Eco-Cement is a new proposed material for the construction industry that uses waste products of multiple industries in its manufacture. It is a composite material of bacteria, urea, calcium carbonate, cement kiln dust, a hydraulic agent, and sand. Through careful evaluation studies, we achieved the qualitative and quantitative identification of the optimal ingredients. The end product utilizes wastes of the cement manufacturing industry, the dairy industry and the poultry growing industry achieving an environmentally desired product.
We used refined stochastic analysis based on a Gaussian model, similar to the one used in analyzing the properties of traditional concrete. Through, detailed stochastic analysis a new, optimized recipe was developed. We correlated this variable with the compressive strength and found significant enhancement at the R-squared value for this correlation (regarding medians). Therefore, we were able to define an optimum value for each variable and thus an optimum recipe using the mechanical properties of the composed material as objective. A simultaneous variation modeling of the design variables (six ingredients) was assumed. The correlation with final strength was initially attempted using traditional regression analysis, but because correlation factors and tests were not satisfactory, an advanced model using neural networks was implemented. Various parameters of the neural network training (determination) were investigated and an optimum one (in terms of mean square error) was found. Finally we used this artificial neural network to formulate the optimum recipe.