基于AI算法的隧道衬砌冷缝检测分类研究  

Research on the Detection and Classification of Cold Joints in Tunnel Lining Based on AI Algorithm

作　　者：冯源 邓立 路景海 邓愿涛 孙武鹏 温先划 朱洪谷 吴佳晔 FENG Yuan;DENG Li;LU Jinghai;DENG Yuantao;SUN Wupeng;WEN Xianhua;ZHU Honggu;WU Jiaye(Sichuan Central Inspection Technology Co.Ltd.,Chengdu 610041,China;Four Bureau of China Railway Group Ltd.,Hefei 236000,China;Yungui Railway Guangxi Co.Ltd.,Nanning 530029,China;China Railway Shanghai Group Co.Ltd.,Shanghai 200071,China;China Railway Nanning Group Co.Ltd.,Nanning 530029,China;Southwest Petroleum University,Chengdu 610500,China)

机构地区：[1]四川升拓检测技术股份有限公司,成都610041 [2]中铁四局集团有限公司,合肥236000 [3]云桂铁路广西有限责任公司,南宁530029 [4]中国铁路上海局集团有限公司,上海200071 [5]中国铁路南宁局集团有限公司,南宁530029 [6]西南石油大学,成都610500

出　　处：《铁道建筑》2025年第1期99-103,共5页Railway Engineering

摘　　要：针对隧道衬砌冷缝,基于冲击弹性波面波的检测技术得到了较为广泛的应用,但其解析难度大,效率低。本文提出一种基于响应函数、核主成分分析法和核极限学习机的人工智能冷缝检测分类模型,实现了数据的自动分析及冷缝分类判别。进而通过对极限学习机关键参数优化方法的选取和改进,提升了模型的预测精度和泛化能力。实际验证结果表明,本文给出的方法具有较高的预测精度,且对冷缝有较低的漏检率,能够有效提升面波法检测衬砌冷缝的解析效率和精度。For the detection of cold joints in tunnel lining,the detection technology based on impact elastic wave surface waves has been widely applied,but it is difficult to be analyzed and has low efficiency.Therefore,this paper proposes an AI cold joint detection and classification model based on response function,kernel principal component analysis,and kernel extreme learning machine,which achieves automatic analysis of data and classification and discrimination of cold joints.Further,by selecting and improving the key parameters optimization method of extreme learning machine,the prediction accuracy and generalization ability of the model are improved.The actual verification results show that,it has high prediction accuracy and low false negative rate for cold joints,which can effectively improve the analysis efficiency and accuracy of surface wave method for detecting cold joints in lining.

关 键 词：隧道衬砌 冷缝 检测 海鸥优化算法 核极限学习机 

分 类 号：U456[建筑科学—桥梁与隧道工程]