A new approach for automated measuring of the melt pool geometry in laser-powder bed fusion

Abstract

Additive manufacturing (AM) offers unique possibilities in comparison to conventional manufacturing processes. For example, complex parts can be manufactured without tools. For metals, the most commonly used AM process is laser-powder bed fusion (L-PBF). The L-PBF process is prone to process disturbances, hence maintaining a consistent part quality remains an important subject within current research. An established indicator for quantifying process changes is the dimension of melt pools, which depends on the energy input and the cooling conditions. The melt pool geometry is normally measured manually in cross sections of solidified welding seams. This paper introduces a new approach for the automated visual measuring of melt pools in cross-sections of parts manufactured by L-PBF. The melt pools are first segmented in the images and are then measured. Since the melt pools have a heterogeneous appearance, segmentation with common digital image processing is difficult, deep learning was applied in this project. With the presented approach, the melt pools can be measured over the whole cross section of the specimen. Furthermore, remelted melt pools, which are only partly visible, are evaluated. With this automated approach, a high number of melt pools in each cross-section can be measured, which allows the examination of trends over the build direction in a specimen and results in better statistics. Furthermore, deviations in the energy input can be estimated via the measured melt pool dimensions.

Research on the calibration algorithm of the GTAW weld pool analytical model designed for time-varying welding speed situation

“…), 𝑣(𝑥) = 𝑊 𝑣 𝑥, 𝑊 𝑣 ∈ ℝ 𝐶 * * 𝐶 (8) where v is the output of another 1×1 convolution. Let𝐶 * = 𝐶 / 8, by adding the input features x to the weighting of the output, we can obtain:…”

“…Although these methods are able to achieve the basic needs of extracting the features of the weld pool surface, their accuracy and efficiency are still need to be improved due to the limitations of algorithms or hardware devices. In recent years, a large number of welding image processing methods based on deep learning technology are proposed which including (i) convolutional neural network (CNN)-based methods [7][8][9][10] ; (ii) recurrent neural network (RNN)-based methods [11][12] ; and (iii) generative adversarial network (GAN)-based methods [13] . These methods not only improve the accuracy and speed of weld pool feature extraction, but also lay the foundation for achieving full automation and intelligence of welding.…”

SOSPCNN: Structurally Optimized Stochastic Pooling Convolutional Neural Network for Tetralogy of Fallot Recognition

“…Pooling is an essential operation in standard convolutional neural networks (CNNs) [13]. Two types of pooling exist.…”

Prediction and validation of melt pool dimensions and geometric distortions of additively manufactured AlSi10Mg

“…), or the anisotropy [7] in material properties. Thermomechanical simulation [8,9], and inherent strain approach [10,11] are two types of finite element (FE)-based simulation techniques frequently used for predicting properties of additively manufactured components ranging from melt pool prediction [12][13][14] to final residual stresses [15] and component distortions. Thermomechanical simulation is a more systematic and sequential approach in which the first step thermal analysis (TA) yields a transient temperature field, which is used as the thermal load to drive the subsequent mechanical analysis (MA) step.…”