<p>The study investigates the interdependency between the wide range of single parameter-based ground motion intensity measures with demand parameters for low-rise reinforced concrete (RC) structures in terms of efficiency and sufficiency. Further to evaluate the the seismic vulnerability of RC structures due to pulse-like (PL) and non-pulse-like (NPL) ground motions, the fragility curves were computed over several IMs, to investigate the likelihood of exceeding the performance levels as a function of ground motion intensity and to identify the optimal demand parameter. Additionally, the Latin hypercube sampling technique was employed to evaluate the ground motion and material uncertainty and to identify the most sensitive material property. To accomplish this goal, two sets of near-fault ground motions consisting of PL and NPL&#xa0;ground motions, compiled from the active tectonic regions worldwide, were utilized to evaluate the seismic response of the structures during a severe seismic event. The results from efficiency, sufficiency, and fragility curves indicate velocity spectrum intensity is the most promising demand parameter. Further, the effect of material uncertainty on the response of structure was found to be higher for PL ground motions with PGA levels above 0.5&#xa0;g. Furthermore, the study examines a feedforward neural network using the Levenberg–Maruardt algorithm to forecast the Pak–Ang damage index for PL ground motions rather than the&#xa0;time-consuming non-linear dynamic analysis.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Exploring the association of ground motion intensity measures and demand parameters with ANN-based predictive modeling and uncertainty analysis

  • Faisal Mehraj Wani,
  • Jayaprakash Vemuri

摘要

The study investigates the interdependency between the wide range of single parameter-based ground motion intensity measures with demand parameters for low-rise reinforced concrete (RC) structures in terms of efficiency and sufficiency. Further to evaluate the the seismic vulnerability of RC structures due to pulse-like (PL) and non-pulse-like (NPL) ground motions, the fragility curves were computed over several IMs, to investigate the likelihood of exceeding the performance levels as a function of ground motion intensity and to identify the optimal demand parameter. Additionally, the Latin hypercube sampling technique was employed to evaluate the ground motion and material uncertainty and to identify the most sensitive material property. To accomplish this goal, two sets of near-fault ground motions consisting of PL and NPL ground motions, compiled from the active tectonic regions worldwide, were utilized to evaluate the seismic response of the structures during a severe seismic event. The results from efficiency, sufficiency, and fragility curves indicate velocity spectrum intensity is the most promising demand parameter. Further, the effect of material uncertainty on the response of structure was found to be higher for PL ground motions with PGA levels above 0.5 g. Furthermore, the study examines a feedforward neural network using the Levenberg–Maruardt algorithm to forecast the Pak–Ang damage index for PL ground motions rather than the time-consuming non-linear dynamic analysis.