Machine learning-driven process of alumina ceramics laser machining

Abstract

Laser machining is a highly flexible non-contact manufacturing technique that has been employed widely across academia and industry. Due to nonlinear interactions between light and matter, simulation methods are extremely crucial, as they help enhance the machining quality by offering comprehension of the inter-relationships between the laser processing parameters. On the other hand, experimental processing parameter optimization recommends a systematic, and consequently time-consuming, investigation over the available processing parameter space. An intelligent strategy is to employ machine learning (ML) techniques to capture the relationship between picosecond laser machining parameters for finding proper parameter combinations to create the desired cuts on industrial-grade alumina ceramic with deep, smooth and defect-free patterns. Laser parameters such as beam amplitude and frequency, scanner passing speed and the number of passes over the surface, as well as the vertical distance of the scanner from the sample surface, are used for predicting the depth, top width, and bottom width of the engraved channels using ML models. Owing to the complex correlation between laser parameters, it is shown that Neural Networks (NN) are the most efficient in predicting the outputs. Equipped with an ML model that captures the interconnection between laser parameters and the engraved channel dimensions, one can predict the required input parameters to achieve a target channel geometry. This strategy significantly reduces the cost and effort of experimental laser machining during the development phase, without compromising accuracy or performance. The developed techniques can be applied to a wide range of ceramic laser machining processes.

Combining Finite Element and Machine Learning Methods to Predict Structures of Architectured Interlocking Ceramics

“…Before using the simulation data to train the ML models, reported in Supporting Information including some repeated runs that were used for averaging, all variables were normalized and scaled in the range [0,1] to balance the weights of different variables with measurements in dissimilar orders. [ 38 ] To predict the dependent variables as a function of the panel design parameters, different ML algorithms were examined. First, linear regression models including the first‐ and the second‐order terms were employed to probe the variables’ linear dependencies (results are not reported) which proved to be not the case here.…”

“…The reported R 2 score in Table 3 is the average R 2 score calculated from 100 random splits (mean ( R 2 ) ± std). [ 38 ] Among the differen t tested ML algorithms, GPR provided the best precision (mean ( R 2 ) > 70%) for all dependent variables with the exception of the heat rate; however, the XGB model predictions for the out‐of‐plane deformation, edge temperature, heat rate, and internal energy meet the required condition (mean ( R 2 ) > 70%) and the same applies to the NN predictions for contact energy, elastic strain energy, edge temperature, and internal energy.…”

Investigation of the thermal analysis of a wavy fin with radiation impact: an application of extreme learning machine

“…The effectiveness of the systems, enhanced learning capabilities, and diverse algorithms of these models make them engineering tools that can forecast and simulate more accurately than conventional mathematical tools. Recently, modelling techniques developed using machine learning have advanced a variety of sectors, from real-world applications to scientific investigation [39][40][41][42]. Thermal dispersion in heat exchangers with wavy-structured fins was widely explored in related investigations.…”

Advanced USP laser process with deep learning and triangular beam shaping for micro Fresnel lenses fabrication

“…In order to reduce the parameter search time for the end-user, we propose to create a learning model to predict the shape of the ablation groove for a given set of laser parameters and laser beam profile. Recent advances in the use of machine/deep learning in laser machining [1][2][3] show that it is possible to predict many machining results from process data. These algorithms are capable of learning the physical principles that govern light-matter interactions, even non-linear principles very difficult to simulate.…”