人工神经网络模型在地震安全评价中的应用
|
摘要
地震影响系数是地震安全评价的一个重要参数 ,其分布受到场地条件和基岩条件的共同影响 ,对场地影响的考虑相对比较成熟 ,对基岩的影响则相对考虑较少 ,难以满足城市和重大工程设防的要求。综合考虑基岩条件和场地条件的地震影响系数计算 ,关键在于确定基岩裂缝、能量、位移、场地等指标的分布与地震影响系数之间的关系 ,由于这种非线性关系需要构建多元非线性模型 ,故通过应用人工神经网络方法 ,得到地震影响系数计算模型。以唐山市为实例 ,进行了计算并分析了计算结果 ,为城市规划和工程建设提出了几点建议
Earthquake affecting coefficient was one of the most important parameters for seismic safety evaluation, the distribution of which was controlled by both the site condition and basement rock condition. More attentions had been paid on the influence of site condition; however, the impact of basement rock condition had not been much considered which could not meet the anti seismic requirements of urban construction. In comprehensive coefficient calculation of both impacts of site condition and basement rock condition, the key point was to determine the distribution of fracture, energy, displacement and site, and their correlation with the seismic impact coefficient. These non linear relations would require setting up multiple elements model, now the neural network method is applied to establish a calculation model for seismic coefficient. Taking Tangshan city as an example, the distribution of earthquake affecting coefficient was calculated and the results analyzed. Finally, some suggestions for city planning and project construction were provided.
Earthquake affecting coefficient was one of the most important parameters for seismic safety evaluation, the distribution of which was controlled by both the site condition and basement rock condition. More attentions had been paid on the influence of site condition; however, the impact of basement rock condition had not been much considered which could not meet the anti seismic requirements of urban construction. In comprehensive coefficient calculation of both impacts of site condition and basement rock condition, the key point was to determine the distribution of fracture, energy, displacement and site, and their correlation with the seismic impact coefficient. These non linear relations would require setting up multiple elements model, now the neural network method is applied to establish a calculation model for seismic coefficient. Taking Tangshan city as an example, the distribution of earthquake affecting coefficient was calculated and the results analyzed. Finally, some suggestions for city planning and project construction were provided.
引文
1 李小军,彭清.不同类别场地地震动参数的计算分析[J].地震工程与工程震动,2001,21(1):29～36
2 黄明利,唐春安,朱万成.岩石破裂过程的数值模拟研究[J].岩石力学与工程学报,2000,19(4):468～471
3 陈艳华,朱庆杰,苏幼坡.基于格里菲斯准则的地下岩体天然裂缝分布的有限元模拟研究[J].岩石力学与工程学报,2003,22(3):364～369
4 朱庆杰,苏幼坡,刘廷权.灾害危险性评价的神经网络预测模型[J],电子与信息学报,2003年,25(增刊):17～22
5 朱庆杰,姜耀俭,陈艳华,苏幼坡.轮南地区奥陶系天然裂缝的发育机理及其对流体活动的控制[J].中国安全科学学报,2003,13(1):53～58
2 黄明利,唐春安,朱万成.岩石破裂过程的数值模拟研究[J].岩石力学与工程学报,2000,19(4):468～471
3 陈艳华,朱庆杰,苏幼坡.基于格里菲斯准则的地下岩体天然裂缝分布的有限元模拟研究[J].岩石力学与工程学报,2003,22(3):364～369
4 朱庆杰,苏幼坡,刘廷权.灾害危险性评价的神经网络预测模型[J],电子与信息学报,2003年,25(增刊):17～22
5 朱庆杰,姜耀俭,陈艳华,苏幼坡.轮南地区奥陶系天然裂缝的发育机理及其对流体活动的控制[J].中国安全科学学报,2003,13(1):53～58
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心