Laser Powder Bed Fusion (LPBF) has revolutionized high-end manufacturing, particularly in aerospace and biomedical fields. However, internal defects such as pores, cracks, and inclusions compromise the structural integrity and service reliability of LPBF components. Laser ultrasonics, a non-contact, broadband non-destructive testing (NDT) method, offers a promising solution for detecting and characterizing these defects. This study systematically investigated laser ultrasonic testing technology for LPBF-fabricated Ti6Al4V using a combined approach of physics-driven simulation modeling and experimental validation. To accurately model material anisotropy, a finite element model was developed that integrated Voronoi algorithm-generated polycrystalline microstructures with orientation-dependent elastic tensors, providing a comprehensive representation of the material’s microstructural heterogeneity. Simulation results revealed that while sub-100-μm defects yield weak ultrasonic scattering signals, the Synthetic Aperture Focusing Technique (SAFT) markedly improves the detection and imaging performance for such small-scale defects. Experimental validation using a laser ultrasonic system identified a 90 μm internal defect in the LPBF Ti6Al4V specimen, though a 75 μm defect was undetectable. This highlights the need for enhanced sensitivity. A signal processing method combining time-truncation principal component analysis (PCA) with targeted noise reduction and SAFT was proposed to reduce high-frequency noise and improve high-resolution imaging, enhancing defect detection accuracy. This study provides theoretical foundations and technical support for high-precision defect detection in metal additive manufacturing components, with significant implications for quality control in high-end equipment manufacturing.

The increasing heterogeneity of multi-agent systems poses significant challenges for jointly training a global model across agents. At the same time, cooperative inference between agents has long been recognized as a powerful mechanism for distributed decision making over networks. Motivated by these observations, we propose a collaboration framework for distributed binary classification over multi-agent networks, where a set of independently trained agents, potentially differing in architecture, feature space, or modality, coordinate their actions during test time to form collective predictions. This coordination is achieved by exchanging local decision statistics through a distributed learning protocol. We develop a theoretical and experimental study of this independent training and cooperative inference paradigm, and examine its performance under different communication budgets and distributed learning rules. We establish classification error guarantees under sufficient, finite-round, and finite-precision communication, together with PAC-style generalization bounds. These results capture the influence of model heterogeneity, network topology, combination policy, and communication constraints on prediction accuracy. Taken together with the experimental results, they reveal both the price of independent training and the benefit of collective prediction for the proposed distributed decision making framework with models learned from data.

Distributed decision-making over graphs involves a group of agents that collaboratively work toward a common objective. In the social learning framework, the agents are tasked to infer an unknown state from a finite set by using a stream of local observations. The probability of decision errors for each agent asymptotically converges to zero at an exponential rate, characterized by the error exponent, which depends on the combination policy employed by the network. This work addresses the challenge of identifying optimal combination policies to maximize the error exponent for the true state while ensuring the errors for all other states converge to zero as well. We derive an upper bound on the achievable error exponent under the social learning rule, and then establish conditions for the combination policy to reach this upper bound. Moreover, we examine the performance loss scenarios when the combination policy is chosen inappropriately. From a geometric perspective, each combination policy induces a weighted nearest neighbor classifier where the weights correspond to the agents' Perron centralities. By implementing an optimized combination policy, we enhance the error exponent, leading to improved accuracy and efficiency in the distributed decision-making process.

Distributed decision-making over networks involves multiple agents collaborating to achieve a common goal. In the social learning process, where agents aim at inferring an unknown state from a stream of local observations, the probability of error in their decisions converges to zero exponentially in the asymptotic regime. The rate of this convergence, known as the error exponent, is influenced by the combination policy employed by the network. This work addresses the challenge of identifying the optimal combination policies to maximize the error exponent. We establish an upper bound on the achievable error exponents by the social learning rule and provide the conditions for the combination policy to reach this upper bound. By implementing the optimized policy, we enhance the error exponent, leading to improved accuracy and efficiency in the distributed decision-making process.

This paper studies the probability of error associated with the social machine learning framework, which involves an independent training phase followed by a cooperative decision-making phase over a graph. This framework addresses the problem of classifying a stream of unlabeled data in a distributed manner. In this work, we examine the classification task with limited observations during the decision-making phase, which requires a non-asymptotic performance analysis. We establish a condition for consistent training and derive an upper bound on the probability of error for classification. The results clarify the dependence on the statistical properties of the data and the combination policy used over the graph. They also establish the exponential decay of the probability of error with respect to the number of unlabeled samples.

With the development of guided waves in the field of non-destructive testing, guided waves have also gradually been applied to the non-destructive testing of cable structures. However, the most commonly used and time-consuming 3D solid model, will cause amount of time in simulation and calculation of the long-distance transmission line, which will further make the difficulty in detecting the defect in long-distance transmission line. To overcome the problem, a low-time-consumed FEM model is established to study the characteristics of the defect of long transmission line. This model degrades from the 3D solid model and agrees good with it, which will greatly reduce the calculation time, especially in studying the frequency domain and time domain response of long-distance transmission line. The results of this proposed model are also verified by the 3D solid model. Based the proposed model, the location, length and radius of the defect can be all successfully predicted by a low calculation time. This study paves the way in the acoustic detection technology of the Long-distance transmission line from a calculation aspect.

In view of the non-destructive and non-contact features, laser ultrasonic (LU) technology has long been the effective method to detect tiny defects for laser powder bed fusion (LPBF) additive manufactured specimens. Of larger concern is the variation and the corresponding mechanism on tested results of LU detection as the property of LPBF additive manufactured specimen is changed. Aiming at the property of surface roughness, this work investigated the propagation characteristics of excited ultrasonic waves in LPBF additive manufactured 316L stainless steel with different surface roughness, as well as the interaction between ultrasonic waves and artificial submillimeter holes. Both numerical simulated and experimental study were conducted. Simulated results revealed that the amplitudes of longitudinal wave (L wave) and its echo wave L1 at the holes exhibited a discernible increase as the surface was coarser. The increase in surface roughness was detrimental to the resolution of defect detection as was expected from the increased amount of noise. LPBF fabrication and the subsequent LU pulse-echo detection were conducted for 316L stainless steel. Both B-scan and C-scan were able to detect the holes with the diameter of 0.6 mm. The speckle phenomenon deriving from the increase in surface roughness emerged, corresponding to the increased ultrasonic signal energy but deteriorated resolution of detected images. It is feasible to optimize LU detected effect by minimize the surface roughness of tested specimens.

As a flexible resource, inverter air conditioners (IACs) have significant potential for demand response. However, due to the diverse characteristics of users, it is challenging to consider users’ internal temperature requirements fairly while controlling IACs power. This paper aims to find a feasible solution to this problem. First, the concept of “comfort state" is utilized to describe the operating state of IACs with different reference temperatures and allowable internal temperature deviations. Then, a distributed leaderless coordinated framework is designed for IACs to ensure the tracking of reference aggregated power and the achieving of fair comfort level sharing. The proposed framework does not require the aggregator to manage all IACs in a centralized manner and also avoids the need of periodic re-issuance of aggregated power deviation. The feasibility of the proposed framework and the impact of time-delay is rigorously analyzed. Besides, a specific implementation method for the proposed framework in the primary frequency support is developed. Numerical studies show that the proposed coordinated framework can effectively mitigate the drastic variations in system frequency when loads fluctuate.

Purpose This paper aims to investigate the moisture diffusion behavior in a system-in-package module systematically by moisture-thermalmechanical-coupled finite element modeling with different structure parameters under increasingly harsh environment. Design/methodology/approach A finite element model for a system-in-package module was built with moisture-thermal-mechanical-coupled effects to study the subsequences of hygrothermal conditions. Findings It was found in this paper that the moisture diffusion path was mainly dominated by hygrothermal conditions, though structure parameters can affect the moisture distribution. At lower temperatures (30°C~85°C), the direction of moisture diffusion was from the periphery to the center of the module, which was commonly found in simulations and literatures. However, at relatively higher temperatures (125°C~220°C), the diffusion was from printed circuit board (PCB) to EMC due to the concentration gradient from PCB to EMC across the EMC/PCB interface. It was also found that there exists a critical thickness for EMC and PCB during the moisture diffusion. When the thickness of EMC or PCB increased to a certain value, the diffusion of moisture reached a stable state, and the concentration on the die surface in the packaging module hardly changed. A quantified correlation between the moisture diffusion coefficient and the critical thickness was then proposed for structure parameter optimization in the design of system-in-package module. Originality/value The different moisture diffusion behaviors at low and high temperatures have seldom been reported before. This work can facilitate the understanding of moisture diffusion within a package and offer some methods about minimizing its effect by design optimization.

In the development of additive manufacturing (AM) technology, laser powder bed fusion (LPBF) is one of the important processing methods. However, the hole defects in the fabricated samples limit the development. Laser ultrasonic (LU) technology plays a major role in the detection of LPBF parts with tiny defects, which has the advantages of non-contact and non-destructive. In this work, the detection of submillimeter internal defects in four typical LPBF alloys by LU technology is studied numerically and experimentally. A multiphysics simulation model of LU detection is established to investigate the propagation characteristics of excited ultrasonic waves in different LPBF alloys and their interaction with submillimeter artificial defects. Simulation results show that the amplitude of longitudinal (L) wave at the defect is the largest in AlSi10Mg alloy, and the amplitude of L wave in the 316L alloy, Ti6Al4V alloy and In718 alloy are very close, but their phases are slightly different. The amplitude of L wave tends to decrease nearly linearly with the increase in defect diameter. Then, four typical LPBF alloys are fabricated and measured by the LU through-transmission detection. The geometric information of artificial holes with a diameter larger than 0.2 mm are clearly characterized by the LU C-scan results, indicating the prominent applicability and feasibility of LU detection on different materials fabricated by LPBF.

Laser powder bed fusion (LPBF) is a cutting-edge metal additive manufacturing technology with promising applications in aerospace, biomedicine, and industrial automation. Its ability to fabricate intricate geometric shapes, achieve high surface precision, and deliver comprehensive results renders it highly advantageous. In LPBF process, several factors such as the interference of laser processing fluctuations, rapid cooling rate, and environmental gas, contribute to the initiation and expansion of defects including pores and cracks, limiting the application of LPBF components. Laser ultrasonic testing (LUT) is a promising non-destructive evaluation method that can accurately detected defects in LPBF manufactured components by non-contact generation and detection. In this paper, a multi-physical field LUT model for the pores and cracks in LPBF additively manufactured Ti6Al4V alloy is proposed. The interaction of ultrasound with defects produces defect reflection echo wave, diffracted wave, and transmission wave. The transmission and diffraction phenomena occur in the longitudinal sound waves at defects. Furthermore, the echo wave reflected at crack exhibited enhanced prominence and an irregular shape. Clear differences were observed in the effects of laser ultrasonic detection on different defect types, pore depths and sizes. Finally, the LUT technology achieved the detection of 90 μm pore defects in LPBF manufactured Ti6Al4V alloy.

Because of rapid heating, cooling, and solidification during metal additive manufacturing (AM), the resulting products exhibit strong anisotropy and are at risk of quality problems from metallurgical defects. The defects and anisotropy affect the fatigue resistance and material properties, including mechanical, electrical, and magnetic properties, which limit the applications of the additively manufactured components in the field of engineering. In this study, the anisotropy of laser power bed fusion 316L stainless steel components was first measured by conventional destructive approaches using metallographic methods, X-ray diffraction (XRD), and electron backscatter diffraction (EBSD). Then, anisotropy was also evaluated by ultrasonic nondestructive characterization using the wave speed, attenuation, and diffuse backscatter results. The results from the destructive and nondestructive methods were compared. The wave speed fluctuated in a small range, while the attenuation and diffuse backscatter results were varied depending on the build direction. Furthermore, a laser power bed fusion 316L stainless steel sample with a series of artificial defects along the build direction was investigated via laser ultrasonic testing, which is more commonly used for AM defect detection. The corresponding ultrasonic imaging was improved with the synthetic aperture focusing technique (SAFT), which was found to be in good agreement with the results from the digital radiograph (DR). The outcomes of this study provide additional information for anisotropy evaluation and defect detection for improving the quality of additively manufactured products.

The continuous online flaws detection of the melt pool plays a significant role in assessing the quality of the printed sample in metal additive manufacturing (AM). However, the continuous detection of melt pool flaws faces great challenges because the suitable continuous acquisition system and real-time processing algorithm for melt pool feature extraction are difficult to establish. In this work, a continuous online flaws detection method combining the photodiode signal and melt pool temperature based on deep learning algorithms in the Laser Powder Bed Fusion (LPBF) is proposed. A multi-signal fusion system is designed to synchronously obtain the photodiode signal and temperature signal of the melt pool. The characteristics of the melt pool photodiode signal under the action of pulsed laser and the characteristics of the melt pool temperature are explored. The aliasing of photodiode signals is found due to the different settings of pulse laser frequency and photodiode sensor acquisition rate in AM. According to the two principles that the melt pool temperature is related to the flaw formation, and accurate flaws identification can be carried out based on the melt pool temperature field, the Back Propagation Neural Network (BPNN), Stacked Sparse AutoEncoder (SSAE), and Long Short-Term Memory (LSTM) are used to build the correlation model between the photodiode signal and the average melt pool temperature, and the flaw detection is carried out through the average melt pool temperature error. Therefore, once the correlation model is established, continuous online flaws detection can be realized only by using photodiode signals. The results demonstrated that a robust correlation can be established between the photodiode signal and the average melt pool temperature through the neural network, and the correlation error can be as low as 2.2 %. The detection of flaws is simplified into a binary classification problem through a reasonable threshold setting, and the LSTM with 74.39 % detection accuracy is more suitable for flaws detection based on the photodiode signal of the melt pool.

Adaptive social learning is a useful tool for studying distributed decision-making problems over graphs. This paper investigates the effect of combination policies on the performance of adaptive social learning strategies. Using large-deviation analysis, it first derives a bound on the steady-state error probability and characterizes the optimal selection for the Perron eigenvectors of the combination policies. It subsequently studies the effect of the combination policy on the transient behavior of the learning strategy by estimating the adaptation time in the low signal-to-noise ratio regime. In the process, it is discovered that, interestingly, the influence of the combination policy on the transient behavior is insignificant, and thus it is more critical to employ policies that enhance the steady-state performance. The theoretical conclusions are illustrated by means of computer simulations.

Significance Metal additive manufacturing technology has several advantages such as efficient formation, short processing cycle, and cost effectiveness. Components with complex spatial structures can be produced via additive manufacturing, thereby overcoming the limitations of traditional manufacturing. Therefore, this technology is favored by the automotive, aerospace, and medical equipment industries. However, possible internal defects, such as lack of fusion, cracks, and holes during the forming process of additive manufacturing, limit its promotion and wide application in the industry. Furthermore, the microstructure of the components changes significantly with the variation in the laser power, process approaches, and scanning parameters during the additive manufacturing process. Furthermore, the stability of the phase and characteristic microstructures are affected by the protective gas, while controlling the surface topography. Therefore, the quality control of metal additive manufacturing products, particularly online monitoring, is of great strategic significance. Several approaches of nondestructive evaluation of flaw inspection and material characterization, such as X-ray computed tomography, fluorescent penetrant inspection, and ultrasonic testing, have attracted much interest. Particularly, ultrasonic testing is one of the most commonly used nondestructive methods for detecting internal defects. Compared to traditional ultrasonic nondestructive testing technology, laser ultrasonic nondestructive testing has the advantages of no-contact, high sensitivity, and suitability for harsh environment, which can realize rapid online monitoring. Progress In this paper, the characteristics of metal additive manufacturing and nondestructive testing on additive manufacturing are briefly introduced, highlighting the fact that applying laser ultrasonic testing on metal additive manufacturing has great strategic significance. Then, two kinds of laser ultrasonic mechanisms are analyzed: thermoelastic mechanism and ablation mechanism. Under the laser ultrasonic simulations and experimentations, the thermoelastic mechanism is chosen without destroying the integrity or performance of the additive manufacturing components. Next, the finite element simulation studies on laser ultrasonic detection are introduced. Based on the finite element method (FEM), the complex models can be processed and the global numerical solution can be obtained by solving heat conduction and thermoelastic equations. Afterward, the principle of laser ultrasonic nondestructive testing and the testing system are introduced, in which common detection methods on laser ultrasonic are listed and briefly explained. Some improving methods on laser ultrasonic testing systems are also discussed. Yan et al. proposed an experimental method of a no-contact all-optical laser ultrasonic detection and built the optical differential detection system using the beam deflection technique, which improved the antinoise ability of the optical path. Finally, the application progress of laser ultrasonic nondestructive testing of metal traditional and additive manufacturing material at domestic and foreign industries is systematically summarized. Moreover, we have analyzed the research progress-both home and abroad-for reference. Studies on laser ultrasonic began earlier abroad. One example is the study by Pierce et al. (1993), who successfully used a pulsed Nd:YAG laser to excite ultrasonic waves in metal aluminum blocks and increased the laser ultrasonic signals by modulating the frequency of laser source, thus excavating the great prospect of laser ultrasonic in nondestructive testing. In comparison, relevant research in China only started in 2006. For instance, Shen et al. detected rectangular metal aluminum blocks with an artificial surface defect with depth of 0.71 mm and width of 2.00 mm by constructing an optical differential detection system based on the beam deflection method, accurately locating the surface defect position. In addition, laser ultrasonic can be used for detecting the defects and measuring other significant parameters or monitoring other important processes, such as characterizing the elastic modulus, measuring residual stress of additive manufacturing alloy parts, and monitoring the changes in additive manufacturing process like recrystallization. Conclusion and Prospect Several studies have shown the feasibility of laser ultrasonic nondestructive testing on metal additive manufacturing. Given that the effect of additive manufacturing process parameters on quality has been widely studied and reported, developing a link between controllable process parameters and the required process characteristics to support feedforward and feedback control is the best way to achieve the goal of its application in future control systems. However, various challenges still exist. For example, the relationships among the parameters must be identified, including the parameters measured from experiments and those important parameters for monitoring and characterizing but cannot be obtained directly. Furthermore, an algorithm for quickly identifying different defects should be explored, and a corresponding feedback control scheme must be established to improve the quality of the additive manufacturing components. Moreover, the feedback data obtained from the entire online monitoring of the additive manufacturing process are expected to be massive and difficult to deal with. Thus, we should find an algorithm that can efficiently process these data, detect the anomalies in real time, and provide corresponding feedback to the closed-loop detection system, thus allowing us to control the manufacturing engineering of the workpiece in real time and ultimately improve the quality of finished products.

The transmission overload after a disturbance poses significant security risk to the power system. Once it happens, an efficient remedial action must be taken to relieve the line overloads as quickly as possible. The line overload mitigation problem becomes more urgent and challenging in the future power grid that integrates distributed and fluctuating renewable energies. In this brief, we propose a distributed corrective control scheme for alleviating the overloads in transmission lines. A linearized AC power flow model is introduced, and the error of linearizion is compensated through a close loop of power flow control. The proposed distributed control framework not only satisfies the need of accommodating distributed resources in future grids, but also guarantees the system security during the line overload mitigation process. The performance of our corrective control scheme is demonstrated in the simulations using the IEEE 14-bus test system.

This paper investigates the effect of combination policies on the performance of adaptive social learning in non-stationary environments. By analyzing the relation between the error probability and the underlying graph topology, we prove that in the slow adaptation regime, combination policies with a uniform Perron eigenvector will provide the smallest steady-state error probability. This result indicates that in terms of learning accuracy, doubly-stochastic combination policies yield optimal performance. Moreover, we estimate the adaptation time of adaptive social learning in the small signal-to-noise regime and show that in this regime, the influence of combination policies on the adaptation time is insignificant.

A numerical model is presented in this article to investigate the interactions between laser generated ultrasonic and the microdefects (0.01 to 0.1 mm), which are on the surface of the laser powder bed fusion additive manufactured 316L stainless steel. Firstly, the influence of the transient sound field and detection positions on Rayleigh wave signals are investigated. The interactions between the varied microdefects and the laser ultrasonic are studied. It is shown that arrival time of reflected Rayleigh (RR) waves wave is only related to the location of defects. The depth can be checked from the feature point Q, the displacement amplitude and time delay of converted transverse (RS) wave, while the width information can be evaluated from the RS wave time delay. With the aid of fitting curves, it is found to be linearly related. This simulation study provides a theoretical basis for quantitative detection of surface microdefects of additive manufactured 316L stainless steel components.

Glioblastoma (GBM) is the most common and aggressive primary malignant brain tumor. The unregulated expression of Claudin-4 (CLDN4) plays an important role in tumor progression. However, the biological role of CLDN4 in GBM is still unknown. This study aimed to determine whether CLDN4 mediates glioma malignant progression, if so, it would further explore the molecular mechanisms of carcinogenesis. Our results revealed that CLDN4 was significantly upregulated in glioma specimens and cells. The inhibition of CLND4 expression could inhibit mesenchymal transformation, cell invasion, cell migration and tumor growth in vitro and in vivo. Moreover, combined with in vitro analysis, we found that CLDN4 can modulate tumor necrosis factor-α (TNF-α) signal pathway. Meanwhile, we also validated that the transforming growth factor-β (TGF-β) signal pathway can upregulate the expression of CLDN4, and promote the invasion ability of GBM cells. Conversely, TGF-β signal pathway inhibitor ITD-1 can downregulate the expression of CLDN4, and inhibit the invasion ability of GBM cells. Furthermore, we found that TGF-β can promote the nuclear translocation of CLDN4. In summary, our findings indicated that the TGF-β/CLDN4/TNF-α/NF-κB signal axis plays a key role in the biological progression of glioma. Disrupting the function of this signal axis may represent a new treatment strategy for patients with GBM.

Background:Detection of circulating tumor cells (CTCs) is a promising technology in tumor management; however, the slow development of CTC identification methods hinders their clinical utility. Moreover, CTC detection is currently challenging owing to major issues such as isolation and correct identification. To improve the identification efficiency of glioma CTCs, we developed a karyoplasmic ratio (KR)-based identification method and constructed an automatic recognition algorithm. We also intended to determine the correlation between high-KR CTC and patients' clinical characteristics.Methods:CTCs were isolated from the peripheral blood samples of 68 glioma patients and analyzed using DNA-seq and immunofluorescence staining. Subsequently, the clinical information of both glioma patients and matched individuals was collected for analyses. ROC curve was performed to evaluate the efficiency of the KR-based identification method. Finally, CTC images were captured and used for developing a CTC recognition algorithm.Results:KR was a better parameter than cell size for identifying glioma CTCs. We demonstrated that low CTC counts were independently associated with isocitrate dehydrogenase (IDH) mutations (p = 0.024) and 1p19q co-deletion status (p = 0.05), highlighting its utility in predicting oligodendroglioma (area under the curve = 0.770). The accuracy, sensitivity, and specificity of our algorithm were 93.4%, 81.0%, and 97.4%, respectively, whereas the precision and F1 score were 90.9% and 85.7%, respectively.Conclusion:Our findings remarkably increased the efficiency of detecting glioma CTCs and revealed a correlation between CTC counts and patients' clinical characteristics. This will allow researchers to further investigate the clinical utility of CTCs. Moreover, our automatic recognition algorithm can maintain high precision in the CTC identification process, shorten the time and cost, and significantly reduce the burden on clinicians.