Determination of the lower calorific and ash values of the lignite coal by using artificial neural networks and multiple regression analysis


Gulec M., GÜLBANDILAR E.

PHYSICOCHEMICAL PROBLEMS OF MINERAL PROCESSING, cilt.55, sa.2, ss.400-406, 2019 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 55 Sayı: 2
  • Basım Tarihi: 2019
  • Doi Numarası: 10.5277/ppmp18149
  • Dergi Adı: PHYSICOCHEMICAL PROBLEMS OF MINERAL PROCESSING
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Sayfa Sayıları: ss.400-406
  • Anahtar Kelimeler: lignite chemical analyses, artificial neural networks, Seyitomer lignite, Tuncbilek lignite, PREDICTION, SUBSIDENCE
  • Eskişehir Osmangazi Üniversitesi Adresli: Evet

Özet

The calorific value of coal varies depending on type of coal and foreign matter content. The calorific value of coal from pits is determined by analyzing moisture, volatile matter, ash and sulfur content in laboratories. This analysis process imposes a burden on businesses both in terms of time and cost. However, calorific value, in particular, can be determined through simpler methods by using ash and moisture values. The aim of this study was to develop a model that reduces the time and labor costs of coal companies by determining the calorific value and ash content of coal with the back-propagation algorithm of artificial neural networks (ANN). The model design was developed based on the data that was obtained from the laboratory analyses of raw coals from the pits of Tuncbilek and Seyitomer mining areas in Turkey. The values of moisture, volatile matter, original ash and sulfur were determined as input variables, and the lower calorific values and ash content were selected as output variables. The lower calorific values (LCV) and Ash estimated by the developed model were compared with the LCV obtained in the laboratory tests and the results showed a correlation. In addition, two different ANN models and multiple regression analysis (MRA) were developed to obtain the single output of the LCV and ash parameters with similar features. As a result, the ANN model and MRA equation models proposed in this study was shown to successfully estimate the LCV and ash content of coals without performing laboratory analyses.