Particle attrition plays an important role throughout the cycles of a circulating fluidised bed (CFB) and a fluidised bed (FB) process, gradually depriving the bed inventory of valuable mass and changing the bed particle size distribution. The mass loss has to be compensated by a make-up stream. For economic and design purposes, attrition cannot be neglected. Although the particles may be efficient catalysts (or reactants), if the compensation for the lost material amounts to very high expenses, the whole process may become uneconomical. It is then clear that the choice of the solids material should take into account its attrition propensity. The main sources of attrition in fluidised bed systems are the jet region, the bubbling bed and the cyclone. It is common practice to predict particle attrition in industrial scale fluidised bed systems by the population balance method, but is it possible to link that prediction with the breakage propensity of a single particle?
This work aims at developing a predictive tool for particle attrition in fluidised and circulating fluidised beds, by attempting to build a path line from the single particle breakage propensity to the attrition occurring in the process. Here, the reference industrial process is the Chemical Looping Combustion (CLC). The CLC is a circulating fluidised bed process under development and as such, the choice of a solids material is critical. A powder of crushed manganese oxide is a candidate material for the CLC process and is used here as test material, as well as its equilibrium equivalent. For simplicity, the two materials are referred to as F-CLC (fresh CLC particles) and E-CLC (equilibrium CLC particles), respectively.
The single particle breakability of F-CLC and E-CLC is assessed by impact tests. The experimental results are then used to correlate the extent of breakage upon impact with the particle size and impact velocity, according to the theoretical model of chipping of Zhang and Ghadiri (2002). Further tests are carried out to unveil the effect of impact angle and number of impacts. The results suggest that E-CLC is highly more inclined to attrition than F-CLC. Moreover, the single particle breakage is found to correlate with the magnitude of the impact velocity and the sin of the angle of impact for both materials.
Recalling the modelling approach of Ghadiri and co-workers, the single particle breakage model, as derived, and the model of surface wear of Archard and Charj (1953) are coupled with CFD-DEM (Computational Fluid Dynamic-Discrete Element Method) simulations to compute the attrition of F-CLC particles in a Stairmand cyclone. Moreover, the same cyclone is used to characterise attrition of F-CLC particles experimentally as a function of particle size, gas inlet velocity and solids loading. Remarkably, the outcomes of the two approaches are found to agree well. A correlation is eventually derived which expresses the extent of attrition in a cyclone as a function of the variables mentioned above. The analysis revealed that the main source of attrition in the cyclone is given by the particle-wall collisions at the opposite section of cyclone inlet, at any operating conditions. Particle-particle collisions and particle sliding against the wall become significant contributors of attrition at high and low solids loading, respectively.
Attrition in the jet region is evaluated at room temperature as the steady state loss rate, using a semi-pilot scale fluidised bed equipped with a porous distributor and a central orifice of variable size. The results of the tests show that jet attrition of F-CLC and E-CLC can be described by two different correlations. The steady state attrition propensity of E-CLC is found to be higher than F-CLC, confirming the outcomes of the impact tests. The analysis on the fines collected on the filter reveals that they are mainly composed by very small particles of about 1 μm.
The correlations of cyclone and jet attrition are implemented in a non-dimensional population balance model (PBM) that simulates attrition in a fluidised bed and a circulating fluidised bed. The latter is composed of a fluidised bed where the recycle of solids is provided by a cyclone. The PBM is validated for the fluidised bed configuration against the experimental PSD (Particle Size Distribution) of F-CLC particles after jet attrition in the fluidised bed. The PBM is eventually used to simulate hypothetical cases of a FB and CFB with low and high single particle breakability as well as low and high superficial velocities to assess the dynamic response of the system in terms of material loss, solids circulation rate, requirements for a make-up and PSD in different regions of the system. The simulations allowed to identify the presence of two subsequent regimes where the loss is firstly dictated by the pre-existing fines of the bed inventory and then by attrition. During the two regimes the mean particle size of the bed inventory increases and decreases, respectively. The PBM reveals that the circulation rate is strongly affected by attrition because of the accumulation of entrained particles which are large enough to be captured by the cyclone and recycled. The loss of material and the need for the make-up stream are found to increase using either larger superficial velocities and/or weaker particles.
