A Buckingham-Pi dimensionless analysis for melt pool stability and defect prediction in additive manufacturing

Laser powder bed fusion (LPBF) is one of the additive manufacturing (AM) techniques and the most studied laser-based AM process for metals and alloys. The optimization of the laser process parameters of LPBF and the prediction of defects, for example, keyholes, cracks, and lack of fusion (LOF), are important for improving the quality of products made with LPBF. Deep learning (DL) is powerful in analyzing complex processes and predicting anomalies; however, much data is generally required for training a DL model. Experimental studies on AM (e.g., LPBF) habitually employ the design of experiments to decrease the number of experiments and save time and costs. Hence, the experimental data are not prepared for DL model creation in most situations. This paper studies the creation of a DL model on a small experimental dataset with unbalanced data and the prediction of the LOF defect of LPBF utilizing the created DL model. Data analytics is mainly conducted based on four DL methods, including Elman neural networks, Jordan neural networks, deep neural networks (DNN) with weights initialized by the deep belief network, and the regular DNN based on four algorithms: “rprop+”, “rprop−”, “sag,” and “slr.” It is shown that the regular DNN after the z-score standardization of the small dataset helps create a more accurate DL model and achieve better analytics and prediction results than the three other DL methods in this paper. The three other DL methods do not work well in the prediction of LOF based on the small dataset (with unbalanced data).

During the laser powder bed fusion (LPBF) additive manufacturing (AM) process, several forces are generated due to the complex interaction between the laser and powder bed, such as vapor recoil force, surface tension force, Marangoni force, gravity, and buoyance force. These forces significantly influence the metallic spatters’ behavior and melt pool stability. The spatters and melt pool stability primarily affect the printed part dimension accuracy, surface quality, and mechanical properties. To investigate spatter behavior and melt pool stability, we recorded the LPBF process with a high-speed camera and analyzed the recorded videos using home-designed Python code. This algorithm can detect the key information about the spatter and melt pool, including the spatter amount, size, initial ejection speed and angle, melt pool size and stability. In addition, this algorithm can identify the newly generated spatters and previous existing spatters to avoid repeated spatter amount counting. We find that under the same process conditions, small spatters have relatively high initial ejection speed, while big spatters have a slow initial ejection speed. In addition, we propose a stability index to quantify the LPBF process stability based on the detected melt pool length change rate. A high stability index indicates a small number of spatters and a more stable melt pool, which can contribute to manufacturing high-quality AM parts. This new image analysis algorithm and melt pool stability index can help to analyze the recorded videos and optimize the LPBF process for printing high-quality parts with fewer spatters and high melt pool stability.

In this article, dynamics and stability of milling operations with cylindrical end mills are investigated. A unified–mechanics–based model, which allows for both regenerative effects and loss–of–contact effects, is presented for study of partial–immersion, high–immersion and slotting operations. Reduced–order models that can be used for certain milling operations such as full–immersion operations and finishing cuts are also presented. On the basis of these models, the loss of stability of periodic motions of the workpiece–tool system is assessed by using Poincare sections and the numerical predictions of stable and unstable motions are found to correlate well with the corresponding experimental observations. Bifurcations experienced by periodic motions of the workpiece–tool system with respect to quasi–static variation of parameters such as axial depth of cut are examined and discussed. For partial–immersion operations, consideration of both time–delay effects and loss–of–contact effects is shown to have a significant influence on the structure of the stability boundaries in the space of spindle speed and axial depth of cut. The sensitivity of system dynamics to multiple–regenerative effects, mode–coupling effects and feed rate is also discussed.

Additive manufacturing (AM), commonly known as 3D printing, has revolutionized the manufacturing landscape by enabling layer-by-layer fabrication of complex geometries from digital models. This paper provides a comprehensive overview of the evolution, current capabilities, and future directions of AM. Beginning with the historical rise of AM, it explores and compares its major technological categories, including material extrusion, vat photopolymerization, powder bed fusion, and directed energy deposition. Each technology is discussed with regard to standard classifications and operational mechanisms. It further examines the crucial role of material properties and selection, emphasizing how polymers, metals, ceramics, and composites influence mechanical performance and application suitability. The paper investigates the deployment of AM across industries such as aerospace, biomedical, automotive, construction, and consumer goods, highlighting transformative applications. Despite its benefits, AM faces challenges such as anisotropic mechanical properties, limited material diversity, high energy consumption, and scalability constraints. Recent advancements leveraging machine learning (ML) or (AI) integration are discussed, particularly in process monitoring, defect prediction, and print quality optimization. ML-integrated process optimization techniques are shown to enhance part performance and production efficiency. Additionally, this study compares AM with subtractive manufacturing (SM), focusing on material utilization, energy efficiency, and production flexibility. A life cycle assessment (LCA) is conducted to evaluate the environmental and economic impacts of AM technologies. Market analysis indicates substantial global growth of the AM industry, fueled by technological maturation and increasing demand for customized solutions. Finally, it projects future research directions, including the development of multi-material printing, integration of AI-driven adaptive systems, sustainable material innovations, and the role of AM in decentralized manufacturing. This holistic analysis affirms AM’s pivotal role in reshaping the future of manufacturing with enhanced sustainability, precision, and design freedom. Overall, this review offers a big-picture view of AM where it stands today and how it’s paving the way for a more innovative, sustainable, and flexible future in manufacturing.

Defect structure process maps for laser powder bed fusion additive manufacturing

Deep learning methods have been widely used and proved effective in defect prediction in Additive Manufacturing (AM) to ensure process stability and part quality. However, the success of deep learning models depends heavily on meticulous training, which usually requires a large homogenous dataset. This poses a challenge for the AM industry where many small- and medium-sized enterprises (SMEs) play a crucial role. On one hand, AM parts are usually customized or one-of-a-kind, and the process settings change frequently, resulting in heterogeneous datasets that usually vary by each print. On the other hand, SMEs are usually constrained by time and budget. As a result, they tend to focus on a limited number of process settings. This often leads to insufficient data collection, making SMEs difficult to properly train deep learning models independently. Therefore, there is a need to learn from the similarities in the physics of AM processes and the defect formation mechanisms, consequently enabling the potential knowledge sharing to learn across different AM scenarios. However, unique challenges in knowledge sharing arise from privacy concerns. Each design or print potentially contains sensitive proprietary information, such as process parameters, part geometries, and quality specifications. Such information may not be shared across different entities within the industry. In this context, Federated Learning (FL) emerges as a promising solution to this data scarcity and privacy challenge. FL is an innovative machine learning method that facilitates collaborative machine learning model training across multiple clients without sharing their locally stored data. In this paper, a FL framework is developed to predict section-wise heat emission, a critical process signature during the Laser Powder Bed Fusion (LPBF) for collaborative knowledge sharing across different manufacturing entities, without direct transfer of sensitive information. The framework learns the relationship between emission readings and their impact on the quality of the printed parts in LPBF across various process parameter scenarios. A Long Short-Term Memory (LSTM) model, tailored with a customized loss function, is developed for each client and trained using their local data respectively. The local models capture the time series properties under the dynamic AM process for each distinct process parameter scenario. The global model aggregates the weights from these individual local models instead of using raw data from each process directly. Three state-of-the-art FL aggregation algorithms (FedAvg, FedProx, and FedAvgM) are employed, and their effectiveness is assessed and compared through a series of experiments. Results show that the FL framework converges fast with a comparable prediction performance, when compared to training local model individually. This work demonstrates the potential of FL-enabled AM modeling and prediction where SMEs can improve their product quality without compromising data privacy.

The method of making parts through additive manufacturing (AM) is becoming more and more widespread due to the possibility of the direct manufacturing of components with complex geometries. However, the technology’s capacity is limited by the appearance of micro-cracks/discontinuities during the layer-by-layer thermal process. The ultrasonic (US) method is often applied to detect and estimate the location and size of discontinuities in the metallic parts obtained by AM as well as to identify local deterioration in structures. The Ti6Al4V (Ti64) alloy prepared by AM needed to acquire a high-quality densification if remarkable mechanical properties were to be pursued. Ultrasonic instruments employ a different type of scanning for the studied samples, resulting in extremely detailed images comparable to X-rays. Automated non-destructive testing with special algorithms is widely used in the industry today. In general, this means that there is a trend towards automation and data sharing in various technological and production sectors, including the use of intelligent systems at the initial stage of production that can exclude defective construction materials, prevent the spread of defective products, and identify the causes of certain instances of damage. Placing the non-destructive testing on a completely new basis will create the possibility for a broader analysis of the primary data and thus will contribute to the improvement of both inspection reliability and consistency of the results. The paper aims to present the C-scan method, using ultrasonic images in amplitude or time-of-flight to emphasize discontinuities of Ti64 samples realized by laser powder-bed fusion (L-PBF) technology. The analysis of US maps offers the possibility of information correlation, mainly as to flaws in certain areas, as well as distribution of a specific flaw in the volume of the sample (flaws and pores). Final users can import C-scan results as ASCII files for further processing and comparison with other methods of analysis (e.g., non-linear elastic wave spectroscopy (NEWS), multi-frequency eddy current, and computer tomography), leading to specific results. The precision of the flight time measurement ensures the possibility of estimating the types of discontinuities, including volumetric ones, offering immediate results of the inspection. In situ monitoring allows the detection, characterization, and prediction of defects, which is suitable for robotics. Detailing the level of discontinuities at a certain location is extremely valuable for making maintenance and management decisions.

Additive manufacturing (AM) is today in the main focus—and not only in commercial production. Products with complex geometry can be built using various AM techniques, which include laser sintering of metal powder. Although the technique has been known for a quite long time, the impact of the morphology of individual powder particles on the process has not yet been adequately documented. This article presents a detailed microscopic analysis of virgin and reused powder particles of MS1 maraging steel. The metallographic observation was performed using a scanning electron microscope (SEM). The particle size of the individual powder particles was measured in the SEM and the particle surface morphology and its change in the reused powder were observed. Individual particles were analyzed in detail using an SEM with a focused ion beam (FIB) milling capability. The powder particles were gradually cut off in thin layers so that their internal structure, chemical element distribution, possible internal defects, and shape could be monitored. Elemental distribution and phase distribution were analyzed using EDS and EBSD, respectively. Our findings lead to a better understanding and prediction of defects in additive-manufactured products. This could be helpful not just in the AM field, but in any metal powder-based processes, such as metal injection molding, powder metallurgy, spray deposition processes, and others.

Additive manufacturing (AM), an enabler of Industry 4.0, recently opened limitless possibilities in various sectors covering personal, industrial, medical, aviation and even extra-terrestrial applications. Although significant research thrust is prevalent on this topic, a detailed review covering the impact, status, and prospects of artificial intelligence (AI) in the manufacturing sector has been ignored in the literature. Therefore, this review provides comprehensive information on smart mechanisms and systems emphasizing additive, subtractive and/or hybrid manufacturing processes in a collaborative, predictive, decisive, and intelligent environment. Relevant electronic databases were searched, and 248 articles were selected for qualitative synthesis. Our review suggests that significant improvements are required in connectivity, data sensing, and collection to enhance both subtractive and additive technologies, though the pervasive use of AI by machines and software helps to automate processes. An intelligent system is highly recommended in both conventional and non-conventional subtractive manufacturing (SM) methods to monitor and inspect the workpiece conditions for defect detection and to control the machining strategies in response to instantaneous output. Similarly, AM product quality can be improved through the online monitoring of melt pool and defect formation using suitable sensing devices followed by process control using machine learning (ML) algorithms. Challenges in implementing intelligent additive and subtractive manufacturing systems are also discussed in the article. The challenges comprise difficulty in self-optimizing CNC systems considering real-time material property and tool condition, defect detections by in-situ AM process monitoring, issues of overfitting and underfitting data in ML models and expensive and complicated set-ups in hybrid manufacturing processes.

Today Industry 4.0 is spreading from developed countries into other developing economies even to small and medium sized enterprises. With a strong infrastructure and fast connection between components of the whole factory, Industry 4.0 will introduce innovations and a more robust, reliable and sustainable production. While Industry 4.0 is having several fundamental technology additive manufacturing has been playing an important role and takes part in today’s Industry 4.0 technology, prominently for automotive vehicle producers. Additive manufacturing has many advantages and tasks such as in reverse engineering, by defect prediction or producing complex shaped parts. Its vitality becomes more important when its implementation is combined with the most recent measurement technology ‘Computer Tomography’. CT can detect complex surfaces or inner structure with X-Ray usage which helps to additive manufacturing parts by inspection which can not be provided by a CMM in similar manner. That study presents a simple sample for the combination of these two technologies. Followingly, voids and GPS analyses after CT measurements of several parts will be introduced. Also, those CT inspections will give an example for a definite stage of Industry 4.0. In conclusion, a brief investigation of those innovative most recent technologies, which are developing in a continuous manner, will be made in the context of Industry 4.0 according to inspection results.

Melt track monitoring in the laser powder bed fusion (LPBF) process is crucial for preventing internal defects in as-printed parts. Uncontrollable melt pool dynamic behavior easily leads to melt track morphology defects. Existing monitoring methods face challenges in balancing modeling accuracy and physical interpretability. Specifically, traditional physics-based models typically require complex monitoring equipment, extensive simulation data, and empirical formulas, resulting in high costs and limited applicability. Meanwhile, conventional data-driven models lack physical constraints, leading to insufficient interpretability, process parameter sensitivity, and poor generalization. To address these challenges, this article proposes a deep Gaussian process-based method for LPBF melt track morphology prediction. The proposed model employs kernel functions in the first layer to learn melt pool evolution patterns and embeds the Rosenthal equation into the second-layer kernel function as a physical constraint, constructing a physically interpretable multilayer Gaussian process framework. Finally, a softmax classifier based on melt track geometric deviation achieves five-category melt track morphology recognition. Multi-condition experimental results demonstrated that the proposed method achieved root mean square errors of 0.069, 0.020, and 0.039 for melt track geometry, outperforming traditional data-driven models in prediction accuracy. The classification accuracy reached 90.76%. Furthermore, the influence of different features on melt track morphology is quantified through time-lagged mutual information analysis and other visualization methods. This study provides an effective solution for achieving quality monitoring and defect prediction in the LPBF process.

RaSP is a method for making rapid and accurate predictions of changes in protein stability that enabled us to calculate ~300 million stability changes for nearly all possible single amino acid changes in the human proteome.

To enhance the performance of metal additive manufacturing (AM) products, it is crucial to accurately predict and control the solidification microstructure and defects formed during laser scanning. Numerical simulations are essential to achieve this goal. In this study, we developed a high-fidelity model and large-scale computational method combining the phase-field and lattice Boltzmann method to accurately predict the thermal fluid flow dynamics in the melt pool, which significantly affects the solidification microstructure and defect formation. The developed computational method reproduced steady-state melt pool dynamics within a realistic computation time using high-resolution grids that were previously unusable. Furthermore, the effects of grid resolution on melt pool dynamics were evaluated. The results demonstrate that the flow behavior in front of the keyhole is crucial for the overall melt pool dynamics, and achieving a high-accuracy flow representation in this region requires a high grid resolution. This method is expected to contribute significantly to the prediction of solidification microstructure and defects in metal AM.

6 - Additive manufacturing: process and microstructure

Additive manufacturing of metals: Microstructure evolution and multistage control