Deciphering and predict corrosion effect, influencing factors and microbial mechanism of sewer concrete corrosion based on extensive data analysis and machine learning.

This study investigated the effect of environmental parameters on microbial-induced concrete corrosion during three corrosion stages. The corrosion effects and influencing factors in three corrosion stages were deeply analyzed and predicted, and the mechanism of microbial corrosion was summarized. When the H₂S concentration was lower than about 15.0 mg/m³, the corrosion was maintained in stage II, which greatly delays the intensification of corrosion. In stage III, the H₂S concentration has the significantly accelerate the corrosion rate. Additionally, incorporating min, max and mean values of input parameters greatly improved the accuracy of machine learning predictions of corrosion rates (R² >0.99). Acidithiobacillus was found to be dominant in the microbial community at H₂S concentrations of 12.0–37.5 mg/m³, and increased temperature promoted the reproduction of Acidithiobacillus. [ABSTRACT FROM AUTHOR]