期刊文献+

基于BP神经网络的周口店遗址裂缝变形监测数据分析与预测

Analysis and prediction of crack deformation monitoring data of Zhoukoudian site based on back propagation neural network
下载PDF
导出
摘要 周口店遗址是世界范围内更新世古人类遗址中内涵较丰富、材料较齐全、较有科研价值的遗址之一,是我国古人类学、旧石器时代考古学和第四纪地质学的科研基地。国家对周口店遗址的长期保护和监测,留存了大量的不同类型的监测数据,但是长期以来对这些监测数据缺乏深层次的数据挖掘和高效利用。基于周口店遗址的地质资料和实地勘察,利用反向传播(back propagation,BP)神经网络方法,运用时间序列预测原理,结合周口店遗址的原始监测数据,对周口店遗址第3地点裂缝变形状况展开了预测研究,并将预测研究分为补全缺失数据和预测未来数据2个方面。结果表明,基于BP神经网络和周口店遗址监测数据所构建的预测模型,经过大量的数据训练,具有较高的预测精度,能够通过预测实现对监测数据缺失部分的补全工作和对监测数据未来发展的预测工作。该研究成果为周口店遗址监测数据的处理提供了一种新方法,丰富了周口店遗址监测中心对大量监测数据的处理手段;此外,构建的周口店遗址第3地点的裂缝变形预警等级结构,对周口店遗址后续预防性保护工作的开展具有重要意义。 Zhoukoudian site is one of the sites with rich connotation,complete materials and scientific research value in the Pleistocene ancient human sites around the world.It is the scientific research base of paleoanthropology,Paleolithic archaeology and Quaternary geology in China.Under the long-term protection and monitoring of Zhoukoudian site by the state,a large number of different types of monitoring data have been retained,but there has been a lack of deep data mining and efficient utilization of these monitoring data for a long time.Based on the geological data and field investigation of Zhoukoudian site,this paper uses the back propagation neural network method and the time series prediction principle,and combines the original monitoring data of Zhoukoudian site to predict the fracture deformation of the third site of Zhoukoudian site.The prediction research is divided into two aspects:completing missing data and predicting future data.The results show that the prediction model based on BP neural network and the monitoring data of Zhoukoudian site has high prediction accuracy after a lot of data training,and can realize the completion of the missing part of the monitoring data and the prediction of the future development of the monitoring data through prediction.The research results provide a new processing method for the monitoring data of Zhoukoudian site,and enrich the processing means of a large number of monitoring data by Zhoukoudian site monitoring center.The early warning system of crack deformation in the third site of Zhoukoudian site is of great significance to the follow-up preventive protection work of Zhoukoudian site.
作者 康凯 孟雨萱 郑健 付前方 许国平 崔德山 KANG Kai;MENG Yuxuan;ZHENG Jian;FU Qianfang;XU Guoping;CUI Deshan(Beijing Zhoukoudian Peking Man Site Management Office,Beijing 102405,China;Beijing Putian Tongda Technology Company,Beijing 102200,China;Beijing Yuanzhen Online Monitoring Technology Company,Beijing 102200,China;Faculty of Engineering,China University of Geosciences(Wuhan),Wuhan 430074,China)
出处 《安全与环境工程》 CAS CSCD 北大核心 2024年第6期169-178,共10页 Safety and Environmental Engineering
基金 国家自然科学基金项目(42277171)。
关键词 裂缝变形 BP神经网络 时间序列预测 周口店遗址 crack deformation back propagation neural network time series prediction Zhoukoudian site
  • 相关文献

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部