Application of artificial neural network to study the performance of jig for beneficiation of non-coking coal

• Published in: Fuel (Guildford) 2012-07, Vol.97, p.151-156
• Main Authors: Panda, Lopamudra, Sahoo, A.K., Tripathy, A., Biswal, S.K., Sahu, A.K.
• Format: Article
• Language: English
• Subjects: ANN; Applied sciences; Artificial neural networks; Ashes; Back propagation; Beneficiation; Coal; Coal mines; Combustible recovery; Energy; Energy. Thermal use of fuels; Exact sciences and technology; Fuels; Jigging; Learning theory; Mathematical models; Neural networks; Non-coking coal
• ISSN: 0016-2361; 1873-7153
• DOI: 10.1016/j.fuel.2012.02.018

► Beneficiation of −3+0.3mm size fraction non-coking coal using jig was studied. ► Effects of particle sizes of bed material and particle sizes of coal were studied. ► Effects of feed rate and water rate on the performance of the jig were studied. ► ANN model was proposed to predict ash and combustible recovery of the concentrate. Non-coking coal is the major resource of energy in India. Apart from its utilization in energy sector, the other major application of this coal is in metallurgical sector. The resource of high quality of non-coking coal is not available as per demand; as a result beneficiation of non-coking coal is now becoming essential. Jigging is one of the economical physical beneficiation processes for Indian high ash non-coking coal. At present scenario in coal washery in India, below 3mm size is not being processed. Attempt has been taken to beneficiate the fine size non-coking coal fractions generated at different sizes of bed materials, feed rates and water rates using laboratory Denver mineral jig. The performance of jig was evaluated in term of Ep and imperfection value. Furthermore artificial neural network (ANN) model has been developed for determining combustible recovery and ash percent of the concentrate. The ANN architecture is made up of three layers (input – hidden – output). A back propagation algorithm was used for training of the ANN model. It has been observed that the predicted values by ANN model are in good agreement with the experimental results.