Paper 14:text

Homogenization is the processing of one type of material so that the inherent fluctuations in respect of quality and/or size distribution are evened out. In blending coal, the aim is to achieve a final product with a well defined composition, from two or more coal types, so that the elements are very well distributed and no large pockets of one type can be identified. In mixing, traces of the individual components can still be identified. There are various reason why blending is necessary. According to Wall et al.1, indigenous coals can be too low in rank and/or high in ash; in these cases, traditional feed coals have some quality deficiencies, which require compensation. This means that power stations have design limitations. Wall et al1. also explains that in a coal fired thermal power station, many parameters interact but if we change an operating condition of the plant to account for one aspect of the coal’s character, this may significantly impact on the area of the plant not considered. Following their example, changing a particular type of coal in a blend to reduce costs may increase the total moisture, resulting in a decrease in flame temperature, which will affect the heat transfer in the furnace and may alter the deposition of ash. The increased gas volume associated with increased moisture, will decrease the operating efficiency of particulate removal systems and any de SOx plant. Therefore, it is very important to cite that the interactions within the plant are very complex, and it would be unreasonable or difficult to ensure an operator that changing one simple operating parameter does not have a negative impact on other areas. Coal blending is one of several options available for reducing sulphur emissions from coal fired power plants. However, decisions about coal blending must deal with uncertainty and variability in coal properties, and with the effect of off design coal characteristics on power plant performance and cost. According to Shih and Frey2, sulphur content, ash content and heating value are considered as normally distributed random variables. The objectives of a model to optimize operation of a plant must include minimizing: (i) the expected (mean) costs of coal blending; and (ii) the variance of coal blending costs. The cost objective function includes coal purchasing cost, ash disposal cost, sulphur removal cost, and fuel switching costs. Chance constraints include several risk measures, such as the probability of exceeding the sulphur emission standard. Finally, homogenization piles are frequently used in order to reduce variability and ensure ore stock. In this context, the most important factors, in respect to pile building can be cited as the type of layer disposition, number of layers and added mass. The last variable can significantly reduce pile variability when a large material amount is homogenized, and variability among the lots decreases (David3 and Parker4). The Deposition sequence of the material also is able to affect grade predictability of each pile, requiring a more careful mining advance forecast. Often, this scheduling is represented by a vector of grades to ore blocks that will be mined along a given activity time, and geostatistical techniques are used to predict them. Kriging technique is a very good estimator, but it is not adequate to predict the uncertainty of a process (Isaaks and Srivastava5). In these cases, it is strongly recommended to employ conditional simulation techniques, which are capable to represent oscillations, and then generate a proper uncertainty model. According to Costa et al.6, on contrary to kriging, geostatistical simulation methods aim to reproduce in situ variability, and the spatial continuity of the input data set. Models used in this way, aim to replicate the spatial BERETTA, F.S., COSTA, J.F.C.L., KOPPE, J.C. and DE SOUZA, V.C. Reducing coal quality variability with properly designed blending piles. Fourth World Conference on Sampling & Blending, The Southern African Institute of Mining and Metallurgy, 2009.