(522a) Computing Blend Time with Mean Age Distribution in Batch Stirred Tank Reactors

Blend time is a basic measure for performance efficiency of a stirred tank reactor. In the last two decades, there have been several occasions that CFD was attempted to compute blend time with RANS model but the reported results have been inconsistent. More recently, LES has been used to compute the time dependent flow solution and tracer history. Good results in blend time predictions have been reported but the huge CPU time demanded by the computation is still an obstacle for the method to be used in industrial reactor design. In this talk, a new method for computing blend time in a stirred tank reactor will be discussed. This method is based on the mean age distribution that can be obtained by solving a conservation equation of passive scalar. Unlike the current method of tracking time history of a passive tracer by solving the time dependent concentration equation, the new method solves only a steady scalar equation for mean age distribution in the reactor. Therefore, the method is extremely efficient numerically. The CPU time consumed by this new method is orders of magnitude shorter than that of the above mentioned current methods. This method makes it possible to predict blend time of an industrial stirred tank reactor on a desk top computer. Different types of impellers will be studied with this new method and the predicted blend time will be compared with experimental results and correlations available in literature. The use of the method for scale-up of stirred tank reactors will also be discussed.