(47cg) CFD Modeling for Prediction and Prevention of Runaway Reaction

Thermal runaway reactions have caused many industrial incidents with dire consequences, including injury and death to humans, negative effects on the environment, and economic losses for companies. The U.S. Chemical Safety and Hazard Investigation Board (CSB) released a report on 167 reactive chemical incidents in the U.S. between 1980 and 2001. About 35% of the incidents were caused by runaway reactions. With the aim to prevent the occurrence of these incidents, it is necessary to have a good understanding of the events that can lead to a runaway reaction.

Mixing operations have an important effect not only in the adequate contact of reactants which affects the performance of the reaction, but also in removing the excess heat generated by the reaction. For exothermic reactions this removal of heat is a crucial operation because when the heat generated exceeds the heat removal, the excess of temperature will increase the heat generation rate, which would produce an auto-acceleration behavior triggering uncontrolled conditions of temperature and pressure inside the reactor.

The mixing inefficiencies lead to non-uniform mass flow distribution which causes inhomogeneous temperature profiles across the reactor. Therefore, it is important to predict this temperature distribution to locate and understand the characteristics of local hot spots that ultimately may provoke a thermal runaway.

Computational Fluid Dynamics (CFD) techniques have been used to identify local and instantaneous values of reactor temperatures and reactant concentrations in mixing tanks to indicate in advance the local hot spots in the reaction mixture. This kind of information is of particular interest to determine situations that can produce a thermal runaway in order to make engineering decisions to avoid them. Specifically, in some cases this information has been used to create an Early Warning Detection System (EWDS) for runaway detection.

The main objective of this work is to discuss the effect of mixing by paddle on the heat accumulation rate, considering the difference between the heat of reaction and the loss of heat by transfer. The results can provide a good understanding of the reaction system using an anchor impeller.

In this work, a time dependent simulation was carried out using a Multiple Reference Frame (MRF) technique; in this method a region away from the impeller is defined as a stationary frame and a region adjacent to the impeller is defined as a rotating frame, but the impeller itself remains stationary with respect to the MRF. Geometry of the tank was created similarly to the Mettler Toledo RC1e reaction calorimeter stirred by an anchor impeller. The simulations were carried out with a three dimensional CFD model using the ANSYS FLUENT 14.0 software.

A moderate to high exothermic reaction, the hydrolysis of acetic anhydride, is used as a model reaction in this work. This is a pseudo first-order reaction, ideal for validating the simulation by experimental results on the dynamic response of a calorimetric reactor. This reaction model is a simple mechanism that has been well studied with much experimental work reported in the literature.

The conclusions drawn from this work will have important implications to elaborate future research by developing a method for early warning detection of runaway reaction and extending this idea to more complex mechanisms of highly exothermic reactions.