Scale-up of Bubbling Fluidized Bed Reactors with Vertical Internals: A New Approach Accounting for Chemistry and Hydrodynamics

The full set of scaling law derived by Glicksman [1] allows the hydrodynamic scale-up of fluidized bed reactors. In case of catalytic reactors, changing the catalyst particle diameter during scale-up may have consequences for the catalyst activity, selectivity and deactivation behavior. For fluidized bed reactors with vertical internals, a new scale-up approach is proposed and tested that helps to avoid this dilemma. INTRODUCTION Since the first fluidized bed reactor was erected in 1921 by Fritz Winkler, a large knowledge about the scale dependencies of mass transfer, residence time etc. in fluidized beds was formed Squires [2]. Usually, a selection of dimensionless numbers, e.g. the full set of scaling relations derived by Glicksman, is applied for scale-up Rudisuli et al. [3]. Today scale-up failures like the well-known that happened in Brownsville (TX, USA) in 1950 can normally be avoided. In Brownsville, a Fischer-Tropsch fluidized bed reactor was directly scaled from an inner diameter of 0.305 m to a 5 m reactor and as consequence of the improper and large scale-up, the residence time and conversion of the reactor was completely changed. The reason was that the bubble diameter in the lab scale reactor exceeded a critical value of 2/3 times the reactor diameter and slugging occurred Hovmand & Davidson [4], whereas in the industrial scale reactor, the bubbling regime was maintained and the bubbles rose much faster leading to a smaller residence time and conversion. Glicksman’s full set of scaling relations is often considered to predict the scale-up of fluidized bed processes. More recent studies however claimed that Glicksman’s scaling laws led to mismatches in different scales, especially for larger gas velocities Sanderson & Rhodes, van Ommen et al. [5, 6]. Foscolo et al. also criticized the lack of a particle pressure term, which shall help to account for homogeneous fluidization [7]. Additionally, in case of catalytic reactors, changing the catalyst particle diameter during scale-up may have consequences for the catalyst activity, selectivity and deactivation behavior. The mass transfer limitations could be changed, leading to a significantly different chemical behavior of the scaled reactor, with respect to selectivity and catalyst deactivation. This paper demonstrates the advantages of the sectoral scale-up approach to scale fluidized bed reactors with vertical internals Rudisuli et al. [8] and tries to define its limitations. To this end, the restrictions of Glicksman’s scale-up law are theoretically derived and a set of experiments, using pressure fluctuation measurements is performed, to study the hydrodynamic behavior with different numbers of vertical tubes. The sectoral scale-up approach can be applied for all reactors with vertical internals, e.g. heat exchanger tubes. It is based on the full set of Glicksman’s scale-up laws; however, the reactor diameter is replaced by the hydraulic diameter of the vertical internals. The new approach leads to a constant catalyst particle size at all scales and is considered as a new possibility to scale both, hydrodynamics and chemistry in a proper way. THEORY In 1984, Glicksman derived the most widely used scale-up approach for fluidized beds based on the conservation of mass and momentum in a nondimensional form Glicksman [1], for both particles and fluid. The derivation of the governing equations was performed assuming an incompressible fluid and omitting all inter-particle forces apart from collision forces. The dimensionless parameters derived by Glicksman are, from left to the right, the Reynolds number, the Froude number, the gas-solid density ratio, the bed geometry ratio, the sphericity of the particles and the particle size distribution, Eq. 1.