Coke property prediction from coal blend properties using ANFIS model

Abstract

The present study is aimed to fulfill the need of a model which will predict the coke properties from coal blend characteristics so that optimisation of coal blends for producing desired quality of stamp charged coke can be achieved with lesser number of carbonisation tests. In this work, the functional relationship between the coal blend properties (ash, volatile matter, average vitrinite reflectance, crucible swelling number, total reactives, vitrinite distribution (V9-V13) and Basicity Index) and the corresponding coke properties such as coke strength after reaction (CSR) and Coke Reactivity Index (CRI) have been mapped using a non-linear system identification technique namely, Adaptive Network-based Fuzzy Inference System (ANFIS). The model has been trained and tested with the help of 67 number of experimental data sets produced using 7 kg carbolite test oven. It has been found that the developed ANFIS model successfully predicts the CSR and CRI. Also the effect of the number of fuzzy rules (M) on the model performance has been investigated to predict the properties and to correlate the above relationship. The developed model could be applied in the area of coal blend design.