TY - JOUR
T1 - Green interlocking paving units
AU - Sojobi, A. O.
AU - Aladegboye, O. J.
AU - Awolusi, T. F.
PY - 2018/6/10
Y1 - 2018/6/10
N2 - Our research produced ultra-lightweight green interlocking paving units through combined usage of sawdust and laterite as partial cement and fine aggregate replacements respectively. Interlocking paving units, water-cured for 90 days at optimum 10% sawdust content, achieved compressive strength of 16.6 MPa and skid resistance of 64.5 PVT, which exceeded the minimum requirements of 3.45–15 MPa and 45PVT respectively. The developed ANN model was classified as very good in terms of prediction efficiency of compressive strength (CS) and skid resistance for different curing media based on Nash & Sutcliffe criterion of efficiency (NSE). However, the prediction efficiency for CS was better than SR in terms of NSE, root-mean-square error (RMSE) and mean-square error (MSE), and was able to capture the interactive effects of considered factors. In line with ANN results, CS and SR increased with curing age but decreased with increasing sawdust content. In terms of durability performance and compressive strength development, the preferable order of curing media is water > laterite > alkaline > acid. Multi-criteria evaluation of ANN models is recommended for better interpretation of model efficiency.
AB - Our research produced ultra-lightweight green interlocking paving units through combined usage of sawdust and laterite as partial cement and fine aggregate replacements respectively. Interlocking paving units, water-cured for 90 days at optimum 10% sawdust content, achieved compressive strength of 16.6 MPa and skid resistance of 64.5 PVT, which exceeded the minimum requirements of 3.45–15 MPa and 45PVT respectively. The developed ANN model was classified as very good in terms of prediction efficiency of compressive strength (CS) and skid resistance for different curing media based on Nash & Sutcliffe criterion of efficiency (NSE). However, the prediction efficiency for CS was better than SR in terms of NSE, root-mean-square error (RMSE) and mean-square error (MSE), and was able to capture the interactive effects of considered factors. In line with ANN results, CS and SR increased with curing age but decreased with increasing sawdust content. In terms of durability performance and compressive strength development, the preferable order of curing media is water > laterite > alkaline > acid. Multi-criteria evaluation of ANN models is recommended for better interpretation of model efficiency.
KW - Artificial neural network
KW - Bulk density
KW - Compressive strength
KW - Curing media
KW - Durability
KW - Elevated temperature
KW - Green interlocking paving units
KW - Laterite
KW - Sawdust
KW - Skid resistance
UR - http://www.scopus.com/inward/record.url?scp=85045275318&partnerID=8YFLogxK
UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-85045275318&origin=recordpage
U2 - 10.1016/j.conbuildmat.2018.04.061
DO - 10.1016/j.conbuildmat.2018.04.061
M3 - RGC 21 - Publication in refereed journal
SN - 0950-0618
VL - 173
SP - 600
EP - 614
JO - Construction and Building Materials
JF - Construction and Building Materials
ER -
