Analysis of L-Valve Performance Feeding Spent Coffee Ground Powders into a Circulating Fluidized Bed

Circulating Fluidized Beds (CFB) are widely used in industrial processes for combustion and gasification due to the high solids throughputs, temperature uniformity and effective gas-solid contact. A common challenge in operating CFBs is to ensure a stable and efficient feeding of the particulate material into the reactor. Solids feeding can be particularly limiting when handling biomass residues. Characteristics such as wide size distribution, variable moisture content, irregular shape and low bulk density may cause poor flowability, therefore the powders are likely to block into feeders and hopper discharge. Using non-mechanical L-valves as feeders to CFB reactors has already proven to be feasible and attractive when processing regular particles, but a quite limited number of studies were conducted to test such feeders with biomasses. In this study, our aim is to evaluate the performance of a non-mechanical L-valve to feed Spent Coffee Ground (SCG) powders into a cold-flow CFB with an overall height of 2 m and an internal diameter of 0.021 m. The L-valve is made of acrylic, with an internal diameter of 0.021 m and a length of 0.22 m. The static pressures throughout the CFB loop (silo, standpipe, L-valve, riser, and cyclone) and the solids mass fluxes were monitored using a data acquisition system from National Instruments^®, model NI-9205, and a digital balance coupled to a solids’ sampler. The performance of the L-valve was assessed through measurements of pressure loss in the valve (ΔP), solids mass flux (G_S) and by filming the gas-solid flow through the acrylic L-valve. The SCG samples were obtained brewing an ordinary Brazilian coffee, and the moisture content of the residue was reduced by keeping the samples in an oven for 24h at 105±2ºC. Two samples of SCGs with different particle-size distributions were prepared: sample A was composed of 1.5 kg of SCG particles retained between the sieve’s apertures of 500 and 300 μm and sample B was composed of 1.35 kg of sample A and 0.15 kg of particles retained between the sieve’s apertures of 300 and 150 μm. An additional sample with moisture content equal to 30% in wet basis, denominated sample C, was prepared by mixing sample A with water in glass flasks, which were sealed and stored at 4°C for 60 h. To cover a wide range of operating conditions with samples A, B, and C, the L-valve performance was evaluated under four aeration velocities (U=0.34, 0.68, 1.02, 1.36 m/s) and three air velocities in the riser (Q=8.8, 10, 11.2 m/s). The assays were carried out in triplicate. The results showed that the aeration velocity overwhelmingly drives the values of Gs. Depending on the aeration velocity, three operating regions could be identified. Under a low aeration rate (U≈0.34 m/s), the packed-bed resistance in the standpipe controlled the transport of solids. A high-pressure drop in the valve and a low solid mass flux were observed in this condition. Under intermediate aeration rates (0.34<U<1.02 m/s) the solids weight in the standpipe was in sync with the buoyancy force generated by aeration and the highest Gs values were observed in this aeration range. Finally, under high aeration rates (U>1.02 m/s), the solids transport was controlled by the gas phase. Turbulent vortices were observed in the base of the standpipe and a diluted solids transport was achieved in the L-valve, resulting in a low-pressure drop. Besides, we observed that the air velocity in the riser had a small effect on the L-valve performance. Depending on the aeration velocity, Gs varied between 1.7 and 3.8 kg/m²s when operating with sample A. With sample B, which has 10% of fines mixed to sample A, the mass flux yielded by the valve was limited to only 1.7 kg/m²s and a stable operation was achieved only for a narrow range of aeration velocities. The presence of fine particles in the sample contributes to increasing the specific surface area and contact among particles. Both the inter-particle attractive forces and mechanical interlocking are enhanced, thus limiting the valve operating range. Using moist samples, on the other hand, had a positive effect on Gs, with values from 2.8 to 24.4 kg/m²s. This behavior can be attributed to a lubricant effect of the water under this saturation level. To burn SCG powders effectively in the soluble coffee industry furnaces, the recommended moisture should vary from 25 and 50%. According to the literature, the lower limit is set to avoid spontaneous combustion and the upper one to preserve burning efficiency. Our results showed that the L-valve did not perform well for samples containing particles of sizes under 300 µm at a mass fraction of 10%. However, it provided stable feeding in a wide range of solid mass fluxes (2.8<Gs<24.4 kg/m²s) for SCG powders with a mean size of 400 µm and for moisture content of 30%. Besides, the solid mass fluxes can be easily varied by adjusting the aeration velocity in the L-valve.