Estimating The Aging Grade Of Heat-Resistant Steel By Using Portable Laser-Induced Breakdown Spectroscopy

Heat-resistant steel is widely used in the construction of various types of industrial pressure-bearing equipment. In this work, laser-induced breakdown spectroscopy (LIBS) was applied to evaluate the aging grade of actual industrial pipe heating surfaces (made by T91 steel) via a self-developed portable LIBS device. First, the pretreatment effect of wavelet threshold denoising (WTD) and multiple scattering correction (MSC) on the LIBS spectrum is discussed. Based on the K-fold-support vector machine-recursive feature elimination (K-SVM-RFE) and support vector machine (SVM), the aging grade model is established using the spectral data of the T91 artificial specimens and of the industrial pipe slices at different aging grades. The influence of the denoising of WTD and the transformation of MSC in the selection of optimal feature subsets using K-SVM-RFE are discussed along with the model's performance. Second, the spectral characteristics of the actual industrial T91 pipe heating surface of coal-fired boilers are investigated, and the spectral data are screened and standardized. Finally, by using the aging grade model to evaluate the aging grade of the industrial T91 pipe based on the calibrated spectral data, an accuracy of 85.17% was achieved. The results demonstrate that the portable LIBS device and the corresponding hybrid model presented here have strong discrimination and robustness for estimating the aging grade of heat-resistant steel.