岩体声发射混沌与智能辨识研究

RESEARCH ON CHAOS AND INTELLIGENT IDENTIFICATION OF ACOUSTIC EMISSION IN ROCK MASS

对岩石试件加载及破坏过程的声发射进行了试验,对工程岩体声发射进行了监测。采用混沌动力学研究了岩体声发射活动规律,计算了工程岩体在变形与破坏过程中不同阶段声发射混沌吸引子。用混沌与神经网络相结合,建立了岩体声发射预测模型。根据岩体声发射各阶段特征,建立了工程岩体稳定性智能辨识模型。研究结果表明,岩体声发射活动存在4个不同的阶段:稳定期、声发射活动初期、声发射活动加剧期和活动反转期。岩体破坏出现在声发射活动反转期,声发射出现反常,混沌吸引子减小,破坏特征呈现。工程应用实践证明,混沌动力学能较好地反映岩体的声发射特征,所建立的预测与智能辨识模型能较好地预测和分析工程岩体稳定性。

Acoustic emission (AE) in rock specimen was tested under loading and breakage, and AE in rock engineering was surveyed. Action laws of AE in rock mass were researched with chaotic kinetics, and chaotic attractors of AE in different stages were calculated during deformation and breakage in rock engineering. By coupling chaos with neural network, a prediction model of AE was established. According to characteristics of AE in different stages, an intelligent identification model for analyzing stability in rock engineering was created. Research results show that there are four different stages in AE activities, i.e., stable period, earlier active period, intensely active period, and reversal period. Breakage occurs in the reversal period with abnormal AE and diminishing value of chaotic attractor. In-situ case verifies that chaotic kinetics can reflect the characteristics of and the intelligent identification model proposed in this paper is able to predict and analyze stability in rock engineering, especially in deep mining.