| The increasing urban wastewater brought by the skyrocketing municipal infrastructure construction has been a pressing issue, posing considerable threat to the safety of concrete structures. As a consequence, improvement of the durability of concrete in the wastewater is of great urgency. The corrosion mechanism of concrete is complicated due to the complexity and interference effects of various factors,which would contribute to the corrosion. The relative low rate and long period of the sewage corrosion should not go unnoticed as well. In order to accelerate the corrosion and shorten the experimental period, sewage with high concentration was prepared. Initially, variation of the concrete mass and strength were studied in the clean water and sewage, respectively. The anti-corrosion performance of the mineral admixtures and antiseptic were also test during the experiments. The second part focused on the influence of biofilms toward the corrosion of concrete. Morphologies of the biofilms were characterized by the Confocal Laser Scanning Microscope. The growth pattern and growth kinetics were studied, macro performance and morphology variation of the concrete were tested and compared in and of the absence of the biofilms, respectively. Variation of the sewage index was monitored during the corrosion process. Key factors which could affect the corrosion of concrete were identified based on the grey correlation. The BP neural network prediction model of sewage-induced concrete corrosion was eventually established.Experimental results showed that mass and strength of the concrete went down after an immersion in sewage for 5 months. There was a mass loss about 1.47 % and an increase of 122.78 μm in surface roughness. The soaked concert samples converted into porous structure, large amount of flake-shaped ettringite crystals with large size were observed in these pores. Porosity of the concrete increased 6.33 %after the corrosion and went up to 32.46 %. Among these anti-corrosion measures mentioned above, addition of mineral admixtures showed better performance thanthe usage of antiseptic. Samples with mineral admixtures additive showed a mass loss of 0.44 % after a 5-month immersion in sewage. Based on previous studies,kinetics models of the biofilms in attachment period and growth period were established. The samples covered with biofilm showed a mass loss about 1.23 %, a calcium hydroxide content of 4.13 % and a porosity of 31.82 % after an immersion in sewage with a COD concentration of 9 000 mg/L for 5 months. The samples with no bilfilm coverage showed a mass loss of 1.67%, a calcium hydroxide content of3.01% and a porosity of 34.13%, accordingly. The corrupted concrete showed lower compressive strength and flexural strength, compared with the sample immersed in clean water. Sample soaked in sewage with no biofilm coverage underwent the highest degree of corrosion, indicating the biofilms could protect the concrete from corrosion to a certain extent. Based on the variation obtained by the dynamic monitoring and the analysis by grey correlation, major factors which would affect the corrosion were identifies as pH value, dissolved oxygen and sulfates. Secondary factors turned out to be COD and suspended solids. The 6-19-1structure BP neural network prediction model was established, in which these above mentioned factors and corrosion time were set as input variants, the compressive and anti-corrosion coefficient were set as output variants. The average prediction error of this model turned out to be around 2.46 %. |
PDF Full Text Request