【作者】 赵同彬；

【导师】 谭云亮；

【作者基本信息】 山东科技大学 ， 采矿工程， 2009， 博士

【摘要】 随着煤炭开采逐步向深部发展,在复杂的应力环境下,深井围岩的流变特性将更加显著,矿山岩层支护控制难度大大增加。本文采用室内试验、理论分析、数值模拟研究为一体的方法,对加锚岩体的蠕变特性和锚固支护变形机理进行了研究。应用RLJW-2000型流变试验机,对加锚前后的深井粉砂岩、红砂岩进行了室内长时蠕变试验。试验结果表明,加锚使岩石的蠕变应力阀值提高30%σc左右,且在各应力水平阶段蠕变量均得到明显控制；加锚后试件的长期强度增加了5%-10%σc破坏后仍具有一定的承载能力。建立了加锚体的蠕变本构模型,推导出了加锚广义开尔文体(B-K)和加锚伯格斯体(B-B)的模型解答,从力学原理的角度说明了锚杆对岩体蠕变变形的控制机理。同时,进行了FLAC软件的二次开发,将2个新的锚固本构模型编译到可调用的动态链接库中,通过数值模型计算表明,锚固本构模型与FLAC锚杆单元作用效果比较相似,并且简化了计算工作量,大大提高了数值软件的运行效率。自主设计了流变拉拔试验装置,在实验室内进行了大量的锚固系统拉拔流变试验,对锚杆与粘结材料界面、基体与粘结材料界面之间的剪应力分布情况进行了同步测试。试验结果表明,锚固界面剪应力分布不均匀,沿轴向呈先增后减的分布形式；两个界面之间的剪应力传递存在不同程度的衰减,但剪应力分布特征基本相近；在长期荷载作用下锚固界面应力的变化与时间相关,影响明显区域主要集中在锚固段的拉拔端和中部；界面失效是锚固系统破坏的主要类型。提出了锚固界面剪应力传递时效性模式,将锚固系统的流变破坏分为粘弹、粘塑、粘脱三个特征阶段。通过试验监测发现,在长期受力状态下,锚杆端头位移经历了弱蠕变和强蠕变两个阶段,锚杆位移的强蠕变与锚固界面的粘脱阶段相对应。较系统的进行了巷道长期稳定性预测研究,采用遗传算法开发了GA-FLAC反分析系统,实现了围岩流变参数的智能反演；采用遗传规划方法开发了GP模型分析系统,对锚固系统失稳破坏的非线性过程进行了模型识别。进化算法的应用提高了反演识别的准确性,从而使深井围岩流变失稳预测结果更加科学和客观。更多还原

【Abstract】 With the gradual development of deep coal mining, the rheological behaviors of deep well’s rock are more significant in a complex stress environment and the mine strata support is more difficult. In this paper, integrated approaches of laboratory tests, theoretical analysis and numerical simulation were adopted to study anchored rock creep properties and anchor support deformation mechanism.Laboratory creep tests of anchored and non-anchored deep pink sandstone, red sandstone were done by using RLJW-2000 rheological testing machine for a long time. The results showed that, bolted rock creep stress threshold was increased by 30% ofσc around, and creep value at each stress level stage was controlled markedly; anchored specimen’s long-term strength was increased by 5%～10% ofσc, it still had a certain load-bearing capacity after destroyed.Anchored body’s creep constitutive model was established, model answers of anchored Generalized Kelvin (B-K) and anchored Burgers (B-B) were derived, the control mechanism of bolted rock creep deformation was illustrated from the view of mechanical rationale. At the same time, FLAC software redevelopment was carried out, in which two new anchored constitutive models were compiled into Dynamic Link Libraries, numerical calculations showed that, the results of anchored constitutive model and bolt unit in FLAC were quite similar, but the calculation workload with the new model was simplified, and the software’s run efficiency was greatly improved.A rheological pull test device was independently designed, a large number of anchored system pulling rheological tests had been done in the laboratory. At the same time, the shear stress distribution on the interface between anchor and bonding material, matrix and bonding material was also investigated. The results showed that, anchorage shear stress distribution on the interface was uneven:first increased and then decreased along the axial direction; then shear stress transmission on the interfaces had different degrees of decay, but the characteristics of shear stress distribution were similar; the anchorage stress on the interface changes were related with time under actions of long-term loads, the significantly affected areas were mainly concentrated in the drawing section and the central section of anchor segments; interface failure was the main type of anchored system damage.Shear stress transmission aging pattern of anchorage interface was proposed, rheological damage of anchored system was divided into three characteristic stages of visco-elastic, visco-plastic and visco-off. Through the testing surveillance, we found that the bolt ends’ displacements were experienced two stages of weak creep and strong creep, bolt displacements’strong creep was corresponded to visco-off phase of anchorage interface.The long-term stability prediction of roadway was carried out, GA-FLAC anti-analysis system was developed by using Genetic Algorithm, the rheological parameters’smart inversion of surrounding rocks was realized; GP model analysis system was developed by using Genetic Programming method, model identification about the non-linear process of anchored system’s instability failure was done. The application of Evolutionary Algorithms improved the accuracy of inversion recognition, made results of deep rock’s rheological instability predictions more scientific and objective.更多还原