Estimation of coal proximate analysis factors and calorific value by multivariable regression method and adaptive neuro-fuzzy inference system (ANFIS)

The proximate analysis is the most common form of coal evaluation and it reveals the quality of a coal sample. It examines four factors including the moisture, ash, volatile matter (VM), and fixed carbon (FC) within the coal sample. Every factor is determined through a distinct experimental procedure under ASTM specified conditions. These determinations are time consuming and require a significant amount of laboratory equipment. The calorific value is one of the most important properties of a solid fuel and its experimental determination requires special instrumentation and highly trained analyst to operate it. This paper develops mathematical and ANFIS models for estimation of two factors of proximate analysis based on the other two factors. Furthermore, the estimation of calorific value of coal samples based on proximate analysis factors is performed using multivariable regression, the Minitab 16 software package, and the ANFIS, Matlab software package. The results indicate that ANFIS is a more powerful tool for estimation of proximate analysis factors and calorific value than multivariable regression method. The following equation estimates the calorific value of coal samples with high precision: Calorific value (btu/lb)= 12204 - 170 Moisture + 46.8 FC - 127 Ash

Keywords:

Coal , Proximate analysis , Calorific value , Data modeling , Regression and ANFIS methods