Investigation Of The Structural Characteristics Of Lime-cement Concrete

ABSTRACT 

This work investigated the structural characteristics of lime cement concrete using 30 selected mix ratios. The properties studied include, compressive strength, flexural strength, splitting tensile strength, shear strength, poisson ratio, modulus of elasticity, and modulus of rigidity.  A total of 360 concrete cube specimen, 360 concrete prototype beam specimen, and 360 concrete cylinder specimen were cast and cured in open water tanks. 3 specimen were cast for each mix proportion. They were then tested in compression, flexure, and splitting tension respectively at 7 days, 14 days, 21 days, and 28 days. Load values obtained from these test were used to determine the other structural properties of the concrete. Materials used in concrete production were the portland cement (PC), hydrated lime (HL), river sand, granite chippings, and water. The highest value of compressive strength recorded from experimental works at 28 days of curing was 30.83N/mm2. This occurred at a water-cement (w/c) ratio of 0.562, having a percentage replacement of PC with hydrated lime of 18.75%. Highest values of flexural strength, splitting tensile strength, shear strength, poisson ratio, modulus of elasticity and modulus of rigidity recorded at 28 days of curing were 5.03N/mm2, 3.725N/mm2, 1.257N/mm2, 0.216, 30.708 x 103 N/mm2 and 13.386 x 103 N/mm2 respectively. Lowest values recorded for compressive strength, flexural strength, split tensile strength, shear strength, poisson ratio, modulus of elasticity and modulus of rigidity recorded at 28 days of curing were 15.12N/mm2, 2.28N/mm2,  2.00N/mm2,  0.569N/mm2,  0.105,  20.264 x 103N/mm2  and  8.803 x 103N/mm2  respectively. A total of 120 data set were generated experimentally for each property studied. 114 sample data of each property were used to teach the artificial neural networks (ANNs) how to accurately predict the structural properties of the lime cement concrete. The remaining 6 sets of data were left out and used to test how well the networks were predicting after being trained. 7 ANN models were created using the neural network toolbox in the Matlab R2014a software. The feed forward back propagation neural network with “trainlm” training function and the mean square error (mse) performance functions were adopted. The end results of the back propagation neural networks were 6-20-1 (6 inputs, 20 neurons in the hidden layer, and 1 output). Maximum percentage error for all networks were generally below 11% while the maximum correlation coefficients were close to 1. The student’s ttest was used to further test the adequacy of the neural network models. The calculated T values for the compressive strength, flexural strength, split tensile strength, shear strength, poisson ratio, modulus of elasticity, and modulus of rigidity neural networks were 1.437, 0.1598, 0.4607, 1.4642, -1.0555, 0.4631, and 1.7069 respectively. They were all less than the 2.065 which is the allowable T value from the statistical table. Therefore, the null hypothesis (Ho) was accepted i.e. there is no significant difference between the neural network models and the experimental results. For lime cement concrete to be used as a structural concrete, PC replacement with hydrated lime must not be up to 30%. Optimum percentage replacement was recorded at 18.75%. Partial replacement of portland cement with hydrated lime was observed to improve the workability of the fresh concrete but reduced the strength of the hardened concrete. The relationship between the structural properties of the lime cement concrete with respect to water cement ratio, showed that the magnitude of each property of concrete increased as water cement ratio increased until the optimum water cement ratios were reached. With the use of the developed ANN models, mix design procedures for lime cement concrete can be carried out with lesser time, and energy requirement since the traditional method of designing mixes by carrying out trial mixes in the laboratory will no longer be required.  

Keywords: Structural properties, concrete, hydrated lime (HL), portland cement (PC), artificial neural network (ANN).