Identification of Stability of Surrounding Rock Based on Rough Set-LVQ Neural Network
Stability identification of surrounding rock is the basis of underground openings construction scheme and support design. The coal gateways are buried deeply with complex geologic conditions and affected badly by mechanical problem, so it is very difficult to confirm the main effect factors of its surrounding rocks stability because of the imperfect knowledge, and effect factors randomicity, fuzziness. In this paper, Rough Set was used to mine surrounding rocks main effect factors from the knowledge expression system based on Information Database of Yan Zhou Coal Gateways. Ultimately, the main effect factors have been sorted out by attribute reduction. Neural networks capability of robustness and fault tolerance could be better enhanced by controlling the mini-confidence and the mini-support, which were used to mine decision rules and eliminate noise and redundancy objects. In the end, the main effect factors and decision rules that were mined could be well utilized to guide neural networks construction and training samples selection. The application result indicates that Rough Set-LVQ neural network model possesses many merits, just as fast training speed, strong capability of fault tolerance, universal approximation, and better capability of generalization and so on. It is demonstrated that the method can be preferably and effectively utilized to settle the problem of identifying the stability of surrounding rock.
stability identification of surrounding rock data mining rough set LVQ neural network
QI Yuliang WANG Tongxu SUN Jun ZHANG Zhenyu WANG Wenbin CHEN Haiming
Key Laboratory of Mine Disaster Prevention and Control , Shandong University of Science and Technolo Key Laboratory of Mine Disaster Prevention and Control , Shandong University of Science and Technolo Department of Geotechnical Engineering, Tongji University, Shanghai, 200092, China
国际会议
The 2007 International Symposium on Safety Science and Technology(2007采矿科学与安全技术国际学术会议)
河南焦作
英文
87-92
2007-04-17（万方平台首次上网日期，不代表论文的发表时间）