地震引起建筑结构损伤可靠性预测仿真

Prediction and Simulation of Damage Reliability of Building Structure Caused by Earthquake

对建筑结构损伤可靠性进行准确预测,可提高建筑结构的安全稳定性。进行建筑结构损伤可靠性预测时,应建立相关的损伤准则,获取损伤建筑结构的可靠性指标,完成建筑结构损伤可靠性预测,但是传统方法通过计算损伤建筑结构因数的平均值和方差,完成可靠性预测,但是不能依据相关的损伤准则,无法建立准确的损伤建筑结构的可靠性指标,降低了可靠性预测的精度。提出一种基于神经网络的地震引起建筑结构损伤可靠性预测方法。上述方法将建筑结构的重点因数提取出来构建损伤准则,将损伤建筑结构的临界值作为损伤准则的安全极限值。再采用BP神经网络-蒙特卡洛计算法,求得损伤准则的安全极限值,最终得出损伤建筑结构的可靠性指标,完成建筑结构损伤可靠性预测。仿真结果证明,所提方法能够为地震损伤建筑结构提供高精度的可靠性指标,进而准确分析建筑结构的可靠性。

It can improve the safety and stability of building structure to predict the reliability of building structure damage accurately. The reliability prediction should build correlative damage criterion and obtain reliability index to complete the prediction. However, traditional method completes the reliability prediction through calculating the average value and variance of damaging building structure factor to complete the prediction, it cannot build accurate relia-bility index of building structure according to correlative damage criterion. Therefore, it reduces the prediction preci- sion. In this paper, a reliability prediction method of building structure damage caused by earthquake based on the neural network is proposed. Firstly, the emphasis factor of building structure is extracted to build damage criterion and the critical value of building structure damage is used as the security limiting value of damage criterion. Then, the BP neural network - Monte Carlo algorithm is used to acquire the security limiting value of damage criterion. Finally, the reliability index of building structure damage is reached and the reliability prediction is completed. The simulation results show that the method mentioned above can provide high-precision reliability index for building structure damage caused by earthquake and then analyze building structure reliability accurately.