(53e) Behaviour of Lumps in Gas-Solid Fluidized-Bed Reactors

Gas-solid fluidized-reactors have been historically implemented in a variety of fields, with applications to the energy, petrochemical and pharmaceutical, to mention just few. More recently, fluidized-bed reactors have also been employed in waste-to-energy conversion plants with CO₂ capture. The widespread use of such reactors is mostly due to their excellent mixing properties, allowing, among other things, for temperature uniformity and for excellent contact between the gas and solid phases. These properties, however, are not guaranteed if lumps or agglomerates are present the fluid bed. This is the case of lumps, that can form, for example, in industrial applications of thermochemical conversion of waste plastic, as this latter tends to melt and embed bed particles. When the formation of these lumps cannot be avoided, a thorough understanding of their behavior is fundamental to limit the disruption they may cause to the reactor’s operation. The aim of this study is to assess the impact of design parameters and operating conditions on the motion characteristics and thermal behavior of such lumps. This knowledge can be helpful in finding the optimal conditions for the prevention of hotspot formation and defluidization.

The experimental setup used in this study consisted of a downscaled pseudo-2D fluidized bed that enabled the direct observation of the simulated lumps via infrared thermography. The fluid bed was operated at ambient temperature and the fluidization velocity was varied between 1U_mf and 10U_mf, ranging from incipient fluidization through bubbling and slugging. The bed material used was rutile sand; 4 different particle mean diameters, 60μm, 100μm, 153μm and 215μm, allowed to assess the effect of the particle size on the motion and heat transfer characteristics of agglomerates. Artificial lumps were produced by inserting heavier lead tracers into pumice beads; changing the tracer’s mass allowed to vary the density of these lumps and better simulate their behaviour in a fluid-bed of sand used for waste plastic treatment. The lumps were heated up and introduced in the fluidized bed. They were then observed as they moved around and exchanged heat with the fluidized bed. X-ray particle tracking (XPT) was used to track the lump position in the fluidized bed. Post-processing of the x-ray images allowed to identify the lump segregation behaviour for different conditions and to obtain the vertical and horizontal dispersion coefficients of the lumps. Infrared thermography (IRT) was used to determine the lump temperature evolution in time. This was done by converting the infrared signal emitted by the simulated lump into a temperature via a calibration curve that had been previously obtained. The heat transfer coefficient between lump and fluidized bed was then obtained through an energy balance.

The effect of the lump density on the lump dispersion coefficients was found to be negligible. Larger fluidization numbers and larger particle sizes, on the other hand, were both found to increase the dispersion coefficients of the lumps following power laws, as shown in Figure 1a,b. The lump segregation behaviour was then investigated. The lump density only plays a role for small fluidization velocities, up to 2U_mf for the bed material with particle size equal to 215 µm. Depending on its density, in fact, the lump was seen to sink at the bottom of the fluid bed, floats at the top or is properly mixed within it, as shown in Table 1a. It was concluded that beyond a threshold value for the fluidization velocity, 2U_mf in the case studied, mixing overcomes segregation, which is due to the difference in size and density between the lump and the bed material. Furthermore, Table 1b shows that a larger particle size allows better lump mixing at smaller fluidization numbers. Once more, the effect of the density of the lump on its heat transfer coefficient with the fluidized bed is negligible, as shown in Figure 2a. On the other hand, an increase in the fluidization velocity caused an increase in the heat transfer coefficient (Figure 2a,b), but only up to a point. The heat transfer coefficient, in fact, then stabilizes onto a constant value as the fluidization velocity keeps increasing. This behaviour agrees with previous results in the literature; this was attributed to the fact that larger gas flow rates improve the mixing and decrease the refreshment rate of bed material around a lump. This mechanism improves the heat transfer but soon saturates, hence the heat transfer follows an asymptotic trend with respect to the fluidization velocity. On the other hand, the heat transfer coefficient trend as a function of the fluidization number strongly depended on the bed material particle size, with the maximum value for each particle size decreasing with the bed material particle size, as shown in Figure 2b and in agreement with previous research. Due to the large size difference between the bed material and the lumps, particle convection was expected to be the main heat transfer mechanism. Once the bed starts bubbling, as the gas flow rate through the fluid bed increases, the excess gas ends in bubbles and the gas superficial velocity through the emulsion phase remains approximately constant. However, the heat transfer coefficient was observed to sharply increase in this range. This was interpreted as proof that particle convection is the dominant heat transfer mechanism. Group A particles presented a first very small increase in the heat transfer coefficient for superficial velocities between the minimum fluidization and the onset of bubbling. In this range of velocities, the whole gas flow rate ends up in the emulsion phase, but this has a negligible effect on the heat transfer coefficient. This is a further proof that particle convection is the main heat transfer mechanism. In fact, as soon as the bed starts bubbling and the mixing within the bed improves, the heat transfer coefficient increases sharply by a factor up to 10 (for average particle diameter equal to 60 µm). Optimal values of fluidization velocity that would guarantee sufficient mixing to prevent segregation, while also maximizing the heat transfer coefficient were found for different configurations. Going beyond these values returns negligible advantages from the points of view of lump mixing and heat transfer. Instead, it may increase the chance of operational problems, such as attrition and elutriation.

While semiempirical correlations for the Nusselt number between larger objects and fluidized bed have already been obtained for gas-convection dominated scenarios, this is not true for cases where particle convection is the main heat transfer mechanism. Hence, a correlation for the Nusselt number as a function of the Reynolds number and of the lump-to-bed-material size ratio was derived (Figure 3):

Nu* = 2 + 0.037 Re_lump^0.38 (d_lump/d_p)^1.14

where the object Reynolds number, Re_obj, was defined with the lump velocity information obtained through XPT.

X-ray particle tracking and infrared thermography were used to characterize the motion characteristics and thermal behaviour of lumps in a fluidized bed. Such lumps can impact the quality of fluidization and cause the formation of hotspots. A thorough knowledge of their behaviour is hence very important for industrial applications of fluidized-bed reactors. It provides, in fact, an insight into how process variables may affect the behaviour of such lumps, helping in preventing the disruption of such reactors.