• PA Owoeye
  • A Usman
  • OS Abejide
Keywords: ANOVA, bridge decks, Flanagan polynomial, Nigerian timber, regression, reliability, stringer spacing


This presentation accounts for established functions of stringer spacing of major Nigerian woods when used as bridge decks. It entails stochastic evaluation of bridge wood decks with absolute regards to the spacing of supporting stringers. A timber bridge deck is modelled on timber stringers in accordance with current specifications as outline in AASHTO LRFD (2010), to represent real life experiment in order to depict the structural behaviour of Nigerian timbers when used for bridge decks. This model is then subjected to some degree of entropy using Advance Second Moment Reliability Assessment (ASMRA) method, which is subsequently analysed using JAVA library with the help of Flanagan polynomial. The concept of ASMRA with Flanagan polynomial returns optimum output values for any input array of data along the input normal distribution curve. Taking Lophira alata (EKKI), Afzelia bipindensis (APA),Chlorophora  exceisa (IROKO) and Mitragyna ciliate (ABURA) to represent N1, N2, N3, and N4 classes of Nigerian timber suitable for bridge decks, convincingly, it was established that, stringer spacing, strength classes, timber thicknesses and width are some of the major factors among others influencing the structural behaviour of Nigerian timber species used for bridge decks. These data are treated as random variables to generate relative optimum values which, are used to predict the relationship between stringer spacing as a dependent variable on other variables using Analysis of Variance (ANOVA) and multiple Regression line analysis. For the considered timber species, it was observed that large percentages, (R2 ≌ 1 that is  100% for EKKI, APA, IROKO and ABURA of the variation in the dependent variable (stringer spacing) is explained by the independent variables (strength classes, timber thicknesses, timber width, plank deck unit weight, unit weight of surfacing material, lane load, thickness of surfacing material  and reliability indices) which are all good fit. Also all the predicted regression lines are reliable and statistically significant as the significance F, (3.59709E-05 for EKKI, 5.58768E-05 for APA, 3.59709E-05 for IROKO and 8.55563E-05 for ABURA) are all less than 0.05. Thus, these established relationships will help in future forecast of stringer spacing for bridge deck design and analysis within acceptable structural reliability indices which are statistically significant (and are all within acceptable values ≤0.05)

Building, Civil & Geotechical Engineering

Journal Identifiers

eISSN: 2467-8821
print ISSN: 0331-8443