Dynamic estimation and reliability control of quality assurance for bored piles
来源期刊：中南大学学报(英文版)2014年第3期
论文作者：ZHAO Wen-yi(赵文艺) XU Zhi-jun(徐志军) ZHENG Jun-jie(郑俊杰) BIAN Xiao-ya(边晓亚)
文章页码：1148 - 1157
Key words：bored piles; quality assurance inspection; reliability control; occurrence probability; Bayesian method
Abstract: A systematic method was proposed to estimate the occurrence probability of defective piles (OPDP) from a site according to quality assurance inspection. The OPDP was firstly suggested as the criterion to weight the performance of a pile foundation. Its prior distribution and updating distribution were deduced to follow Beta distributions. To calibrate the OPDP, a dynamic estimation model was established according to the relationships between prior mean and variance and updating mean and variance. Finally, a reliability-control method dealing with uncertainties arising from quality assurance inspection was formalized to judge whether all the bored piles from a site can be accepted. It is exemplified that the OPDP can be substantially improved when more definite prior information and sampling formation become available. For the example studied herein, the Bayesian estimator of updating variance for OPDP is reduced from 0.0037 to 0.0014 for the first inspection, from 0.0014 to 0.0009 for the second inspection, and with less uncertainty by incorporating experience information.
ZHAO Wen-yi(赵文艺)1, XU Zhi-jun(徐志军)2, ZHENG Jun-jie(郑俊杰)1, BIAN Xiao-ya(边晓亚)1
(1. School of Civil Engineering & Mechanics, Huazhong University of Science and Technology,
Wuhan 430074, China;
2. School of Civil Engineering and Architecture, Henan University of Technology, Zhengzhou 450001, China)
Abstract:A systematic method was proposed to estimate the occurrence probability of defective piles (OPDP) from a site according to quality assurance inspection. The OPDP was firstly suggested as the criterion to weight the performance of a pile foundation. Its prior distribution and updating distribution were deduced to follow Beta distributions. To calibrate the OPDP, a dynamic estimation model was established according to the relationships between prior mean and variance and updating mean and variance. Finally, a reliability-control method dealing with uncertainties arising from quality assurance inspection was formalized to judge whether all the bored piles from a site can be accepted. It is exemplified that the OPDP can be substantially improved when more definite prior information and sampling formation become available. For the example studied herein, the Bayesian estimator of updating variance for OPDP is reduced from 0.0037 to 0.0014 for the first inspection, from 0.0014 to 0.0009 for the second inspection, and with less uncertainty by incorporating experience information.
Key words:bored piles; quality assurance inspection; reliability control; occurrence probability; Bayesian method
