• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2018, Vol. 54 ›› Issue (10): 167-172.doi: 10.3901/JME.2018.10.167

• 运载工程 • 上一篇    下一篇

基于PSO-BP神经网络的盾构刀具配置研究

牛江川1, 韩利涛1, 李素娟2, 郭京波1, 刘进志1   

  1. 1. 石家庄铁道大学机械工程学院 石家庄 050043;
    2. 石家庄铁道大学信息科学与技术学院 石家庄 050043
  • 收稿日期:2017-08-28 修回日期:2018-02-05 出版日期:2018-05-20 发布日期:2018-05-20
  • 通讯作者: 牛江川(通信作者),男,1977年出生,博士,副教授,硕士研究生导师。主要研究方向为计算机辅助设计、机械系统动力学。E-mail:menjc@163.com
  • 基金资助:
    国家自然科学基金(51275321)、国家863计划(2012AA041803)和河北省教育厅自然科学青年基金(QN2014151)资助项目。

Research on Shield Cutting Tool Configuration Based on PSO-BP Neural Network

NIU Jiangchuan1, HAN Litao1, LI Sujuan2, GUO Jingbo1, LIU Jinzhi1   

  1. 1. School of Mechanical Engineering, Shijiazhuang Tiedao University, Shijiazhuang 050043;
    2. School of Information Science and Technology, Shijiazhuang Tiedao University, Shijiazhuang 050043
  • Received:2017-08-28 Revised:2018-02-05 Online:2018-05-20 Published:2018-05-20

摘要: 盾构刀具在盾构机掘进过程中起着关键的作用,其配置选型是否合理决定着工程的成败。为了对盾构刀具进行合理的配置,根据盾构刀具的配置原则,针对盾构刀具配置的地质适用性,在粒子群优化算法(Particle swarm optimization,PSO)与神经网络算法(Back propagation,BP)的基础上提出智能配置方法。建立地质条件与盾构刀具类型之间的关系模型,并利用成功的盾构施工案例作为样本数据对该模型进行训练,训练后可以利用模型智能推荐盾构刀具配置方案。利用工程案例进行测试分析,将测试结果与实际配置方案进行对比,并与BP神经网络测试结果进行比较。测试结果表明,基于PSO-BP神经网络算法不但能够很好地实现盾构刀具配置方案的合理推荐,并且在计算精度和训练时间两个方面PSO-BP神经网络算法比BP神经网络算法都有显著提高。

关键词: BP神经网络, 盾构刀具, 粒子群优化算法, 智能选型

Abstract: Shield cutting tools play a key role in the process of shield machine tunneling, and its reasonable configuration selection affects the success of the project. For the reasonable configuration and the geological applicability of shield cutting tools, an intelligent configuration method based on PSO-BP neural network hybrid algorithm is put forward, where the configuration principle of shield cutting tools is considered. The successful shield construction cases are used as sample data to train the relationship model, which is established between geological conditions and the types of shield cutting tools. And the trained model can achieve intelligent recommendation for the configuration scheme of shield cutting tools. The trained model is tested by engineering case, and the test result is compared with the actual configuration scheme. The test results show that, the PSO-BP neural network algorithm can not only achieve the reasonable recommendation on configuration scheme of shield cutting tools, but also it has significant improvement compared with the BP neural network algorithm in two aspects of calculation accuracy and training time.

Key words: BP neural network, intelligent selection, particle swarm optimization, shield cutting tool

中图分类号: