Advances in Manufacturing

BRMPNet: bidirectional recurrent motion planning networks for generic robotic platforms in smart manufacturing

Bo-Han Feng, Bo-Yan Li, Xin-Ting Jiang, Qi Zhou, You-Yi Bi

2025, 13(3): 477-492. doi:10.1007/s40436-024-00529-6

摘要 ( 9 )

PDF (259KB) ( 2 )

参考文献 | 相关文章 | 多维度评价

In the era of Industry 4.0, robot motion planning faces unprecedented challenges in adapting those high-dimension dynamic working environments with rigorous real-time planning requirements. Traditional sampling-based planning algorithms can find solutions in high-dimensional spaces but often struggle with achieving the balance among computational efficiency, real-time adaptability, and solution optimality. To overcome these challenges and unlock the full potential of robotic automation in smart manufacturing, we propose bidirectional recurrent motion planning network (BRMPNet). As an imitation learning-based approach for robot motion planning, it leverages deep neural networks to learn the heuristics for approximate-optimal path planning. BRMPNet employs the refined PointNet++ network to incorporate raw point-cloud information from depth sensors and generates paths with a bidirectional strategy using long short-term memory (LSTM) network. It can also be integrated with traditional sampling-based planning algorithms, offering theoretical assurance of the probabilistic completeness for solutions. To validate the effectiveness of BRMPNet, we conduct a series of experiments, benchmarking its performance against the state-of-the-art motion planning algorithms. These experiments are specifically designed to simulate common operations encountered within generic robotic platforms in smart manufacturing such as mobile robots and multi-joint robotic arms. The results demonstrate BRMPNet’s superior performance on key metrics including solution quality and computational efficiency, suggesting the promising potential of learning-based planning in addressing complex motion planning challenges.

The full text can be downloaded at https://link.springer.com/article/10.1007/s40436-024-00529-6

Separation of fringe patterns in fast deflectometric measurement of transparent optical elements based on neural network-assisted fast iterative filtering method

Ting Chen, Pei-De Yang, Xiang-Chao Zhang, Wei Lang, Yu-Nuo Chen, Min Xu

2025, 13(3): 493-510. doi:10.1007/s40436-024-00509-w

摘要 ( 9 )

PDF (258KB) ( 1 )

参考文献 | 相关文章 | 多维度评价

Transparent optical elements play a significant role in optical imaging and sensing, and the form qualities of these elements are critical to the functionalities of opto-electrical equipment. Therefore, rapid measurement of advanced transparent optical devices is urgently needed. Deflectometry, as a commonly used measurement method, has broad applications in form measurement. However, there are some challenges in the reflective deflectometric measurement of transparent elements, such as fringe superposition, low reflectivity, and non-uniform backgrounds, which severely affect the measurement accuracy. To address these issues, a single-frame fringe separation method is proposed for the deflectometric measurement of transparent elements. A fast iterative filtering method is utilized for coarse fringe separation and a convolutional neural network is adopted to solve the information leakage and incomplete fringe separation. The construction of the neural network involves improving and refining the filtering method to achieve precise separation of fringes. The proposed method achieves fringe separation and forms reconstruction of the upper and lower surfaces. Through simulations and experiments, the effectiveness and robustness of the proposed method are demonstrated, and the measurement accuracy can achieve 65 nm root-of-mean-squared-error (RMSE).

The full text can be downloaded at https://link.springer.com/article/10.1007/s40436-024-00509-w

Machine learning-based extraction of mechanical properties from multi-fidelity small punch test data

Zheng-Ni Yang, Jie Zou, Li Huang, Rui Yang, Jing-Yi Zhang, Chao Tong, Jing-Yu Kong, Zhen-Fei Zhan, Qing Liu

2025, 13(3): 511-524. doi:10.1007/s40436-024-00540-x

摘要 ( 8 )

PDF (256KB) ( 2 )

参考文献 | 相关文章 | 多维度评价

The extraction of mechanical properties plays a crucial role in understanding material behavior and predicting performance in various applications. However, the traditional methods for determining these properties often involve complex and time-consuming tests, which may not be practical in certain situations. To address this challenge, we developed a novel machine learning methodology that leveraged multi-fidelity datasets obtained from small punch test (SPT) experiments. SPT is a simple technique in which a localized load is applied to a small specimen, and the resulting deformation is measured. By analyzing the load-displacement data obtained from the SPT, valuable insights into the mechanical properties of the material can be obtained. In this study, we developed a multi-fidelity model capable of predicting the mechanical properties of steel and aluminum alloys. The proposed model considers variations in the material thickness and can effectively predict the mechanical properties of materials with different thicknesses, accommodating practical scenarios in which material samples exhibit varying thicknesses owing to different applications or manufacturing processes. In constructing our model, we synergistically incorporated low-fidelity finite element method (FEM) data and high-fidelity experimental data to predict the material properties. This integration enabled us to optimize and bolster the accuracy of our predictions, thereby facilitating a comprehensive and dependable characterization of the mechanical behavior of the material. By leveraging the advantages of SPT and incorporating multi-fidelity modeling techniques, our approach offers a practical and efficient solution for extracting mechanical properties. The ability to predict the properties of steel and aluminum alloys and materials with varying thicknesses enhances the versatility and applicability of our model in real-world scenarios.

The full text can be downloaded at https://link.springer.com/article/10.1007/s40436-024-00540-x

AI-enabled intelligent cockpit proactive affective interaction: middle-level feature fusion dual-branch deep learning network for driver emotion recognition

Ying-Zhang Wu, Wen-Bo Li, Yu-Jing Liu, Guan-Zhong Zeng, Cheng-Mou Li, Hua-Min Jin, Shen Li, Gang Guo

2025, 13(3): 525-538. doi:10.1007/s40436-024-00519-8

摘要 ( 10 )

PDF (247KB) ( 1 )

参考文献 | 相关文章 | 多维度评价

Advances in artificial intelligence (AI) technology are propelling the rapid development of automotive intelligent cockpits. The active perception of driver emotions significantly impacts road traffic safety. Consequently, the development of driver emotion recognition technology is crucial for ensuring driving safety in the advanced driver assistance system (ADAS) of the automotive intelligent cockpit. The ongoing advancements in AI technology offer a compelling avenue for implementing proactive affective interaction technology. This study introduced the multimodal driver emotion recognition network (MDERNet), a dual-branch deep learning network that temporally fused driver facial expression features and driving behavior features for non-contact driver emotion recognition. The proposed model was validated on publicly available datasets such as CK+, RAVDESS, DEAP, and PPB-Emo, recognizing discrete and dimensional emotions. The results indicated that the proposed model demonstrated advanced recognition performance, and ablation experiments confirmed the significance of various model components. The proposed method serves as a fundamental reference for multimodal feature fusion in driver emotion recognition and contributes to the advancement of ADAS within automotive intelligent cockpits.

The full text can be downloaded at https://link.springer.com/article/10.1007/s40436-024-00519-8

An AI-assistant health state evaluation method of sensing devices

Le-Feng Shi, Guan-Hong Chen, Gan-Wen Chen

2025, 13(3): 539-551. doi:10.1007/s40436-024-00517-w

摘要 ( 7 )

PDF (248KB) ( 1 )

参考文献 | 相关文章 | 多维度评价

The health states of sensing devices have a long-reaching influence on many smart application scenarios, such as smart energy and intelligent manufacturing. This paper proposes an ensemble methodology of the health-state evaluation of sensing devices, based on artificial intelligence (AI) technologies, which firstly takes into the operational characteristics, then designs a method of scenario identification to extract the typical scenarios, and subsequently puts forth a specific health-state evaluation. This method could infer the causalities of faulty devices effectively, which provides the interpretable basis for the health-state evaluation and enhances the evaluation accuracy of the health states. The suggested method has the promising potential to support the efficiently fine management of sensing devices in smart age.

The full text can be downloaded at https://link.springer.com/article/10.1007/s40436-024-00517-w

Pears classification by identifying internal defects based on X-ray images and neural networks

Ning Wang, Sai-Kun Yu, Zheng-Pan Qi, Xiang-Yan Ding, Xiao Wu, Ning Hu

2025, 13(3): 552-561. doi:10.1007/s40436-024-00512-1

摘要 ( 11 )

PDF (248KB) ( 1 )

参考文献 | 相关文章 | 多维度评价

In order to increase the sales and profitability, it is essential to classify the pears according to the external morphology (including shape, color and luster) and internal defects that can be quantitatively detected by various approaches. However, the existing classification methods concentrate mainly on the external quality rather than the internal defects. Therefore, this investigation develops an efficient and accurate classification method that can identify the internal sclerosis and bruises by combining the X-ray non-destructive testing and the convolutional neural network. Initially, the relations between the characteristics of the internal defects, i.e., internal sclerosis and bruises, and the grayscale features of the X-ray images are analyzed to provide the experimental data and demonstrate the theoretical foundations. Then, the X-ray images are processed by resolution reduction, feature enhancement and gradient reconstruction to improve the training efficiency and classification precision. Finally, the 18-layer residual network (ResNet-18) is optimized and trained to identify the internal bruises and sclerosis and classify the pears based on the identification results. It is found that the overall accuracy can reach 96.67% for identifying the bruised and sclerotic pears. The proposed method could also be applied to other fruits for defects identification and quality classification.

The full text can be downloaded at https://link.springer.com/article/10.1007/s40436-024-00512-1

Surface softening mechanism based on microstructure analyses under ultrasonic impact condition for Ti-17 titanium alloy

Chang-Feng Yao, Wen-Hao Tang, Liang Tan, Min-Chao Cui, Yun-Qi Sun, Tao Fan, Xu-Hang Gao

2025, 13(3): 562-583. doi:10.1007/s40436-024-00525-w

摘要 ( 8 )

PDF (382KB) ( 1 )

参考文献 | 相关文章 | 多维度评价

Ultrasonic impact significantly influences the mechanical properties and flow stress of Ti-17 titanium alloy. In this study, compression tests on Ti-17 titanium alloy were conducted under ultrasonic impact conditions, varying ultrasonic amplitudes and compression rates. The flow stress, surface elemental content, microhardness, and microstructure of Ti-17 titanium alloy were tested, and the softening mechanism of Ti-17 titanium alloy under ultrasonic impact conditions was investigated. The results indicate that the softening mechanism of Ti-17 titanium alloy involved ultrasonic softening combined with stress superposition. Ultrasonic impact leads to a higher distribution of grain orientation differences, alters the distribution of small-angle grain boundaries, and changes the distribution of surface phases, resulting in a reduced density of α phases. The geometrically necessary dislocation density at the surface increases, and the average grain size decreases from 2.91 μm to 2.73 μm. The Brass-type texture essentially disappears, transforming mainly into a Copper-type texture {112}<11-1>, with the maximum pole density decreasing from 73.98 to 39.88.

The full text can be downloaded at https://link.springer.com/article/10.1007/s40436-024-00525-w

Allowance distribution and parameters optimization for high-performance machining of low rigidity parts in multistage machining processes

Hao Sun, Sheng-Qiang Zhao, Fang-Yu Peng, Rong Yan, Xiao-Wei Tang

2025, 13(3): 584-605. doi:10.1007/s40436-024-00520-1

摘要 ( 8 )

PDF (254KB) ( 1 )

参考文献 | 相关文章 | 多维度评价

There are a large number of low rigidity parts in the aerospace field, and how to achieve high-performance manufacturing in their multistage machining processes has received increasing attention. Optimizing the distribution of machining allowance and machining parameters is one of the most convenient ways to improve the machining performance of these parts. In this paper, firstly, considering the machining accuracy and machining efficiency comprehensively, the error efficiency cooperation coefficient of low rigidity parts during machining is established. Based on the semi-parametric regression theory and measured data, the machining error transfer factor within the cooperation coefficient is calibrated. Secondly, the machining optimization strategy based on the Bayesian framework is proposed, and the optimization of multiple machining parameters is realized with the goal of minimizing the error efficiency cooperation coefficient. Finally, the optimization software of machining processes of low rigidity parts for engineering application is developed. In the verification experiments of blade parts, the error efficiency cooperation coefficient is reduced to 0.032 1 after optimization, and the average improvement of machining errors of all measured points is 14.31 μm. Besides, the above method is applied to low rigidity shaft parts, and the effectiveness of the proposed method is further verified.

The full text can be downloaded at https://link.springer.com/article/10.1007/s40436-024-00520-1

Study on the mechanism of burr formation in ultrasonic vibration-assisted honing 9Cr18MoV valve sleeve

Peng Wang, Chang-Yong Yang, Ying-Ying Yuan, Yu-Can Fu, Wen-Feng Ding, Jiu-Hua Xu, Yong Chen

2025, 13(3): 606-619. doi:10.1007/s40436-024-00516-x

摘要 ( 8 )

PDF (258KB) ( 2 )

参考文献 | 相关文章 | 多维度评价

The precision, lifespan, and stability of the electro-hydraulic servo valve sleeve are significantly impacted by the edge burrs that are easily created when honing the valve sleeve. The existing deburring process mainly rely on manual operation with high cost and low efficiency. This paper focuses on reducing the burr size during the machining process. In this paper, a single-scratch test with a finite element simulation model is conducted to study the mechanism of burr generation. The tests were carried out under ultrasonic vibration and non-ultrasonic vibration conditions to explore the effect of ultrasonic vibration on burrs. Besides, a honing experiment is conducted to verify the conclusions. The results at various cutting parameters are analyzed, and the mechanism of burr generation is revealed. The stiffness lacking of the workpiece edge material is the main reason for the burr generation. The cutting depth shows a significant effect on burr size while the cutting speed does not. The inhibition mechanism of ultrasonic vibration on burrs is also revealed. The separation of the burr stress field under ultrasonic vibration and the higher bending hinge point is the reason for burr fracturing. The re-cutting effect of ultrasonic vibration reduces the burr growth rate. The results of the honing experiment verified these conclusions and obtained a combination of honing parameters to minimize the burr growth rate.

The full text can be downloaded at https://link.springer.com/article/10.1007/s40436-024-00516-x

The effect of the slope angle and the magnetic field on the surface quality of nickel-based superalloys in blasting erosion arc machining

Lin Gu, Ke-Lin Li, Xiao-Ka Wang, Guo-Jian He

2025, 13(3): 620-633. doi:10.1007/s40436-024-00523-y

摘要 ( 8 )

PDF (373KB) ( 1 )

参考文献 | 相关文章 | 多维度评价

Electrical arc machining (EAM) is an efficient process for machining difficult-to-cut materials. However, limited research has been conducted on sloped surface machining within this context, constraining the further application for complex components. This study conducts bevel machining experiments, pointing out that the surface quality becomes unsatisfactory with the increasing bevel angle. The discharge condition is counted and analyzed, while the flow field and the removed particle movement of the discharge gap are simulated, demonstrating the primary factor contributing to the degradation of surface quality, namely the loss of flushing. This weakens both the plasma control effect and debris evacuation, leading to the poor discharge condition. To address this issue, the magnetic field is implemented in blasting erosion arc machining (BEAM). The application of a magnetic field effectively regulates the arc plasma, enhances debris expulsion, and significantly improves the discharge conditions, resulting in a smoother and more uniform sloped surface with a reduced recast layer thickness. This approach provides the possibility of applying BEAM to complex parts made of difficult-to-cut materials in aerospace and military industries.

The full text can be downloaded at https://link.springer.com/article/10.1007/s40436-024-00523-y

Effect of annealing and strain rate on the microstructure and mechanical properties of austenitic stainless steel 316L manufactured by selective laser melting

Zhi-Ping Zhou, Zhi-Heng Tan, Jin-Long Lv, Shu-Ye Zhang, Di Liu

2025, 13(3): 634-654. doi:10.1007/s40436-024-00528-7

摘要 ( 10 )

PDF (268KB) ( 1 )

参考文献 | 相关文章 | 多维度评价

New insights are proposed regarding the α′-martensite transformation and strengthening mechanisms of austenitic stainless steel 316L fabricated using selective laser melting (SLM-ed 316L SS). This study investigates the effects of annealing on the microstructural evolution, mechanical properties, and corrosion resistance of SLM-ed 316L SS specimens. The exceptional ultimate tensile strength (807 MPa) and good elongation (24.6%) of SLM-ed 316L SS was achieved by SLM process and annealing treatment at 900 ℃ for 1 h, which was attributed to effective dislocation strengthening and grain boundary strengthening. During tensile deformation, annealed samples exhibited deformation twinning as a result of the migration from high-angle grain boundaries to low-angle grain boundaries, facilitating the α′-martensite transformation. Consequently, a deformation mechanism model is proposed. The contribution of dislocation strengthening (~61.4%) is the most important strengthening factor for SLM-ed 316L SS annealed 900 ℃ for 1 h, followed by grain boundary strengthening and solid solution strengthening. Furthermore, the corrosion resistance of SLM-ed 316L SS after annealing treatment is poor due to its limited re-passivation ability.

The full text can be downloaded at https://link.springer.com/article/10.1007/s40436-024-00528-7

Mechanical properties and failure mechanisms of self-piercing riveted aluminum alloys with different edge distances

Jin-Rui Duan, Chao Chen

2025, 13(3): 655-667. doi:10.1007/s40436-024-00541-w

摘要 ( 7 )

PDF (259KB) ( 1 )

参考文献 | 相关文章 | 多维度评价

Self-piercing riveting (SPR) is widely used in thin-walled structures and the automotive industry to join aluminum alloy sheets. Lightweight vehicles are a common trend in the automotive industry. To further reduce vehicle weight and ensure the strength of the AA5052 aluminum alloy thin-sheet joint, the optimization of the amount of material used in the joint should be considered. The effect of the riveting position on the joint strength was investigated using riveting methods with different edge distances. Five edge distances (4.5, 6.5, 8.5, 10.5 and 12.5 mm) along the longitudinal direction were used in the investigations. In addition, a shear test was performed to analyze the mechanical properties of the joint. The results showed that as the edge distance decreased, the damage pattern of the joint changed from rivet pulling out of the plate to tearing at the upper plate edge, and as the edge pitch increased, the lap shear strength gradually increased. The minimum edge distance required to meet the deformation strength of the joint was 8.5 mm. This study provides a reference for reducing the amount of joint material, achieving lightweight production of automobiles, and failure repair of joints.

The full text can be downloaded at https://link.springer.com/article/10.1007/s40436-024-00541-w

Temperature evolution prediction for laser directed energy deposition enabled by finite element modelling and bi-directional gated recurrent unit

Kai-Xiong Hu, Kai Guo, Wei-Dong Li, Yang-Hui Wang

2025, 13(3): 668-687. doi:10.1007/s40436-024-00511-2

摘要 ( 8 )

PDF (259KB) ( 4 )

参考文献 | 相关文章 | 多维度评价

In the laser-directed energy deposition (L-DED) process, achieving an efficient temperature evolution prediction of molten pools is critical but challenging. To resolve this issue, this study presents an innovative approach that integrates a high-fidelity finite element (FE) model and an effective machine-learning model. Firstly, a high-fidelity FE model for the L-DED process was developed and subsequently validated through an experimental examination of the cross-sectional geometries of the molten pools and temperature fields of the substrate. Then, a Bi-directional gated recurrent unit (Bi-GRU) was formulated to predict the temperature evolution of the molten pools during L-DED. By training the Bi-GRU model using datasets generated from the FE model, it was deployed to efficiently predict the temperature evolution of the manufactured multi-layer single-bead walls. The results demonstrated that, in terms of the average mean absolute error, this approach outperformed other approaches designed based on the gated recurrent unit (GRU) model, long short-term memory model, and recurrent neural network models by 26.7%, 52.1%, and 65.2%, respectively. The results also showed that the prediction time required by this approach, once trained, was significantly reduced by five orders of magnitude compared with the FE model. Therefore, this approach accurately predicts the temperature evolution of multi-layer single-bead walls in a computationally efficient manner. This approach is a promising solution for supporting the real-time control of the L-DED process in industrial applications.

The full text can be downloaded at https://link.springer.com/article/10.1007/s40436-024-00511-2

当期目录