【作者】 代高飞；

【导师】 尹光志；

【作者基本信息】 重庆大学 ， 采矿工程， 2002， 博士

【摘要】 冲击地压是矿井开采中发生的一种动力现象，它严重威胁着矿井的安全生产。随着矿井开采活动逐渐向深部延伸，冲击地压将越来越严重。许多学者对冲击地压的发生机理、预测预报和防治等进行了大量的研究，但目前仍然有许多问题没有得到根本解决。由于煤岩体材料的非线性特征，决定了冲击地压系统是一个高度复杂的非线性系统。在该系统中，非线性和不确定性是其主要特征。采用传统的确定性理论难以从本质上解释冲击地压系统复杂的非线性行为。而现代非线性科学的分叉与混沌理论、突变理论、自组织理论、神经网络方法等为这一难题的解决提供了新的思路和方法。因此，采用现代非线性科学的相关理论来研究岩石的失稳破坏和冲击地压有望取得新的突破。本文基于前人的研究基础，对煤岩进行了无损伤CT检测实时实验，采用分叉与混沌理论、突变理论、自组织理论和神经网络方法对煤岩的非线性动力学特征和冲击地压进行了研究，主要工作如下：1） 采用无损伤CT检测实时实验对单轴压缩荷载作用下煤岩破坏全过程损伤扩展特性进行了实验研究，并对其分叉、混沌和自组织特征进行了分析。2） 根据煤岩无损伤CT细观实时实验结果和损伤变量的定义，提出了单轴压缩荷载作用下煤岩损伤演化方程和损伤本构模型。3） 提出了冲击地压的粘滑失稳机理。采用单状态变量本构模型和双状态变量本构模型对煤岩失稳和冲击地压的非线性动力学行为和演化过程进行了研究，得到了冲击地压系统的相图，探讨了冲击地压及其演化过程的混沌特征。4） 根据砚石台煤矿冲击地压的实际情况，通过建立相应的地质力学模型，采用现代非线性科学的突变理论对冲击地压的发生过程进行了突变分析，得出了系统发生冲击地压时受力的临界值及顶底板变形量和能量释放的表达式。5） 根据砚石台煤矿的实际情况，采用分形理论和自组织理论对冲击地压进行了分析。研究结果表明，可根据分形维数对冲击地压的强度进行分析；冲击地压发生的规模与频率符合幂律规则，具有明显的自组织特征。6） 采用人工神经网络和遗传算法相结合的方法，结合砚石台煤矿的实际情况，对冲击地压的预测预报进行了研究，开发了BPAGA TOOLS预测软件。实际应用表明，该方法具有较高的可信度，从而为冲击地压的预测预报探索了一条新的途径。更多还原

【Abstract】 Rockburst is an abrupt dynamics phenomenon which endangers coal mine. With the increase of excavation depth, it is more and more severity. The key problems still exist though lots of studies were made by scholars and technicians. Because of the nonlinear behavior of coal and rock in mine, rockburst is a greatly complex system and the essence of rockburst are nonlinear characteristics. So it is difficult for traditional theories to explain and solve these problems. On the other hand, new theories and methods to settle these matters were applied by modern nonlinear science based on self-organization, fractal, catastrophe, bifurcation, chaos and artificial neural network. Therefore, it is possible to obtain greater breakthrough for the research of rockburst with modern nonlinear science. According to the CT real-time testing, self-organization, fractal, catastrophe, chaos and artificial neural network were used in this thesis. The main contents of this thesis are as follows: 1) The CT real-time testing of coal under uniaxial compression was carried out. The behavior of microscopic was analyzed. The characteristics of bifurcation, chaos and self-organization in the testing were obtained.2) According to the results of the CT real-time microscopic damage propagation testing of coal and the define of damage variable, the constitutive model and evolution function of coal under uniaxial compression were put forward. 3) The mechanism of stick-slip to rockburst was put forward in this thesis. The phase plots and evolution of rockburst system were analyzed particularly with the Single State Variable Model (SSVM) and Tow State Variables Model (TSVM). The characteristics of chaos in rockburst and the dynamics behavior of failure in coal were studied completely in this paper. 4) On the basis of practice instances in Yanshitai coal mine, the process of rockburst is studied with catastrophe in the modern nonlinear science through the establishment of geology model and mechanics model of rockburst in Yanshitai coal mine. The stress threshold when rockburst took place, the distortion value and the expression of energy in rockburst were obtained.5) In view of the realistic conditions in Yanshitai coal mine of Nan Tong mining bureau, rockburst is researched carefully with factual and self-organization. The research shows: fractal dimension is an useful method to rockburst, the scale and the frequency of<WP=6>rockburst accord with Power Law which has obvious self-organization characteristics.6) The new method of rockburst prediction were put out on the basis of BP artificial neural network and genetic algorithm in the realistic conditions in Yanshitai mine. The object-oriented software of BPAGA TOOLS was designed which achieved better results in rockburst prediction. It was a new approach with higher reliability to ruckburst prediction更多还原