Lignite Quality Estimation Using Artificial Neural Networks (ANN) and Adaptive Neuro-Fuzzy Inference Systems (ANFIS)

摘要

Recent advances in Artificial Neural Networks (ANN) and Adaptive Neuro-Fuzzy Inference Systems (ANFIS) have provided a new approach to the estimation of related quality characteristics, such as heating value, ash content and moisture of coals used for power generation. The basic strategy for developing ANN and ANFIS models for the prediction of missing quality values, is to train the models using an existing quality data set and an appropriate learning method. Statistical analysis results of the estimated values, showed that ANN and ANFIS are not only more accurate than the widely used regression models, but also tends to reproduce the variability of the initial data, while regression models generate a smooth representation of reality.