Deep Autoencoder Model Research Articles

A Two-Stage Transfer Learning-Based Deep Learning Approach for Production Progress Prediction in IoT-Enabled Manufacturing

In make-to-order manufacturing enterprises, accurate production progress (PP) prediction is an important basis for dynamic production process optimization and on-time delivery of orders. The implementation of Internet of Things (IoT) makes it possible to take real-time production state as an important factor affecting PP. In the IoT-enabled workshop, a two-stage transfer learning-based prediction method using both historical production data and real-time state data is proposed to solve the problem of low-prediction accuracy and poor generalization performance caused by insufficient data of target order. The deep autoencoder (DAE) model with transfer learning is designed to extract the generalized features of target order in the first stage, which uses bootstrap sampling to avoid over fitting. The deep belief network (DBN) model with transfer learning is constructed to fit the nonlinear relation for PP prediction in the second stage. A real case from an IoT enabled machining workshop is taken to validate the performance of the proposed method over the other methods such as DBN, deep neural network.

SPONTANEOUS CATEGORIZATION AND SELF-LEARNING WITH DEEP AUTOENCODER MODELS

У цій роботі автор досліджував обробку інформації у моделях глибокого автокодування. Було продемонстровано що непідконтрольне навчання автокодерам певного класу може призвести до появи компактного та структурованого внутрішнього представлення простору вхідних даних, що може бути співвіднесено з категоріями вищого рівня. Була запропонована і продемонстрована практичнa можливість виявити та виміряти цю формується інформаційну структуру шляхом застосування непідконтрольного кластеризації щільності в просторі активації фокусного прихованого шару моделі. На основі отриманих висновків запропонований новий підхід до навчання моделей нейронних мереж, що базується на структурах виникаючих у в неконтрольованом інформаційному середовищі навчання, який є ітеративним, керованим навколишнім середовищем, вимагає мінімального нагляду та з подібністю до вивчення біологічних систем i також дає хороші результати класифікації при навчанні нових концепцій вищого рівня навіть при мінімальному наборі маркованих даних. На закінчення надається обговорення теоретичних основ спонтанної категоризації в системах самонавчання.

Cyber Physical Security Analytics for Anomalies in Transmission Protection Systems

Protection systems are one of the most critical components in the transmission system and are becoming more digital with ongoing automation. These digital systems are prone to vulnerabilities/attacks, and exploitation of these vulnerabilities may cause major impacts on the electric grid performance. Multiple alarms reported in the control center could be a result of the faults (expected operations) or failures in the protection system (anomalies/ unexpected operation). Situational awareness gained through sensors such as a phasor measurement unit (PMU) and data acquired through the cyber system provide an opportunity to develop continuous cyber-physical monitoring of the system. Note that relay data are not reported in the control center continuously. This paper presents a cyber-physical data analytics based technique to monitor transmission protection system and detect malicious activity. Initially, continuous monitoring of PMU data is utilized for data anomaly detection, which includes bad or missing data using long short-term memory (LSTM). Then, PMU data of interest are utilized for failure diagnosis, using a semisupervised deep autoencoder model. In this research, cyber anomalies are modeled by manipulating the setting/logic design of protective devices, and a ridge regression based classifier with a feature engineering pipeline is used to detect cyber anomalies. The results from the deep autoencoder model and ridge regression based classifier are then utilized for detailed investigation to find the root causes of the observed events assisted by the cyber log data from the protection devices. The algorithm is validated using a real-time simulation of the IEEE test system with industrial hardware relays and PMUs in the loop. Data analytics algorithm running on server utilizes these real-time data continuously for anomaly detection and classification for the developed use cases.

A High Generalizable Feature Extraction Method Using Ensemble Learning and Deep Auto-Encoders for Operational Reliability Assessment of Bearings

Feature extraction is a major challenge in operational reliability assessment, which requires techniques and prior knowledge. Deep auto-encoder (DAE) is a popular deep learning method and is widely used in feature extraction. However, low generalization ability and structure parameters design are still the major problems of DAE for operational reliability assessment. To overcome the two problems, an ensemble DAE is proposed for operational reliability assessment. Firstly, different structure parameters are employed to design a series of DAEs for feature learning from the measured data. Secondly, a feature ensemble strategy is designed to enhance the generalization ability of the DAE model, in which the features learned by different DAEs are clustered to remove the irrelevant DAEs and select the more general feature subset. Finally, the operational reliability indicator is defined by the Euclidean distance of the selected features and the operational reliability model is developed. The proposed method is utilized to analyze the experimental bearings and the results indicate that the proposed method is effective for operational reliability assessment.

Unsupervised deep feature embeddings for speaker diarization

Speaker diarization aims to determine ?who spoke when?? from multispeaker recording environments. In this paper, we propose to learn a set of high-level feature representations, referred to as feature embeddings, from an unsupervised deep architecture for speaker diarization. These sets of embeddings are learned through a deep autoencoder model when trained on mel-frequency cepstral coefficients (MFCCs) of input speech frames. Learned embeddings are then used in Gaussian mixture model based hierarchical clustering for diarization. The results show that these unsupervised embeddings are better compared to MFCCs in reducing the diarization error rate. Experiments conducted on the popular subset of the AMI meeting corpus consisting of 5.4 h of recordings show that the new embeddings decrease the average diarization error rate by 2.96%. However, for individual recordings, maximum improvement of 8.05% is acquired.

Elevator Fault Detection Using Profile Extraction and Deep Autoencoder Feature Extraction for Acceleration and Magnetic Signals

In this paper, we propose a new algorithm for data extraction from time-series data, and furthermore automatic calculation of highly informative deep features to be used in fault detection. In data extraction, elevator start and stop events are extracted from sensor data including both acceleration and magnetic signals. In addition, a generic deep autoencoder model is also developed for automated feature extraction from the extracted profiles. After this, extracted deep features are classified with random forest algorithm for fault detection. Sensor data are labelled as healthy and faulty based on the maintenance actions recorded. The remaining healthy data are used for validation of the model to prove its efficacy in terms of avoiding false positives. We have achieved above 90% accuracy in fault detection along with avoiding false positives based on new extracted deep features, which outperforms results using existing features. Existing features are also classified with random forest to compare results. Our developed algorithm provides better results due to the new deep features extracted from the dataset when compared to existing features. This research will help various predictive maintenance systems to detect false alarms, which will in turn reduce unnecessary visits of service technicians to installation sites.

Optimizing neural networks for medical data sets: A case study on neonatal apnea prediction.

The neonatal period of a child is considered the most crucial phase of its physical development and future health. As per the World Health Organization, India has the highest number of pre-term births [1], with over 3.5 million babies born prematurely, and up to 40% of them are babies with low birth weights, highly prone to a multitude of diseases such as Jaundice, Sepsis, Apnea, and other Metabolic disorders. Apnea is the primary concern for caretakers of neonates in intensive care units. The real-time medical data is known to be noisy and nonlinear and to address the resultant complexity in classification and prediction of diseases; there is a need for optimizing learning models to maximize predictive performance. Our study attempts to optimize neural network architectures to predict the occurrence of apneic episodes in neonates, after the first week of admission to Neonatal Intensive Care Unit (NICU). The primary contribution of this study is the formulation and description of a set of generic steps involved in selecting various model-specific, training and hyper-parametric optimization algorithms, as well as model architectures for optimal predictive performance on complex and noisy medical datasets. The data used for the study being inherently complex and noisy, Kernel Principal Component Analysis (PCA) is used to reduce dataset dimensionality for the analysis such as interpretations and visualization of the dataset. Hyper-parametric and parametric optimization, in different categories, are considered, including learning rate updater algorithms, regularization methods, activation functions, gradient descent algorithms and depth of the network, based on their performance on the validation set, to obtain a holistically optimized neural network, that best model the given complex medical dataset. Deep Neural Network Architectures such as Deep Multilayer Perceptron's, Stacked Auto-encoders and Deep Belief Networks are employed to model the dataset, and their performance is compared to the optimized neural network obtained from the parametric exploration. Further, the results are compared with Support Vector Machine (SVM), K Nearest Neighbor, Decision Tree (DT) and Random Forest (RF) algorithms. The results indicate that the optimized eight layer Multilayer Perceptron (MLP) model, with Adam Decay and Stochastic Gradient Descent (AUC 0.82) can outperform the conventional machine learning models, and perform comparably to the Deep Auto-encoder model (AUC 0.83) in predicting the presence of apnea in neonates. The study shows that an MLP model can undergo significant improvements in predictive performance, by the proposed step-wise optimization. The optimized MLP is proved to be as accurate as deep neural network models such as Deep Belief Networks and Deep Auto-encoders for noisy and nonlinear data sets, and outperform all conventional models like Support Vector Machine (SVM), Decision Tree (DT), K Nearest Neighbor and Random Forest (RF) algorithms. The generic nature of the proposed step-wise optimization provides a framework to optimize neural networks on such complex nonlinear datasets. The investigated models can help neonatologists as a diagnostic tool.

A novel speech emotion recognition algorithm based on wavelet kernel sparse classifier in stacked deep auto-encoder model

Since the contextual information has an important impact on the speaker’s emotional state, how to use emotion-related context information to conduct feature learning is a key problem. The existing speech emotion recognition algorithms achieve the relatively high recognition rate; these algorithms are not very good application to the real-life speech emotion recognition systems. Therefore, in order to address the abovementioned issues, a novel speech emotion recognition algorithm based on improved stacked kernel sparse deep model is proposed in this paper, which is based on auto-encoder, denoising auto-encoder, and sparse auto-encoder to improve the Chinese speech emotion recognition. The first layer of the structure uses a denoising auto-encoder to learn a hidden feature with a larger dimension than the dimension of the input features, and the second layer employs a sparse auto-encoder to learn sparse features. Finally, a wavelet-kernel sparse SVM classifier is applied to classify the features. The proposed algorithm is evaluated on the testing dataset, which contains the speech emotion data of spontaneous, non-prototypical, and long-term. The experimental results show that the proposed algorithm outperforms the existing state-of-the-art algorithms in speech emotion recognition.

Performance degradation assessment of slewing bearings based on deep auto-encoder and optimized particle filtering

Due to the low speed and heavy load conditions of slewing bearings, extracting of effective features for fault diagnosis and prediction is difficult but crucial. Moreover, challenges such as large data volumes, unlabeled and multi-source bring more difficulties for advanced prognosis and health management methods. To solve these problems, a novel method for performance degradation assessment of bearings based on raw signals is proposed. In this methodology, a combination of deep auto-encoder (DAE) algorithm and particle filter algorithm is utilized for feature extraction and remaining useful life (RUL) prediction. First, the raw vibration signal is employed to train parameters of a restricted Boltzmann machine to build the DAE model. Through encoding and decoding multi-source data, root mean square error of reconstruction error between the raw signal and reconstructed signal is employed to detect incipient faults of slewing bearings. Then, degradation trend model is established by particle filtering to predict RUL of bearings. The effectiveness of proposed method is validated using simulated and experimental vibration signals. Results illustrate that proposed method can evaluate the performance degradation process and RUL of slewing bearings.

Unsupervised abnormality detection through mixed structure regularization (MSR) in deep sparse autoencoders.

The purpose of this study is to introduce and evaluate the mixed structure regularization (MSR) approach for a deep sparse autoencoder aimed at unsupervised abnormality detection in medical images. Unsupervised abnormality detection based on identifying outliers using deep sparse autoencoders is a very appealing approach for computer-aided detection systems as it requires only healthy data for training rather than expert annotated abnormality. However, regularization is required to avoid overfitting of the network to the training data. We used coronary computed tomography angiography (CCTA) datasets of 90 subjects with expert annotated centerlines. We segmented coronary lumen and wall using an automatic algorithm with manual corrections where required. We defined normal coronary cross section as cross sections with a ratio between lumen and wall areas larger than 0.8. We divided the datasets into training, validation, and testing groups in a tenfold cross-validation scheme. We trained a deep sparse overcomplete autoencoder model for normality modeling with random structure and noise augmentation. We assessed the performance of our deep sparse autoencoder with MSR without denoising (SAE-MSR) and with denoising (SDAE-MSR) in comparison to deep sparse autoencoder (SAE), and deep sparse denoising autoencoder (SDAE) models in the task of detecting coronary artery disease from CCTA data on the test group. The SDAE-MSR achieved the best aggregated area under the curve (AUC) with a 20% improvement and the best aggregated Average Precision (AP) with a 30% improvement upon the SAE and SDAE (AUC: 0.78 to 0.94, AP: 0.66 to 0.86) in distinguishing between coronary cross sections with mild stenosis (stenosis grade<0.3) and coronary cross sections with severe stenosis (stenosis grade>0.7). The improvements were statistically significant (Mann-Whitney U-test, P<0.001). Similarly, The SDAE-MSR achieved the best aggregated AUC (AP) with an 18% (18%) improvement upon the SAE and SDAE (AUC: 0.71 to 0.84, AP: 0.68 to 0.80). The improvements were statistically significant (Mann-Whitney U-test, P<0.05). Deep sparse autoencoders with MSR in addition to explicit sparsity regularization term and stochastic corruption of the input data with Gaussian noise have the potential to improve unsupervised abnormality detection using deep-learning compared to common deep autoencoders.

IAI-DSAE: A Computational Method for Adenosine to Inosine Editing Site Prediction

As an important post-transcriptional modification, adenosine-to-inosine RNA editing generally occurs in both coding and noncoding RNA transcripts in which adenosines are converted to inosines. Accordingly, the diversification of the transcriptome can be resulted in by this modification. It is significant to accurately identify adenosine-to-inosine editing sites for further understanding their biological functions. Currently, the adenosine-to-inosine editing sites would be determined by experimental methods, unfortunately, it may be costly and time consuming. Furthermore, there are only a few existing computational prediction models in this field. Therefore, the work in this study is starting to develop other computational methods to address these problems. Given an uncharacterized RNA sequence that contains many adenosine resides, can we identify which one of them can be converted to inosine, and which one cannot? To deal with this problem, a novel predictor called iAI-DSAE is proposed in the current study. In fact, there are two key issues to address: one is ‘what feature extraction methods should be adopted to formulate the given sample sequence?’ The other is ‘what classification algorithms should be used to construct the classification model?’ For the former, a 540-dimensional feature vector is extracted to formulate the sample sequence by dinucleotide-based auto-cross covariance, pseudo dinucleotide composition, and nucleotide density methods. For the latter, we use the present more popular method i.e. deep spare autoencoder to construct the classification model. Generally, ACC and MCC are considered as the two of the most important performance indicators of a predictor. In this study, in comparison with those of predictor PAI, they are up 2.46% and 4.14%, respectively. The two other indicators, Sn and Sp, rise at certain degree also. This indicates that our predictor can be as an important complementary tool to identify adenosine-toinosine RNA editing sites. For the convenience of most experimental scientists, an easy-to-use web-server for identifying adenosine-to-inosine editing sites has been established at: http://www.jci-bioinfo.cn/iAI-DSAE, by which users can easily obtain their desired results without the need to go through the complicated mathematical equations involved. It is important to identify adenosine-to-inosine editing sites in RNA sequences for the intensive study on RNA function and the development of new medicine. In current study, a novel predictor, called iAI-DSAE, was proposed by using three feature extraction methods including dinucleotidebased auto-cross covariance, pseudo dinucleotide composition and nucleotide density. The jackknife test results of the iAI-DSAE predictor based on deep spare auto-encoder model show that our predictor is more stable and reliable. It has not escaped our notice that the methods proposed in the current paper can be used to solve many other problems in genome analysis.

Classifier Adaptive Fusion: Deep Learning for Robust Outdoor Vehicle Visual Tracking

Deep auto-encoder (DAE) models have been successfully used in object tracking due to its strong capability of feature representation. However, single deep auto-encoder model would not be robust enough to represent the appearance model of outdoor vehicle for its harsh working environment, such as illumination variation, occlusion, cluttered background and so on. In this paper, a novel multiple-DAE-based tracking approach, that is, classifier adaptive fusion for robust outdoor vehicle visual tracking approach is proposed under particle filter framework. Firstly, two deep auto-encoders are offline trained by gray-scale image and gradient image of the raw training images, respectively to obtain the stronger feature representation of gray-scale image and gradient image. Secondly, two classifiers are constructed using the encoder of the two well-trained deep auto-encoders and the output of the each classifier is used to compute the confidence of the corresponding particles. Finally, the confidence output of the two classifiers is fused and applied in online tracking, where, the fusion weight of the each classifier is computed according to the distribution of particles represented by different classifier. Extensive tracking experiments conducted on visual tracking benchmark (VTB) show that the proposed tracking algorithm outperforms 9 popular tracking algorithms in the challenge scenes of outdoor vehicle tracking such as illumination variation, occlusion, cluttered background and scale variation.

An Intelligent Scheme for Continuous Authentication of Smartphone Using Deep Auto Encoder and Softmax Regression Model Easy for User Brain

The smartphones are becoming more popular these days because of having more technologies in a smaller and slim-sized device. The people can get Internet connectivity and some other facilities, such as communication and storing data, with these smaller devices. It is necessary for securing personal data on smartphone devices with authentication techniques. In recent years, most of the static authentication techniques, such as password, patterns, and fingerprint, were used for securing smartphones. Most of the studies are focusing on a new method called continuous authentication for improving the security of smartphones. The continuous authentication techniques differ from other static authentication techniques by authenticating the smartphone user periodically. It uses behavioral features of users for authenticating the user. Continuous authentication is easy for the user's brain, as a user need not to memorize anything. In this paper, a continuous authentication technique is designed with the Deep Auto Encoder and Softmax Regression techniques (DAE-SR). The DAE-SR achieved 0.950% and 0.970% of accuracy on predicting the users on different states (walking and sitting) of users.

Software Abstractions for Large-Scale Deep Learning Models in Big Data Analytics

The goal of big data analytics is to analyze datasets with a higher amount of volume, velocity, and variety for large-scale business intelligence problems. These workloads are normally processed with the distribution on massively parallel analytical systems. Deep learning is part of a broader family of machine learning methods based on learning representations of data. Deep learning plays a significant role in the information analysis by adding value to the massive amount of unsupervised data. A core domain of research is related to the development of deep learning algorithms for auto-extraction of complex data formats at a higher level of abstraction using the massive volumes of data. In this paper, we present the latest research trends in the development of parallel algorithms, optimization techniques, tools and libraries related to big data analytics and deep learning on various parallel architectures. The basic building blocks for deep learning such as Restricted Boltzmann Machines (RBM) and Deep Belief Networks (DBN) are identified and analyzed for parallelization of deep learning models. We proposed a parallel software API based on PyTorch, Hadoop Distributed File System (HDFS), Apache Hadoop MapReduce and MapReduce Job (MRJob) for developing large-scale deep learning models. We obtained about 5-30% reduction in the execution time of the deep auto-encoder model even on a single node Hadoop cluster. Furthermore, the complexity of code development is significantly reduced to create multi-layer deep learning models.

Fast and scalable distributed deep convolutional autoencoder for fMRI big data analytics.

In recent years, analyzing task-based fMRI (tfMRI) data has become an essential tool for understanding brain function and networks. However, due to the sheer size of tfMRI data, its intrinsic complex structure, and lack of ground truth of underlying neural activities, modeling tfMRI data is hard and challenging. Previously proposed data modeling methods including Independent Component Analysis (ICA) and Sparse Dictionary Learning only provided shallow models based on blind source separation under the strong assumption that original fMRI signals could be linearly decomposed into time series components with corresponding spatial maps. Given the Convolutional Neural Network (CNN) successes in learning hierarchical abstractions from low-level data such as tfMRI time series, in this work we propose a novel scalable distributed deep CNN autoencoder model and apply it for fMRI big data analysis. This model aims to both learn the complex hierarchical structures of the tfMRI big data and to leverage the processing power of multiple GPUs in a distributed fashion. To deploy such a model, we have created an enhanced processing pipeline on the top of Apache Spark and Tensorflow, leveraging from a large cluster of GPU nodes over cloud. Experimental results from applying the model on the Human Connectome Project (HCP) data show that the proposed model is efficient and scalable toward tfMRI big data modeling and analytics, thus enabling data-driven extraction of hierarchical neuroscientific information from massive fMRI big data.

Unsupervised Novelty Detection Using Deep Autoencoders with Density Based Clustering

Novelty detection is a classification problem to identify abnormal patterns; therefore, it is an important task for applications such as fraud detection, fault diagnosis and disease detection. However, when there is no label that indicates normal and abnormal data, it will need expensive domain and professional knowledge, so an unsupervised novelty detection approach will be used. On the other hand, nowadays, using novelty detection on high dimensional data is a big challenge and previous research suggests approaches based on principal component analysis (PCA) and an autoencoder in order to reduce dimensionality. In this paper, we propose deep autoencoders with density based clustering (DAE-DBC); this approach calculates compressed data and error threshold from deep autoencoder model, sending the results to a density based cluster. Points that are not involved in any groups are not considered a novelty; the grouping points will be defined as a novelty group depending on the ratio of the points exceeding the error threshold. We have conducted the experiment by substituting components to show that the components of the proposed method together are more effective. As a result of the experiment, the DAE-DBC approach is more efficient; its area under the curve (AUC) is shown to be 13.5 percent higher than state-of-the-art algorithms and other versions of the proposed method that we have demonstrated.

Unsupervised Deep Feature Learning for Remote Sensing Image Retrieval

Due to the specific characteristics and complicated contents of remote sensing (RS) images, remote sensing image retrieval (RSIR) is always an open and tough research topic in the RS community. There are two basic blocks in RSIR, including feature learning and similarity matching. In this paper, we focus on developing an effective feature learning method for RSIR. With the help of the deep learning technique, the proposed feature learning method is designed under the bag-of-words (BOW) paradigm. Thus, we name the obtained feature deep BOW (DBOW). The learning process consists of two parts, including image descriptor learning and feature construction. First, to explore the complex contents within the RS image, we extract the image descriptor in the image patch level rather than the whole image. In addition, instead of using the handcrafted feature to describe the patches, we propose the deep convolutional auto-encoder (DCAE) model to deeply learn the discriminative descriptor for the RS image. Second, the k-means algorithm is selected to generate the codebook using the obtained deep descriptors. Then, the final histogrammic DBOW features are acquired by counting the frequency of the single code word. When we get the DBOW features from the RS images, the similarities between RS images are measured using L1-norm distance. Then, the retrieval results can be acquired according to the similarity order. The encouraging experimental results counted on four public RS image archives demonstrate that our DBOW feature is effective for the RSIR task. Compared with the existing RS image features, our DBOW can achieve improved behavior on RSIR.

An efficient compression of ECG signals using deep convolutional autoencoders

Background and objectiveAdvances in information technology have facilitated the retrieval and processing of biomedical data. Especially with wearable technologies and mobile platforms, we are able to follow our healthcare data, such as electrocardiograms (ECG), in real time. However, the hardware resources of these technologies are limited. For this reason, the optimal storage and safe transmission of the personal health data is critical. This study proposes a new deep convolutional autoencoder (CAE) model for compressing ECG signals. MethodsIn this paper, a deep network structure of 27 layers consisting of encoder and decoder parts is designed. In the encoder section of this model, the signals are reduced to low-dimensional vectors; and in the decoder section, the signals are reconstructed. The deep learning approach provides the representations of the low and high levels of signals in the hidden layers of the model. Hence, the original signal can be reconstructed with minimal loss. Very different from traditional linear transformation methods, a deep compression approach implies that it can learn to use different ECG records automatically. ResultsThe performance was evaluated on an experimental data set comprising 4800 ECG fragments from 48 unique clinical patients. The compression rate (CR) of the proposed model was 32.25, and the average PRD value was 2.73%. These favourable observation suggest that our deep model can allow secure data transfer in a low-dimensional form to remote medical centers. We present an effective compression approach that can potentially be used in wearable devices, e-health applications, telemetry and Holter systems.

Reservoir Inflow Forecast Using a Clustered Random Deep Fusion Approach in the Three Gorges Reservoir, China

Reservoir inflow forecast plays a crucial part in programming, development, operation, and management of water resource systems. To better reveal the complex properties of daily reservoir inflow, a clustered deep fusion (CDF) approach is proposed in this paper. First, variational mode decomposition (VMD) is used to decompose the daily reservoir inflow series into multiple modes, which are clustered into different sets by fuzzy c-means according to the Xie-Beni index in view of attribute domain. In each cluster, a deep autoencoder model (DAE) is developed for deep representations of the attributes in the deep domain. DAE outputs are finally fused at the synthesis domain into the forecasting results using random forest (RF). In this way, the inflow time series may be successively observed in the attribute domain, deep domain, and synthesis domain, which results in a clearer understanding of reservoir inflow trend. The present approach is modeled and evaluated using historical data collected from the Three Gorges Reservoir, China. For comparison, two kinds of learning patterns—deep learning (VMD-DAE-RF and DAE) and shallow learning (feed-forward neural network, least-squares support regression, and RF)—are applied to the same case. The results indicate that the proposed CDF model outperforms all comparison models in terms of mean absolute percentage error (6.174%), root mean-square error (1,077.428 m3/s), and correlation coefficient criteria (0.987). Thus, it is concluded that deep learning in the cluster fusion architecture is more promising.

Wind Turbine Blade Breakage Monitoring With Deep Autoencoders

Monitoring wind turbine blade breakages based on supervisory control and data acquisition (SCADA) data is investigated in this research. A preliminary data analysis is performed to demonstrate that existing SCADA features are unable to present irregular patterns prior to occurrences of blade breakages. A deep autoencoder (DA) model is introduced to derive an indicator of impending blade breakages, the reconstruction error (RE), from SCADA data. The DA model is a neural network of multiple hidden layers organized symmetrically. In training DA models, the restricted Boltzmann machine is applied to initialize weights and biases. The back-propagation method is subsequently employed to further optimize the network structure. Through examining SCADA data, we observe that the trend of RE will shift by the blade breakage. To effectively detect RE shifts through online monitoring, the exponentially weighted moving average control chart is deployed. The effectiveness of the proposed monitoring approach is validated by blade breakage cases collected from wind farms located in China. The computational results prove the capability of the proposed monitoring approach in identifying impending blade breakages.