土工基础

›› 2016, Vol. 30 ›› Issue (2): 223-226.

• 工程实录 • 上一篇    下一篇

BP神经网络在工业厂房变形预测中的应用

石 荔，黄 涛   

  1. （中冶集团武汉勘察研究院有限公司，武汉 430080）
  • 收稿日期:2015-12-30 出版日期:2016-04-20 发布日期:2016-04-27
  • 作者简介:石 荔（1988-），女，硕士研究生，助理工程师，研究方向为地质勘察等。 

Industrial Building Deformation Predictions by BP Neural Network

SHI Li, HUANG Tao    

  1. （Wuhan SurveyGeotechnical Research Institute Co. Ltd. Of MCC, Wuhan 430080）
  • Received:2015-12-30 Online:2016-04-20 Published:2016-04-27

摘要： 对于正在使用的工业厂房的安全控制，日常维护和监控起着重要的作用。结合既有的监测数据，利用数值分析软件MATLAB和BP神经网络基本原理建立实时分析模型，即以现有监测数据为训练样本，对神经网络进行训练，利用训练后的网络进行预测。以某钢厂500 t废酸水处理车间监测数据为例，利用BP神经网络进行预测，预测结果和实际位移数据相吻合，结果表明了神经网络对监测数据进行处理和预测的可行性，对监测工作起到指导和预警作用。

关键词: 监测数据, BP神经网络, MATLAB

Abstract: The safety control, scheduled maintenance and monitoring of the industrial buildings under their normal operation are the very important tasks. This paper presents a case history of the comparison of the deformation data between the predicted and observed. The existing monitored data were used as the training samples for the neural network system. After the training, the deformations of the operational buildings could be predicted by the neural network. The deformation monitoring data on a 500tonne waste water treatment shop were used by the neural network and the future deformation of the shop was predicted. The results between the monitoring and predicted are consistent.  

Key words: Monitoring data, BP Neural Network, MATLAB