Desulphurisation of cement flue gases

Abstract: Flue gas desulphurisation (FGD) is a crucial method to minimise the SO2 emissions from industrial processes. The FGD system utilise an alkaline sorbent to remove SO2 from the flue gases. Calcium in form of limestone is a commonly used sorbent where gypsum is produced asa by-product. However, the limestone reactivity, along with impurities within the sorbent, can significantly influence the effectiveness of SO2 removal and the quality of the by-products. At Heidelberg materials Cement Sverige an intermediate product, raw meal (RM) 8, is used assorbent in the FGD and gypsum is used as setting retarder in the cement. The aim with this project is to examine if raw meal 7 or A-sten is a better sorbent than raw meal 8 with respect to consumption rate, gypsum quality and its effect on the cement properties, and economic viability. To accomplish this a theoretical study was performed along with data analysis. RM 7 and A-sten are both purer than RM 8 but the raw material cost for RM 7 is about 1.24 times the raw material cost for RM 8 and the production costs and transportation costs are greater for RM 7 than for RM 8. The raw material cost for A-sten is about 0.45 times the cost for RM 8 but is not produced on site. All sorbents contain magnesium which can react with sulphur and precipitate as epsomite or hexahydrite, which also acts as retarders, where epsomite retard the cement setting time significantly compared to gypsum. To calculate the consumption rate of raw meal and A-sten into the scrubber three different methods were used. The difference between the methods lies in the consideration of how the calcium and magnesium species in the sorbent react with sulphur, with all calcium and magnesium reacting with sulphur in method 1, method 2 take the mass fraction of sulphurcontaining species into consideration, and method 3 incorporating mole fractions of calcium and magnesium species as well as sulphur from the sorbent. The pH and SO2 emissions were analysed for two different time periods where the first analysed period shows a correlation between low pH and high SO2 emissions, while the latter analysed period lacks a clear pH-SO2 correlation. A correlation between a lower pH and a lower percentage of MgSO4⸱6H2O and MgSO4⸱7H2O in the slurry could be made. The calculated rawmeal flow rate for RM 8 is between 520 and 554 kg/h, depending on the method used. To achieve the same desulphurisation efficiency with RM 7 the flow was calculated to 499-538kg/h and 427-464 kg/h for A-sten. Given the comparable mass flow rates of RM 8 and RM 7 it is advisable to retain RM 8 as asorbent in the scrubber regardless of the higher magnesium content due the higher cost associated with RM 7. However, the epsomite content in the slurry should be considered when optimising sulphur in cement production. Using A-sten as sorbent would minimise the rawmaterial costs and result in purer gypsum slurry with a lower epsomite content. Operating with a purer sorbent can also enhance the efficiency of the FGD process, leading to lower SO2 emissions. The calculations in the report assume that calcium and magnesium in the different sorbents react similarly, further analysis of their reactivity is recommended for more accurate results.