Search

Remaining useful life (RUL) prediction has been increasingly considered in many industrial fields for the reliability and safety of their systems. As a data analysis tool of deep learning, deep convolutional neural network (CNN) shows great potential for RUL prediction. This paper proposes an intelligent RUL prediction method based on a double-CNN model architecture. Given the powerful feature extraction capability of CNN, the proposed method is fed with original vibration signals with no need to resort to any feature extractor, which can also retain the useful information in maximum. The prediction includes two stages: first, incipient fault point is identified by the first CNN model and a proposed “3/5” principle; then, the second CNN model is constructed for RUL prediction. In practice, RULs of identical components are different from each other, which poses a major challenge in RUL prediction. To overcome this problem, an intermediate reliability variable is first calculated in this paper, instead of directly predicting the RUL value. Then, a mapping algorithm is proposed to map reliability to RUL. To demonstrate the effectiveness of the proposed method, data of four tests of bearing degradation are utilized for RUL prediction. Compared with state-of-the-art methods, the proposed method shows higher prediction accuracy and robustness. The prediction results and evaluation indexes demonstrated the effectiveness and superiority of the proposed method. [ABSTRACT FROM AUTHOR]

Time series classification is an important research topic in machine learning and data mining communities, since time series data exist in many application domains. Recent studies have shown that machine learning algorithms could benefit from good feature representation, explaining why deep learning has achieved breakthrough performance in many tasks. In deep learning, the convolutional neural network (CNN) is one of the most well-known approaches, since it incorporates feature learning and classification task in a unified network architecture. Although CNN has been successfully applied to image and text domains, it is still a challenge to apply CNN to time series data. This paper proposes a tensor scheme along with a novel deep learning architecture called multivariate convolutional neural network (MVCNN) for multivariate time series classification, in which the proposed architecture considers multivariate and lag-feature characteristics. We evaluate our proposed method with the prognostics and health management (PHM) 2015 challenge data, and compare with several algorithms. The experimental results indicate that the proposed method outperforms the other alternatives using the prediction score, which is the evaluation metric used by the PHM Society 2015 data challenge. Besides performance evaluation, we provide detailed analysis about the proposed method. [ABSTRACT FROM AUTHOR]

This paper proposes a novel intelligent fault diagnosis method to automatically identify different health conditions of wind turbine (WT) gearbox. Unlike traditional approaches, where feature extraction and classification are separately designed and performed, this paper aims to automatically learn effective fault features directly from raw vibration signals while classify the type of faults in a single framework, thus providing an end-to-end learning-based fault diagnosis system for WT gearbox without additional signal processing and diagnostic expertise. Considering the multiscale characteristics inherent in vibration signals of a gearbox, a new multiscale convolutional neural network (MSCNN) architecture is proposed to perform multiscale feature extraction and classification simultaneously. The proposed MSCNN incorporates multiscale learning into the traditional CNN architecture, which has two merits: 1) high-level fault features can be effectively learned by the hierarchical learning structure with multiple pairs of convolutional and pooling layers; and 2) multiscale learning scheme can capture complementary and rich diagnosis information at different scales. This greatly improves the feature learning ability and enables better diagnosis performance. The proposed MSCNN approach is evaluated through experiments on a WT gearbox test rig. Experimental results and comprehensive comparison analysis with respect to the traditional CNN and traditional multiscale feature extractors have demonstrated the superiority of the proposed method. [ABSTRACT FROM AUTHOR]

Human pose recovery in videos is usually conducted by matching 2-D image features and retrieving relevant 3-D human poses. In the retrieving process, the mapping between images and poses is critical. Traditional methods assume this mapping relationship as local joint detection or global joint localization, which limits recovery performance of these methods since this two tasks are actually unified. In this paper, we propose a novel pose recovery framework by simultaneously learning the tasks of joint localization and joint detection. To obtain this framework, multiple manifold learning is used and the shared parameter is calculated. With them, multiple manifold regularizers are integrated and generalized eigendecomposition is utilized to achieve parameter optimization. In this way, pose recovery is boosted by both global mapping and local refinement. Experimental results on two popular datasets demonstrates that the recovery error has been reduced by 10%–20%, which proves the performance improvement of the proposed method. [ABSTRACT FROM PUBLISHER]

This paper introduces a new application of deep belief network (DBN) as an emerging artificial neural network for recognizing and classifying different partial discharge (PD) patterns. Phase resolved PD (PRPD) technique with different window intervals is used to manipulate three PD types, including corona, surface, and internal discharges measured in a high-voltage lab. Four approaches are proposed for extracting features from the raw measured data. In the first approach, the DBN is used as both a feature extractor and a PD classifier. The other three approaches extract discriminatory features using statistical and vector-norm-based operators to train a DBN classifier. The impact of the various phase windows through the PRPD method on the performance of the trained classifier is evaluated to obtain the best window interval. It is shown that the deep architectures are capable of learning important distinguishable features from PD data without any data preprocessing. This eliminates time-consuming feature extraction processes that produce the handcrafted features. Based on a comparison analysis, when the input data are corrupted by noise levels or no feature extraction technique is used to preprocess the data, the proposed approach outperforms other techniques, such as artificial neural networks, fuzzy logic classifiers, and support vector machines. [ABSTRACT FROM AUTHOR]

Tool condition monitoring is critical in ultraprecision manufacturing in order to optimize the performance of the overall process, while maintaining the desired part quality. Recently, deep learning has been successfully applied to numerous classification tasks in manufacturing, often to forecast part quality. In this paper, a novel deep learning data-driven modeling framework is presented, which includes a fusion of multiple stacked sparse autoencoders for tool condition monitoring in ultraprecision machining. The proposed computational framework consists of two main structures. First, a training model that is designed with the ability to process multiple parallel feature spaces to learn the lower-level features. Second, a feature fusion structure that is used to learn the higher-level features and associations to tool wear. To achieve this learning structure, a modified loss function is utilized that enhances the feature extraction and classification tasks. A dataset from a real manufacturing process is used to demonstrate the performance of the proposed framework. Experimental results and simulations show that the proposed method successfully classifies the ultraprecision machining case study with over 96% accuracy, while also outperforming comparable methodologies. [ABSTRACT FROM AUTHOR]

Bearing remaining useful life (RUL) prediction plays a crucial role in guaranteeing safe operation of machinery and reducing maintenance loss. In this paper, we present a new deep feature learning method for RUL estimation approach through time frequency representation (TFR) and multiscale convolutional neural network (MSCNN). TFR can reveal nonstationary property of a bearing degradation signal effectively. After acquiring time-series degradation signals, we get TFRs, which contain plenty of useful information using wavelet transform. Owing to high dimensionality, the size of these TFRs is reduced by bilinear interpolation, which are further regarded as inputs for deep learning models. Here, we introduce an MSCNN model structure, which keeps the global and local information synchronously compared to a traditional convolutional neural network (CNN). The salient features, which contribute for RUL estimation, can be learned automatically by MSCNN. The effectiveness of the presented method is validated by the experiment data. Compared to traditional data-driven and different CNN-based feature extraction methods, the proposed method shows enhanced performance in the prediction accuracy. [ABSTRACT FROM AUTHOR]

In this paper, a data-driven fault diagnosis model dealing with chemical imbalanced data streams is investigated. Different faults occur with varied frequencies by continuous arrival in chemical plants, while this issue has been hardly addressed in developing a diagnosis model. A novel incremental imbalance modified deep neural network (incremental-IMDNN) is proposed to promote the fault diagnosis to the imbalanced data stream. The first step in designing the incremental-IMDNN is the employment of an imbalance modified method combined with active learning for the extraction and generation of the most valuable information keeping in view the model feedback. DNN is utilized as a basic diagnosis model to excavate potential information. Then for the continuous arrival of new fault modes, DNN is promoted in an incremental hierarchical way. Unlike the traditional model that trained on a static snapshot of data, this model inherits the existing knowledge and hierarchically expands the diagnosis model by the similarity of faults. Similar faults that are judged by fuzzy clustering merge into a superclass, and every submodel shares the same architecture that is prevalent in previous research, which can be trained in parallel. We validate the performance of the proposed method in a Tennessee Eastman (TE) dataset, and the simulation results indicate that the proposed incremental-IM-DNN is better than the existing methods and possesses significant robustness and adaptability in chemical fault diagnosis. [ABSTRACT FROM AUTHOR]

The key challenge of intelligent fault diagnosis is to develop features that can distinguish different categories. Because of the unique properties of mechanical data, predetermined features based on prior knowledge are usually used as inputs for fault classification. However, proper selection of features often requires expertise knowledge and becomes more difficult and time consuming when volume of data increases. In this paper, a novel deep learning network (LiftingNet) is proposed to learn features adaptively from raw mechanical data without prior knowledge. Inspired by convolutional neural network and second generation wavelet transform, the LiftingNet is constructed to classify mechanical data even though inputs contain considerable noise and randomness. The LiftingNet consists of split layer, predict layer, update layer, pooling layer, and full-connection layer. Different kernel sizes are allowed in convolutional layers to improve learning ability. As a multilayer neural network, deep features are learned from shallow ones to represent complex structures in raw data. Feasibility and effectiveness of the LiftingNet is validated by two motor bearing datasets. Results show that the proposed method could achieve layerwise feature learning and successfully classify mechanical data even with different rotating speed and under the influence of random noise. [ABSTRACT FROM PUBLISHER]

Unmanned aerial vehicles are an essential component in the realization of Industry 4.0. With drones helping to improve industrial safety and efficiency in utilities, construction, and communication, there is an urgent need for drone-based intelligent applications. In this paper, we develop a unified framework to simultaneously detect and count vehicles from drone images. We first explore why the state-of-the-art detectors fail in highly dense drone scenes, which provides more appropriate insights. Then, we propose an effective loss to push the anchors toward matching the ground-truth boxes as much as possible, specifically designed for scale-adaptive anchor generation. Inspired by attention mechanisms in the human visual system, we maximize the mutual information between object classes and features by combining bottom-up cues with top-down attention mechanisms specifically designed for feature extraction. Finally, we build a counting layer with a regularized constraint related to the number of vehicles. Extensive experiments demonstrate the effectiveness of our approach. For both tasks, our proposed method achieves state-of-the-art results on all four challenging datasets. In particular, our results reduce error by a larger factor than previous methods. [ABSTRACT FROM AUTHOR]

Nowadays digital surveillance systems are universally installed for continuously collecting enormous amounts of data, thereby requiring human monitoring for the identification of different activities and events. Smarter surveillance is the need of this era through which normal and abnormal activities can be automatically identified using artificial intelligence and computer vision technology. In this paper, we propose a framework for activity recognition in surveillance videos captured over industrial systems. The continuous surveillance video stream is first divided into important shots, where shots are selected using the proposed convolutional neural network (CNN) based human saliency features. Next, temporal features of an activity in the sequence of frames are extracted by utilizing the convolutional layers of a FlowNet2 CNN model. Finally, a multilayer long short-term memory is presented for learning long-term sequences in the temporal optical flow features for activity recognition. Experiments 1 https://github.com/Aminullah6264/Activity%5fRec%5fML-LSTM. are conducted using different benchmark action and activity recognition datasets, and the results reveal the effectiveness of the proposed method for activity recognition in industrial settings compared with state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Efficient and accurate planetary gearbox fault diagnosis is the key to enhance the reliability and security of wind turbines. Therefore, an intelligent and integrated approach based on deep belief networks (DBNs), improved logistic Sigmoid (Isigmoid) units, and impulsive features is proposed in this paper. The vanishing gradient problem is an inherent drawback of conventional Sigmoid units, and it usually occurs in the backpropagation process of DBNs, resulting in that the training is considerably slowed down and the classification rate is reduced. To solve this problem, Isigmoid units are designed to combine the merits of unsaturation from leaky rectified linear (LReL) units. The results of handwritten digit recognition experiments show the superiority of Isigmoid over Sigmoid on convergence speed and classification accuracy. Since impulses contain much useful fault information, especially for early failures, an integrated approach using the optimized Morlet wavelet transform, kurtosis index, and soft-thresholding is applied to extract impulse components from original signals to improve the diagnosis accuracy. Then, the features extracted from original signals and impulsive signals are employed to train and test the DBNs with Isigmoid, Sigmoid, and LReL units for comparison. Finally, the results of planetary gearbox fault diagnosis show that Isigmoid has higher comprehensive performance than conventional sigmoid and LReL. [ABSTRACT FROM AUTHOR]

In addition to diagnosing a wiring of the vehicle in operation, it is also very important to find wire mismatches during the assembly process. In this paper, we propose a new method combining time–frequency-domain reflectometry and deep learning to verify that the wire is connected to the proper port of the underhood electrical center. Considering the frequency characteristics of each wire (black, blue, red, and yellow), we develop an optimization signal design algorithm. Using the time–frequency cross correlation (TFCC), the reflected signal generated at the impedance discontinuities is acquired and converted into the Wigner–Ville distribution image. Through the proposed algorithm, the existing images are converted into new images, which are easy to distinguish among groups. The new images are used as input of the convolutional neural network and trained to learn the feature of each group. The lengths, compensation filters, and the port information to be connected to each wire are stored in the filter bank. If the distance derived using the TFCC is different from the stored length, the wire is considered defective, and the acquired signal is restored by the compensation filter designed by the overcomplete wavelet transform method. Experimental results demonstrate the effectiveness of the proposed method for detecting the wire mismatch and fault location. [ABSTRACT FROM AUTHOR]

In this paper, we design an automatic pearl classification machine, composed of four parts: feeding mechanism, delivering mechanism, vision-based detection device, and classification mechanism. Pearls can be delivered to the detection device one by one, where multiview images of each pearl can be collected. A novel multistream convolutional neural network (MS-CNN) is developed to cope with these multiview images, with each stream processing an image of particular viewing angle and different streams sharing part of weights to fuse high-order features without losing too much diversity. Using the machine, we collect 52 500 multiview images for 10 500 pearls, i.e., each pearl has five images of top, left, right, main, and rear views. These pearls were labeled by the experienced professionals in advance, and grouped into two classes with rough rules and seven classes with fine rules. Experimental results show that, compared with the support vector machine and backpropagation neural network, our MS-CNN behaves much better in both classification tasks, obtaining 92.14% and 91.24% accuracies. Moreover, the visualization of activations of convolutional kernels suggests that MS-CNN, imitating the manual process, can indeed recognize relatively complex features. These results indicate the potential value of our machine in the pearl industry. [ABSTRACT FROM PUBLISHER]

Regular inspection of nuclear power plant components is important to guarantee safe operations. However, current practice is time consuming, tedious, and subjective, which involves human technicians reviewing the inspection videos and identifying cracks on reactors. A few vision-based crack detection approaches have been developed for metallic surfaces, and they typically perform poorly when used for analyzing nuclear inspection videos. Detecting these cracks is a challenging task since they are tiny, and noisy patterns exist on the components’ surfaces. This study proposes a deep learning framework, based on a convolutional neural network (CNN) and a Naïve Bayes data fusion scheme, called NB-CNN, to analyze individual video frames for crack detection while a novel data fusion scheme is proposed to aggregate the information extracted from each video frame to enhance the overall performance and robustness of the system. To this end, a CNN is proposed to detect crack patches in each video frame, while the proposed data fusion scheme maintains the spatiotemporal coherence of cracks in videos, and the Naïve Bayes decision making discards false positives effectively. The proposed framework achieves a 98.3% hit rate against 0.1 false positives per frame that is significantly higher than state-of-the-art approaches as presented in this paper. [ABSTRACT FROM PUBLISHER]

Data-driven soft sensors have been widely utilized in industrial processes to estimate the critical quality variables which are intractable to directly measure online through physical devices. Due to the low sampling rate of quality variables, most of the soft sensors are developed on small number of labeled samples and the large number of unlabeled process data is discarded. The loss of information greatly limits the improvement of quality prediction accuracy. One of the main issues of data-driven soft sensor is to furthest exploit the information contained in all available process data. This paper proposes a semisupervised deep learning model for soft sensor development based on the hierarchical extreme learning machine (HELM). First, the deep network structure of autoencoders is implemented for unsupervised feature extraction with all the process samples. Then, extreme learning machine is utilized for regression through appending the quality variable. Meanwhile, the manifold regularization method is introduced for semisupervised model training. The new method can not only deeply extract the information that the data contains, but learn more from the extra unlabeled samples as well. The proposed semisupervised HELM method is applied in a high–low transformer to estimate the carbon monoxide content, which shows a significant improvement of the prediction accuracy, compared to traditional methods. [ABSTRACT FROM PUBLISHER]

This paper deals with a new transformation and fusion of digital input patterns used to train and test feedforward neural network for a wound-rotor three-phase induction machine windings short-circuit diagnosis. The single type of short-circuits tested by the proposed approach is based on turn-to-turn fault which is known as the first stage of insulation degradation. Used input/output data have been binary coded in order to reduce the computation complexity. A new procedure, namely addition and mean of the set of same rank, has been implemented to eliminate the redundancy due to the periodic character of input signals. However, this approach has a great impact on the statistical properties on the processed data in terms of richness and of statistical distribution. The proposed neural network has been trained and tested with experimental signals coming from six current sensors implemented around a setup with a prime mover and a 5.5 kW wound-rotor three-phase induction generator. Both stator and rotor windings have been modified in order to sort out first and last turns in each phase. The experimental results highlight the superiority of using this new procedure in both training and testing modes. [ABSTRACT FROM PUBLISHER]

We demonstrate accurate spatio-temporal gait data classification from raw tomography sensor data without the need to reconstruct images. This is based on a simple yet efficient machine learning methodology based on a convolutional neural network architecture for learning spatio-temporal features, automatically end-to-end from raw sensor data. In a case study on a floor pressure tomography sensor, experimental results show an effective gait pattern classification F-score performance of 97.88 $\pm$ 1.70%. It is shown that the automatic extraction of classification features from raw data leads to a substantially better performance, compared to features derived by shallow machine learning models that use the reconstructed images as input, implying that for the purpose of automatic decision-making it is possible to eliminate the image reconstruction step. This approach is portable across a range of industrial tasks that involve tomography sensors. The proposed learning architecture is computationally efficient, has a low number of parameters and is able to achieve reliable classification F-score performance from a limited set of experimental samples. We also introduce a floor sensor dataset of 892 samples, encompassing experiments of 10 manners of walking and 3 cognitive-oriented tasks to yield a total of 13 types of gait patterns. [ABSTRACT FROM PUBLISHER]

Fused magnesium furnace (FMF) is the vital equipment for magnesia refractory production. The melting process of FMF is subject to melting temperature, magnesite quality and composition, and much more dynamic factors, which are prone to abnormal conditions such as overheating, abnormal exhausting, semi-molten, and affect the stability and safety of the production seriously. Due to the difficulties of effectively monitoring and accurately identifying the abnormal conditions with the existing methods, this article proposes a novel method for identifying the abnormal conditions of FMF based on deep learning and multi-information fusion. First, the Wasserstein distance based generative adversarial networks with gradient penalty is developed to solve the problem of obtaining overheating condition image samples with obvious visual characteristics. Then, the image-based deep convolutional neural network model is established to recognize the abnormal condition images and extract image features whose training set is the combination of the generated images and the original images. Finally, based on multisource information fusion, the traditional process current data are fused with the extracted image features, and a support vector machine model is established to complete the classification and identification of the FMF working conditions. Various experiments on actual industrial data show that the proposed method is effective and advanced. [ABSTRACT FROM AUTHOR]

A novel hybrid feature-reduction methodology is proposed as a contribution to the induction motor fault classification, to improve the classification rate of the current waveform events related to varieties of induction machine faults. This methodology relies on the combination of a feature-extraction technique based on the smoothed ambiguity plane designed for maximizing the separability between classes using Fisher's discriminant ratio, with the feature-selection technique, based on the proposed error-probability model to select an optimal number of the extracted features. This model depends on two parameters, namely, the smoothing kernel used to derive the features and the distance measurement. The proposed methodology is validated experimentally on a 5.5-kW induction motor test bench, and their performances are compared with the classification algorithm based on neural networks with sigmoid and wavelets in hidden neurons, known as a flexible tool for learning and recognizing system faults. The results obtained show an accurate classification independent from the load level. [ABSTRACT FROM PUBLISHER]

Slow feature analysis (SFA) is an unsupervised learning method that extracts the latent variables from a time series dataset based on the temporal slowness aspect. Neural networks, owing to their ability to model complex nonlinearities, can be used to extract slow features (SFs) from a dataset obtained from a complicated process. Siamese neural networks can be used for this purpose. Siamese neural networks have a provision of handling two samples at a time and this aspect helps in extracting SFs. For supervised learning applications, the extracted SFs should help predict the outputs. In this article, we present two approaches that extract SFs using Siamese neural networks. The output relevance aspect is brought into feature extraction as a regularization term in the objective function of the Siamese neural network. Such regularization improves the performance of the neural network model. The proposed approaches are implemented on three datasets to demonstrate their effectiveness. [ABSTRACT FROM AUTHOR]

Uncertainty modeling of renewable energy sources, load demand, electricity price, etc. create a high volume of data in smart grids. Accordingly, in this article, a precise forecasting method based on a deep learning concept with microclustering (MC) task is presented. The MC method is structured based on hybrid unsupervised and supervised clustering tasks by K-means and Gaussian support vector machine, respectively. In the proposed method, the input data sequence is clustered by the MC task, and then the forecasting process is employed. By applying the MC, input data in each hour are categorized into different groups, and a distinctive forecasting unit is allocated to each one. In this way, more clusters and forecasting networks are earmarked for the hours with higher fluctuation rates. The bi-directional long short-term memory (B-LSTM), which is one of the newest recurrent artificial neural networks, is proposed as the forecasting unit. The B-LSTM has bidirectional memory—feedforward and feedback loops—that helps us to investigate both previous and future hidden layers data. The optimal number of clusters in each hour is determined based on the Davies–Bouldin index. To evaluate the performance of the proposed method, in this study, three forecasting tasks including the wind speed, load demand, and electricity price are studied in different periods using the Ontario province, Canada, data set. The results are compared with other benchmarking methods to verify the robustness and effectiveness of the proposed method. In fact, the proposed method, which is equipped with the MC technique and B-LSTM networks, significantly promotes the forecasting results, especially in spike points. [ABSTRACT FROM AUTHOR]

Industrial process data are naturally complex time series with high nonlinearities and dynamics. To model nonlinear dynamic processes, a long short-term memory (LSTM) network is very suitable for soft sensor model development. However, the original LSTM does not consider variable and sample relevance for quality prediction. In order to overcome this problem, a spatiotemporal attention-based LSTM network is proposed for soft sensor modeling, which can, not only identify important input variables that are related to the quality variable at each time step, but also adaptively discover quality-related hidden states across all time steps. By taking the spatiotemporal quality-relevant interactions into consideration, the prediction performance can be improved for the soft sensor model. The effectiveness and flexibility of the proposed model is demonstrated on an industrial hydrocracking process to predict the initial boiling points of heavy naphtha and aviation kerosene. [ABSTRACT FROM AUTHOR]

Fault diagnosis is vital in manufacturing system, since early detections on the emerging problem can save invaluable time and cost. With the development of smart manufacturing, the data-driven fault diagnosis becomes a hot topic. However, the traditional data-driven fault diagnosis methods rely on the features extracted by experts. The feature extraction process is an exhausted work and greatly impacts the final result. Deep learning (DL) provides an effective way to extract the features of raw data automatically. Convolutional neural network (CNN) is an effective DL method. In this study, a new CNN based on LeNet-5 is proposed for fault diagnosis. Through a conversion method converting signals into two-dimensional (2-D) images, the proposed method can extract the features of the converted 2-D images and eliminate the effect of handcrafted features. The proposed method which is tested on three famous datasets, including motor bearing dataset, self-priming centrifugal pump dataset, and axial piston hydraulic pump dataset, has achieved prediction accuracy of 99.79%, 99.481%, and 100%, respectively. The results have been compared with other DL and traditional methods, including adaptive deep CNN, sparse filter, deep belief network, and support vector machine. The comparisons show that the proposed CNN-based data-driven fault diagnosis method has achieved significant improvements. [ABSTRACT FROM PUBLISHER]