14th International Conference on Applications of Statistics and Probability in Civil Engineering
This paper presents a reliability-based study on the duration of load (DOL) effect in timber members under wind loads. First, an auto regressive and moving average (ARMA) time-series model was established to simulate 50 years wind speed record in three Canadian cities. The model parameters were calibrated for each city based on historical data from the GEM-LAM platform of Environment Canada. Comparisons on the statistics and probability distribution of hourly wind speeds and yearly maximum wind speeds were conducted to validate the use of ARMA model. Subsequently, reliability analysis was carried out to evaluate the short-term and long-term reliability of timber members under simulated wind loads. The process considered the randomness of wind loads and the damage accumulation behaviour of timber. Finally, recommended values of the load duration factor under wind loads were obtained and compared with the values in the current Canadian design standard.