This paper presents a probabilistic approach for studying the reliability of cementless hip prostheses in the presence of mechanical uncertainties and its application to the investigation of the influence of bone-impl...This paper presents a probabilistic approach for studying the reliability of cementless hip prostheses in the presence of mechanical uncertainties and its application to the investigation of the influence of bone-implant interface properties. The non-linear deterministic model of the bone-implant coupled system and its finite element implementation are described, and the proposed reliability analysis is exposed. It is demonstrated that the distribution (uniform, truncated Gaussian and truncated lognormal distribution) of the two chosen parameters and the truncation lengths have a minor influence on the Hasofer-Lind index. This index logically increases as the failure threshold increases. FORM and SORM approximations are compared with the results obtained using a crude Monte-Carlo method for the estimation of failure probability. The performance of three Monte-Carlo methods is studied in terms of the necessary number of FE calculations. The method based on the Directional Simulation (DS) technique is efficient and less time-consuming. The validity and operational capacity of the proposed approach would not be compromised by an increase in the number of uncertain parameters.展开更多
文摘This paper presents a probabilistic approach for studying the reliability of cementless hip prostheses in the presence of mechanical uncertainties and its application to the investigation of the influence of bone-implant interface properties. The non-linear deterministic model of the bone-implant coupled system and its finite element implementation are described, and the proposed reliability analysis is exposed. It is demonstrated that the distribution (uniform, truncated Gaussian and truncated lognormal distribution) of the two chosen parameters and the truncation lengths have a minor influence on the Hasofer-Lind index. This index logically increases as the failure threshold increases. FORM and SORM approximations are compared with the results obtained using a crude Monte-Carlo method for the estimation of failure probability. The performance of three Monte-Carlo methods is studied in terms of the necessary number of FE calculations. The method based on the Directional Simulation (DS) technique is efficient and less time-consuming. The validity and operational capacity of the proposed approach would not be compromised by an increase in the number of uncertain parameters.