Date of Award


Document Type

Dissertation - Restricted

Degree Name

Doctor of Philosophy (PhD)


Civil, Construction, and Environmental Engineering

First Advisor

Keith F. Faherty

Second Advisor

S. Vinnakota

Third Advisor

Thomas H. Wenzel

Fourth Advisor

Stephen M. Heinrich

Fifth Advisor

David W. Green


This study was undertaken in an effort to derive allowable stresses for three wood species grown in Rwanda. Wood properties will vary between species, between locations within a species and even within a single tree. This variability must be well understood in order to establish design values for wood and wood composites. Estimates of the mean and the near-minimum values of small, clear samples of wood were obtained using the international testing standards (ISO). An indepth review of current sampling techniques used in wood engineering strongly suggests that the random technique of sampling and its variations remains a reliable technique which, when designed properly, will yield accurate estimates. Several samples of the three timber species chosen by using the random technique of sampling were tested for both the physical (density, specific gravity, swelling and shrinkage characteristics) and the mechanical properties (bending strength and stiffness, compressive strength parallel to grain, tensile strength parallel to grain and shear strength parallel to grain) using universal testing machines available both at the National University of Rwanda (UNR) and at Marquette University. Unadjusted for end use data were obtained in the unseasoned and at 12 percent moisture content. A randomized two factor factorial analysis was performed in an effort to quantify the effects of a particular species and the forest location from where it was sampled on the value of the property mean. The conclusion reached was that not enough evidence existed to support or reject the hypothesis that forest location and species have any effect on the strength values of the wood sampled. The effects of each factor individually and the interaction effect were not significant at the 5% point level. Using the Australian and the Andean Pact Structural Grading Rules and data obtained in this study, a grading rule is proposed for structural lumber in Rwanda. This Grading Rule describes three structural grades based upon the tolerance limits established on the strength-reducing factors. Allowable design values are derived for the three timber species for bending, compression, shear, tension and modulus of elasticity using small, clear data and a number of adjustment factors. Furthermore, a species independent stress system is proposed with the primary objective of reducing the volume of information that in the past faced the engineer and other wood end-users. Regression and correlation analysis performed to obtain the relationships between the primary property (MOR) and the secondary properties suggests that strength grouping is desirable not only for its marketing and technological flexibilities, but also for the optimization gained in the structural use of the timber resource. Based upon the findings of this study, a number of recommendations are made for future research in wood engineering in Rwanda.



Restricted Access Item

Having trouble?