Experimental and ANN-Based Evaluation of Mechanical Properties and Self-Healing Capabilities of Bacterial Concrete with Fly Ash Addition

The development of sustainable, long-lasting concrete materials has attracted significant interest due to the growing need for durable infrastructure. This paper examines the effects of multiplying calcite-precipitating bacteria and fly ash on the mechanical characteristics and repair capabilities of concrete, as well as the use of Artificial Neural Network (ANN) modelling to forecast compressive strength. The mixes were made of M30-grade concrete, with cement substituted with fly ash at proportions of 0- 30 and a bacterial solution of Sporosarcina pasteurii at 10- 30. Six mixes (M0-M5) were prepared, with M0 being the control mix with no fly ash or bacteria. The experimental findings showed that a moderate addition of fly ash and a bacterial solution had a significant positive effect on the mechanical properties of concrete. The best mix (M2) with 15 per cent fly ash and 10 per cent bacterial solution had compressive strength of 24.3 Mpa, 31.2 Mpa and 38.6 Mpa after 7, 14 and 28 days, respectively; this is an improvement of about 11 per cent over the control mix (34.7 Mpa) at 28 days. Equally, flexural strength was enhanced to 5.1 MPa as compared to 4.6 MPa (control mix), and the split tensile strength increased to 3.6 MPa as compared to 3.2 MPa at 28 days. The analysis of crack healing revealed that the performance of the self-healing increased markedly with decreasing crack width, and its maximum healing efficiency of 70% was attained for the optimum mix. An Artificial Neural Network (ANN) model was applied in order to further test the predictive power of the developed system. The model showed strong predictive capability, with R2 values ranging from 0.9216 to 0.9864 for predicting compressive strength across the training, validation, and testing datasets. The highest prediction accuracy was achieved with the 28-day strength model, demonstrating the consistency of the developed ANN framework. Overall, the results suggest that the combination of calcite-precipitating bacteria and 15% fly ash has a significant impact on enhancing the mechanical performance and self-healing capacity of concrete. In addition, the ANN model provided accurate forecasts of compressive strength, indicating that bio-based self-healing concrete technology with machine learning has the potential to create a sustainable and durable construction material.