Search

Deep learning has impacted defect prediction in additive manufacturing (AM), which is important to ensure process stability and part quality. However, its success depends on extensive training, requiring large, homogeneous datasets--remaining a challenge for the AM industry, particularly for small- and medium-sized enterprises (SMEs). The unique and varied characteristics of AM parts, along with the limited resources of SMEs, hamper data collection, posing difficulties in the independent training of deep learning models. Addressing these concerns requires enabling knowledge sharing from the similarities in the physics of the AM process and defect formation mechanisms while carefully handling privacy concerns. Federated learning (FL) offers a solution to allow collaborative model training across multiple entities without sharing local data. This article introduces an FL framework to predict section-wise heat emission during laser powder bed fusion (LPBF), a vital process signature. It incorporates a customized long short-term memory (LSTM) model for each client, capturing the dynamic AM process's time-series properties without sharing sensitive information. Three advanced FL algorithms are integrated--federated averaging (FedAvg), FedProx, and FedAvgM--to aggregate model weights rather than raw datasets. Experiments demonstrate that the FL framework ensures convergence and maintains prediction performance comparable to individually trained models. This work demonstrates the potential of FL-enabled AM modeling and prediction where SMEs can improve their product quality without compromising data privacy. [ABSTRACT FROM AUTHOR]

The polymer‐based acetabular cup prosthesis, a vital component of hip replacement surgery, significantly contributes to the recovery of patients afflicted with osteoarthritic conditions. Nevertheless, the current clinical usage of polymer acetabular cup prostheses commonly encounters the challenge of balancing wear resistance and oxidation resistance, significantly impacting both their lifespan and patients′ quality of life. Consequently, researchers have persistently enhanced the attributes of polymer acetabular cup prosthetic materials. These enhancements, including irradiation and filler modifications, are intended to concurrently bolster both the wear and oxidation resistance of the prosthesis materials. This comprehensive approach aims to address wear‐associated clinical complications like osteolysis and oxidative brittleness, ultimately extending their in vivo service life. For this reason, this paper retrospectively discusses the progress of research on the modification of polymer acetabular cup prosthesis materials for high wear and oxidation resistance and explores potential design methods for optimising artificial acetabular cup materials, with a view to providing new ideas for extending the service life of artificial joint implants. [ABSTRACT FROM AUTHOR]

Melt pool dynamics in metal additive manufacturing (AM) is critical to process stability, microstructure formation, and final properties of the printed materials. Physics-based simulation, including computational fluid dynamics (CFD), is the dominant approach to predict melt pool dynamics. However, the physics-based simulation approaches suffer from the inherent issue of very high computational cost. This paper provides a physics-informed machine learning method by integrating the conventional neural networks with the governing physical laws to predict the melt pool dynamics, such as temperature, velocity, and pressure, without using any training data on velocity and pressure. This approach avoids solving the nonlinear Navier-Stokes equation numerically, which significantly reduces the computational cost (if including the cost of velocity data generation). The difficult-to-determine parameters' values of the governing equations can also be inferred through data-driven discovery. In addition, the physics-informed neural network (PINN) architecture has been optimized for efficient model training. The data-efficient PINN model is attributed to the extra penalty by incorporating governing PDEs, initial conditions, and boundary conditions in the PINN model. [ABSTRACT FROM AUTHOR]

Powder bed fusion (PBF) is an additive manufacturing process in which laser heat liquefies blown powder particles on top of a powder bed, and cooling solidifies the melted powder particles. During this process, the laser beam heat interacts with the powder causing thermal emission and affecting the melt pool. This paper aims to predict heat emission in PBF by harnessing the strengths of recurrent neural networks. Long short-term memory (LSTM) networks are developed to learn from sequential data (emission readings), while the learning is guided by process physics including laser power, laser speed, layer number, and scanning patterns. To reduce the computational efforts on model training, the LSTM models are integrated with a new approach for down-sampling the pyrometry raw data and extracting useful statistical features from raw data. The structure and hyperparameters of the LSTM model reflect several iterations of tuning based on the training on the pyrometer readings data. Results reveal useful knowledge on how raw pyrometer data should be processed to work the best with LSTM, how physics features are informative in predicting overheating, and the effectiveness of physics-guided LSTM in emission prediction. [ABSTRACT FROM AUTHOR]

Emerging laser precision machining, particularly using pulsed lasers, enlightens the innovation and functionalization of cemented carbides. These backbone materials of the tooling industry are usually considered difficult to machine or shape using conventional mechanical approaches. The coating of cemented carbide tools, deemed to improve their mechanical and thermal properties, is a common supplementary surface treatment prior to their application. This work aims to study the synergic effects of nanosecond laser ablation and coating deposition on the surface, as well as the mechanical integrity of cemented carbides. In this regard, two plain WC–Co grades with different metallic binder content (10%wtCo and 15%wtCo) were first processed by a short-pulsed nanosecond laser. Subsequently, an AlTiN film was physically vapor-deposited on the laser-processed surfaces. The resulting surface integrity was assessed in terms of topographical, morphological, and microstructural changes. Mechanical integrity was evaluated in terms of indentation and sliding contact responses using Vickers hardness and scratch tests, respectively, the latter including frictional, penetrating, and sliding performances under selected surface processing conditions. In general, the nanosecond laser ablation proved to be beneficial for the mechanical integrity of coated cemented carbides in most studied cases, as it increased surface hardness, reduced penetration depth, and hindered damage during sliding. This was the case despite a slight increase in surface roughness, as well as minor morphological and microstructural changes at the coating–substrate interface, discerned. [ABSTRACT FROM AUTHOR]

Abstract Background Clinical prognostic parameters of liver metastasis from pancreatic adenocarcinoma have not been specifically identified.This study is to explore the risk factors of liver metastasis in advanced pancreatic adenocarcinoma (PDAC) patients in China. Methods A multicenter cohort study was conducted to explore whether liver metastasis in locally advanced and metastatic PDAC could be reflected by some common laboratory indexes. We collected 1787 advanced PDAC patients from three participating hospitals between 2004 and 2014. The associations between some laboratory indexes and risks of liver metastases were analyzed. Results Results have shown that 87% of stage IV patients developed synchronous liver metastasis. Primary tumor location (body/tail vs. head/neck, OR 0.55, 95% CI 0.36-0.83), primary tumor diameter (≥20 mm vs.

In order to study the influence of semiconductors on the magnetic properties and giant magnetoimpedance effect of FINEMET ribbon, titanium dioxide coating layer with different thickness was sputtered onto the free surface of the FINEMET ribbon by RF magnetron sputtering to prepare the FINEMET/TiO2 composite ribbons. The morphology, magnetic properties, and giant magnetoimpedance of the FINEMET/TiO2 composite ribbons were analyzed. The results show that the GMI ratio of composite ribbons first increases and then decreases with the increase of TiO2 layer thickness (0 ~ 150 nm). When the thickness of TiO2 thin film is 100 nm, the GMI ratio reaches the maximum 57.3%, which indicates that a certain thickness of TiO2 thin film can significantly improve the GMI effect. The result can be explained by the combined result of electromagnetic interaction and stress between TiO2 thin film and the FINEMET ribbon. [ABSTRACT FROM AUTHOR]

The strong adherence (stiction) of adjacent surfaces is a major design concern in microelectromechanical system. The present research concerns the elastic-plastic adhesion of a rough surface contacting with a rigid flat. Rough surface is characterized by fractal geometry using a two-variable Weierstrass-Mandelbrot function [The Fractal Geometry of Nature (Freeman, New York, 1982)]. The microcontact model of single asperity is established in terms of fractal parameters, and the plastic adhesion model is also developed with the Dugdale approximation to consider the adhesive interaction within and outside the contact area. Then according to the plastic flow criterion, the Maguis-Dugdale transition theory [Adhesion of Adhesion and Rapture of Elastic Solids (Springer-Verlag, New York, 2005)] and the plastic adhesion model are used to solve the elastic-plastic adhesive interaction for the two approaching surfaces by incorporating the fractal surface model. Simulations of the adherence force versus the approaching distance of surfaces are performed. The necessity of considering the plastic deformation in microsized interface adhesion is also validated by comparing the result of the presented model with that of the Morrow model [J. Tribol 127, 206 (2005)]. The influence of fractal parameters on adherence force shows that controlling the surface topography is a readily means to control the adherence force for the elastic-plastic adhesion of rough surfaces, which is due to the intermolecular adhesive interactions. [ABSTRACT FROM AUTHOR]

We simulate by the Monte Carlo method the dynamic hysteresis and hysteresis scaling of two-dimensional uniaxially anisotropic Heisenberg model submitted to a time-oscillating magnetic field h of frequency f and amplitude h0. It is revealed that the hysteresis area as a function of f and h0 exhibits well-defined power law behaviors over the low-f and high-f regimes as long as h0 is high enough. The power law exponents for the low-f regime depend on the uniaxially anisotropic factor K, but the exponents for the high-f regime are universal. The single-peaked hysteresis dispersions at different h0 follow both the empirical scaling approach and the single-variable scaling hypothesis, respectively, demonstrating the existence of a single characteristic time for spin reversal mode given the amplitude h0. However, the scaling may be broken when h0 is low, due to the coexistence of spin reversal and spin tilting resonances. [ABSTRACT FROM AUTHOR]

The frequency dispersion of magnetic hysteresis for uniaxially anisotropic Heisenberg model in two-dimensional square lattice is studied by Monte Carlo simulation and mean-field calculation in order to investigate the high frequency spin-tilting resonance. It is observed that besides the spin reversal resonance which usually occurs under ac magnetic field of low frequency and high amplitude, the hysteresis dispersion may exhibit a second peak which occurs in the regime of high frequency and low amplitude and depends on the uniaxial anisotropy. It is argued that the second peak originates from the spin tilting in resonance with the ac magnetic field, which generates hysteresis far from saturation and is essentially different from the spin reversal resonance peak. The high frequency power-law scaling of the hysteresis is demonstrated for both the spin reversal mode and spin-tilting mode. [ABSTRACT FROM AUTHOR]