Nowadays, the fine chemical industry requires increasingly faster time-to-market as well as economically efficient and safe processes. In addition, the growing product variety needs more versatile production plants able to produce from small amounts up to several hundred tons per year. As a consequence, the time devoted to development is limited, and the production often takes place in multipurpose plants. This implies that a given process can run in different reactors making the process development a long and complicated task. As a matter of fact, the behaviour of a reactive system changes with scale and also changes from one equipment to another which often leads to difficulties in controlling a reactor at the industrial scale. In most cases, chemical incidents are caused by loss of control and wrong assessment of the thermal potential, resulting in runaway reactions or deviations in different units. These incidents could be foreseen and avoided or at least decreased with an appropriate process development and risk assessment. All processes should be optimised to achieve a fair productivity and still remain inherently safe. Such process characterization aims at providing a better understanding of reaction pathways and thermal behaviour. To reach this goal, save resources and effort, the real system dynamics can be reproduced by models where parameters are estimated from experimental calorimetric data. The dynamic behaviour of the reaction system can be represented by two models: 1) The reaction kinetics, based on an assumption of the reaction scheme considering different reactions occurring along the process course and 2) The reactor dynamics highlighting how the heating/cooling system controls the reaction course by adjusting its temperature and the adopted feed profile strategy. To acquire enough knowledge on the studied reaction system, several experiments have to be performed in a multiscale based approach to maximize the disturbances and explore the overall reaction kinetics. Nevertheless, even with this method, the investigation can be very long due to the increasingly complex models, involving a large number of parameters and amount of experiments. Therefore, a new approach to solve these issues has to be developed. The proposed approach is summarized in three steps: 1. Reaction Kinetic Investigation: Planning and minimize the number of experiments in a statistically optimal way to cover the experimental space efficiently and using a numerical method to extract the reaction kinetics based on the power rate law. 2. Reactor Dynamic Investigation: Planning the experiments in order to obtain information about the dynamic behaviour of the jacket together with its temperature controller and then use a numerical method to extract the parameters representing the reactor. 3. Risk assessment and Optimization: Combine the reaction kinetics and reactor dynamics to shape a Process Implementation model serving to predict the behaviour of the overall system and identify the optimal operating conditions the process inherently safe. This research establishes a new process scale-up methodology, regarding the assessment of the thermal potential. The results demonstrate that a multiscale approach joined with modelling and inherent safety, lead to a better understanding of the system. This approach also notably helps to optimise the operating conditions making the process inherently safe and remaining economically favourable.
EPFL2017
