| Abstract: | Distinguishing the difficulty degree of top coal caving was a precondition of the popularization and application of the roadway sub-level caving in steep seam. Because of complexity and uncertainty of the coal seam, the expression of influence factors was difficulty with exact data. According to the fuzzy and uncertainty of influence factors, triangular fuzzy membership functions were adopted to carry out the factors ambiguity, of which the factors not only have the consistency of semantic meaning, but also dissolve sufficiently expert knowledge. Based on the properties and structures of fasART fuzzy neural networks of fuzzy logic system and practical needs, a simplified fasART model was put forward, stability and reliability of the network were improved, the deficiency of learning samples and uncertainty of the factors were better treated. The method is of effective and practical value was identified by experiments. |
