The Effect of Saline Solution Conditioning on South African Fly Ash Resistivity

Abstract

Fly ash resistivity is an important characteristic that not only affects the electrical charge rate of fly ash particles but also the ultimate performance of electrostatic precipitators used for fly ash separation and collection. South African fly ash is highly resistive, thus limiting the collection efficiency of electrostatic precipitators at several coal-fired power generation plants in South Africa. Several flue gas conditioning studies (e.g., duct injection of sulphur trioxide, SO3) have shown to be effective in reducing fly ash resistivity to subsequently improve electrostatic precipitator performance. An earlier industrial pilot testing campaign involving the injection of a reject saline solution, containing primarily sodium sulphates and chlorides, into the boiler of one of Eskom’s coal-fired power station units gave a subsequent reduction in particulate matter emissions. This therefore highlights the possibility of using a waste stream, in this case a reject saline solution from an industrial reverse osmosis plant, as a suitable and cost-effective alternative to other flue gas conditioning agents such as SO3. An assay of the reject saline solution from the reverse osmosis plant showed the presence of other ionic species (e.g., Ca, K, Fe, etc.), which, although present in low concentrations, could also contribute to an observed reduction in particulate matter emissions. The observed improvement in particulate matter emissions on industrial scale therefore necessitated further investigation, to elucidate the effect of saline solution conditioning on fly ash resistivity. Fly ash samples from two Eskom coal-fired power generation stations (PS-1 and PS-2) were sampled and conditioned (treated) with reverse osmosis plant saline solution in different weight ratios by total wet mass of saline solution (5, 8, 10, 20, 30 and 35 wt.% for PS-1; 20, 30 and 35 wt.% for PS-2). The selection of these mixing ratios was based on typical flue gas volumetric flow rates (500–680 Nm3/s) under varying load factors (generated load as a percentage of the maximum output possible from the power generating unit), as well as saline solution injection rates used earlier during an industrial pilot testing campaign (5 and 25 m3/hr). After drying at ambient conditions, samples were crushed to obtain homogenous fine powders. Physiochemical characterisation of the unconditioned and conditioned fly ash samples was carried out with various techniques (proximate analysis, thermogravimetric analysis, particle size distribution analysis, X-ray fluorescence and quantitative evaluation of materials by scanning electron microscopy). Ash resistivity was measured in an experimental resistivity rig. The latter was performed under dry and humid (7 wt.% moisture) conditions for both ascending and descending temperature increments, in the range 90–330 C. The observed measurements were found to be within the acceptable error ratio as stipulated by the IEEE Standard 548 of 1984. Under dry conditions, the resistivity measurements for the ascending and descending runs followed the same trends for ash samples of both the power stations (PS-1 and PS-2). The resistivity measurements for temperatures less than 240 C (range representative of surface resistivity) exceeded the optimum operating range for electrostatic precipitators (1  1011 Ω.cm in the temperature range 115–150 C). Whereas the average resistivity of the dry ash samples, over the entire measurement range, decreased monotonously with increasing temperature, the average resistivity of the 7 wt.% ambient moisture ash samples (unconditioned and conditioned) exhibited an eccentric curve with a local maximum in the range 130–180 C for both the ascending and descending temperature measurements. For temperatures less than 210 C, initial ascending resistivity measurements for the conditioned ash samples (under dry conditions) were found to be substantially lower than the descending runs, while subsequent repeats of the ascending runs showed a similar trend to the results obtained for the descending runs. The observed disparity for the initial ascending runs could be attributed to volatile and soluble chemical species (e.g., carbonaceous impurities not detected with the saline solution assay) adsorbed onto the ash particles and then removed when heating the ash samples. Furthermore, with moisture injection, it was determined that the fly ash resistivity for both PS-1 and PS-2 decreased with an increase in saline solution loading (a subsequent increase in sodium sulphate/chloride species concentration). Ash samples conditioned with a higher ratio of saline solution (greater than 20 wt.%) gave resistivity results within the desired ash resistivity range for optimum electrostatic precipitator performance. An assessment of the Bickelhaupt model was performed to determine the ability of the model to predict the resistivity of the ash when conditioned with the varying concentrations of saline solution samples. The original Bickelhaupt model (1979) offered a good prediction for samples with the higher conditioning agent concentrations by mass when compared with the lower concentration conditioned samples. The Bickelhaupt model for sodium-depleted ashes showed an improvement for the volume resistivity of both power station conditioned ash samples. A modification to the original Bickelhaupt model provided a suitable prediction for the lower and higher concentration ash samples for PS-1 however the same formula was not as suitable for PS-2 ash samples although it did offer a better prediction than the original Bickelhaupt formula.