ARTIFICIAL INTELLIGENCE METHODS FOR CORRELATING LASER-INDUCED BREAKDOWN SPECTROSCOPY (LIBS) MEASUREMENTS WITH DEGREE OF SENSITIZATION (DOS) VALUES TO DETERMINE THE SENSITIZATION OF AN ALLOY
Methods and systems for determining sensitization of an alloy includes correlating laser-induced breakdown spectroscopy (LIBS) measurements with degree of sensitization (DoS) values to determine the sensitization of an alloy. Sensitization is characterized by new phase precipitates preferably along the grain boundaries (GBs). In an embodiment, the method includes the features of (1) selective chemical etching of the new phase precipitate of an alloy to induce quantitative chemical composition change, correlated with the DoS values, on the alloy surface, (2) LIBS measurements to semi-quantitatively probe the chemical composition change on the etched surface due to selective chemical etching, (3) establishing calibration models by correlating the LIBS spectra with the DoS using artificial intelligence (AI) algorithms/approaches to determine a sensitization of an alloy.
1 . A method for determining sensitization of an alloy by correlating laser-induced breakdown spectroscopy (LIBS) measurements with the degree of sensitization (DoS) using artificial intelligence (AI), the method comprising:
selective chemical etching of the new phase precipitate of an alloy to induce quantitative chemical composition change on a surface of the alloy, wherein the selective chemical etching is conducted using one or more alloy etchants;
measuring LIBS spectra using a LIBS system to semi-quantitatively probe the chemical composition change on the etched surface of the alloy, wherein the LIBS spectra are measured using a single laser pulse or multiple laser pulses; and
determining sensitization of the alloy using an artificial intelligence (AI) algorithm, wherein the AI algorithm correlates the LIBS spectra with the DoS.
2 . The method according to claim 1 , wherein the alloy is an aluminum alloy or a steel.
3 . The method according to claim 1 , wherein a new phase is formed by migration of specific atoms in a crystalline material and is different from the homogeneous alloy.
4 . The method according to claim 1 , wherein the one or more alloy etchants include one or more of nitric acid, Keller's reagent, and ammonium persulfate.
5 . The method according to claim 1 , further comprising surface polishing the surface of the alloy before the selective chemical etching.
6 . The method according to claim 5 , wherein the surface polishing includes one of sanding, ultrasonic polishing, lapping, sandblasting, rumbling and tumbling.
7 . The method according to claim 1 , wherein the LIBS system includes a single-pulsed laser beam as a plasma excitation source, or two or more pulsed laser beams as the plasme excitation source.
8 . The method according to claim 1 , wherein the measuring includes detecting a plasma emission during a defined gating interval, with synchronization of each laser pulse.
19 . The method according to claim 1 , wherein the measuring includes detecting a plasma emission during an integrated time period, without synchronization of each laser pulse.
10 . The method according to claim 1 , further including implementing a LIBS signal enhancement approach, wherein the LIBS signal enhancement approach includes one of spatial confinement, magnetic confinement, flame enhancement, and argon gas enhancement.
11 . The method according to claim 1 , wherein the DoS conforms to the standards set by ASTM international standard G67-18 or ASTM international standard G108-94 (2015).
12 . The method according to claim 1 , wherein the determining sensitization includes providing the quantitative information of the DoS within a range or with a specific value of an alloy.
13 . The method according to claim 1 , wherein the artificial intelligence algorithm includes one of principal component-discrimination function analysis (PC-DFA), discriminant analysis, partial least squares regression analysis, partial least squares discriminant analysis, a k-nearest neighbors algorithm, an artificial neural network, a soft independent modeling of class analogy, a support vector machine algorithm, and classification and regression trees.
14 . The method according to claim 1 , wherein the artificial intelligence algorithm includes one of statistical learning, computer intelligence, and soft computing.
15 . A method for determining sensitization of an alloy, the method comprising:
selective chemical etching a surface of the alloy, wherein the selective chemical etching is performed using one or more alloy etchants;
measuring laser-induced breakdown spectroscopy (LIBS) spectra of the etched surface of the alloy using a LIBS system to semi-quantitatively probe a chemical composition change of a new phase precipitate of the alloy on the etched surface of the alloy due to the selective chemical etching; and
correlating the LIBS spectra with the degree of sensitization (DoS) of the alloy to thereby determine a sensitization of the alloy.
16 . The method of claim 15 , wherein the LIBS system includes a single pulse laser source or a multiple pulse laser source.
17 . The method of claim 16 , wherein the LIBS system further includes one or more compact or bulk spectrometers, and optical components for light delivery and collection.
18 . The method according to claim 15 , wherein the alloy is an aluminum alloy a steel.
19 . The method according to claim 15 , further comprising surface polishing the surface of the alloy before the selective chemical etching.
20 . The method according to claim 15 , wherein the correlating includes using an artificial intelligence (AI) algorithm to determine the sensitization of the alloy.
21 . The method of claim 20 , wherein the artificial intelligence algorithm includes one of principal component-discrimination function analysis (PC-DFA), discriminant analysis, partial least squares regression analysis, partial least squares discriminant analysis, a k-nearest neighbors algorithm, an artificial neural network, a soft independent modeling of class analogy, a support vector machine algorithm, and classification and regression trees.
22 . A laser-induced breakdown spectroscopy (LIBS) system for determining sensitization of an alloy, the system comprising:
a pulsed laser source configured to emit laser pulses directed at a sample to induce a plasma on a surface of the sample, the sample including a selectively chemical etched surface of an alloy;
a spectrometer configured to measure LIBS spectra of the etched surface of the alloy to semi-quantitatively probe a chemical composition change of a new phase precipitate of the alloy on the etched surface of the alloy; and
one or more processors, configured to correlate the LIBS spectra with a degree of sensitization (DoS) of the alloy to thereby determine a sensitization of the alloy.
23 . The system of claim 22 , wherein the one or more processors are configured to correlate using an artificial intelligence (AI) algorithm to determine the sensitization of the alloy.
24 . The system of claim 23 , wherein the artificial intelligence algorithm includes one of principal component-discrimination function analysis (PC-DFA), discriminant analysis, partial least squares regression analysis, partial least squares discriminant analysis, a k-nearest neighbors algorithm, an artificial neural network, a soft independent modeling of class analogy, a support vector machine algorithm, and classification and regression trees