孙光华，齐育珠，刘志义，等. 基于k-均值聚类的红砂岩循环加卸载下破坏前兆研究［J］.河南理工大学学报（自然科学版），doi:10.16186/j.cnki.1673-9787.2024070082

SUN G H, QI Y Z, LIU Z Y,et al.Study of damage precursors under cyclic loading and unloading in red sandstone based on K-means clustering［J］.Journal of Henan Polytechnic University（Natural Science），doi: 10.16186/j.cnki.1673-9787.2024070082

基于k-均值聚类的红砂岩循环加卸载下破坏前兆研究(网络首发)

孙光华^1,2,3，齐育珠^1,2，刘志义^1,2,3，韩凯明^1,2，郭将^1,2，吴柄纬^1,2

(1.华北理工大学 矿业工程学院，河北 唐山 063210；2.华北理工大学 唐山市矿业开采与安全技术重点实验室，河北 唐山 063210；3.华北理工大学 河北省矿业开发与安全技术实验室，河北 唐山 063210)

摘要: [目的] 工程实践中岩石常处于循环加卸载状态，针对探究循环加卸载下红砂岩的破坏前兆问题， [方法] 进行了循环加卸载下红砂岩声发射力学监测试验，采用k-均值聚类算法方法，进行了声发射特征参数聚类分析研究岩石损伤破裂模式。[结果] 结果表明：声发射事件主要发生在循环加载中，卸载阶段声发射事件发生极少并进入“间歇期”，随着循环次数积累峰值应力逐步增高，声发射事件数量突增并进入“活跃期”；循环加卸载初期声发射b值最大，随着循环次数的增加，声发射b值逐级降低并稳定在0.75~1.25区间内，在岩石进入破裂阶段前声发射b值出现骤降，由于每次循环加载开始前声发射b值也出现大幅度骤降现象，岩石却并未破裂，将声发射b值骤降作为岩石破坏前兆的判断依据存在局限性，需结合其他现象研究岩石破裂前兆；根据k-均值聚类将岩石损伤破裂过程的声发射信号分为3类即聚类1、聚类2和聚类3，每个聚类对应不同损伤阶段。循环加卸载末期，聚类3声发射事件能量陡然上升，反映出加载末期砂岩破裂尺度明显增大直至破坏，可以作为岩石的破坏前兆特征。[结论] 分析发现聚类3声发射信号在岩石弹性变形阶段零星出现，随着循环积累，在岩石完全破坏阶段前声发射信号持续密集出现，将声发射b值和聚类信号结合分析来提升预测岩石损伤破裂的准确率。

关键词: 循环加卸载；声发射；b值分析；k-均值聚类；岩石破坏前兆

中图分类号：TD853

doi: 10.16186/j.cnki.1673-9787.2024070082

基金项目: 河北省自然科学基金资助项目（E2024209129）国家自然科学基金资助项目（52204134）

收稿日期：2024-07-24

修回日期：2024-09-20

网络首发日期：2024-11-01

Study of damage precursors under cyclic loading and unloading in red sandstone based on K-means clustering

SUN Guanghua^1,2,3, QI Yuzhu^1,2, LIU Zhiyi^1,2,3, HAN Kaiming^1,2, GUO Jiang^1,2, WU Bingwei^1,2

(1.School of Mining Engineering, North China University of Science and Technology, Tangshan 063210, China；2. School of Mining Engineering, Tangshan City Key Laboratory of Mining Development and Security Technology, Tangshan 063210, China；3. School of Mining Engineering, Hebei Provincial Mining Development and Safety Technology Laboratory, Tangshan 063210, China)

Abstract: [Objective] In engineering practice, rocks are often in the state of cyclic loading and unloading, and in order to explore the problem of failure precursors of red sandstone under cyclic loading and unloading, [Methods] The acoustic emission mechanical monitoring test of red sandstone under cyclic loading and unloading was carried out, and the cluster analysis of acoustic emission characteristic parameters was carried out by using the k-means clustering algorithm to study the rock damage and fracture mode. [Results] The results show that the AE events mainly occur during cyclic loading, and the AE events in the unloading stage occur very rarely and enter the "intermittent period", and the number of AE events increases suddenly and enters the "active phase" with the accumulation of the number of cycles. The B value of AE is the largest in the initial stage of cyclic loading and unloading, and with the increase of the number of cycles, the B value of AE decreases step by step and stabilizes in the range of 0.75~1.25, and the B value of AE decreases sharply before the rock enters the rupture stage. According to the k-means clustering, the acoustic emission signals during rock damage and rupture were divided into three categories, namely cluster 1, cluster 2 and cluster 3, and each cluster corresponded to different damage stages. At the end of the cyclic loading and unloading stage, the energy of the cluster 3 acoustic emission event increased sharply, reflecting the obvious increase of the fracture scale of the sandstone at the end of loading until the failure, which can be used as a precursor to the failure of the rock. [Conclusion] The analysis shows that the acoustic emission signals of cluster 3 appear sporadically in the elastic deformation stage of rocks, and with the accumulation of cycles, the acoustic emission signals continue to appear intensively before the complete failure stage of rocks.

Key words: cyclic loading and unloading; acoustic emission; b-value analysis; cluster analysis; precursors of rock destruction

CLC: