Petroleum coke is an interesting fuel for FBC boilers. Worldwide about 40 Mt/a are produced with over 18 Mt/a being high sulphur (>3.5% sulphur) and over 88% of that production is in North America [Zierold, 1994], In the past, petroleum coke was burnt primarily in cement kilns. However, this situation is rapidly changing [Swain, 1991]. The primary driving force for petroleum coke use is price. Costs can be in the order of $0.2/MBtu for high sulphur petroleum coke compared with perhaps $2.0 MBtu for natural gas, while coal can cost five to eight times as much as coke depending on location and individual circumstances [Richardson and Taibi, 1993]. As part of its mandate CANMET carried out a major review of critical issues for petroleum coke firing in FBC boilers [Anthony, 1994] and identified agglomeration or fouling of heat transfer surfaces, and blockage of the return leg as a significant problem when burning 100% petroleum coke. In some cases the problem can be very severe and in one case in Japan the boiler owners changed from petroleum coke to anthracite after a year's effort to resolve this issue. It was also clear from interviewing plant managers and operators of several boilers that the problem was believed to be associated with high levels of vanadium in the petroleum coke, and probably due to the existence of eutectic mixtures containing the low melting point compound V{sub}2O{sub}5 [Zierold and Voyles, 1993]. In consequence, a widespread practice was the addition of Mg compounds at various levels, either in the form of MgO (available in the U.S. as "Fuelpro"), kaolin or dolomite, all of which were reported to have some beneficial effect [Anthony, 1992 and 1994]. In order to study this phenomenon, ashes and deposits were obtained from an industrial CFBC boiler which had particularly severe fouling in the return leg, with deposits of a metre deep building up over a period weeks to months (Fig. 1). As a first step it was decided to carry out a detailed series of analyses to obtain phase composition of the samples, i.e., identify the chemical species in the deposits. It was hoped that these analyses would identify any low melting point components that might be responsible for agglomeration or sintering phenomena and further that they would also be useful in helping to interpret the scanning electron microscopy and microprobe studies to follow. In all, ten samples were analyzed from the CFBC boiler (Table 1).
展开▼
