Optimizing Biomineralization Technique on Silty Sand: Advanced Pump Flow Method with Artificial Intelligence Approach and Statistical Validation

This study aims to investigate an efficient approach to improving the engineering properties of silty sand using the biomineralization technique through pump flow. This study has employed varying flow rates (0.5 ml/min to 2 ml/min), biocementation solution concentration (0.25 M to 1 M), treatment cycles (7, 14, and 28 days), and pH. Unconfined compressive strength (UCS) test has been employed to assess strength stiffness and investigate failure modes. Maximum compressive strength has been found to be 764.40 kPa, and calcite deposition of 14.85 weight percentage. Key parameters, including flow rate, molar concentration of the cementation solution, number of treatment cycles, and bacterial density, are optimized to achieve successful biomineralization in this study. The aci1d-digestion method has been performed to investigate the precipitation and distribution of calcium carbonate which is further confirmed using microanalysis. Then, data sets of observations and results are explored statistically using ANOVA and MLRA. Artificial neural network (ANN) has validated complex bio-geo-chemical reactions of varying independent variables and the outcomes of dependent variables experimental results with RMSE = 9.8094 and R2 = 0.9999. Statistical analysis of this experimental study data sets and ANN approach show successful optimization of the biomineralization technique through the pump flow method with varying parameters.