Fuzzy neural network for chain pillar design in longwall coal mining

In this paper, an approach employing fuzzy neural networks (FNN) is proposed for designing a longwall chain pillar system. The major elements related to chain pillar design are characterized by fuzzy membership functions and neural networks. The FNN uses a rule base of propositions in the form of IF A THEN B and is unique in that it is capable of learning and storing some associations. A defuzzification algorithm is taken to give the final decision for the design variables. Test data show a good agreement between FNN and ALPS results.