Neural Modeling and Experimental Investigation of the Erosion Characteristics of Boiler Grade Steels Impacted by Fly Ash

Abstract

In this study, experimental investigations on solid particle erosion have been carried out for two typical boiler grade steels used for fabricating boiler components, namely, 0.22C steel and 1.25Cr-Mo-V steel. The erosion tests were conducted using an air jet erosion testing facility conforming to ASTM Standard G76 international testing method. Impingement angle of the air-particle stream on the target material was varied from 30 to 75, and the particle impact velocity was also varied from 40 to 150 m/s. The surface morphology of the eroded samples was investigated by SEM to elucidate the erosion mechanism under different particle impact conditions. Further, an artificial neural network (ANN) model has been developed to predict the erosion behavior of these boiler grade steels impacted by boiler fly ash for more realistic characterization of erosion potential using pertinent data. An efficient network training optimization algorithm has been employed for faster convergence and better predictions. The ANN predictions of erosion rate are found to be in excellent agreement with the actual measured data as reported in the literature. It has been observed from the neural model results that the erosion potential of actual boiler fly ash with various levels of silica content is considerably higher than the synthetic erodent (alumina particles) usually used in the experiments to mimic fly ash erosion behavior under simulated conditions.