Online Program Home
My Program

Abstract Details

Activity Number: 493 - New Statistical Methods for Lumber Analytics
Type: Invited
Date/Time: Wednesday, August 1, 2018 : 10:30 AM to 12:20 PM
Sponsor: SSC
Abstract #326606
Title: Bayesian Parametric Models Without Likelihoods: Assessing Accumulated Damage in Forest Products
Author(s): Samuel WK Wong* and Jim Zidek and Chun-Hao Yang
Companies: University of Florida and University of British Columbia and University of Florida
Keywords: approximate Bayesian computation; accumulated damage models; duration of load; lumber reliability

The long-term reliability of lumber is an important consideration for its use in construction. Empirical investigations since the 1950s have shown that forest products weaken over time when subjected to sustained loads, and this effect must be considered to ensure the safety of wood-based structures. This led to the development of damage models that enable the strength of lumber under future loadings to be characterized, and these models have informed the setting of standards for structural safety since the 1980s. The formulation of the classic damage models is via an ODE that describes the rate of damage accumulation in a structural member over time, and a likelihood-based analysis is not straightforward. Recently, there is renewed interest in damage modeling with the development of new engineered lumber products such as cross laminated timber and strand-based wood composites. In this talk, I will present an approximate Bayesian approach to calibrate these damage models from experimental data and generate reliability assessments. Some alternative ways to model damage are also suggested.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2018 program