Calibration Of On-Line Ash Analyzers Using Adaptive Neuro-Fuzzy Inference Systems (ANFIS)

This paper proposes a new approach for calibration of dual gamma-energy transmission (DU-ET) ash analyser for improved online analysis of run-of-mine lignite. An adaptive neuro-fuzzy inference system (ANFIS) was developed to model the relationship between the intensities measured by the scintillation counter of the analyzer and the measured ash. Samples collected from Megalopolis lignite mines were measured for their ash content by using standard laboratory methods as well as by DUET. Obtained data was divided into training, testing and validation sets. A first order Sugeno type ANFIS was used in conjunction with early stop training. The calibration performance obtained by this approach was significantly better compared with that obtained by standard calibration methods based on multiple linear regression analysis.