Search

Film cooling is an effective technology for improving the thrust-to-weight ratio of an aero-engine. The axial orientation of film cooling holes is one of the key factors determining the film cooling effect. However, the machining quality of cooling holes is still carried out by manual inspection, whose accuracy and consistency cannot be guaranteed. To address this problem, this paper proposes an axial orientation extraction algorithm based on laser point cloud processing. Firstly, the blade surface is scanned by a line laser sensor and a blade point cloud is obtained. Secondly, a single film hole can be segmented from the blade point cloud based on scanning line fitting residual. Thirdly, an improved Gaussian map is to complement the arc belt obtained by Gaussian map to obtain a complete circle along an error-sensitive direction. Finally, the axial orientation is estimated by random sample consensus with an initial value. The accuracy and consistency of the algorithm are verified by comparing the extraction result by the algorithm with those by the Geomagic software. The result shows that the axial accuracy extracted by the algorithm can reach 0.405°. [ABSTRACT FROM AUTHOR]

Freehand sketching on a paper is commonly found to be used as it is a natural and intuitive communication channel for portraying ideas. Additional strokes are often drawn over the existing ones to make the sketch clearer. For its usefulness in subsequent steps, a sketch is transformed to be a 3D CAD model that is typically performed manually by a designer due to a lack of technology support. Geometrical reconstruction from an offline scanning of a sketch has been a challenge because enriched information obtained during sketching is reduced to a batch of data points. This paper presents an algorithm for identifying a single-line drawing from a paper-based overtraced sketch as it is the first important step towards geometrical reconstruction. Key activities in this approach are thick-line sketch creation, contour boundary extraction, segments identification, and line drawing creation. The algorithm has been successfully implemented on LabVIEW and tested with several samples. [ABSTRACT FROM AUTHOR]

In order to realize real-time and precise monitoring of the tool wear in the milling process, this paper presents a tool wear predictive model based on the stacked multilayer denoising autoencoders (SMDAE) technique, the particle swarm optimization with an adaptive learning strategy (PSO-ALS), and the least squares support vector machine (LSSVM). Cutting force and vibration information are adopted as the monitoring signals. Three steps make up the unique feature extraction and fusion method: multi-domain features extraction, principal component analysis (PCA)-based dimension reduction, and SMDAE-based dimension increment. As a novel feature representation learning approach, the SMDAE technique is utilized to fuse the PCA-based fusion features to enrich the effective information by increasing the dimension, thus helping polish up the predictive performance of the proposed model. PSO-ALS is used to obtain the optimal parameters for LSSVM, simplifying the problem and increasing the population diversity. Twelve sets of milling experiments are conducted to demonstrate the reliable performance of the proposed model. The experimental results show that the presented model is superior to models such as PSO-LSSVM in predictive performance, and the SMDAE technique effectively improves the prediction accuracy of the established model. The findings of this paper offer theoretical guidelines for monitoring milling tool wear in real industrial situations. [ABSTRACT FROM AUTHOR]

Rolling bearings typically operate under time-varying conditions, which present challenges for fault diagnosis due to the presence of modulation effect and noise component in the bearing vibration signal. This paper introduces a novel feature extraction method that combines complementary ensemble empirical mode decomposition (CEEMD), Teager-Kaiser energy operator (TKEO), and self-organizing map (SOM) to enhance the identification accuracy of bearing characteristics by removing unwanted components from the vibration signal. Firstly, CEEMD and TKEO are used to generate demodulated instantaneous energies of the raw vibration signal and each decomposed mode, which represent bearing features. Next, relevant features are extracted by computing similarity distances using the SOM network. The feature signal is then reconstructed, and its power spectrum is used to evaluate the visualization of the characteristic frequencies of rolling bearings. The efficiency of the proposed method is confirmed through the analysis of simulated and experimental vibration signals collected from two types of bearings with different geometries. The results demonstrate that the proposed method can effectively suppress noise, extract characteristics, and identify the fault condition of rolling bearings compared to previous studies. [ABSTRACT FROM AUTHOR]

The robotic assembly of screws is the basic task for the automation assembly of complex equipment. However, a complete robotic assembly framework is difficult to be designed due to the integration of multiple technologies to achieve efficient and stable operations. In this paper, a robotic assembly workflow is proposed, which mainly consists of a feature extraction stage, a grasping stage, and an installation stage. In the feature extraction stage, a feature extraction algorithm consisting of a semantic segmentation network and an object classification module is designed. The semantic segmentation network segments the areas of multiple categories' objects, and the object classification module selects an appropriate target object. The grasping stage and installation stage involve the position alignment of the objects. A position alignment method is developed based on image-based visual servoing using the point features extracted from the segmented areas. The experiments are conducted on a real robot. The alignment errors in grasping stage are less 0.53 mm. The assemblies for a M6-sized screw in ten experiments are successful. The experiment results verify the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

Machining is a versatile field in the manufacturing industry. In milling operations, tool wear is considered the most critical factor affecting the surface quality of the milled piece. Furthermore, the gradual tool wear impacts the milling process, leading to significant downtime, which has serious financial consequences. Unavoidably, a sustainable and reliable condition monitoring system must be developed to reduce the risk of downtime and enhance production quality. The deployment of prognostic and health management (PHM) solutions is becoming increasingly important. It is regarded as one of the main levers for monitoring tool wear status. In this paper, a novel methodology is proposed for extracting pertinent health indicators (HIs) that reflect the degradation behavior of a set of milling cutters and estimating their remaining useful lives (RULs). First, a new time-frequency signal-analysis approach, titled empirical wavelet packet decomposition (EWPD), is proposed to scrutinize the data collected via multi-sensor acquisition. This technique provides a new segmentation of the signal's Fourier spectrum, distributed on levels, to investigate a broader variety of frequency bands and enhance the traditional segmentation structure's performance. Second, a new health indicator is designed based on an innovative selection of the time-domain features computed for each frequency band over each level. Finally, the long short-term memory (LSTM) network is used to estimate the RUL of each cutter. A comparison between the suggested processing method and the wavelet packet transform (WPT) is made to support the hypothesis regarding the effectiveness of the proposed technique. Experimental outcomes seem to be satisfying. [ABSTRACT FROM AUTHOR]

Penetration recognition is the critical technology to improve manufacturing quality and automation level in arc welding. In this paper, recent research advances concerning penetration recognition based on machine learning are comprehensively reviewed, focused on signals feature extraction and application of machine learning in this field. Three types of signals originating from arc welding, namely, weld pool image, arc sound, and arc voltage, are used to analyze the correlation with the penetration state and extract effective features. Next, the following contents briefly introduce some machine learning methods including conventional machine learning and deep learning, and emphatically summarize their application and performance in penetration recognition. Notably, the deep learning method possesses higher classification accuracy and better generalization ability in penetration recognition due to its deeper architecture and effective learning capacity. In the end, some challenges and tendencies, involving multi-sensor information fusion, data-driven recognition, model development incorporating attention mechanism, lightweight model deployment, and real-time control, are presented for further research. This paper is an attempt to provide a reference source and some guidance for researchers who use machine learning methods for penetration monitoring. [ABSTRACT FROM AUTHOR]

Image segmentation can be used to monitor some unconventional operations to ensure the quality of the assembled product in the product assembly process. The current supervised image segmentation method requires much time for labeling. And, the assembly segmentation feature is difficult to extract from color images, due to the influence of color, illumination, and texture factors. To solve these problems, this paper proposes a domain adaptive segmentation method for mechanical assembly with normal images as input. The method involves an adversarial learning stage and a self-training stage. The adversarial learning stage includes a segmentation network and a discriminator network. In the segmentation network, transformer and PPM layers are added to improve the segmentation accuracy of small parts. Meanwhile, the discriminator network is used to align the output features for the source and target data. In the self-training stage, the segmentation accuracy of the part edges is further improved by fine-tuning the parameter model. In addition, to ensure the effectiveness of assembly monitoring, this study establishes a domain adaptive semantic segmentation dataset based on normal image, and evaluate the proposed method. Compared with other domain adaptation methods, our two-stage training method has clearer edge processing performance, and the mIoU index reaches 83.67%, about 3.5% higher than the existing advanced method. In general, our network structure is more suitable for mechanical assembly segmentation and monitoring. [ABSTRACT FROM AUTHOR]

In this study, aiming at problems that the existing surface roughness prediction models fail to consider, the time-varying characteristics of the rough grinding process, and the difficulty in feature selection, a real-time surface roughness prediction model based on multi-sensor signal fusion is developed. Firstly, features are extracted from the time domain, frequency domain, and time–frequency domain of force and vibration signal, and then features that do not reflect the time-varying characteristics of the machining process are eliminated. Finally, a multi-sensor signal fusion method is proposed based on the principal component analysis (PCA). Results show that fused features are capable of retaining the physical meaning of original features, and achieving stable and high-precision prediction of surface roughness when they are input into the BP neural network (BPNN). Specifically, the mean absolute percentage error (MAPE) and maximum percentage error (MPE) of a single sample achieved using the fused signal from the first stage are 3.25% and 8.77%, respectively, while those from the second stage are 2.77% and 6.80%, respectively. These values are lower than those obtained from a single force signal (11.2% and 28.4%, 9.87% and 36.64%) or a single vibration signal (13.68% and 36.5%, 8.13% and 28.76%). Additionally, it is found that the surface roughness at different stages of grinding processing is dominated by different factors. For example, during the first stage, the fourth principal component PC4 is the redundant information for the model. However, this feature significantly affects the performance of the model during the second stage. Overall, this paper lays the foundation for understanding mechanisms of time-varying surface roughness in an actual grinding process and realizing accurate monitoring. [ABSTRACT FROM AUTHOR]

The superior performance of machining large parts with the free-form surface is demonstrated by a five-axis gantry machine tool composed of a gantry mainframe and a milling head. A turntable is used to adjust the position of the workpiece rather than a linear guideway when machining impellers or integral turbines. Furthermore, collisions between adjacent blades are almost unavoidable in the machining region, which is relatively narrow for milling heads. The majority of existing methods attempt to solve the problem by changing the tool's tilt angle. However, when an end-mill cutter is employed, this method will deteriorate surface quality and reduce efficiency. On the condition that the milling head has three rotary axes (TRA), a novel NC code generation method with a fixed tool orientation vector for narrow space surfaces is proposed in this paper. When the geometrical model, considered a supplementary condition, is combined with the kinematics model, an inverse kinematics solution to a general multi-axis gantry machine tool with a TRA milling head is obtained. The problem of multi-axis coupling in the inverse kinematics model is solved by combining the iteration method and geometrical model. In the narrow space for a milling head, the redundant degree of freedom, provided by a machine tool equipped with a TRA milling head, enables adjusting the posture of the milling head without changing the tool orientation vector. Feature points are extracted by preprocessing points selected from machined parts and the TRA milling head to choose the feasible posture and simulate the machining process. Collisions in the narrow space are successfully avoided throughout the cutting process. The NC code generation method is verified on CAM software and cut experiments, proven effective in this paper. [ABSTRACT FROM AUTHOR]

A severe tool wear is often encountered during the process of turning/milling difficult-to-cut materials like Inconel 718. To protect the cutting process from the tool failure, the countermeasure is currently limited to frequent tool changes. Such preventive solution increases not only the machine downtime but also the manufacturing cost. Therefore, there is a strong demand for real-time tool life detection in production lines. In this study, we proposed a stacked LSTM encoder-decoder model for tool wear anomaly detection. Our model only requires "normal" datasets for model training and yields the anomaly score based on the deviation between input and output sequences. This aspect is especially valuable for the actual production lines where the operating conditions are optimized and anomalies are rare. The proposed model is also an end-to-end learning architecture and requires no pre-processing for feature extraction and selection. This characteristic enables real-time processing, which is significant for anomaly detection. In this paper, we first validated our model using artificially generated sinusoidal waveforms and demonstrated its high performance in detecting deviations in amplitude, frequency, waveform bias, and added noise components. The proposed model was then implemented into an actual turning process for Inconel 718 to detect the tool wear in real-time. After being trained by AE (Acoustic Emission) or audio signals captured in the "normal" cutting stage, the model can derive appropriate anomaly scores well matched with the cutting stages of "normal," "transitional" and "anomalous" corresponding to the tool wear. Finally, deep insight into law/rule-based, data-based, and model-based approaches are discussed. [ABSTRACT FROM AUTHOR]

Surface roughness is an important indicator for shaft grinding, and its prediction is always a core issue in practice. The surface roughness of each part for a large shaft would be different with same processing parameters, which leads to traditional processing parameter-based prediction being ineffective; thus, the surface roughness prediction of large shaft grinding is more difficult. This paper proposes a surface roughness prediction method for large shaft grinding based on deep learning, and three kinds of process signals are selected as the inputs, including spindle current, vibration, and acoustic emission. An experiment considering the contribution degree of each processing parameter is designed to generate the dataset on a large shaft grinding platform, and the appropriate surface roughness intervals are divided. The multiple process signals are fused through an attentional CNN-LSTM architecture, in which the CNN is used to extract the features of the process signals, the LSTM deals with the sequential output of CNN, and the self-attention mechanism is utilized to realize the automatic weight allocation. Moreover, the model degradation problem after long-term use of the grinding platform is solved by the parameters fine-tuning approach. The experimental results show that the proposed method achieves the best accuracy than all comparative items, and the model performance can be recovered through parameters fine-tuning; thus, the proposed method is effective. [ABSTRACT FROM AUTHOR]

Machining of Inconel 718 causes rapid tool failure, which affects the tooling cost and dimension tolerance of the components. Literature attributed flank wear as the dominant failure criterion that determines tool life during the milling operation. Flank wear width (VB) can be measured using digital microscopes or predicted in-process by a machine vision-based tool condition monitoring (MV-TCM) system. In MV-TCM, geometric and textural features are extracted from the flank wear region to represent VB progression. However, the leading cutting edge, where flank wear is measured, experiences progressive chipping and built-up edge when machining Inconel 718. These failure modes distract pixel distribution and luminosity, creating complex flank wear features on the leading cutting edge. Nevertheless, the wear region extracted from the side cutting edge shows a consistent change in features that can be used to predict flank wear progression under such failure modes. In addition, the scale-invariant fractal dimension can complement the geometric parameters, improving the reliability of features used to predict flank wear. This paper presents a multi-layer perceptron neural network (MLPNN) that was trained using a synergy of geometric and fractal features extracted from the side cutting edge of square inserts to predict flank wear progression during face milling of Inconel 718. The MLPNN shows an accuracy of 95.5% and a mean absolute percentage error of 1.099% during MV-TCM. The paper shows the potential of applying an in-process MV-TCM expedited by spatial binary features to estimate flank wear progression when milling Inconel 718. [ABSTRACT FROM AUTHOR]

Among the diverse challenges in machining processes, chatter has a significant detrimental effect on surface quality and tool life, and it is a major limitation factor in achieving higher material removal rate. Early detection of chatter occurrence is considered a key element in the milling process automation. Online detection of chatter onset has been continually investigated over several decades, along with the development of new signal processing and machining condition classification approaches. This paper presents a review of the literature on chatter detection in milling, providing a comprehensive analysis of the reported methods for sensing and testing parameter design, signal processing and various features proposed as chatter indicators. It discusses data-driven approaches, including the use of different techniques in the time–frequency domain, feature extraction, and machining condition classification. The review outlines the potential of using multiple sensors and information fusion with machine learning. To conclude, research trends, challenges and future perspectives are presented, with the recommendation to study the tool wear effects, and chatter detection at dissimilar milling conditions, while utilization of considerable large datasets—Big Data—under the Industry 4.0 framework and the development of machining Digital Twin capable of real-time chatter detection are considered as key enabling technologies for intelligent manufacturing. [ABSTRACT FROM AUTHOR]

During the gear hobbing machine process, the hob performance degradation assessment is significant for optimizing the tool changing frequency and improving machining efficiency. It is challenging to recognize hob wear state through vibration analysis and signal processing. Especially in the condition of intense vibration and noise interference, extracting signal features that reflect the wear state is challenging work. This paper proposes a feature extraction method called Cyclic Statistical Energy (CSE) to obtain the hob wear characteristics by tracking the sensitive frequency band. For the method, the vibration signal model of hob spindle is established first based on the vibration-generation mechanism of the machining process. Then, an index E is proposed to track the wear resonance frequency band of the vibration signal. Furthermore, the cyclic statistical order is analyzed and discussed. The hob performance degradation assessment can be realized by analyzing the energy index E with time variation. The experimental setup has been designed, and the two tests have been conducted in the production line: (1) impact hammer test and (2) hob whole life cutting test. The impact hammer test is set to obtain the modal parameters of the hob. Based on this, the whole life cutting test is designed and realized by which the vibration data of whole life with different wear states can be acquired. The proposed method CSE is successfully applied into online industry experiments test. The results show that the proposed method can give a more accurate performance degradation assessment curve for tool condition monitoring compared with traditional methods such as Wavelet Packet Decomposition (WPD). [ABSTRACT FROM AUTHOR]

Tool wear in CNC milling is a gradual process which significantly affects product quality. Left unmonitored, it could increase risks of tool breakage, leading to losses due to scrap and equipment damage. A modular neural network (MNN), the dissociation artificial neural network (Dis-ANN), was proposed in this paper for tool wear prediction. The Dis-ANN consists of a modular structure constructed out of parallel ANN modules (referred to as the dissociation unit), connected to an intermediary. The output of each ANN module is dependent on input feature vectors formed from the concatenation of both previous and current feature values, allowing each module to account for feature trends. Dis-ANN was validated using the Slot Milling Dataset (collected in the University of Malaya workshop) and the 2010 PHM Data Challenge Dataset. The Slot Milling Dataset contains data in the form of images of machined workpiece surfaces and acoustic signals during milling. In order to account for uneven lighting in each workpiece surface image, image features were extracted by processing grey-level co-occurrence matrix–based texture descriptors of different non-overlapping sections within the same image. For model validation using the 2010 PHM Data Challenge Dataset, Dis-ANN was validated using features extracted from the dataset under two conditions — with and without the addition of random feature noise. Results showed Dis-ANN was better at learning complex non-linear relationships while showing higher robustness to input feature noise compared to linear regression, support vector regression, and monolithic ANN. Furthermore, the modular design of Dis-ANN facilitated model architecture optimization to minimize network redundancy. [ABSTRACT FROM AUTHOR]

As fault identification algorithms for rotating machinery based on deep learning are developing rapidly, convolutional neural networks (CNNs) have been attracting extensive attention due to their feature extraction capabilities. However, most of the current CNN–based fault identification models can only evaluate one aspect of the fault, and the recognition accuracy is low in the case of a complex fault. To achieve various and complex fault diagnoses, this paper proposes a multi-tasking atrous convolution neural network (MACNN). First, the network introduces sub-modules such as atrous convolutional layer, batch normalization processing, and PReLU activation function to improve the efficiency of down-sampling and better realize fault feature extraction. Secondly, a parallel multi-independent output layer and a special loss function corresponding to the multitasking structure are proposed, which make the network better in solving the problem of multi-dimensional fault assessment. Finally, based on the MACNN, experiments on rotating machinery fault diagnosis are carried out. In the experiment, three kinds of bearing fault positions and four kinds of depths are identified together, and the accuracy can reach more than 99%, which has obvious advantages over other neural network methods such as artificial neural network (ANN), traditional three-layer convolutional neural network (3L-CNN), and ACNN without multi-task learning. [ABSTRACT FROM AUTHOR]

In many industrial applications, it is possible to approximate the shape of mechanical parts with geometric primitives such as spheres, boxes, and cylinders. This information can be used to plan robotic grasping and manipulation procedures. The work presented in this paper investigated the use of the state-of-the-art PointNet deep neural network for primitive shape recognition in 3D scans of real-life objects. To obviate the need of collecting a large set of training models, it was decided to train PointNet using examples generated from artificial geometric models. The motivation of the study was the achievement of fully automated disassembly operations in remanufacturing applications. PointNet was chosen due to its suitability to process 3D models, and ability to recognise objects irrespective of their poses. The use of simpler shallow neural network procedures was also evaluated. Twenty-eight point cloud scenes of everyday objects selected from the popular Yale-CMU-Berkeley benchmark model set were used in the experiments. Experimental evidence showed that PointNet is able to generalise the knowledge gained on artificial shapes, to recognise shapes in ordinary objects with reasonable accuracy. However, the experiments showed some limitations in this ability of generalisation, in terms of average accuracy (78% circa) and consistency of the learning procedure. Using a feature extraction procedure, a multi-layer-perceptron architecture was able to achieve nearly 83% classification accuracy. A practical solution was proposed to improve PointNet generalisation capabilities: by training the neural network using an error-corrupted scene, its accuracy could be raised to nearly 86%, and the consistency of the learning results was visibly improved. [ABSTRACT FROM AUTHOR]

The particle size of fixed abrasive pad (FAP) is an important factor affecting its machining performance and surface condition. The surface condition of consolidated FAP is closely related to the tribological characteristics. However, the relationship between tribological characteristics of FAP and surface condition has not been studied in depth. Therefore, the mapping relationship between friction characteristic signal and FAP surface condition is established in this paper, and the influence of FAP abrasive particle size on the friction signal characteristic in the lapping process of quartz glass is explored. Four FAPs (W3.5, W7, W10, and W14) were used for the lapping experiments and tribological experiments on quartz glass, and the friction signals in each processing stage were processed and feature signals were extracted. Moreover, the material removal rate, debris particle size distribution and micro morphology, friction coefficient, surface micro morphology of workpiece, and FAP at each processing stage were analyzed. The experimental results showed that there is a close relationship between energy of wavelet packet frequency band 8 and marginal spectrum amplitude and cutting performance of FAP surface abrasive particles. Their values increase with the increase of the abrasive particle size. There is a close relationship between energy of wavelet packet frequency band 7 and cutting performance of FAP surface matrix. Their value decreases with the increase of abrasive particle size. This study provides a theoretical basis for FAP online monitoring. [ABSTRACT FROM AUTHOR]

A reliable material removal rate (MRR) prediction method significantly optimizes the grinding surface quality and improves the processing efficiency for robotic abrasive belt grinding. Using worn-belt image features to predict MRR is a direct and reliable method; however, this method is rarely reported at present. This paper proposes an MRR prediction method for Inconel 718 grinding based on the abrasive belt image analysis and categorical boosting (CatBoost) algorithm. During belt grinding, four wear types of abrasive belts, namely fracture, adhesion, rubbing wear, and fall-off, are identified and analyzed. Under various grinding parameters, the experimental MRR rapidly decreases at first, then in a gradual manner. For an effective evaluation of belt wear severity, cutting grain area ratio, color moments, and texture features are extracted from belt images. MRR and abrasive belt image features are strongly correlated after normalization. All image features are taken into account for MRR prediction model training. Verification experiments indicate that the predicted data is in good agreement with the experimental data. The maximum absolute error, mean absolute error, root mean square error, and determination coefficient of the MRR prediction model are 0.17 μm, 0.4 μm, 0.2 μm, and 99.42%, respectively, which are superior to those of other popular machine learning algorithms. In this study, we present a comprehensive understanding of the relationship between MRR and abrasive belt characteristics, as well as demonstrate the feasibility of accurately predicting MRR using the CatBoost algorithm. [ABSTRACT FROM AUTHOR]

In the deformation machining (DM) approach, the machining and incremental forming operations are combined to utilize the strengths of the two processes to produce monolithic components with thin structures using a single setup. DM offers a high degree of flexibility to manufacture monolithic components in comparison to conventional techniques. Some freeform monolithic products are difficult to machine with the 3-axis machining approach due to restricted entry of the tool in certain portions of the geometry. This paper presents a novel double-sided deformation machining (DSDM) technique to manufacture freeform double-sided monolithic components. As a test case, a double-sided monolithic impeller-shaped structure with freeform blades is manufactured with the proposed technique. For this, first, a blank is machined to create the required structure and blades of desired thickness. These machined structures are used as preforms for the incremental forming (bending) of the blades to the final shape. As the existing commercial toolpath development packages are not compatible with the DM process, a novel feature-based double-sided combined machining-forming toolpath is developed for the DSDM operation. Here, the blades are deformed separately and sequentially. This avoids the collision issue and achieves a faster forming operation. To evaluate the quality of the monolithic components developed, a few response parameters, like the process forces, the geometrical accuracy in terms of springback, and the surface roughness, are reported. [ABSTRACT FROM AUTHOR]

During machining processes, accurate prediction of cutting tool wear is prominent to prevent ineffective tool utilisation and significant resource waste. Tool wear conditions and progression involve complex physical mechanisms, and a promising approach is to deploy heterogeneous sensors and design a deep learning algorithm to conduct real-time tool wear monitoring and precious prediction. To tackle the challenge of deep learning algorithms in processing complex signals from heterogeneous sensors, in this paper, a systematic methodology is designed to combine signal de-noising, feature extraction, feature optimisation and deep learning-based prediction. In more details, the methodology is comprised of the following three steps: (i) signal de-noising is carried out by a designed Hampel filter-based method to eradicate random spikes and outliers in the signals for raw data quality enhancement; (ii) features extracted from heterogeneous sensors in the time and frequency domains are optimised using designed recursive feature elimination and cross-validation (RFECV)-based and Isomap-based methods; (iii) a convolutional neural networks (CNN) algorithm is devised to process the optimised features to implement tool wear prediction. In this paper, a case study showed that 80% features were reduced from the originally extracted features and 86% prediction accuracy was achieved based on the developed methodology. The presented methodology was benchmarked with several main-stream methodologies, and the superior performance of the methodology over those comparative methodologies in terms of prediction accuracy was exhibited. [ABSTRACT FROM AUTHOR]

Technical failures in machines are major sources of unplanned downtime in any production and result in reduced efficiency and system reliability. Despite the well-established potential of Machine Learning techniques in condition-based maintenance (CBM), the lack of access to failure data in production machines has limited the development of a holistic approach to address machine-level CBM. This paper presents a practical approach for failure mode prediction using multiple sensors installed in a bearing ring grinder for process control as well as condition monitoring. Bearing rings are produced in a set of 7 experimental runs, including 5 frequently occurring production failures in the critical subsystems. An advanced data acquisition setup, implemented for CBM in the grinder, is used to capture information about each individual grinding cycle. The dataset is pre-processed and segmented into grinding cycle stages before time and frequency domain feature extraction. A sensor ranking algorithm is proposed to optimize feature selection for failure classification and the installation cost. Random forest models, benchmarked as best performing classifiers, are trained in a two-step classification framework. The presence of failure mode is predicted in the first step and the failure mode type is identified in the second step using the same feature set. Defining the feature set in the failure detection step improves the predictor generalization with the classifiers' performance accuracy of 99 % on the test dataset. The presented approach demonstrates an efficient failure mode classification by selecting crucial sensors resulting in a cost-effective CBM implementation in a bearing ring grinder. [ABSTRACT FROM AUTHOR]

Skin parts are widely used in the modern aerospace industry. The remaining wall thickness of pockets on a skin part must be precisely controlled to reduce overall weight and ensure strength. The in-process dimensional accuracy forecast of remaining wall thickness is urgently required for process efficiency improvement and compensation. In this paper, the spindle power signal of pocket milling is used for the in-process dimensional accuracy forecast. Based on deep learning mothed, the correlation between dimensional accuracy and features extracted from power signal is established. To improve the applicability of the forecast model, a deep transfer learning model based on pockets clustering and feature reconstruction is proposed. The pockets of a skin part are divided into different types through clustering. Then, the trained deep learning network for feature extraction of one type of pocket is transferred to a new type by feature reconstruction. The results indicate that the deep transfer learning model can achieve in-process dimensional accuracy forecast for remaining wall thickness with high precision. [ABSTRACT FROM AUTHOR]

Tool wear prediction is a challenging aspect of the milling machine as the cutting tool is responsible for the accuracy and precision of the final machined product. The accuracy of tool wear prediction depends on how well-established data are provided to the prediction model. The appropriate and optimal features need to be selected from the features to provide appropriate data to the model. The implementation of proper feature selection methods substantially reduces the complexity of the raw data for analysis and transfers vital information to the prediction model. In this paper, the IEEE NUAA Ideahouse tool wear dataset was used for the study of tool wear prediction. Different feature selection methods, namely the Pearson correlation coefficient (PCC) and random forest (RF), along with hybrid methods such as PCC with RF and principal component analysis (PCA) with PCC, were used for the feature selection. The selected features obtained from the feature selection techniques were then applied to the different variants of the long short-term memory (LSTM) model for tool wear prediction. The best feature selection method was chosen based on the performance metrics such as the root mean square error (RMSE), R-squared (R2), and mean absolute percentage error (MAPE). The hybrid feature selection method of RF, PCA with PCC used together with the encode-decoder LSTM model was found to have the best accuracy with R2, RMSE, and MAPE values of 0.97, 0.036, and 5.79, respectively, for tool wear prediction compared to the other feature selection methods and LSTM variants. [ABSTRACT FROM AUTHOR]

Chatter is a self-excited vibration that reduces the quality of the workpiece surface and accelerates tool wear, it is a phenomenon that is difficult to monitor and control. In order to achieve high-performance manufacturing, a reliable online monitoring system for chatter needs to be developed. This paper proposes a chatter identification method based on optimized variational modal decomposition (OVMD) and fuzzy entropy to achieve online monitoring of chatter accurately. The method embeds the whale optimization algorithm (WOA) into the VMD method, and the chattering features are extracted for both simulated and experimental signals to achieve the optimal decomposition of the force signal in a rapid way. In order to solve the problem that sample entropy and approximate entropy are not sensitive to chatter detection, fuzzy entropy is introduced to monitor the chatter feature components. Experimental results show that the fuzzy entropy is particularly sensitive to the occurrence of chatter in specific embedding dimensions, and this method can monitor the dynamics of milling with higher sensitivity and stability. [ABSTRACT FROM AUTHOR]

Recently, with the advance in information technology, pure data-driven approaches such as machine learnings have been widely applied in status diagnosis. However, the accuracy of those predictions strongly relies on the original data, which largely depends on the selected sensors and signal features. Furthermore, for unsupervised machine learning schemes, although it could avoid the concern of labeling in training, it lacks a quantified evaluation of the prediction results. These concerns significantly limit the effectiveness of modern machine learning and thus should be investigated. Meanwhile, ball bearings are fundamental key machine elements in rotating machinery and their condition monitoring should be critical for both quality control and longevity assessment. In this paper, by utilizing ball bearing failure diagnosis as the main theme, the flow of feature selection and evaluation, as well as the evaluation flow for multiple failure diagnosis, is developed for accessing the status of bearings in their imbalance, lubrication, and grease contamination levels based on unsupervised machine learning. The experimental results indicated that with proper feature selection, the failure identification could be more definite. Finally, a novel model based on the second norm to quantify the classification level of each cluster in hyperspace is proposed as the measure for unsupervised machine learning as the basis for performance evaluation and optimization of unsupervised machine learning schemes and should benefit related machine reliability evaluation studies and applications. [ABSTRACT FROM AUTHOR]

The bearing fault diagnosis plays an important role to reduce catastrophic failures and ensure the continuity of running machines to avoid heavy economic loss. The vibration signals of rolling bearings are often nonlinear and nonstationary; it is difficult to extract sensitive features and diagnose faults by traditional signal processing methods. To solve this problem, a novel intelligent fault-diagnosis approach based on whale optimization algorithm grey wolf optimization-variational mode decomposition (WOAGWO-VMD) algorithm and the marine predators algorithm optimization-least squares support vector machine (MPA-LSSVM) is proposed in this paper. Firstly, hybrid algorithm WOAGWO is used to optimize the parameters of VMD and obtain the optimal combination (K; α). Then, the optimized VMD algorithm is utilized to decompose the vibration signal of the rolling bearing into several intrinsic mode functions, and a new sensitive indicator is created to select the components containing the most information. For these components, the dispersion entropy feature, permutation entropy feature, and singular value feature are extracted to form the multi-feature vector. Finally, the feature vectors obtained are input to the MPA-LSSVM for diagnosis and identification. The validity and strength of the proposed method is verified by experimental data under different bearing conditions. The results have shown that the proposed method can extract the fault feature information of 16 bearing signals of different fault types effectively and identify them accurately. [ABSTRACT FROM AUTHOR]

In milling processing, the wear state of the tool has an essential influence on the processing quality. The machining process is not continuous in the cycloid milling process, and the signal of the empty tool part increases the difficulty of identifying the tool wear state. At present, most of the researchers use experimental data and cut out the signal of the empty tool part in the signal by data processing. However, it will affect the original signal to a certain extent and destroy the confidential information in the original signal. A novel method using variational auto-encoder (VAE) for tool wear status identification is proposed. Due to VAE has structural characteristics that reduce the dimensionality of high-dimensional data to lower dimensionality. This requires VAE to find and learn the significant features, which are hidden in the complex raw data. In this paper, the signals of the empty tool do not need to be cut out; the effective value of the three-phase current signals obtained in the real processing is converted into the form of three-dimensional color images. VAE is applied to extract features from the image samples and then realize the classification of different wear states of the tool. A large number of comparison experiments are conducted, and the result shows that the presented method has a better recognition performance for the actual processing data. It is more suitable for the recognition of tool wear status in the actual milling process. [ABSTRACT FROM AUTHOR]

As a complex thermo-physical process, the plasma arc welding (PAW) is easy to be unstable due to external interferences. Weld quality monitoring is important for intelligent robot PAW welding. Due to the different instability mechanisms, it is difficult to obtain high adaptivity and accuracy with features extracted in a single time window. In this paper, a novel feature extraction method based on sliding multiscale windows is proposed to improve model accuracy and calculation speed. A group of windows with different time widths are established to extract multiscale information. Windows slide throughout welding process and are synchronized on the timeline for feature correlation. The welding current and arc voltage are processed to extract features inside windows, including signal denoising by discrete wavelet transform (DWT) and dimension reduction by primary components analysis (PCA). Based on the feature vectors extracted from multiscale-windows, support vector machine (SVM) with radial basis function (RBF) kernel is used. The best window width is determined automatically by model training. The proposed method is used to predict weld quality for PAW in the field of shipbuilding. The results show that the model with multiscale feature extraction is helpful to improve prediction precision and recall ratio. [ABSTRACT FROM AUTHOR]

Effective cutting tool life prediction is significant for ensuring processing quality and improving production efficiency. Data-driven prediction methods have been widely used. However, traditional methods assume that there are high-quality sensor data, whereas, in practice, factors such as poor installation of sensors and environmental interference often lead to poor-quality data, leading to unreliable analysis results and incorrect decisions. Thus, in this paper, a sensor data anomaly detection and correction method is proposed. It mainly includes four parts: data preprocessing, abnormal data detection, correction of detected abnormal data, and tool life prediction and evaluation. First, the raw condition monitoring data are preprocessed for feature extraction and health index (HI) construction. Second, the HIs of historical training samples are clustered based on the dynamic time warping (DTW) algorithm, and the abnormal data are detected based on error calculation with a preset error threshold. Third, the detected abnormal data are optimized via similarity matching using k-nearest neighbors with dynamic time warping (KNN-DTW). Finally, the optimized data are used for tool life prediction and evaluation. The proposed method has been tested on real data acquired from a turbine factory. The comparison results show that the prediction effect can be significantly improved after adopting the proposed method, which verifies the necessity of sensor data anomaly detection and correction. [ABSTRACT FROM AUTHOR]

A safe and reliable avionics system is an important guarantee for airplanes when accomplishing all kinds of tasks. With the ever more comprehensive and intelligentized development of avionics system, the layout of cable brackets becomes more complicated, which results in the quick and reliable assembly, and inspection of cable brackets is becoming more and more demanding. Nowadays, the detection and recognition of cable brackets in aircraft assembly scenes still rely on manual labor to compare the real assembly scene with the designed two-dimensional drawings or graphs. With the increase in number and types of cable brackets, the recognition process in multi-assembly scenes becomes intricate, time-wasting, and highly mistakable. Aiming at the cable bracket recognition problems, this paper proposes a low retrain complexity hybrid model for adaptive recognition of cable brackets. Firstly, this paper builds a weight sharing feature extraction model using improved loss function and multi-scale ensemble convolutional neural network according to the analysis of cable brackets. Secondly, the support vector machine is used to classify the feature vector extracted by pre-trained feature extraction model. By this way, the hybrid model could be retrained quickly for new types of cable brackets never seen in multiple assembly scenes, under condition that bracket number and types are variable. At last, the performances of other feature extraction methods such as SIFT, HOG, and multi-layer perceptron (MLP) are evaluated for contrast with the proposed method. As shown in the results, our method obtains higher recognition accuracy. [ABSTRACT FROM AUTHOR]

Tool wear and faults will affect the quality of machined workpiece and damage the continuity of manufacturing. The accurate prediction of remaining useful life (RUL) is significant to guarantee the processing quality and improve the productivity of automatic system. At present, the most commonly used methods for tool RUL prediction are trained by history fault data. However, when researching on new types of tools or processing high value parts, fault datasets are difficult to acquire, which leads to RUL prediction a challenge under limited fault data. To overcome the shortcomings of above prediction methods, a deep transfer reinforcement learning (DTRL) network based on long short-term memory (LSTM) network is presented in this paper. Local features are extracted from consecutive sensor data to track the tool states, and the trained network size can be dynamically adjusted by controlling time sequence length. Then in DTRL network, LSTM network is employed to construct the value function approximation for smoothly processing temporal information and mining long-term dependencies. On this basis, a novel strategy of Q-function update and transfer is presented to transfer the deep reinforcement learning (DRL) network trained by historical fault data to a new tool for RUL prediction. Finally, tool wear experiments are performed to validate effectiveness of the DTRL model. The prediction results demonstrate that the proposed method has high accuracy and generalization for similar tools and cutting conditions. [ABSTRACT FROM AUTHOR]

The high dynamic range existing in arc welding with high energy density challenges most of the industrial cameras, causing badly exposed pixels in the captured images and bringing difficulty to the feature detection from internal weld pool. This paper proposes a novel monitoring method called adaptive image fusion, which increases the amount of information contained in the welding images and can be realized on the common industrial camera with low cost. It combines original images captured rapidly by the camera into one fused image, and the setting of these images is based on the real-time analysis of realistic scene irradiance during the welding process. Experiments are carried out to find out the operating window for the adaptive image fusion method, providing the rules for getting a fused image with as much as information as possible. The comparison between the imaging with or without the proposed method proves that the fused image has a wider dynamic range and includes more useful features from the weld pool. The improvement is also verified by extracting both the internal and external features of weld pool within a same fused image with proposed method. The results show that the proposed method can adaptively expand the dynamic range of visual monitoring system with low cost, which benefits the feature extraction from the internal weld pool. [ABSTRACT FROM AUTHOR]

With the development of manufacturing industry and technology, traditional manual welding technology is gradually unable to meet the need for industrial mass production in the field of fusion welding and additive manufacturing. As a result, an automatic welding method using robots to replace manual welding is needed. This paper studies the additive weld seam tracking technology based on laser vision and designs a welding seam tracking system. The images of linear structure light which reflect welding seam information are collected by vision sensor. The structure light extraction algorithm window is selected under the guidance of the over-exposure characteristics of all kinds of noises. The ERFNet network is applied for the structure light and its corresponding feature point extraction. The accurate center line of structure light is extracted from strong background noise and the feature point of weld seam is obtained through regression. It realizes the online path planning and deviation correction of the weld seam tracking in real-time. The proposed algorithm is demonstrated by the weld feature extraction experiment and welding seam tracking experiment based on groove additive task. It shows that the offset is within one pixel and the distance is within 0.25 mm between the welding feature points extracted by ERFNet and the manually marked welding points. The proposed algorithm has the performance of high robustness, strong adaptability and can meet the practical welding requirements. [ABSTRACT FROM AUTHOR]

Sensor signals acquired in industrial equipment contain rich information which can be analyzed to facilitate effective monitoring of equipment, early detection of system anomalies, quick diagnosis of fault root causes, and intelligent system design and control. In many mechatronic systems, multiple signals are acquired by different sensor channels (i.e., multi-channel data) which can be represented by high-order arrays (tensorial data). The multi-channel data has a high-dimensional and complex cross-correlation structure. It is crucial to develop a method that considers the interrelationships between different sensor channels. This paper proposes a new equipment monitoring approach based on uncorrelated multilinear discriminant analysis that can effectively model the multi-channel data to achieve a superior monitoring and fault diagnosis performance compared to other competing methods. The proposed method is applied directly to the high-dimensional tensorial data. Features are extracted and combined with multivariate control charts to achieve real-time fault detection of equipment. The effectiveness of the proposed method in quick detection of equipment faults is demonstrated with both the simulation and a real-world case study. [ABSTRACT FROM AUTHOR]

Tool wear monitoring and prediction are of great importance for machining precision and surface integrity. In order to ensure good quality of the machining components, tool wear should be monitored and predicted in time. However, the traditional methods greatly depend on feature selection and extraction, and accuracy and generalization are limited. In this paper, a novel model based on deep learning was proposed to monitor and predict tap tool wear. Firstly, 1D CNN was designed to extract features from the vibration data, and then the residual block with the dilated convolution was specially developed to accept the features from 1D CNN. After that, the fully connected neural networks were designed to predicate tool wear. Moreover, to verify the superiority and generalization of the proposed method, the taping experiment from a real engine cylinder head production line was conducted. During the tapping process, the vibration signal and tool wear were collected. And then the training experiment was conducted with the collected data. The prediction results with the proposed model were compared with those of the state-of-the-art algorithms. The compared results showed that the proposed model was more robust and accurate for tool wear prediction. [ABSTRACT FROM AUTHOR]

In this paper, an autonomous seam recognition and feature extraction method for multi-pass welding based on laser stripe edge guidance network is proposed to overcome the interference of strong reflection, spatter, and arc noise in actual welding environment. Firstly, the laser stripe edge guidance network consisting of modified VGGnet, progressive laser stripe feature extraction, non-local laser stripe edge feature extraction, one-to-one guidance module, and multi-feature fusion module is introduced to recognize the laser stripe under heavy arc noises. Afterwards, the gray centroid method is adopted to obtain the thinning laser stripe. Aiming at extracting the position of feature points, the least square method and non-uniform rational B-splines with second derivative are utilized. Finally, experiments and analysis show that our proposed method performs favorable in terms of effectiveness, flexible, accuracy, and robustness, which could meet the actual welding requirements. Besides, the maximum error and maximum root mean square error for feature extraction are 4.7 pixel and 1.78 pixel, respectively. [ABSTRACT FROM AUTHOR]

Accurate tool condition monitoring (TCM) is essential for the development of fully automated milling processes. However, while considerable research has been conducted in industrial and academic settings, the complexity of milling processes continues to complicate the implementation of TCM. This paper presents a review of the state-of-the-art methods employed for conducting TCM in milling processes. The review includes three key components: (1) sensors, (2) feature extraction, and (3) monitoring models for the categorization of cutting tool states in the decision-making process. In addition, the primary strengths and weaknesses of current practices are presented for these three components. Finally, this paper concludes with a list of recommendations for future research. [ABSTRACT FROM AUTHOR]

In the modern manufacturing industry, the welding quality is one of the key factors which affect the structural strength and the comprehensive quality of the products. It is an important part to establish the standard of welding quality detection and evaluation in the process of production management. At present, the detection technologies of welding quality are mainly performed based on the 2D image features. However, due to the influence of environmental factors and illumination conditions, the welding quality detection results based on grey images are not robust. In this paper, a novel welding detection system is established based on the 3D reconstruct technology for the arc welding robot. The shape from shading (SFS) algorithm is used to reconstruct the 3D shapes of the welding seam and the curvature information is extracted as the feature vector of the welds. Furthermore, the SVM classification method is adopted to perform the evaluation task of welding quality. The experimental results show that the system can quickly and efficiently fulfill the detection task of welding quality, especially with good robustness for environmental influence cases. Meanwhile, the method proposed in this paper can well solve the weakness issues of conventional welding quality detection technologies. [ABSTRACT FROM AUTHOR]

Chatter is a kind of self-excited vibration, which will result in very poor quality and dimensional accuracy on the machined surface, even a harmful effect on machining operation. A simple, reliable, and accurate chatter identification algorithm is crucial for taking control strategy before it is fully developed. This paper proposes a novel frequency-domain search (FDS) algorithm to identify the chatter during milling. An identification feature based on vibration base frequency is extracted according to the FDS algorithm, where the complicated signal processing algorithms are not needed before feature extraction. Compared with most of the existing identification features, the introduced feature does not need to set thresholds according to different machining conditions. Meanwhile, the feature extraction only needs a small amount of data to guarantee the timeliness of identification. Hammer test and milling experiments with various cutting parameters are carried out, and both force signal and vibration signal in the experiments are utilized to validate the effectiveness of the proposed algorithm. The results show that the proposed algorithm can identify the milling chatter accurately, whether using force signal or vibration signal, even slight chatter in the initial machining stage can be identified. Furthermore, the research reveals that dominant chatter frequencies appear around the natural frequency of the spindle-tool system but do not exactly equal to its natural frequency. The chatter frequencies are found to be determined by the combination of natural characteristics of the system and cutting conditions. [ABSTRACT FROM AUTHOR]

The internal crack defect is prone to appear in the inner part of casting products during the production process due to the influence of casting craft and on-site environment. In order to ensure the internal quality of the casting products, nondestructive examination and defect detection technology should be used to detect the internal crack of the casting product. The existing defect detection technologies have some problems such as poor generalization and low accuracy. Therefore, an internal crack defect detection method based on the Relief algorithm and Adaboost-SVM is proposed in this paper. Firstly, casting image is preprocessed by grayscale transformation, bilateral filtering, and adaptive image segmentation. Secondly, HOG feature, invariant moment feature, and LBP feature are extracted, and sensitive feature set is selected by Relief algorithm. Finally, the Adaboost-SVM is used to construct the internal crack detection model to realize the crack detection with high generalization and accuracy. The effectiveness of the method is verified by the casting image dataset collected in the actual industrial field. The experimental result reveals that the proposed method could not only extract sensitive feature set but also has better classification performance and generalization ability than other common classifiers. [ABSTRACT FROM AUTHOR]

Tool wear monitoring is deemed as an essential technology of the intelligent manufacturing to guarantee the processing quality and improve the machining efficiency. In this paper, a prediction model based on adaptive particle swarm optimization (APSO) algorithm and least squares support vector machine (LS-SVM) algorithm is proposed for the recognition of drill wear. Cutting force signal and vibration signal are used for tool wear monitoring. And these signals are preprocessed through wavelet threshold de-noising algorithm. Multiple signal feature extraction methods are carried out to process the sample data related to drill wear status. The mean absolute error of the tool wear recognition model is 0.91%, better than the standard LS-SVM algorithm under the same condition. [ABSTRACT FROM AUTHOR]

Safety and fault diagnosis of rotating machinery play an important role in industrial systems. The reliability of the diagnosis performances is mainly linked to the signal processing tools used in the analysis phase. In this paper, a new method is proposed for the classification of rolling defects combining the optimized wavelet multi-resolution analysis (OWMRA), principal component analysis (PCA), and neuro-fuzzy. The OWMRA is performed to decompose the measured vibration signals in different frequency bands and extract additional information about the defect. The decomposition levels obtained from optimized WMRA are then used as the input of the PCA method. The extraction of individual feature sets including time domain features is generated to disclose health conditions of the bearing. Hence, the proposed classification method of neuro-fuzzy based on multi-scale principal component analysis (MSPCA) applies to real signals to analyze several types of defects on the bearings' ball, outer race and inner race with variable fault diameter; load; and motor speed. The obtained results prove the reliability of the proposed diagnosis method to classify three types of bearing faults, and to give a better classification with greater efficiency compared to the application of individual classifiers or the artificial neural networks (ANN) alone. Finally, The effectiveness of our approach has been proven in terms of the successful classification rate compared on the one hand with two classification algorithms and two sets of features and on the other hand with the time domain method based on different choices of features. [ABSTRACT FROM AUTHOR]

Laser triangulation and image processing are two major methods for surface quality inspection. The former has the advantage of three-dimensional information, while the latter can provide more detailed characteristics. However, there are few researches on effective integration of the two methods. In this paper, a lab scale laser braze quality inspection system utilizing information from laser triangulation scanning and optical image acquisition was able to inspect an automotive laser braze at a scanning speed of 1.5 m/min with 100% accuracy. The process started with defect zoning to sort out braze areas with and without defect based on the joint surface profile reconstructed from laser triangulation. Feature extraction was then carried out on the images of defect-containing braze areas, in which the location, size, geometry, and texture features of the defects were extracted from the gray-scale images, and since these features were extracted from only defect-containing areas, a lot of computational cost was saved, which reduced time consumed greatly. For defect classification, in the training stage the extracted features were processed by the principal component analysis and the first three principal components were calculated for 5 typical defects: surface cavity, pore, hump, burn-through, and lack-of-braze. In the test stage, the first three principal components of the images of interest were calculated and compared with the established principal component values of the 5 defect types for defect classification. [ABSTRACT FROM AUTHOR]

Assembly validation is a key part of product design. Current methods, such as physical prototyping are time-consuming and do not offer immediate validation results. Assembly motion simulation systems have been proposed as a solution to this problem. However, widespread adoption of such systems is hindered due to their ties to proprietary computer aided design (CAD) software or expensive and often cumbersome hardware. Recently, virtual/augmented reality (VR/AR) technologies and simulation have been heralded as two of the key enabling factors of Industry 4.0. Collective interests in these technologies by industry and community have brought many low-cost software and hardware tools to the market, which opens a gateway to achieving assembly validation at a much lower cost. This paper presents an assembly validation system that is independent of CAD packages, interoperable and implemented using relatively low-cost and commercially available hardware and software tools. The system features intuitive bare-hand manipulation of part models through a virtual hand model that tracks the hands. Collision detection and physics modelling allow for hand-part and part-part interactions to be natural, thus validating assembly interactions. An assembly feature extraction algorithm has also been implemented to analyse the planar face features of the part models to detect possible mating assembly features between parts concerned. A constraint management system considers identified mating features and determines the allowable motion of parts once constraints are applied and removed. Pulling force is used to facilitate the removal of constraints. [ABSTRACT FROM AUTHOR]

Sensors are being widely used in many industrial practices and contain rich information which can be analyzed to detect system anomalies. The outputs of sensors are time-ordered data known as waveform signals, which are also called profiles. Many monitoring methods only focus on a single profile. However, multiple profiles are recorded by different sensor channels in many processes. It is crucial to study methods to analyze the multi-channel profiles. In this paper, uncorrelated multilinear discriminant analysis is suggested for fault detection and diagnosis. Then, the algorithm combined with tensor-to-tensor projection is proposed to make it get better performance in improving the accuracy of detection and reducing the fluctuation of the results in analyzing multi-channel profiles. The proposed method is applied directly to the multi-channel profiles. Discriminative and uncorrelated features are extracted, which are then fed into classifiers to identify different fault types. The effectiveness of the developed method is demonstrated by using both the simulation and a real-world case study. The real profiles in the case study are from a sensor fusion application in multiple forging operation processes, where multi-channel profiles are monitored to detect the faults of missing parts. [ABSTRACT FROM AUTHOR]

The robustness of the image processing algorithm is very important based on vision sensor in robotic seam tracking, which will directly affect the accuracy of weld seam shaping quality. Especially in GMAW (Gas Metal Arc Welding), there is a lot of strong noise image. This paper studies an algorithm for the several weld seam images with strong noise in robotic GMAW, such as the atypical weld seam, the strong arc light and the large spatter. Based on a purpose-built visual sensing system, the fast image segmentation, the feature area recognition of the convolutional neural network (CNN), and the feature search technique are used to identify the weld seam features accurately in the algorithm. The selection range of the threshold is increased from 0.5 × 107 to 0.9 × 107 by using the proposed algorithm, which reduces the difficulty of parameter adjustment and increases the stability of seam tracking system. And, the accuracy of the CNN model was 98.0% for the atypical weld seam identification. To evaluate the robustness of the proposed algorithm, the accuracy is verified using experiments on two typical strong noise images. The experiments show that the average error of feature extraction accuracy is 0.26 mm and 0.29 mm. The results show that the proposed algorithm can extract the feature of weld seam image with strong noise accurately and effectively. [ABSTRACT FROM AUTHOR]

In this paper, we propose a solder joint defect type classification method for automatic optical inspection machines in the manufacturing system of printed circuit boards. The inspection procedure for the solder joint defect type classification consists of an offline stage, which sets the optimal feature extraction region, and an online stage which classifies a defect type. In the offline stage, we use an optimization technique, namely, the genetic algorithm, to optimize the feature extraction region. In this stage, the optimal feature extraction region for defect type classification is constructed automatically. In the online stage, feature extraction regions are used to segment the solder joint image after component image acquirement. We then extract various color features from the segmented feature extraction regions. Next, we use support vector machine, which is one of the machine learning model's method to classify the solder joint defect type. To evaluate the performance of the proposed method, ten types of solder joint defects were used in an experiment. The experimental results verified the effectiveness of the method in terms of the recognition rate, and its convenience. [ABSTRACT FROM AUTHOR]

The safety of industrial installations requires real-time monitoring of the occurrence of defects in induction machines that are widely used in this field. The implementation of this type of system typically needs to process a large amount of data provided by sensors and thus necessitates high computing mass, which complicates sensor utilization in real time. In this paper, we propose a hierarchical recurrent neural network by stacking two long short-term memory layers to form a single end-to-end network. Trained to establish complex temporal relations in raw time series signals. Those signals are directly provided by the sensors without any preprocessing or hand engineered features extraction. To train the network, we use the stator currents of a three-phase induction motor captured in a steady state. The currents represent several operation modes, which comprise the healthy and failed states with several types of mechanical defects, electrical defects, and combinations thereof. The experimental results were obtained using data from a real test bed to demonstrate the robustness and speed of the proposed approach for real-time monitoring of the operating status of an induction motor. [ABSTRACT FROM AUTHOR]