Search

Accurate wind power prediction is critical for efficient grid management and the integration of renewable energy sources into the power grid. This study presents an effective deep-learning approach that improves short-term wind power forecasting accuracy. The method incorporates a Variational Autoencoder (VAE) with a self-attention mechanism applied in both the encoder and decoder. This empowers the model to leverage VAE's strengths in time-series modeling and nonlinear approximation while focusing on the most relevant features within the wind power data. The effectiveness of this approach is evaluated through a comprehensive comparison with eight established deep learning methods, including Recurrent Neural Networks (RNNs), Long Short-Term Memory (LSTM) networks, Bidirectional LSTMs (BiLSTMs), Convolutional LSTMs (ConvLSTMs), Gated Recurrent Units (GRUs), Stacked Autoencoders (SAEs), Restricted Boltzmann Machines (RBMs), and vanilla VAEs. Real-world data from five wind turbines in France and Turkey is used for the evaluation. Five statistical metrics are employed to quantitatively assess the prediction performance of each method. The results indicate that the SA-VAE model consistently outperformed other models, achieving the highest average R2 value of 0.992, demonstrating its superior predictive capability compared to existing techniques.

Precise traffic flow prediction is a central component of advancing intelligent transportation systems, providing essential insights for optimizing traffic management, reducing travel times, and alleviating congestion. This study introduces an efficient deep learning approach that synergistically integrates the benefits of wavelet-based denoising and Recurrent Neural Networks (RNNs). This integrated methodology is introduced to effectively capture the inherent nonlinearity and temporal dependencies in time series traffic data. Specifically, Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) are introduced to address the challenges associated with accurately forecasting traffic flow. To enhance prediction quality, traffic flow data is preprocessed using exponential smoothing and wavelet-based filtering as denoising filters, effectively eliminating outliers. The effectiveness of the proposed techniques is evaluated using traffic measurements collected from diverse highway locations across California, including the Old Bayshore highway, situated south of Interstate 880 (I880), and the Ashby Ave highway, positioned west of Interstate 80 (I80) in the San Francisco Bay Area. The results obtained through integrating both architectures, including LSTM and GRU, within the wavelet transform-based filter demonstrate an enhancement in forecasting performance. Specifically, LSTM with wavelet-based denoising using Symlet and Haar wavelets achieved high prediction performance with an average R2 of 0.982 and 0.9811, respectively.

Drunk driving poses a significant threat to road safety, necessitating effective detection methods to enhance preventive measures and ensure the well-being of road users. Recognizing the critical importance of identifying drunk driving incidents for public safety, this paper introduces an effective semi-supervised anomaly detection strategy. The proposed strategy integrates three key elements: Independent Component Analysis (ICA), Kantorovitch distance (KD), and double Exponentially Weighted Moving Average (DEWMA). ICA is used to handle non-gaussian and multivariate data, while KD is used to measure the dissimilarity between normal and abnormal events based on ICA features. The DEWMA is applied to KD charting statistics to detect changes in data and uses a nonparametric threshold to improve sensitivity. The primary advantage of this approach is its ability to perform anomaly detection without requiring labeled data. The study also used XGBoost for the later calculation of the SHAP (SHapley Additive exPlanations) values to identify the most important variables for detecting drunk driving behavior. The approach was evaluated using publicly available data from gas and temperature sensors, as well as digital cameras. The results showed that the proposed approach achieved an F1-score of 98% in detecting the driver’s drunk status, outperforming conventional PCA-based and ICA-based methods.

Desertification detection is a crucial step to improve the management of affected areas and aid in mitigating the negative impacts of desertification. This study proposes a semi-supervised approach that uses Landsat imagery and radiometric data to detect desertification. The approach involves extracting radiometric data, which is used as an indicator to identify the thematic type and desertification evolution over time. Four anomaly detection techniques, including One-Class Support Vector Machine (OCSVM), Isolation Forest, Elliptic Envelope, and Local Outlier Factor, are trained on radiometric data from non-desertified regions to detect abnormal events related to desertification. These semi-supervised techniques use unlabeled data during training and only require desertification-free data, making them practical. The study was conducted in the arid region around Biskra, Algeria, which is a well-known area strongly affected by desertification. The OCSVM method achieved the highest detection accuracy of 95.40% in comparison to other methods and studies. Furthermore, to enhance result reliability, the Bootstrap technique was employed to generate 95% confidence intervals for five evaluation metrics.

Accurately predicting key features in WWTPs is essential for optimizing plant performance and minimizing operational costs. This study assesses the potential of various machine learning models for predicting the inflow to anoxic sludge reactors. Firstly, it conducts a comprehensive evaluation of diverse machine learning models, including k-Nearest Neighbors (kNN), Random Forest (RF), XGBoost, CatBoost, LightGBM, and Decision Tree Regression (DTR), for predicting the flow into the Anoxic section under various weather conditions (dry, rainy, and stormy). Secondly, the study introduces parsimonious models guided by variable importance from the XGBoost algorithm. Furthermore, the study employs SHAP (SHapley Additive exPlanations) to elucidate model predictions, providing insights into the contribution of each feature. Data from the COST Benchmark Simulation Model (BSM1) is used to verify the investigated models' effectiveness. Each dataset consists of 14 days of influent data at 15-minute intervals, with 80% of the data used for model training. Results show that ensemble learning methods, particularly CatBoost and XGBoost, demonstrate satisfactory predictive results for Anoxic section flow in the presence of increased variability under rainy and stormy conditions. Notably, the CatBoost and XGBoost models achieve average Mean Absolute Percentage Error values of 1.33% and 1.59%, outperforming the other methods.

The accurate modeling and forecasting of Battery State of Health (SOH) is crucial for ensuring reliable performance and longevity of lithium-ion batteries. This article introduces a data-driven approach for SOH prediction using Gaussian Process Regression (GPR), selected for its ability to model complex data relationships and capture prediction uncertainty without relying on future load information. Recognizing the effect of reversible capacity recoveries on prediction accuracy, this work employs the ISEA-RWTH 48 Li-ion Cells dataset, deliberately devoid of such recoveries prior to training the GPR model. The GPR model was evaluated and compared with Support Vector Regression (SVR) using the publicly available dataset. First and second End of Life (EOL) scenarios were considered, relevant to primary and secondary battery applications. The results demonstrated the GPR model's superior performance. Particularly, mid-life and late-life predictions displayed better accuracy with GPR, showcasing higher R2 values and lower MAPE values (e.g., mid-life prediction: GPR's average R2 = 0.99, SVR's = 0.9789; GPR's average MAPE = 0.1916, SVR's = 1.3028). Moreover, GPR exhibited the ability to quantify uncertainty in capacity degradation and forecast first and second EOL instances effectively (e.g., mid-life predictions had 1.7 cycle error at 1st EOL and 8.9 cycle error at 2nd EOL). The research also offers valuable insights into the application of machine learning methods for predicting the health degradation of lithium-ion batteries.

Solar photovoltaic (PV) systems have become a vital renewable energy source, witnessing rapid global demand. Nevertheless, these systems are susceptible to faults and anomalies that can deteriorate performance and yield significant consequences. Hence, this paper is dedicated to reviewing recent advancements in monitoring, modeling, and fault detection methods for PV systems. It encompasses diverse PV system types, including grid-connected, stand-alone, and hybrid configurations, and delves into the latest data acquisition and monitoring techniques. The review also discusses various performance modeling approaches, including empirical, analytical, and numerical models, highlighting the significance of model validation and calibration. Furthermore, it provides a comprehensive analysis of model-based fault detection techniques. Overall, this paper underscores the pivotal role of fault detection in PV systems and offers a thorough comprehension of available techniques vital for enhancing system management and maintenance.

The evolution of smart grids has led to technological advances and a demand for more efficient and sustainable energy systems. However, the deployment of communication systems in smart grids has increased the threat of cyberattacks, which can result in power outages and disruptions. This paper presents a semi-supervised hybrid deep learning model that combines a Gated Recurrent Unit (GRU)-based Stacked Autoencoder (AE-GRU) with anomaly detection algorithms, including Isolation Forest, Local Outlier Factor, One-Class SVM, and Elliptical Envelope. Using GRU units in both the encoder and decoder sides of the stacked autoencoder enables the effective capture of temporal patterns and dependencies, facilitating dimensionality reduction, feature extraction, and accurate reconstruction for enhanced anomaly detection in smart grids. The proposed approach utilizes unlabeled data to monitor network traffic and identify suspicious data flow. Specifically, the AE-GRU is performed for data reduction and extracting relevant features, and then the anomaly algorithms are applied to reveal potential cyberattacks. The proposed framework is evaluated using the widely adopted IEC 60870-5-104 traffic dataset. The experimental results demonstrate that the proposed approach outperforms standalone algorithms, with the AE-GRU-based LOF method achieving the highest detection rate. Thus, the proposed approach can potentially enhance the cybersecurity in smart grids by accurately detecting and preventing cyberattacks.

Abstract Precise and efficient ozone ( $$\hbox {O}_{3}$$ O 3 ) concentration prediction is crucial for weather monitoring and environmental policymaking due to the harmful effects of high $$\hbox {O}_{3}$$ O 3 pollution levels on human health and ecosystems. However, the complexity of $$\hbox {O}_{3}$$ O 3 formation mechanisms in the troposphere presents a significant challenge in modeling $$\hbox {O}_{3}$$ O 3 accurately and quickly, especially in the absence of a process model. Data-driven machine-learning techniques have demonstrated promising performance in modeling air pollution, mainly when a process model is unavailable. This study evaluates the predictive performance of nineteen machine learning models for ozone pollution prediction. Specifically, we assess how incorporating features using Random Forest affects $$\hbox {O}_{3}$$ O 3 concentration prediction and investigate using time-lagged measurements to improve prediction accuracy. Air pollution and meteorological data collected at King Abdullah University of Science and Technology are used. Results show that dynamic models using time-lagged data outperform static and reduced machine learning models. Incorporating time-lagged data improves the accuracy of machine learning models by 300% and 200%, respectively, compared to static and reduced models, under RMSE metrics. And importantly, the best dynamic model with time-lagged information only requires 0.01 s, indicating its practical use. The Diebold-Mariano Test, a statistical test used to compare the forecasting accuracy of models, is also conducted.

Wastewater treatment plants (WWTPs) are energy-intensive facilities that play a critical role in meeting stringent effluent quality regulations. Accurate prediction of energy consumption in WWTPs is essential for cost savings, process optimization, regulatory compliance, and reducing carbon footprint. This paper introduces an efficient approach for predicting energy consumption in WWTPs, leveraging deep learning models, data augmentation, and feature selection. Specifically, Spline Cubic interpolation enriches the dataset, while the Random Forest model identifies important features. The study investigates the impact of lagged data to capture temporal dependencies. Comparative analysis of five deep learning models on original and augmented datasets from Melbourne WWTP demonstrates substantial performance improvement with augmented data. Incorporating lagged energy consumption data further enhances accuracy, providing valuable insights for effective energy management. Notably, the Long Short-Term Memory (LSTM) and Bidirectional Gated Recurrent Unit (BiGRU) models achieve Mean Absolute Percentage Error (MAPE) values of 1.36% and 1.436%, outperforming state-of-the-art methods.

Effective fault detection in chemical processes is of utmost importance to ensure operational safety, minimize environmental impact, and optimize production efficiency. To enhance the monitoring of chemical processes under noisy conditions, an innovative statistical approach has been introduced in this study. The proposed approach, called Multiscale Principal Component Analysis (PCA), combines the dimensionality reduction capabilities of PCA with the noise reduction capabilities of wavelet-based filtering. The integrated approach focuses on extracting features from the multiscale representation, balancing the need to retain important process information while minimizing the impact of noise. For fault detection, the Kantorovich distance (KD)-driven monitoring scheme is employed based on features extracted from Multiscale PCA to efficiently detect anomalies in multivariate data. Moreover, a nonparametric decision threshold is employed through kernel density estimation to enhance the flexibility of the proposed approach. The detection performance of the proposed approach is investigated using data collected from distillation columns and continuously stirred tank reactors (CSTRs) under various noisy conditions. Different types of faults, including bias, intermittent, and drift faults, are considered. The results reveal the superior performance of the proposed multiscale PCA-KD based approach compared to conventional PCA and multiscale PCA-based monitoring methods.

Favipiravir (FP) and ebselen (EB) belong to a diverse class of antiviral drugs known for their significant efficacy in treating various viral infections. Utilizing molecular dynamics (MD) simulations, machine learning, and van der Waals density functional theory, we accurately elucidate the binding properties of these antiviral drugs on a phosphorene single-layer. To further investigate these characteristics, this study employs four distinct machine learning models—Random Forest, Gradient Boosting, XGBoost, and CatBoost. The Hamiltonian of antiviral molecules within a monolayer of phosphorene is appropriately trained. The key aspect of utilizing machine learning (ML) in drug design revolves around training models that are efficient and precise in approximating density functional theory (DFT). Furthermore, the study employs SHAP (SHapley Additive exPlanations) to elucidate model predictions, providing insights into the contribution of each feature. To explore the interaction characteristics and thermodynamic properties of the hybrid drug, we employ molecular dynamics and DFT calculations in a vacuum interface. Our findings suggest that this functionalized 2D complex exhibits robust thermostability, indicating its potential as an effective and enabled entity. The observed variations in free energy at different surface charges and temperatures suggest the adsorption potential of FP and EB molecules from the surrounding environment.

Photovoltaic (PV) systems, as critical components of the power grid, have become increasingly reliant on standard Information Technology (IT) computing and network infrastructure for their operation and maintenance. However, this dependency exposes PV systems to heightened vulnerabilities and the risk of cyber-attacks. In recent times, the number of reported cyber-attacks targeting PV systems has increased significantly. This paper provides an overview of the cybersecurity challenges faced by PV systems, emphasizing their susceptibility to anomalies and cyber threats. It highlights the urgency of implementing robust cybersecurity measures to protect the integrity and reliability of PV installations. By understanding and addressing these challenges, stakeholders can ensure the resilience and secure integration of PV systems within the power grid infrastructure.

Fault detection is crucial in maintaining reliability, safety, and consistent product quality in chemical engineering processes. Accurate fault detection allows for identifying anomalies, signaling deviations from the system’s nominal behavior, ensuring the system operates within desired performance parameters, and minimizing potential losses. This paper presents a novel semi-supervised data-based monitoring technique for fault detection in multivariate processes. To this end, the proposed approach merges the capabilities of Principal Component Analysis (PCA) for dimensionality reduction and feature extraction with the Kolmogorov–Smirnov (KS)-based scheme for fault detection. The KS indicator is computed between the two distributions in a moving window of fixed length, allowing it to capture sensitive details that enhance the detection of faults. Moreover, no labeling is required when using this fault detection approach, making it flexible in practice. The performance of the proposed PCA–KS strategy is assessed for different sensor faults on benchmark processes, specifically the Plug Flow Reactor (PFR) process and the benchmark Tennessee Eastman (TE) process. Different sensor faults, including bias, intermittent, and aging faults, are considered in this study to evaluate the proposed fault detection scheme. The results demonstrate that the proposed approach surpasses traditional PCA-based methods. Specifically, when applied to PFR data, it achieves a high average detection rate of 98.31% and a low false alarm rate of 0.25%. Similarly, when applied to the TE process, it provides a good average detection rate of 97.27% and a false alarm rate of 6.32%. These results underscore the efficacy of the proposed PCA–KS approach in enhancing the fault detection of high-dimensional processes.

Abstract This study aims to develop an assumption-free data-driven model to accurately forecast COVID-19 spread. Towards this end, we firstly employed Bayesian optimization to tune the Gaussian process regression (GPR) hyperparameters to develop an efficient GPR-based model for forecasting the recovered and confirmed COVID-19 cases in two highly impacted countries, India and Brazil. However, machine learning models do not consider the time dependency in the COVID-19 data series. Here, dynamic information has been taken into account to alleviate this limitation by introducing lagged measurements in constructing the investigated machine learning models. Additionally, we assessed the contribution of the incorporated features to the COVID-19 prediction using the Random Forest algorithm. Results reveal that significant improvement can be obtained using the proposed dynamic machine learning models. In addition, the results highlighted the superior performance of the dynamic GPR compared to the other models (i.e., Support vector regression, Boosted trees, Bagged trees, Decision tree, Random Forest, and XGBoost) by achieving an averaged mean absolute percentage error of around 0.1%. Finally, we provided the confidence level of the predicted results based on the dynamic GPR model and showed that the predictions are within the 95% confidence interval. This study presents a promising shallow and simple approach for predicting COVID-19 spread.

Solar Photovoltaic (PV) systems represent key and transformative technology at the forefront of the global shift towards sustainable energy solutions [...]

Efficient detection of sensor faults in wind turbines is essential to ensure the reliable operation and performance of these renewable energy systems. This paper presents a novel semi-supervised data-based monitoring technique for fault detection in wind turbines using SCADA (supervisory control and data acquisition) data. Unlike supervised methods, the proposed approach does not require labeled data, making it cost-effective and practical for wind turbine monitoring. The technique builds upon the Independent Component Analysis (ICA) approach, effectively capturing non-Gaussian features. Specifically, a dynamic ICA (DICA) model is employed to account for the temporal dynamics and dependencies in the observed signals affected by sensor faults. The fault detection process integrates fault indicators based on I2d, I2e, and squared prediction error (SPE), enabling the identification of different types of sensor faults. The fault indicators are combined with a Double Exponential Weighted Moving Average (DEWMA) chart, known for its superior performance in detecting faults with small magnitudes. Additionally, the approach incorporates kernel density estimation to establish nonparametric thresholds, increasing flexibility and adaptability to different data types. This study considers various types of sensor faults, including bias sensor faults, precision degradation faults, and freezing sensor faults, for evaluation. The results demonstrate that the proposed approach outperforms PCA and traditional ICA-based methods. It achieves a high detection rate, accurately identifying faults while reducing false alarms. It could be a promising technique for proactive maintenance, optimizing the performance and reliability of wind turbine systems.

This study presents an effective data-driven anomaly detection scheme for drunk driving detection. Specifically, the proposed anomaly detection approach amalgamates the desirable features of the t-distributed stochastic neighbor embedding (t-SNE) as a feature extractor with the Isolation Forest (iF) scheme to detect drivers’ drunkenness status. We used the t-SNE model to exploit its capacity in reducing the dimensionality of nonlinear data by preserving the local and global structures of the input data in the feature space to obtain good detection. At the same time, the iF scheme is an effective and unsupervised tree-based approach to achieving good detection of anomalies in multivariate data. This approach only employs normal events data to train the detection model, making them more attractive for detecting drunk drivers in practice. To verify the detection capacity of the proposed t-SNE-iF approach in reliably detecting drivers with excess alcohol, we used publically available data collected using a gas sensor, temperature sensor, and a digital camera. The overall detection system proved a high detection performance with AUC around 95%, demonstrating the proposed approach’s robustness and reliability. Furthermore, compared to the Principal Component Analysis (PCA), Incremental PCA (IPCA), Independent component analysis (ICA), Kernel PCA (kPCA), and Multi-dimensional scaling (MDS)-based iForest, EE, and LOF detection schemes, the proposed t-SNE-based iF scheme offers superior detection performance of drunk driver status.

Accurate modulation identification of the received signals is undoubtedly a central component in multiple-input multiple-output (MIMO) communication systems, facilitating the demodulation task. This study presents a flexible and semi-supervised deep learning-driven strategy for automatic modulation identification. To this end, the multiclass classification problem is treated as multiple binary discrimination problems to address modulation identification challenges. Here, we merge the features extraction ability of the Generative Adversarial Network (GAN) model and the semi-supervised anomaly detection scheme, the one-class Support Vector Machine (1SVM). Essentially, a single GAN-based 1SVM detector is trained using training data of each class, with the samples of that class as inlier and all other samples as anomalies (i.e., one-vs.-rest). The 1SVM is trained using the features learned by the GAN model. A dataset consisting of three digital modulations (i.e., BFSK, CPFSK, and PAM4) and three analog modulations (i.e., AM-DSB, AM-SSB, and WB-FM), widely used in wireless communications systems, is employed to demonstrate the performance of the considered deep learning-based methods. Compared to Restricted Boltzmann Machine (RBM) and Deep Belief Network (DBN)-based 1SVM, the conventional GAN, DBN, and RBM with softmax layer as discriminator layer, the proposed GAN-based 1SVM detector offers superior discrimination performance of modulation types by achieving an averaged accuracy of 0.951 and F1-Score of 0.954. Results also showed that the GAN-1SVM detector dominates the state-of-the-art modulation classification techniques.

Wastewater treatment plants (WWTPs) are energy-intensive facilities that fulfill stringent effluent quality norms. Energy consumption prediction in WWTPs is crucial for cost savings, process optimization, compliance with regulations, and reducing the carbon footprint. This paper evaluates and compares a set of 23 candidate machine-learning models to predict WWTP energy consumption using actual data from the Melbourne WWTP. To this end, Bayesian optimization has been applied to calibrate the investigated machine learning models. Random Forest and XGBoost (eXtreme Gradient Boosting) were applied to assess how the incorporated features influenced the energy consumption prediction. In addition, this study investigated the consideration of information from past data in improving prediction accuracy by incorporating time-lagged measurements. Results showed that the dynamic models using time-lagged data outperformed the static and reduced machine learning models. The study shows that including lagged measurements in the model improves prediction accuracy, and the results indicate that the dynamic K-nearest neighbors model dominates state-of-the-art methods by reaching promising energy consumption predictions.

This study introduces a new method for identifying COVID-19 infections using blood test data as part of an anomaly detection problem by combining the kernel principal component analysis (KPCA) and one-class support vector machine (OCSVM). This approach aims to differentiate healthy individuals from those infected with COVID-19 using blood test samples. The KPCA model is used to identify nonlinear patterns in the data, and the OCSVM is used to detect abnormal features. This approach is semi-supervised as it uses unlabeled data during training and only requires data from healthy cases. The method’s performance was tested using two sets of blood test samples from hospitals in Brazil and Italy. Compared to other semi-supervised models, such as KPCA-based isolation forest (iForest), local outlier factor (LOF), elliptical envelope (EE) schemes, independent component analysis (ICA), and PCA-based OCSVM, the proposed KPCA-OSVM approach achieved enhanced discrimination performance for detecting potential COVID-19 infections. For the two COVID-19 blood test datasets that were considered, the proposed approach attained an AUC (area under the receiver operating characteristic curve) of 0.99, indicating a high accuracy level in distinguishing between positive and negative samples based on the test results. The study suggests that this approach is a promising solution for detecting COVID-19 infections without labeled data.

Favipiravir (FP) and Ebselen (EB) belong to a broad range of antiviral drugs that have shown active potential as medications against many viruses. Employing molecular dynamics simulations and machine learning (ML) combined with van der Waals density functional theory, we have uncovered the binding characteristics of these two antiviral drugs on a phosphorene nanocarrier. Herein, by using four different machine learning models (i.e., Bagged Trees, Gaussian Process Regression (GPR), Support Vector Regression (SVR), and Regression Trees (RT)), the Hamiltonian and the interaction energy of antiviral molecules in a phosphorene monolayer are trained in an appropriate way. However, training efficient and accurate models for approximating the density functional theory (DFT) is the final step in using ML to aid in the design of new drugs. To improve the prediction accuracy, the Bayesian optimization approach has been employed to optimize the GPR, SVR, RT, and BT models. Results revealed that the GPR model obtained superior prediction performance with an R2 of 0.9649, indicating that it can explain 96.49% of the data’s variability. Then, by means of DFT calculations, we examine the interaction characteristics and thermodynamic properties in a vacuum and a continuum solvent interface. These results illustrate that the hybrid drug is an enabled, functionalized 2D complex with vigorous thermostability. The change in Gibbs free energy at different surface charges and temperatures implies that the FP and EB molecules are allowed to adsorb from the gas phase onto the 2D monolayer at different pH conditions and high temperatures. The results reveal a valuable antiviral drug therapy loaded by 2D biomaterials that may possibly open a new way of auto-treating different diseases, such as SARS-CoV, in primary terms.

The accurate land cover change detection is critical to improve the landscape dynamics analysis and mitigate desertification problems efficiently. Desertification detection is a challenging problem because of the high degree of similarity between some desertification cases and like-desertification phenomena, such as deforestation. This article provides an effective approach to detect deserted regions based on Landsat imagery and variational autoencoder (VAE). The VAE model, as a deep learning-based model, has gained special attention in features extraction and modeling due to its distribution-free assumptions and superior nonlinear approximation. Here, a VAE approach is applied to spectral signatures for detecting pixels affected by the land cover change. The considered features are extracted from multitemporal images and include multispectral information, and no prior image segmentation is required. The proposed method was evaluated on the publicly available remote sensing data using multitemporal Landsat optical images taken from the freely available Landsat program. The arid region around Biskra in Algeria is selected as a study area since it is well-known that desertification phenomena strongly influence this region. The VAE model was evaluated and compared with restricted Boltzmann machines, deep learning model, and binary clustering algorithms, including Agglomerative, BIRCH, expected maximization, k-mean clustering algorithms, and one-class support vector machine. The comparative results showed that the VAE consistently outperformed the other models for detecting changes to the land cover, mainly deserted regions. This study also showed that VAE outperformed the state-of-the-art algorithms.

Forecasting the different types of emergency department (ED) demands (patient flows) in hospital systems much aids ED managers in looking into various options to appropriately allocating the restricted resources available per patient attendance. Deep learning networks have recently gained great success in modeling time-dependent in time series data. Thus, this work advocates the use of deep learning-driven models for patient flows forecasting. Notably, we examine and compare seven deep learning models, Deep Belief Network (DBN), Restricted Boltzmann machines (RBM), Long Short Term Memory (LSTM), Gated recurrent unit (GRU), combined GRU and convolutional neural networks (CNN-GRU), LSTM-CNN, and Generative Adversarial Network based on Recurrent Neural Networks (GAN-RNN), to forecast patient flow in a hospital emergency department. We introduce a forecaster layer as output for each model to enable traffic flow forecasting. Patient flow data from different ED services, including biology, radiology, scanner, and echography, in Lille regional hospital in France, is used as a case study in assessing the considered forecasting models. Four metrics of effectiveness are adopted for evaluating and comparing the forecasting methods. The results show the promising performance of deep learning models for ED patient flow forecasting compared to shallow methods (i.e., ridge regression and support vector regression). In addition, the results highlighted the superior performance of the DBN compared to the other models by achieving an averaged mean absolute percentage error of around 4.097% and R2 of 0.973.

A monitoring mechanism is vital for detecting malicious attacks against cyber systems. Detecting denial of service (DOS) and distributed DOS (DDOS) is one of the most important security challenges facing network technologies. This paper introduces a reliable detection mechanism based on the continuous ranked probability score (CRPS) statistical metric and exponentially smoothing (ES) scheme for enabling efficient detection of DOS and DDOS attacks. In this regard, the CRPS is used to quantify the dissimilarity between a new observation and the distribution of normal traffic. The ES scheme, which is sensitive in detecting small changes, is applied to CRPS measurements for anomaly detection. Moreover, in CRPS-ES approach, a nonparametric decision threshold computed via kernel density estimation is used to suitably detect anomalies. Tests on three publically available datasets proclaim the efficiency of the proposed mechanism in detecting cyber-attacks.

Wind power is one of the most potential energies and the major available renewable energy sources. Precisely predicting wind power production is essential for the management and the integration of wind power in a smart grid. The goal of this study is to predict wind power production with sufficient accuracy based on various factors using ensemble learning-based methods that consider the time-dependent nature of the wind power measurements. Essentially, the ensemble learning methods combine multiple learners to obtain an enhanced prediction performance in comparison to conventional standalone learners. In addition, they reduce the overall prediction error and have the capacity to merge various models. At first, this paper investigates the prediction capability of the well-known ensemble approaches Boosted Trees, Random Forest, and Generalized Random Forest for wind power prediction. We compared the prediction performance of these ensemble models to two frequently used prediction methods: Gaussian process regression, and Support Vector Regression. Experimental measurements recorded every ten minutes from actual wind turbines located in France and Turkey are used to test the prediction efficiency of the studied models. Experimental results have shown that the ensemble methods can predict wind power production with high accuracy compared to the standalone models. Furthermore, the findings clearly reveal that the lagged variables contribute significantly to the ensemble models, and permits constructing more parsimonious models.

Accurately detecting Parkinson's disease (PD) at an early stage is certainly indispensable for slowing down its progress and providing patients the possibility of accessing to disease-modifying therapy. Towards this end, the premotor stage in PD should be carefully monitored. An innovative deep-learning technique is introduced to early uncover whether an individual is affected with PD or not based on premotor features. Specifically, to uncover PD at an early stage, several indicators have been considered in this study, including Rapid Eye Movement and olfactory loss, Cerebrospinal fluid data, and dopaminergic imaging markers. A comparison between the proposed deep learning model and twelve machine learning and ensemble learning methods based on relatively small data including 183 healthy individuals and 401 early PD patients shows the superior detection performance of the designed model, which achieves the highest accuracy, 96.45% on average. Besides detecting the PD, we also provide the feature importance on the PD detection process based on the Boosting method.

The accurate forecast of wastewater treatment plant (WWTP) key features can comprehend and predict the plant behavior to support process design and controls, improve system reliability, reduce operational costs, and endorse optimization of overall performances. Deep learning technologies as proven data-driven soft-sensors should be developed for WWTP applications to tackle the process of non-linearity and the dynamic nature of environmental data. This study adopts deep learning-based models as soft-sensors to forecast WWTP key features, such as influent flow, influent temperature, influent biochemical oxygen demand (BOD), effluent chloride, effluent BOD, and power consumption. We constructed six deep learning models derived from long short-term memory (LSTM) and gated recurrent unit (GRU), namely traditional LSTM and GRU, the exponentially smoothed LSTM, and the adaptive version of LSTM and smoothed LSTM. The employment of a smoothed LSTM technique is expected to reduce the outlier effect and to improve forecasting accuracy. Meanwhile, the usage of adaptive deep models will enhance the capabilities of the LSTM to quickly and accurately follow the trend of future data. We compared the performance of these models with Bi-directional LSTM (BiLSTM) and the seasonal decomposition using local regression. The historical records from a coastal municipal WWTP in Saudi Arabia are used to verify the investigated models' effectiveness. The proposed models provide promising forecasting results but require no assumptions on the data distributions. In terms of efficiency, GRU based models converge faster than LSTM based models. In terms of accuracy, the LSTM soft-sensor shows overall the optimal result for all key features followed by the exponentially-smoothed GRU and LSTM. By contrast, the adaptive models achieved the lowest forecasting performance compared to the other models. These findings will benefit practitioners to achieve data-driven WWTP management.

This paper introduces an unsupervised deep learning-driven scheme for mental tasks’ recognition using EEG signals. To this end, the Multichannel Wiener filter was first applied to EEG signals as an artifact removal algorithm to achieve robust recognition. Then, a quadratic time-frequency distribution (QTFD) was applied to extract effective time-frequency signal representation of the EEG signals and catch the EEG signals’ spectral variations over time to improve the recognition of mental tasks. The QTFD time-frequency features are employed as input for the proposed deep belief network (DBN)-driven Isolation Forest (iF) scheme to classify the EEG signals. Indeed, a single DBN-based iF detector is constructed based on each class’s training data, with the class’s samples as inliers and all other samples as anomalies (i.e., one-vs.-rest). The DBN is considered to learn pertinent information without assumptions on the data distribution, and the iF scheme is used for data discrimination. This approach is assessed using experimental data comprising five mental tasks from a publicly available database from the Graz University of Technology. Compared to the DBN-based Elliptical Envelope, Local Outlier Factor, and state-of-the-art EEG-based classification methods, the proposed DBN-based iF detector offers superior discrimination performance of mental tasks.

The capacity of photovoltaic solar power installations has been boosted last years by reaching a new record with 175 GWdc of newly installed solar power in 2021. To guarantee reliable performances of photovoltaic (PV) plants and maintain target requirements, faults have to be reliably detected and diagnosed. A method for an effective and reliable fault diagnosis of PV plants based on the behavioral model and performance analysis under the LabVIEW environment is presented in this paper. Specifically, the first phase of this study consists of the behavioral modeling of the PV array and the inverter in order to estimate the electricity production and analyze the performance of the 9.54 kWp Grid Connected PV System (GCPVS). Here, the results obtained from the empirical models were validated and calibrated by experimental data. Furthermore, a user interface for modeling and analyzing the performance of a PV system under LabVIEW has been designed. The second phase of this work is dedicated to the design of a simple and efficient diagnostic tool in order to detect and recognize faults occurring in the PV systems. Essentially, the residuals obtained using the parametric models are analyzed via the performance loss rates (PLR) of four electrical indicators (i.e., DC voltage, DC current, DC power, and AC power). To evaluate the proposed method, numerous environmental anomalies and electrical faults affecting the GCPVS were taken into account. Results demonstrated the satisfactory prediction performance of the considered empirical models to predict the considered variables, including DC current, DC power, and AC power with an R2 of 0.99. Moreover, the obtained results show that the detection and recognition of faults were successfully achieved.

Fault detection is a necessary component to perform ongoing monitoring of photovoltaic plants and helps in their safety, maintainability, and productivity with the desired performance. In this study, an innovative technique is introduced by amalgamating Latent Variable Regression (LVR) methods, namely Principal Component Regression (PCR) and Partial Least Square (PLS), and the Triple Exponentially Weighted Moving Average (TEWMA) statistical monitoring scheme. The TEWMA scheme is known for its sensitivity to uncovering changes of small magnitude. Nevertheless, TEWMA can only be utilized for monitoring single variables and ignoring the correlation among monitored variables. To alleviate this difficulty, the LVR methods (i.e., PCR and PLS) are used as residual generators. Then, the TEWMA is applied to the obtained residuals for fault detection purposes, where the detection threshold is computed via kernel density estimation to improve its performance and widen its applicability in practice. Real data with different fault scenarios from a 9.54 kW photovoltaic plant has been used to verify the efficiency of the proposed schemes. Results revealed the superior performance of the PLS-TEWMA chart compared to the PLS-TEWMA chart, particularly in detecting anomalies with small changes. Moreover, they have almost comparable performance for large anomalies.

Over the past few years, there has been a significant increase in the interest in and adoption of solar energy all over the world. However, despite ongoing efforts to protect photovoltaic (PV) plants, they are continuously exposed to numerous anomalies. If not detected accurately and in a timely manner, anomalies in PV plants may degrade the desired performance and result in severe consequences. Hence, developing effective and flexible methods capable of early detection of anomalies in PV plants is essential for enhancing their management. This paper proposes flexible data-driven techniques to accurately detect anomalies in the DC side of the PV plants. Essentially, this approach amalgamates the desirable characteristics of ensemble learning approaches (i.e., the boosting (BS) and bagging (BG)) and the sensitivity of the Double Exponentially Weighted Moving Average (DEWMA) chart. Here, we employ ensemble learning techniques to exploit their capability to enhance the modeling accuracy and the sensitivity of the DEWMA monitoring chart to uncover potential anomalies. In the ensemble models, the values of parameters are selected with the assistance of the Bayesian optimization algorithm. Here, BS and BG are adopted to obtain residuals, which are then monitored by the DEWMA chart. Kernel density estimation is utilized to define the decision thresholds of the proposed ensemble learning-based charts. The proposed monitoring schemes are illustrated via actual measurements from a 9.54 kW PV plant. Results showed the superior detection performance of the BS and BG-based DEWMA charts with non-parametric threshold in uncovering different types of anomalies, including circuit breaker faults, inverter disconnections, and short-circuit faults. In addition, the performance of the proposed schemes is compared to that of BG and BS-based DEWMA and EWMA charts with parametric thresholds.

To operate wastewater treatment plants (WWTPs) with optimized efficiency, influent conditions (ICs) as initial states of inflow fed to WWTPs were monitored to identify potential anomalies that would trigger adverse events or system crash. To employ voluminous measurements for data-driven decisions, the non-linear, non-Gaussian, non-stationary, auto-correlated, cross-correlated, hetero-skedastic, case-specific nature of multivariate environmental datasets must be considered. This research proposed kernel machine learning models, the kernel principal components analysis based one-class support vector machine (KPCA-OCSVM) with various kernels, to learn anomaly-free training set then classify the testing set. A seven-years multivariate ICs time series was introduced with exploratory analysis performed to reveal temporal behaviors and statistical properties. KPCA with polynomial kernels sufficiently output representative features, based on which OCSVM with Gaussian kernels sensitively and specifically identified anomalies in ICs that were previously omitted by WWTP operators. The proposed kernel algorithms surpassed previous linear PCA-based K-nearest-neighbors models, and improved outcomes with limited increase in computation cost. Without requiring linear, Gaussian, stationary, independent, and homo-skedastic qualities from data, the proposed flexible environmental data science approach could be transferred, rebuilt, and tuned conveniently for ICs from different WWTPs.

Aggregation is a vital behavior when performing complex tasks in most of the swarm systems, such as swarm robotics systems. In this paper, three new aggregation methods, namely the distance-angular, the distance-cosine, and the distance-Minkowski k-nearest neighbor (k-NN) have been introduced. These aggregation methods are mainly built on well-known metrics: the cosine, angular, and Minkowski distance functions, which are used here to compute distances among robots' neighbors. Relying on these methods, each robot identifies its k-nearest neighborhood set that will interact with. Then, in order to achieve the aggregation, the interactions sensing capabilities among the set members are modeled using a virtual viscoelastic mesh. Analysis of the results obtained from the ARGoS simulator shows a significant improvement in the swarm aggregation performance compared to the conventional distance-weighted k-NN aggregation method. Also, the aggregation performance of the methods is reported to be robust to partially faulty robots and accurate under noisy sensors.

Recently, increasing interest in managing pedestrian and bicycle flows has been demonstrated by cities and transportation professionals aiming to reach community goals related to health, safety, and the environment. Precise forecasting of pedestrian and bicycle traffic flow is crucial for identifying the potential use of bicycle and pedestrian infrastructure and improving bicyclists’ safety and comfort. Advances in sensory technology enable collecting massive traffic flow data, including road traffic, bicycle, and pedestrian traffic flow. This paper introduces a novel deep hybrid learning model with a fully guided-attention mechanism to improve bicycles and pedestrians’ traffic flow forecasting. Notably, the proposed approach extends the modeling capability of the Variational Autoencoder (VAE) by merging a long short-term memory (LSTM) model with the VAE’s decoder and using a self-attention mechanism at multi-stage of the VAE model (i.e., decoder and before data resampling). Specifically, LSTM improves the VAE decoder’s capacity in learning temporal dependencies, and the guided-attention units enable selecting relevant features based on the self-attention mechanism. This proposed deep hybrid learning model with a multi-stage guided-attention mechanism is called GAHD-VAE. Proposed methods were validated with traffic measurements from six publicly available pedestrian and bicycle traffic flow datasets. The proposed method provides promising forecasting results but requires no assumptions that the data are drawn from a given distribution. Results revealed that the GAHD-VAE methodology can efficiently enhance the traffic forecasting accuracy and achieved better performance than the deep learning methods VAE, LSTM, gated recurrent units (GRUs), bidirectional LSTM, bidirectional GRU, convolutional neural network (CNN), and convolutional LSTM (ConvLSTM), and four shallow methods, linear regression, lasso regression, ridge regression, and support vector regression.

Wind power represents a promising source of renewable energies. Precise forecasting of wind power generation is crucial to mitigate the challenges of balancing supply and demand in the smart grid. Nevertheless, the major difficulty in wind power is its high fluctuation and intermittent nature, making it challenging to forecast. This study aims to develop efficient data-driven models to accurately forecast wind power generation. Crucially, the main contributions of this work are listed in the following major elements. Firstly, we investigate the performance of enhanced machine learning models to forecast univariate wind power time-series data. Specifically, we employed Bayesian optimization (BO) to optimally tune hyperparameters of the Gaussian process regression (GPR), Support Vector Regression (SVR) with different kernels, and ensemble learning (ES) models (i.e., Boosted trees and Bagged trees) and investigated their forecasting performance. Secondly, dynamic information has been incorporated in their construction to further enhance the forecasting performance of the investigated models. Specifically, we introduce lagged measurements to enable capturing time evolution into the design of the considered models. Furthermore, more input variables (e.g., wind speed and wind direction) are used to further improve wind prediction performance. Actual measurements from three wind turbines in France, Turkey, and Kaggle are used to verify the efficiency of the considered models. The results reveal the benefit of considering lagged data and input variables to better forecast wind power. The results also showed that the optimized GPR and ensemble models outperformed the other machine learning models.

The accurate modeling and forecasting of the power output of photovoltaic (PV) systems are critical to efficiently managing their integration in smart grids, delivery, and storage. This paper intends to provide efficient short-term forecasting of solar power production using Variational AutoEncoder (VAE) model. Adopting the VAE-driven deep learning model is expected to improve forecasting accuracy because of its suitable performance in time-series modeling and flexible nonlinear approximation. Both single- and multi-step-ahead forecasts are investigated in this work. Data from two grid-connected plants (a 243 kW parking lot canopy array in the US and a 9 MW PV system in Algeria) are employed to show the investigated deep learning models’ performance. Specifically, the forecasting outputs of the proposed VAE-based forecasting method have been compared with seven deep learning methods, namely recurrent neural network, Long short-term memory (LSTM), Bidirectional LSTM, Convolutional LSTM network, Gated recurrent units, stacked autoencoder, and restricted Boltzmann machine, and two commonly used machine learning methods, namely logistic regression and support vector regression. The results of this investigation demonstrate the satisfying performance of deep learning techniques to forecast solar power and point out that the VAE consistently performed better than the other methods. Also, results confirmed the superior performance of deep learning models compared to the two considered baseline machine learning models.

Detecting faults induction machines is crucial for a safe operation of these machines. The aim of this paper is to present a statistical fault detection methodology for the detection of faults in three-phase wound rotor induction machines (WRIM). The proposed fault detection approach is based on the use of principal components analysis (PCA). However, conventional PCA-based detection indices, such as the T2 and the Q statistics, are not well suited to detect small faults because these indices only use information from the most recent available samples. Detection of small faults is one of the most crucial and challenging tasks in the area of fault detection and diagnosis. In this paper, a new statistical system monitoring strategy is proposed for detecting changes resulting from small shifts in several variables associated with WRIM. The proposed approach combines modeling using PCA modeling with the exponentially weighted moving average (EWMA) control scheme. In the proposed approach, EWMA control scheme is applied on the ignored principal components to detect the presence of faults. The performance of the proposed method is compared with those of the traditional PCA-based fault detection indices. The simulation results clearly show the effectiveness of the proposed method over the conventional ones, especially in the presence of faults with small magnitudes.