Flash calcination of limestone in a bench-scale sorbent activation process (SAP) unit

Sorbent Activation Process;flue gas desulfurization;mercury capture;flash calcination;high surface area;calcium-based sorbent

Sugiyono, Ivan ; Rood, Mark J. ; Rood ; Mark J. ; Rostam-Abadi ; Massoud

关键词: Sorbent Activation Process;    flue gas desulfurization;    mercury capture;    flash calcination;    high surface area;    calcium-based sorbent;   

Others  :  https://www.ideals.illinois.edu/bitstream/handle/2142/30886/SUGIYONO_IVAN.pdf?sequence=1&isAllowed=y

美国|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

Coal-fired power plants produce 40 % of the total electricity in the United States. Theflue gas generated from burning coal contains air pollutants including sulfur oxides (SOx),hydrochloric acid (HCl) and elemental and ionic mercury (Hgo and Hg2+). A process option toremove these pollutants from the flue gas is by injection of sorbents downstream of a boiler andup-stream of a particulate control device. Activated carbon (AC) is a suitable sorbent to capturevapor-phase mercury and calcium-based sorbents such as quicklime (CaO) and hydrated lime(Ca(OH)2) are suitable sorbents to capture SOx and HCl. This research addresses producingquicklime by a novel process to remove SOx and HCl from flue gas streams. Quicklime iscommercially prepared by thermal decomposition of limestone (CaCO3) in a rotary kiln. Thesurface area of commercial quicklime, a key parameter of reactivity, is typically < 2 m2/g.Therefore, increasing the surface area of quicklime in a cost-effective process would enhance itseffectiveness as a sorbent for control of combustion-generated air pollutants.Illinois State Geological Survey (ISGS), a division of the Prairie Research Institute at theUniversity of Illinois at Urbana Champaign (UIUC), and Electric Power Research Institute(EPRI), Palo Alto, CA, have developed a patent-pending Sorbent Activation Process (SAP)technology for on-site production and direct injection of quicklime into flue gas generated bycoal fired power plants (US Patent Application 20,110,223,088). This process is an extension ofa similar patented process for on-site production of activated carbon (AC) to remove vapor-phasemercury emissions in the flue gas (US Patents 6, 451, 094 and 6,558,454). SAP utilizes anentrained-flow reactor in which sorbent (AC or quicklime) particles are subjected to a < 5 secondresidence time during their production. On-site production of quicklime could help lower theproduction cost of quicklime sorbent for dry sorbent injection (DSI) applications.In this research, a bench-scale SAP unit (2 kg/hr limestone feed rate) was used to preparequicklime from two limestone samples. The impacts of particle size, surface morphologies oflimestone, and operating parameters of SAP including temperature profile, and residence time onthe product quicklime were investigated. SAP experiments were designed to provide engineeringdata and guidelines for operating a pilot-scale (20 kg/hr limestone feed rate) and designing a fullscaleSAP units (135 kg/hr limestone feed rate) currently being tested at a coal-fired power plantin the United States. Additionally, kinetic information about calcination of the two limestone samples was obtained from the analysis of non-isothermal decomposition measured bythermogravimetric analysis (TGA) method. Furthermore, the kinetic information was used topredict limestone calcination in SAP.Lime sorbents prepared in SAP contained between 20 and 80 wt % calcium oxide(balance calcium carbonate) and had surface areas ranging between 5 and 12 m2/g depending onoperation conditions employed. Non-isothermal TGA experiments were analyzed by several dataanalysis approaches including Coats-Redfern, Criado linearization and DTG-curve fittingmethod using DTG-SIM software to obtain the kinetic parameters (activation energy, frequencyfactor, and reaction order) for thermal decomposition (calcination) of the two limestone samples.The values of the kinetic parameters were in good agreement with those previously reported inthe literature. The kinetic models predicted the experimental TGA calcination in N2 with lessthan 10% deviation. However, only the Coats-Redfern-based kinetic model predicted the TGAcalcinations in CO2 data with less than 10 % deviation. The kinetic parameters were used topredict limestone conversions in an ideal flash calciner and in SAP. Ideal flash calciner assumedisothermal condition throughout the reactor while the later one used the actual temperatureprofiles in SAP to predict limestone conversion at different CO2 partial pressures. The impact ofmass and heat transfer limitation, lime sintering phenomenon, and particle size distribution oflimestone/lime were not included in the model. The experimental limestone conversions werehigher than those predicted by the models.Based on the results from SAP experiments and model predictions, it was concluded thatthe actual temperature of limestone particle was likely much higher than the gas temperaturemeasured in SAP. Future work should include: 1) installation of additional thermocouples tocontinuously monitor both axial and radial temperature profiles in the SAP, 2) an understandingof the flow pattern and hydrodynamic inside the SAP to better estimate gas-gas and gas-solidmixing, 3) testing several size-graded limestone samples to evaluate the impact of particle sizeon limestone calcination, 4) calibrating the propane and combustion air flow rates to obtain moreaccurate readings, 5) quantify the extent of particle deposition in SAP, 6) measure gas phaseconcentrations of CO, CO2, O2, NOx, and hydrocarbons (HCs), and verify those measuredvalues, and 7) incorporate mass and heat transports effects in the model to better predictcalcination performance of limestone in bench-, pilot-, and full-scale SAPs.