关键词: 劣化, 碱激发剂, 冻融, 硫酸盐, 微观, 人工神经网络

Abstract: To study the deterioration of alkali-excited fly ash concrete under sulfate environment, this paper studies the variations in mass loss, ultrasonic wave velocity loss, and compressive strength loss of fly ash concrete, by examining the specimens of different alkali admixtures (0%, 5%, 8%, and 10%) under the coupling of freeze-thaw cycling and different sulfate contents. We determine deterioration patterns through experimental tests using electron microscopy, EDS energy spectroscopy, and XRD diffraction, and develop an artificial neural network prediction model of deterioration patterns based on the test data. The results show adding alkaline exciters improves the durability of fly ash concrete significantly at the optimal alkali admixture of 8%. The deterioration of the specimens shows a linear positive correlation with the number of freeze-thaw cycles. In terms of mass loss, the specimens with different dosages show two stages, smooth declining and accelerated declining; the latter stage appears earlier for the specimens without the alkali exciter. In terms of the loss of ultrasonic wave velocity and the loss of compressive strength, the specimens doped with alkali exciters show only a steady declining stage. Microscopic analysis reveals that the intensified deterioration of alkali-excited fly ash concrete is caused by gradual generation of gypsum and calcovanadate, and our artificial neural network prediction model has high accuracy.

Key words: deterioration, alkali-excited, freeze and thaw, sulfate, microscopic, artificial neural network