...
  • 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Pure and Applied Geophysics >Ground Motion Prediction Model Using Artificial Neural Network
【24h】

Ground Motion Prediction Model Using Artificial Neural Network

机译：使用人工神经网络接地运动预测模型

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Abstract This article focuses on developing a ground motion prediction equation based on artificial neural network (ANN) technique for shallow crustal earthquakes. A hybrid technique combining genetic algorithm and Levenberg–Marquardt technique is used for training the model. The present model is developed to predict peak ground velocity, and 5% damped spectral acceleration. The input parameters for the prediction are moment magnitude ( M ~(w)), closest distance to rupture plane ( R ~(rup)), shear wave velocity in the region ( V ~(s30)) and focal mechanism ( F ). A total of 13,552 ground motion records from 288 earthquakes provided by the updated NGA-West2 database released by Pacific Engineering Research Center are utilized to develop the model. The ANN architecture considered for the model consists of 192 unknowns including weights and biases of all the interconnected nodes. The performance of the model is observed to be within the prescribed error limits. In addition, the results from the study are found to be comparable with the existing relations in the global database. The developed model is further demonstrated by estimating site-specific response spectra for Shimla city located in Himalayan region.
机译：摘要本文侧重于开发基于人工神经网络（ANN）技术的地面运动预测方程进行浅地震。结合遗传算法和Levenberg-Marquardt技术的混合技术用于训练模型。开发了本模型以预测峰地速度，5％阻尼光谱加速度。预测的输入参数是矩幅度（m〜（w）），最接近的破裂平面（R〜（Rup）），区域中的剪切波速度（V〜（S30））和焦点机制（F）。由太平洋工程研究中心发布的更新的NGA-WIST2数据库提供的288个地震中共有13,552个地面运动记录，以开发模型。所考虑的模型的ANN架构由192个未知数组成，包括所有互连节点的权重和偏差。观察到模型的性能在规定的错误限制范围内。此外，发现该研究的结果与全球数据库中的现有关系相当。通过估计位于喜马拉雅地区的Shimla城市的场地特定响应光谱，进一步证明了开发的模型。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号