Defect formation is a critical challenge for powder-based metal additive manufacturing (AM). Current understanding on the three important issues including formation mechanism, influence and control method of metal AM defects should be updated. In this review paper, multi-scale defects in AMed metals and alloys are identified and for the first time classified into three categories, including geometry related, surface integrity related and microstructural defects. In particular, the microstructural defects are further divided into internal cracks and pores, textured columnar grains, compositional defects and dislocation cells. The root causes of the multi-scale defects are discussed. The key factors that affect the defect formation are identified and analyzed. The detection methods and modeling of the multi-scale defects are briefly introduced. The effects of the multi-scale defects on the mechanical properties especially for tensile properties and fatigue performance of AMed metallic components are reviewed. Various control and mitigation methods for the corresponding defects, include process parameter control, post processing, alloy design and hybrid AM techniques, are summarized and discussed. From research aspect, current research gaps and future prospects from three important aspects of the multi-scale AM defects are identified and delineated. 

Multi-scale defects in powder-based additively manufactured metals and alloys

Compared with traditional manufacturing methods, laser additive manufacturing (LAM) has incomparable and huge advantages when preparing functionally gradient materials. However, due to the large temperature gradient and the change of material composition, the residual stress of the structure can be very large, which will seriously affect the mechanical properties of the structure, and cause the structure to bend and crack. In this paper based on the finite element method, the residual stress of TC4/Inconel718 materials widely used in aerospace is studied. And it mainly focuses on the influence of key parameters such as laser energy and material composition of transition layers. The results show that the laser power has a positive correlation with the residual stress of the structure and the change of the residual stress of the contact surface is relatively gentle when the components of the two gradient materials are close. The results obtained are in good agreement with the experiment. The research in this article has reference value for understanding, measuring and reasonably controlling the residual stress of functionally gradient materials by LAM. 

Finite element analysis for residual stress of TC4/Inconel718 functionally gradient materials produced by laser additive manufacturing

The laser powder bed fusion (L-PBF) method of additive manufacturing (AM) is increasingly used in various industrial manufacturing fields due to its high material utilization and design freedom of parts. However, the parts produced by L-PBF usually contain such defects as crack and porosity because of the technological characteristics of L-PBF, which affect the quality of the product. Laser ultrasonic testing (LUT) is a potential technology for on-line testing of the L-PBF process. It is a non-contact and non-destructive approach based on signals from abundant waveforms with a wide frequency-band. In this study, a method of LUT for on-line inspection of L-PBF process was proposed, and a system of LUT was established approaching the actual environment of on-line detection to evaluate the method applicability for defects detection of L-PBF parts. The detection results of near-surface defects in L-PBF 316L stainless steel parts show that the crack-type defects with a sub-millimeter level within 0.5 mm depth can be identified, and accordingly, the positions and dimensions information can be acquired. The results were verified by X-ray computed tomography, which indicates that the present method exhibits great potential for on-line inspection of AM processes.

/pdf/laser-ultrasonic-testing-for-near-surface-defects-inspection-atd5r6x6sn.pdf

Laser ultrasonic testing for near-surface defects inspection of 316L stainless steel fabricated by laser powder bed fusion

Influences of processing parameters and heat treatment on microstructure and mechanical behavior of Ti-6Al-4V fabricated using selective laser melting

In this work, Fe-based coatings were prepared by laser cladding and the effects of multi-step laser remelting (MLR) on phase components, microstructure evolution, hardness and wear behavior were investigated. The as-cladded coating exhibits a complex phase structure and no obvious variation is observed when remelted for 1–3 times. The as-cladded coating comprises three structures including Mo-rich dendrites, γ -(Fe, Cr) matrix and several kinds of intermetallic compounds . The Mo-rich dendrites were gradually refined to nanoscale size grains with the increase of remelting times. The Vickers hardness and wear resistance of the Fe-based coatings improves firstly and then aggravates with the MLR process. The highest hardness of 1487.4 HV 0.1 and the smallest wear rate of (0.21 ± 0.01) × 10 −6 mm 3 /(N·m) is obtained when the coating is remelted twice. The improved wear performance mainly comes from the existence of hard intermetallic compounds and the refinement of Mo-rich dendrites. Despite a slight compensation of hardness and wear resistance, the coating remelted for three times also exhibits an excellent wear performance due to the microstructure homogenization and the absence of fatigue damage. This work broadens the opportunities to the microstructure modification and property optimization for laser-cladded coatings and presents further understanding on laser manufacturing technology. • Multiphase Fe-based coatings with super-high Vickers hardness and excellent wear resistance were prepared by laser cladding. • Multi-step laser remelting process can refine Mo-rich dendrites to nanoscale size, thereby promoting microstructure homogeneity and release lattice stress. • The coating remelted twice has the highest hardness of 1487.4 HV 0.1 and the lowest wear rate of (0.21 ± 0.01) × 10 −6 mm 3 /(N·m). • Remelting for three times can prevent the coating from fatigue wear damage. 

Dendrite refinement and wear performance enhancement in laser-cladded Fe-based coatings after multi-step laser remelting

Research on the influence of heat treatment on residual stress of TC4 alloy produced by laser additive manufacturing based on laser ultrasonic technique

DONGWANG, Center for Speech and Language Technologies, Tsinghua University, China QINGYANG HONG, School of Informatics, Xiamen University, China LANTIAN LI, Center for Speech and Language Technologies, Tsinghua University, China WENQIANG DU, Center for Speech and Language Technologies, Tsinghua University, China YANG ZHANG, Shenzhen International Graduate School, Tsinghua University, China TAO JIANG, Xiamen TalentedSoft Co. Ltd, China HUI BU, Beijing Shell Shell Technology Co. Ltd, China XIN XU, Beijing Shell Shell Technology Co. Ltd, China

CNSRC 2022 Evaluation Plan

Evaluation trials are used to probe performance of automatic speaker veriﬁcation (ASV) systems. In spite of the clear importance and impact, evaluation trials have not been seriously treated in research and engineering practice. This paper ﬁrstly presents a theoretical analysis on evaluation trials and highlight-s potential bias with the most popular cross-pairing approach used in trials design. To interpret and settle this problem, we deﬁne the concept of trial conﬁg and C-P map derived from it. The C-P map measures the performance of an ASV system on various trial conﬁgs in a 2-dimensional map. On the map, each location represents a particular trial conﬁg and its corresponding color represents the system performance. Experiments conducted on representative ASV systems show that the proposed C-P map offers a powerful evaluation toolkit for ASV performance analysis and comparison. The source code for C-P map has been release at https://gitlab.com/csltstu/sunine .

C-P Map: A Novel Evaluation Toolkit for Speaker Verification

With the emergence of deep learning, the performance of automatic speech recognition (ASR) systems has remarkably improved. Especially for resource-rich languages such as English and Chinese, commercial usage has been made feasible in a wide range of applications. However, most languages are low-resource languages, presenting three main difficulties for the development of ASR systems: (1) the scarcity of the data; (2) the uncertainty in the writing and pronunciation; (3) the individuality of each language. Uyghur, Kazakh, and Kyrgyz as examples are all low-resource languages, involving clear geographical variation in their pronunciation, and each language possesses its own unique acoustic properties and phonological rules. On the other hand, they all belong to the Altaic language family of the Altaic branch, so they share many commonalities. This paper presents an overview of speech recognition techniques developed for Uyghur, Kazakh, and Kyrgyz, with the purposes of (1) highlighting the techniques that are specifically effective for each language and generally effective for all of them and (2) discovering the important factors in promoting the speech recognition research of low-resource languages, by a comparative study of the development path of these three neighboring languages.

/pdf/automatic-speech-recognition-for-uyghur-kazakh-and-kyrgyz-an-3kbbkad1.pdf

Automatic Speech Recognition for Uyghur, Kazakh, and Kyrgyz: An Overview

Deep learning has significantly boosted the performance improvement of automatic speech recognition (ASR) with the cooperation of large amounts of data resources. For minority languages, however, there are almost no large-scale data resources, limiting the development of ASR technologies in these languages. In this paper, we publish a free Kirghiz speech database accompanied by associated language resources. The entire database involves 128 h of speech data from 163 speakers and corresponding transcriptions. To our knowledge, this is the largest Kirghiz speech database that is dedicated to the ASR task and is publicly free so far. In addition, we also provide several baseline systems based on Kaldi and WeNet to demonstrate how these public data resources can be used to facilitate the Kirghiz ASR research. This publication is a part of the M2ASR project, and all the resources can be downloaded at the project webpage.

/pdf/m2asr-kirghiz-a-free-kirghiz-speech-database-and-accompanied-16yhv84z.pdf

Wenqiang Du

Papers

Research on the influence of heat treatment on residual stress of TC4 alloy produced by laser additive manufacturing based on laser ultrasonic technique

CNSRC 2022 Evaluation Plan

C-P Map: A Novel Evaluation Toolkit for Speaker Verification

Automatic Speech Recognition for Uyghur, Kazakh, and Kyrgyz: An Overview

M2ASR-KIRGHIZ: A Free Kirghiz Speech Database and Accompanied Baselines