Abstract:Laser-induced breakdown spectroscopy (LIBS) has the potential to serve as a valuable tool in the field of metal failure detection. The grain size grade represents a critical indicator for the measurement of metal ageing, and is therefore crucial for the estimation of failure. The present study focuses on the 12Cr1MoV steel, a material with a wide range of industrial applications, and records both spectral and acoustic data pertaining to laser ablation. An innovative modelling approach was proposed, which combines acoustic information with LIBS spectra for the evaluation of heat-resistant steel grain size, thereby improving the ability to identify grain size grades. Initially, it was revealled that the acoustic energy did not exhibit significant downward trend with the continuous laser ablation, but the acoustic energy fluctuations became more intense. Based on this characteristic, only the spectral and acoustic data generated by the initial 40 laser pulse ablations were selected for retention in the subsequedn modelling process. Two data fusion strategies were proposed to obtain combined data, employing linear discriminant analysis (LDA) and support vector machine (SVM) classification models, and utilising three types of data (spectral data, acoustic spectrum data, and spectral and acoustic spectral combination data). The first strategy is to combine the dimensionality reduction-processed spectral data with the dimensionality reduction-processed acoustic spectral data to obtain combined data. The second strategy is to combine spectral data with acoustic spectral data and then perform dimensionality reduction to obtain combined data. The performance of data fusion models based on the first data fusion strategy is superior to models trained with a single data type (spectral data or acoustic spectral data), achieving a classification accuracy of 92.29%. The second data fusion strategy yielded unsatisfactory results due to the significant difference in dimensions between spectral data (7797 dimensions) and acoustic spectral data (200 dimensions). To address this, a modification was proposed by carrying out spectral feature screening on spectra data using recursive feature elimination (RFE) before data fusion and studying the impact of the remaining variables from RFE on the final grain size estimation model. The results showed that the model achieved the highest classification accuracy of 98.85% when the spectral data was reduced to 2500 dimensions. This study illustrates the effectiveness of integrating spectral and acoustic spectral data for enhanced metal assessment.

Determination of austenitic steel alloys composition using laser-induced breakdown spectroscopy (LIBS) and machine learning algorithms

Selection of Atomic Emission Lines on the Mutual Identification of Austenitic Stainless Steels with a Combination of Laser-induced Breakdown Spectroscopy (LIBS) and Partial-least-square Regression (PLSR)

Quantitative Analysis of Steel Alloy Elements Based on LIBS and Deep Learning of Multi-Perspective Features

Multielement simultaneous quantitative analysis of trace elements in stainless steel via full spectrum laser-induced breakdown spectroscopy

A Quantitative Analysis Method Assisted by Image Features in Laser-Induced Breakdown Spectroscopy.

Accurate prediction analysis of steel alloy elements by femtosecond laser-ablation spark-induced breakdown spectroscopy and out-of-bag random forest regression

Enhanced Accuracy for the Determination of Chromium in Steel by Laser-Induced Breakdown Spectroscopy (LIBS) and Extreme Random Tree (ERT)

Determination of Cobalt in Low-Alloy Steels Using Laser-Induced Breakdown Spectroscopy Combined with Laser-Induced Fluorescence

Optical Scanning Analysis of Static Samples by Compact Laser Induced Breakdown Spectroscopy

Microstructure classification of steel samples with different heat-treatment processes based on laser-induced breakdown spectroscopy (LIBS)

Spectral Interference Elimination and Self-Absorption Reduction in Laser-Induced Breakdown Spectroscopy Assisted with Laser-Stimulated Absorption

Defect identification of metal additive manufacturing parts based on laser-induced breakdown spectroscopy and machine learning

In situ elemental analysis and failures detection during additive manufacturing process utilizing laser induced breakdown spectroscopy

Investigation of Excitation Interference in Laser-Induced Breakdown Spectroscopy Assisted with Laser-Induced Fluorescence for Chromium Determination in Low-Alloy Steels

Stability and Accuracy Improvement of Element Analysis in Steel Alloys Using Polarization-Resolved Laser-Induced Breakdown Spectroscopy

Assessment of metal grain size of 12Cr1MoV steel by LIBS coupled with acoustic wave information

Effect of laser beam shaping on the determination of manganese and chromium elements in steel samples using laser-induced breakdown spectroscopy

Quantitative Analysis of Steel Samples Using Laser-Induced Breakdown Spectroscopy with an Artificial Neural Network Incorporating a Genetic Algorithm.

Classification of steel based on laser-induced breakdown spectroscopy combined with restricted Boltzmann machine and support vector machine

Combination of Internal Standard and Dominant Factor PLS for Improving Long-term Stability of LIBS Measurements

Artificial intelligence-based process for metal scrap sorting