No abstract

No abstract

Low grade iron ores with impurity gangue minerals containing silica and alumina must be upgraded to an acceptable level of iron content. Concentrates, due to their fine sizes, are not suitable to be directly charged to the iron-making processes such as the blast furnace or the DR-plant. Hence, an agglomeration technique should be applied to fine concentrate. The most commonly employed one is pelletizing in iron ore industry. In pelletizing, iron ore, water and a binder are balled in a mechanical disc or drum to produce agglomerates. Bentonite is the most widely used binder. However, it is considered as an impurity due to its high SiO 2 and Al 2 O 3 content. Many researchers have investigated different binders, mostly of organic origin, in pursuit of finding a viable alternative binder to bentonite. Organic binders were found to yield good quality green and dry pellets. However, they fail to impart enough strength to the pre-heated and fired pellets as a result of insufficient slag bonding. Boron compounds free of silica and alumina are thought to be a potential solution to overcome the lack of slag forming constituents encountered with organic binders as they are known for their low melting temperatures as well as for their ability to also lower the melting temperatures of silicates. A few researchers have investigated the use of boron compounds in iron ore agglomeration and found promising results which have been covered in this paper.

The use of conventional bentonite binder is favorable in terms of mechanical and metallurgical pellet properties, however, because of its acid constituents bentonite is considered as impurity especially for iron ores with high acidic content. Therefore, alternative binders to bentonite have been tested. Organic binders are the most studied binders and they yield pellets with good wet strength; they fail in terms of preheated and fired pellet strengths. This study was conducted to investigate how insufficient pellet strengths can be improved when organic binders are used as binder. The addition of a low-melting temperature and slag bonding/strength increasing constituent (free in acidic contents) into pellet feed was proposed. Addition of boron compounds such as colemanite, tincal, borax pentahydrate, boric acid together with organic binders such as CMC, starch, dextrin and some organic based binders, into iron oxide pellet was tested. Wet and thermally treated pellet physical-mechanical qualities (balling - moisture content - size - shape - drop number - compressive strengths - porosity - dustiness) were determined. The results showed that good quality wet, dry, preheated and fired pellets can be produced with combined binders (an organic binder plus a boron compound) when compared with bentonite-bonded pellets. While organic binders provided sufficient wet and dry pellet strengths, the boron compounds provided the required preheated and fired pellet strengths at even lower firing temperature. Especially, the contribution of boron compound addition is most pronounced for hematite pellets which do not have strengthening mechanism through oxidation like magnetite pellets during firing. Therefore, addition of boron compound is beneficial to recover the low physical-mechanical qualities of pellets produced with organic binders through slag bonding mechanism. Furthermore, lowering the firing temperature thanks to low-melting boron compounds will be cost-effective for firing part of the pelletizing plants.

Gross calorific value (GCV) of coal was predicted by using as-received basis proximate analysis data. Two main objectives of the study were to develop prediction models for GCV using proximate analysis variables and to reveal the distinct predictors of GCV. Multiple linear regression (MLR) and artifcial neural network (ANN) (multilayer perceptron MLP, general regression neural network GRNN, and radial basis function neural network RBFNN) methods were applied to the developed 11 models created by different combinations of the predictor variables. By conducting 10fold cross-validation, the prediction accuracy of the models has been tested by using R 2 , RM SE , M AE , and M AP E. In this study, for the first time in the literature, for a single dataset, maximum number of coal samples were utilized and GRNN and RBFNN methods were used in GCV prediction based on proximate analysis. The results showed that moisture and ash are the most discriminative predictors of GCV and the developed RBFNN-based models produce high performance for GCV prediction. Additionally, performances of the regression methods, from the best to the worst, were RBFNN, GRNN, MLP, and MLR.