Search Results (64)

A rapid dynamic security assessment (DSA) is crucial for online preventive and restoration decision-making. The deep learning-based DSA models have high efficiency and accuracy. However, the complex model structure and high training cost make them hard to update quickly. This paper proposes a dynamic security partition assessment method, aiming to develop accurate and incrementally updated DSA models with simple structures. Firstly, the power grid is self-adaptively partitioned into several local regions based on the mean shift algorithm. The input of the mean shift algorithm is a symmetric electrical distance matrix, and the distance metric is the Chebyshev distance. Secondly, high-level features of operating conditions are extracted based on the stacked denoising autoencoder. The symmetric electrical distance matrix is modified to represent fault locations in local regions. Finally, DSA models are constructed for fault locations in each region based on the radial basis function neural network (RBFNN) and Chebyshev distance. An online incremental updating strategy is designed to enhance the model adaptability. With the simulation software PSS/E 33.4.0, the proposed dynamic security partition assessment method is verified in a simplified provincial system and a large-scale practical system in China. Test results demonstrate that the Chebyshev distance can improve the partition quality of the mean shift algorithm by approximately 50%. The RBFNN-based partition assessment model achieves an accuracy of 98.96%, which is higher than the unified assessment with complex models. The proposed incremental updating strategy achieves an accuracy of over 98% and shortens the updating time to 30 s, which can meet the efficiency of online application. Full article

To improve the reliability and maintainability of the nuclear safety-class digital control system (DCS), this paper conducts a study on the fault prediction of critical components in the output circuit of the nuclear safety-class signal conditioning module. To address the issue of insufficient feature extraction for the minor offset fault feature and the low accuracy of fault prediction, a predictive model based on stacked denoising autoencoder (SDAE) feature extraction and an improved gated recurrent unit (GRU) is proposed. Therefore, fault simulation modeling is performed for critical components of the signal output circuit to obtain fault datasets of critical components, and the SDAE model is used to extract fault features. The fault prediction model based on GRU is established, and the number of hidden layers, the number of hidden layer nodes, and the learning rate of the GRU model are optimized using the adaptive gray wolf optimization algorithm (AGWO). The prediction performance evaluation metrics include the root mean square error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), and absolute error (EA), which are used for evaluating the prediction results of models such as the AGWO-GRU model, recurrent neural network (RNN) model, and long short-term memory network (LSTM). The results show that the GRU model optimized by AGWO has a better prediction accuracy (errors range within 0.01%) for the faults of the circuit critical components, and, moreover, can accurately and stably predict the fault trend of the circuit. Full article

Aiming at the problem that it is difficult to effectively characterize the operation status of hydropower units with a single vibration signal feature under the influence of multiple factors such as water–machine–electricity coupling, a multidimensional fusion feature extraction method for hydroelectric units based on time–frequency analysis and unsupervised learning models is proposed. Firstly, the typical time–domain and frequency–domain characteristics of vibration signals are calculated through amplitude domain analysis and Fourier transform. Secondly, the time–frequency characteristics of vibration signals are obtained by combining the complementary ensemble empirical mode decomposition and energy moment calculation methods to supplement the traditional time–domain and frequency–domain characteristics, which have difficulty in comprehensively reflecting the correlation between nonlinear non–stationary signals and the state of the unit. Finally, in order to overcome the limitations of shallow feature extraction relying on artificial experience, a Stacked Denoising Autoencoder is used to adaptively mine the deep features of vibration signals, and the extracted features are fused to construct a multidimensional feature vector of vibration signals. The proposed multidimensional information fusion feature extraction method is verified to realize the multidimensional complementarity of feature attributes, which helps to accurately distinguish equipment state types and provides the foundation for subsequent state identification and trend prediction. Full article

The paper proposes an anomaly detection method for UAVs based on wavelet decomposition and stacked denoising autoencoder. This method takes the negative impact of noisy data and the feature extraction capabilities of deep learning models into account. It aims to improve the accuracy of the proposed anomaly detection method with wavelet decomposition and stacked denoising autoencoder methods. Anomaly detection based on UAV flight data is an important method of UAV condition monitoring and potential abnormal state mining, which is an important means to reduce the risk of UAV flight accidents. However, the diversity of UAV mission scenarios leads to a complex and harsh environment, so the acquired data are affected by noise, which brings challenges to accurate anomaly detection based on UAV data. Firstly, we use wavelet decomposition to denoise the original data; then, we used the stacked denoising autoencoder to achieve feature extraction. Finally, the softmax classifier is used to realize the anomaly detection of UAV. The experimental results demonstrate that the proposed method still has good performance in the case of noisy data. Specifically, the Accuracy reaches 97.53%, the Precision is 97.50%, the Recall is 91.81%, and the F1-score is 94.57%. Furthermore, the proposed method outperforms the four comparison models with more outstanding performance. Therefore, it has significant potential in reducing UAV flight accidents and enhancing operational safety. Full article

With the development of artificial intelligence and deep learning, deep neural networks have become an important method for predicting the remaining useful life (RUL) of lithium-ion batteries. In this paper, drawing inspiration from the transformer sequence-to-sequence task’s transformation capability, we propose a fusion model that integrates the functions of the stacked denoising autoencoder (SDAE) and the Transformer model in order to improve the performance of RUL prediction. Firstly, the health factors under three different conditions are extracted from the measurement data as model inputs. These conditions include constant current and voltage, random discharge, and the application of principal component analysis (PCA) for dimensionality reduction. Then, SDAE is responsible for denoising and feature extraction, and the Transformer model is utilized for sequence modeling and RUL prediction of the processed data. Finally, accurate prediction of the RUL of the four battery cells is achieved through cross-validation and four sets of comparison experiments. Three evaluation metrics, MAE, RMSE, and MAPE, are selected, and the values of these metrics are 0.170, 0.202, and 19.611%, respectively. The results demonstrate that the proposed method outperforms other prediction models in terms of prediction accuracy, robustness, and generalizability. This provides a new solution direction for the daily life prediction research of lithium-ion batteries. Full article

The reliability and stability of differential protection in power transformers could be threatened by several types of inferences, including magnetizing inrush currents, current transformer saturation, and overexcitation from external faults. The robustness of deep learning applications employed for power system protection in recent years has offered solutions to deal with several disturbances. This paper presents a method for detecting internal faults in power transformers occurring simultaneously with inrush currents. It involves utilizing a data window (DW) and stacked denoising autoencoders. Unlike the conventional method, the proposed scheme requires no thresholds to discriminate internal faults and inrush currents. The performance of the algorithm was verified using fault data from a typical Korean 154 kV distribution substation. Inrush current variation and internal faults were simulated and generated in PSCAD/EMTDC, considering various parameters that affect an inrush current. The results indicate that the proposed scheme can detect the appearance of internal faults occurring simultaneously with an inrush current. Moreover, it shows promising results compared to the prevailing methods, ensuring the superiority of the proposed method. From sample N–3, the proposed DNN demonstrates accurate discrimination between internal faults and inrush currents, achieving accuracy, sensitivity, and precision values of 100%. Full article

Measuring the similarity between two trajectories is fundamental and essential for the similarity-based remaining useful life (RUL) prediction. Most previous methods do not adequately account for the epistemic uncertainty caused by asynchronous sampling, while others have strong assumption constraints, such as limiting the positional deviation of sampling points to a fixed threshold, which biases the results considerably. To address the issue, an uncertain ellipse model based on the uncertain theory is proposed to model the location of sampling points as an observation drawn from an uncertain distribution. Based on this, we propose a novel and effective similarity measure metric for any two degradation trajectories. Then, the Stacked Denoising Autoencoder (SDA) model is proposed for RUL prediction, in which the models can be first trained on the most similar degradation data and then fine-tuned by the target dataset. Experimental results show that the predictive performance of the new method is superior to prior methods based on edit distance on real sequence (EDR), longest common subsequence (LCSS), or dynamic time warping (DTW) and is more robust at different sampling rates. Full article

Internet of Things (IoT)-assisted skin cancer recognition integrates several connected devices and sensors for supporting the primary analysis and monitoring of skin conditions. A preliminary analysis of skin cancer images is extremely difficult because of factors such as distinct sizes and shapes of lesions, differences in color illumination, and light reflections on the skin surface. In recent times, IoT-based skin cancer recognition utilizing deep learning (DL) has been used for enhancing the early analysis and monitoring of skin cancer. This article presents an optimal deep learning-based skin cancer detection and classification (ODL-SCDC) methodology in the IoT environment. The goal of the ODL-SCDC technique is to exploit metaheuristic-based hyperparameter selection approaches with a DL model for skin cancer classification. The ODL-SCDC methodology involves an arithmetic optimization algorithm (AOA) with the EfficientNet model for feature extraction. For skin cancer detection, a stacked denoising autoencoder (SDAE) classification model has been used. Lastly, the dragonfly algorithm (DFA) is utilized for the optimal hyperparameter selection of the SDAE algorithm. The simulation validation of the ODL-SCDC methodology has been tested on a benchmark ISIC skin lesion database. The extensive outcomes reported a better solution of the ODL-SCDC methodology compared with other models, with a maximum sensitivity of 97.74%, specificity of 99.71%, and accuracy of 99.55%. The proposed model can assist medical professionals, specifically dermatologists and potentially other healthcare practitioners, in the skin cancer diagnosis process. Full article

In this study, a novel intelligent inverter fault diagnosis approach based on a stacked denoising autoencoder–generative adversarial network–long short-term memory (SDAE-GAN-LSTM) under an imbalanced sample is proposed for a three-phase permanent-magnet synchronous motor (PMSM) drive system. The proposed method can address the problem of unbalanced fault data samples and improve the accuracy of fault classification. Concretely speaking, firstly, the stacked denoising autoencoder (SDAE) is pre-trained to obtain the optimum decoder network. Afterward, a new generator of generative adversarial networks (GANs) is designed to generate high-quality samples by migrating the pre-trained optimal decoder network to the hidden layer and output layer of the generator of GANs. Additionally, a new model of long short-term memory (LSTM) based on the second discriminator of the GANs is presented for fault diagnosis. The generator of GANs is cross-trained using the reconstruction error gained by SDAE and the fault diagnosis error obtained by LSTM, resulting in the generation of high-quality samples for fault discrimination. Simulation and experimental results demonstrate the effectiveness of the proposed fault diagnosis approach, and the average fault identification accuracy reaches 98.63%. Full article

In recent years, extreme weather has occurred frequently, and the risk of heavy rainfall and flooding faced by the people has risen. It is therefore an urgent requirement to carry out applied research on heavy rainfall and flooding risk assessment. We took Henan Province, where a major flood disaster occurred in 2021, as an example to analyze the impact factors of urban flooding and conduct a risk assessment. Indicators were first selected from population, housing, and the economy, and correlation analysis was used to optimize the indicator system. Then, a deep clustering network model based on a stacked denoising autoencoder (SDAE) was constructed, the feature information implied in the disaster indicators was abstracted into potential features through the coding and decoding of the network, and a small number of potential features were used to express the complex relationship between the disaster indicators. The results of the study show that the high-risk areas of flood damage in Henan Province in 2021 account for 2.3%, the medium-risk areas account for 9.4%, and the low-risk areas account for 80.3%. These evaluation results are in line with the actual situation in Henan Province, and the division of the grade in some areas is more reasonable compared with the entropy weighting method, which is a commonly used method of disaster assessment. The new model does not need to calculate weights to cope with changes in indicators and disaster conditions. The research results can provide scientific reference for urban flood risk management, disaster prevention and mitigation, and regional planning. Full article

Community question answering (CQA), with its flexible user interaction characteristics, is gradually becoming a new knowledge-sharing platform that allows people to acquire knowledge and share experiences. The number of questions is rapidly increasing with the open registration of communities and the massive influx of users, which makes it impossible to match many questions to suitable question answering experts (noted as experts) in a timely manner. Therefore, it is of great importance to perform expert recommendation in CQA. Existing expert recommendation algorithms only use data from a single platform, which is not ideal for new CQA platforms with sparse historical interaction and a small number of questions and users. Considering that many mature CQA platforms (source platforms) have rich historical interaction data and a large amount of questions and experts, this paper will fully mine the information and transfer it to new platforms with sparse data (target platform), which can effectively alleviate the data sparsity problem. However, the feature composition of questions and experts in different platforms is inconsistent, so the data from the source platform cannot be directly transferred for training in the target platform. Therefore, this paper proposes feature-alignment-based cross-platform question answering expert recommendation (FA-CPQAER), which can align expert and question features while transferring data. First, we use the rating predictor composed by the BP network for expert recommendation within the domains, and then the feature matching of questions and experts between two domains by similarity calculation is achieved for the purpose of using the information in the source platform to assist expert recommendation in the target platform. Meanwhile, we train a stacked denoising autoencoder (SDAE) in both domains, which can map user and question features to the same dimension and align the data distributions. Extensive experiments are conducted on two real CQA datasets, Toutiao and Zhihu datasets, and the results show that compared to the other advanced expert recommendation algorithms, this paper’s method achieves better results in the evaluation metrics of MAE, RMSE, Accuracy, and Recall, which fully demonstrates the effectiveness of the method in this paper to solve the data sparsity problem in expert recommendation. Full article

Hepatocellular Carcinoma (HCC) is the most frequent malignant liver tumor and the third cause of cancer-related deaths worldwide. For many years, the golden standard for HCC diagnosis has been the needle biopsy, which is invasive and carries risks. Computerized methods are due to achieve a noninvasive, accurate HCC detection process based on medical images. We developed image analysis and recognition methods to perform automatic and computer-aided diagnosis of HCC. Conventional approaches that combined advanced texture analysis, mainly based on Generalized Co-occurrence Matrices (GCM) with traditional classifiers, as well as deep learning approaches based on Convolutional Neural Networks (CNN) and Stacked Denoising Autoencoders (SAE), were involved in our research. The best accuracy of 91% was achieved for B-mode ultrasound images through CNN by our research group. In this work, we combined the classical approaches with CNN techniques, within B-mode ultrasound images. The combination was performed at the classifier level. The CNN features obtained at the output of various convolution layers were combined with powerful textural features, then supervised classifiers were employed. The experiments were conducted on two datasets, acquired with different ultrasound machines. The best performance, above 98%, overpassed our previous results, as well as representative state-of-the-art results. Full article

Malfunctions in relay protection devices are predominantly caused by current transformer (CT) saturation which produces distortion in current measurements and disturbances in power system protection. The development of deep learning in power system protection is on the rise recently because of its robustness. This study presents a CT saturation detection where the secondary current becomes distorted. The proposed scheme offers a wide range of saturation detection and consists of a moving-window technique and stacked denoising autoencoders. Moreover, Bayesian optimization was used to minimize the difficulty of determining neural network structure for the proposed approach. The performance of the algorithm was evaluated for a-g faults on 154 kV and 345 kV overhead transmission line in South Korea. The waveform variation has been generated by PSCAD for different scenarios that heavily influence CT saturation. Moreover, a comparative analysis with other methods demonstrated the superiority of the proposed DNN method. With the proposed algorithm to detect CT saturation, it significantly yielded high accuracy and precision for CT saturation detection which were approximately 99.71% and 99.32%, respectively. Full article

Artificial intelligence (AI), in particular deep learning, has proven to be efficient in medical diagnosis. This paper introduces a new hybrid deep learning model for pneumonia diagnosis based on chest CT scans. At the core of the model, a Gaussian mixture is combined with the expectation-maximization algorithm (EMGMM) to extract the regions of interest (ROI), while a convolutional denoising autoencoder (DAE) and deep restricted Boltzmann machine (DRBM) are combined for the classification. In order to prevent the model from learning trivial solutions, stochastic noises were added as an input to the unsupervised learning phase. The dataset used in this work is a publicly available dataset of chest X-rays for pneumonia on the Kaggle website; it contains 5856 images with 1583 normal cases and 4273 pneumonia cases, with an imbalance ratio (IR) of 0.46. Several operations including zooming, flipping, shifting and rotation were used in the augmentation phase to balance the data distribution across the different classes, which led to enhancing the IR value to 0.028. The computational analysis of the results show that the proposed model is promising as it provides an average accuracy value of 98.63%, sensitivity value of 96.5%, and specificity value of 94.8%. Full article

Indoor positioning is the basic requirement of future positioning services, and high-precision, low-cost indoor positioning algorithms are the key technology to achieve this goal. Different from outdoor maps, indoor data has the characteristic of uneven distribution and close correlation. In areas with low data density, in order to achieve a high-precision positioning effect, the positioning time will be correspondingly longer, but this is not necessary. The instability of WiFi leads to the introduction of noise when collecting data, which reduces the overall performance of the positioning system, so denoising is very necessary. For the above problems, a positioning system using the DBSCAN algorithm to segment regions and realize regionalized positioning is proposed. DBSCAN algorithm not only divides the dataset into core points and edge points, but also divides part of the data into noise points to achieve the effect of denoising. In the core part, the dimensionality of the data is reduced by using stacking auto-encoders (SAE), and the localization task is accomplished by using a deep neural network (DNN) with an adaptive learning rate. At the edge points, the random forest (RF) algorithm is used to complete the localization task. Finally, the proposed architecture is verified on the UJIIndoorLoc dataset. The experimental results show that our positioning accuracy does not exceed 1.5 m with a probability of less than 87.2% at the edge point, and the time is only 32 ms; the positioning accuracy does not exceed 1.5 m with a probability of less than 98.8% at the core point. Compared with indoor positioning algorithms such as multi-layer perceptron and K Nearest Neighbors (KNN), good results have been achieved. Full article