LAUSR

Abstract Indonesian people traditionally use Gigantochloa apus bamboo as structural materials. For modern building design, structural grading should be conducted on bamboo culm to get its design characteristic values. Non-destructive assesment on each culm condition, dimension, and geometric were conducted to predict the strength and capacity. Wall density ( ρ w ) and Modulus of Elasticity measured by fixed load deflection using Panter machine ( E p ) have moderate correlation with apparent and true modulus of elasticity ( E app , E true ) and Modulus of Rupture (MOR). Additional predictors in multiple regression improved the adj- R 2 , but it was not reliable enough for estimating E app , E true , and MOR of G. apus culm, thus bamboo structural grading should better use capacity grading rather than strength grading. Diameter ( D ) is a potential grading parameter that provides strong evidence to predict flexural capacity. Non destructive test which deals with measuring linear mass ratio ( q ) is a simple method that refers to dimension and density, and it has strong correlation with flexural rigidity ( EI app , EI true ) and capacity ( M max ). The measurements of both D and q are proper indicating variables for predicting the capacity of G. apus. The grade of structural bamboo could be classified by D , q, and combination of both measurement based on ISO 22156 and confident band method. Confident band method resulted in a more conventional value than that from ISO 22156, thus it proved to be safer and more reliable. The combination of linear mass and square of diameter ( qD 2 ) was the best predictor for estimating the stiffness, while qD was the best one for estimating M max . In capacity grading, additional measurement of eccentricity ( E c ), culm density ( ρ c ) , wall density ( ρ w ) , moisture content ( M c ) , and ovality ( O v ) predictors significantly improved the model rather than a single predictor, but fixed load bending stiffness measured using panter machine ( EI p ), taper ( t a ), and out of straightness ( s o ) addition into the model did not give significant contribution.