process, a significant engineering problem is to study the operation and control of this batch process. The first stage of study in this paper is to process the experimental data, then to build the models based on the experimental operation of every batch with the artificial neural networks. These models, termed operation models, can reflect the operational characteristics of every operation process. But in the further simulation study, it was found that these models could not be successfully applied. The principal reason for this difficulty is that the experimental data used to model the reaction are too sparse. To solve this problem, we perform the second stage to build an intrinsic model. The ldquogray boxrdquo model, which is able to describe the characteristic of the batch process, is called an intrinsic model. Then in the third stage of this paper, we propose a multiple time scaled control strategy based on the intrinsic model. The major result of the simulation study suggests that the control system holds promise in engineering applications.