A neuro-fuzzy system for ash property prediction

This paper proposes a neuro-fuzzy approach to solve the problem of predicting the property of ashes originated from combustion processes for electric generation. The adopted approach uses fuzzy logic for modeling as well as neural networks for extraction and optimization of the fuzzy model via learning from data. A two-phase learning strategy is adopted to determine the structure and parameters of the predictive model on the basis of experimental data derived from observation of different combustion processes. Preliminary results on the application of the proposed neuro-fuzzy approach are also reported.