Simulation of Correlated Nonstationary Lumber Properties

A stochastic model has been developed for simulating correlated nonstationary spatial variation of modulus of elasticity (MOE) and parallel to grain compressive strength $\left({\sigma }_c\right)$. Two grades of $38\times 89\mathrm{mm}$ spruce‐pine‐fir machine stress‐rated lumber, 2400f‐2.0E and 1650f‐1.5E, have been tested to obtain within‐member MOE and ${\sigma }_c$ profiles. Each piece, 4.88‐m long, has been cut into 152‐mm segments where within member ${\sigma }_c$ values have been obtained. Test results have revealed that the within‐member MOE and ${\sigma }_c$ profiles are correlated and nonstationary. The data have been used to develop and verify a stochastic model. Trend‐removal techniques, spectral representation method, bivariate standard normal simulation model, and moving‐average model have been used to generate within‐member MOE and ${\sigma }_c$ profiles. The correlation structure and nonstationary characteristics of the MOE and strength profiles have been preserved. Good agreement has been obtained between model predictions and test data.