STRENGTH PREDICTION AND OPTIMIZATION OF ANIMAL BONE ASH- METAKAOLIN CEMENT BLENDS VIA RESPONSE SURFACE METHODOLOGY

Abstract

This paper tries to investigate the effect of blending animal bone ash (ABA) and
metakaolin (MK) on the mortar compressive strength of ternary cement blends. The
mortar strengths of the ternary cement blends are influenced by several parameters
such as blending ratio, replacement level and curing age. The mortar strength of
the ternary cement blend was optimized via response surface methodology using
central composite design and Box and Behnken design. Both models were compared
to obtain the best strength prediction. Results showed that both models
experimental data for the prediction of the mortar strength satisfied second order
polynomial regression. The mortar strength of ABA-MK-PLC experienced an increase
as the blending ratio between 0.2 -0.5 and curing age between 3 -60 days whereas
a slight reduction in the mortar strength was observed as the replacement level was
increased between 7.5 – 12.5 wt.%. This slight diminution observed despite the high
replacement level further proves that pozzolanic activities occurred. The analysis of
variance (ANOVA) indicated that the curing age significantly affected the response
for both CCD and BBD models respectively with the best mortar strength prediction
via CCD. The ANOVA revealed that predicted R2 (R2pred) and R2 for CCD was very close
in comparison with BBD and thus most suitable to describe the mortar compressive
strength of the cement blends with R2pred of 0.9107. The interactive effect of the
blending ratio, replacement level and curing age were also investigated and
optimized. The optimum condition for the prediction of the mortar strength of the
ternary cement blend were blending ratio 0.423, replacement level 7.565 wt.%, and
curing age 60 days.