Search Results (274)

High-voltage diodes, as key devices in power electronic systems, have important significance for system reliability and preventive maintenance in terms of storage life prediction. In this paper, we propose a hybrid modeling framework that integrates the Long Short-Term Memory Network (LSTM) and Transformer structure, and is hyper-parameter optimized by the Improved Artificial Bee Colony Algorithm (IABC), aiming to realize the high-precision modeling and prediction of high-voltage diode storage life. The framework combines the advantages of LSTM in time-dependent modeling with the global feature extraction capability of Transformer’s self-attention mechanism, and improves the feature learning effect under small-sample conditions through a deep fusion strategy. Meanwhile, the parameter type-aware IABC search mechanism is introduced to efficiently optimize the model hyperparameters. The experimental results show that, compared with the unoptimized model, the average mean square error (MSE) of the proposed model is reduced by 33.7% (from 0.00574 to 0.00402) and the coefficient of determination (R²) is improved by 3.6% (from 0.892 to 0.924) in 10-fold cross-validation. The average predicted lifetime of the sample was 39,403.3 h, and the mean relative uncertainty of prediction was 12.57%. This study provides an efficient tool for power electronics reliability engineering and has important applications for smart grid and new energy system health management. Full article

In the aluminum electrolysis production process, the traditional cell control method based on cell voltage and series current can no longer meet the goals of energy conservation, consumption reduction, and digital-intelligent transformation. Therefore, a new digital cell control technology that is centrally dependent on various process parameters has become an urgent demand in the aluminum electrolysis industry. Among them, the real-time online measurement of alumina concentration is one of the key data points for implementing such technology. However, due to the harsh production environment and limitations of current sensor technologies, hardware-based detection of alumina concentration is difficult to achieve. To address this issue, this study proposes a soft-sensing model for alumina concentration based on a long short-term memory (LSTM) neural network optimized by a weighted average algorithm (WAA). The proposed method outperforms BiLSTM, CNN-LSTM, CNN-BiLSTM, CNN-LSTM-Attention, and CNN-BiLSTM-Attention models in terms of predictive accuracy. In comparison to LSTM models optimized using the Grey Wolf Optimizer (GWO), Harris Hawks Optimization (HHO), Optuna, Tornado Optimization Algorithm (TOC), and Whale Migration Algorithm (WMA), the WAA-enhanced LSTM model consistently achieves significantly better performance. This superiority is evidenced by lower MAE and RMSE values, along with higher R² and accuracy scores. The WAA-LSTM model remains stable throughout the training process and achieves the lowest final loss, further confirming the accuracy and superiority of the proposed approach. Full article

To combat microphone eavesdropping on devices like smartphones, ultrasonic-based methods offer promise due to human inaudibility and microphone nonlinearity. However, existing systems suffer from low jamming efficiency, poor energy utilization, and weak robustness. Based on these problems, this paper proposes a novel ultrasonic-based jamming algorithm called the Time–Frequency Mosaic (TFM) technique, which can be used for anti-eavesdropping. The proposed TFM technique can generate short-time, frequency-coded jamming signals according to the voice frequency characteristics of different speakers, thereby achieving targeted and efficient jamming. A jamming prototype using the Time–Frequency Mosaic technique was developed and tested in various scenarios. The test results show that when the signal-to-noise ratio (SNR) is lower than 0 dB, the text Word Error Rate (WER) of the proposed method is basically over 60%; when the SNR is 0 dB, the WER of the algorithm in this paper is on average more than 20% higher than that of current jamming algorithms. In addition, when the jamming system maintains the same distance from the recording device, the algorithm in this paper has higher energy utilization efficiency compared with existing algorithms. Experiments prove that in most cases, the proposed algorithm has a better jamming effect, higher energy utilization efficiency, and stronger robustness. Full article

This paper proposes an Adam-optimized Deep Belief Networks (Adam-DBNs) denoising method for throat-attached piezoelectric signals. The method aims to process mechanical vibration signals captured through polyvinylidene fluoride (PVDF) sensors attached to the throat region, which are typically contaminated by environmental noise and physiological noise. First, the short-time Fourier transform (STFT) is utilized to convert the original signals into the time–frequency domain. Subsequently, the masked time–frequency representation is reconstructed into the time domain through a diagonal average-based inverse STFT. To address complex nonlinear noise structures, a Deep Belief Network is further adopted to extract features and reconstruct clean signals, where the Adam optimization algorithm ensures the efficient convergence and stability of the training process. Compared with traditional Convolutional Neural Networks (CNNs), Adam-DBNs significantly improve waveform similarity by 6.77% and reduce the local noise energy residue by 0.099696. These results demonstrate that the Adam-DBNs method exhibits substantial advantages in signal reconstruction fidelity and residual noise suppression, providing an efficient and robust solution for throat-attached piezoelectric sensor signal enhancement tasks. Full article

Standalone photovoltaic (PV) systems offer a viable path to decentralized energy access but face limitations during periods of low solar irradiance. While batteries provide short-term storage, their capacity constraints often restrict the use of surplus energy, highlighting the need for long-duration solutions. Green hydrogen, generated via proton exchange membrane (PEM) electrolyzers, offers a scalable alternative. This study proposes a stochastic energy management framework that leverages a Markov decision process (MDP) to coordinate PV generation, battery storage, and hydrogen production under variable irradiance and uncertain load demand. The strategy dynamically allocates power flows, ensuring system stability and efficient energy utilization. Real-time weather data from Goiás, Brazil, is used to simulate system behavior under realistic conditions. Compared to the conventional perturb and observe (P&O) technique, the proposed method significantly improves system performance, achieving a $99.9\%$ average efficiency (vs. 98.64%) and a drastically lower average tracking error of 0.3125 (vs. 9.8836). This enhanced tracking accuracy ensures faster convergence to the maximum power point, even during abrupt load changes, thereby increasing the effective use of solar energy. As a direct consequence, green hydrogen production is maximized while energy curtailment is minimized. The results confirm the robustness of the MDP-based control, demonstrating improved responsiveness, reduced downtime, and enhanced hydrogen yield, thus supporting sustainable energy conversion in off-grid environments. Full article

This research evaluates the performance of six metaheuristic algorithms in the active and reactive power management of wind turbines (WTs) integrated into an AC microgrid (MG). The population-based genetic algorithm (PGA) is proposed as the primary optimization strategy and is rigorously compared against five benchmark techniques: Monte Carlo (MC), particle swarm optimization (PSO), the JAYA algorithm, the generalized normal distribution optimizer (GNDO), and the multiverse optimizer (MVO). This study aims to minimize, through independent optimization scenarios, the operating costs, power losses, or CO₂ emissions of the microgrid during both grid-connected and islanded modes. To achieve this, a coordinated control strategy for distributed generators is proposed, offering flexible adaptation to economic, technical, or environmental priorities while accounting for the variability of power generation and demand. The proposed optimization model includes active and reactive power constraints for both conventional generators and WTs, along with technical and regulatory limits imposed on the MG, such as current thresholds and nodal voltage boundaries. To validate the proposed strategy, two scenarios are considered: one involving 33 nodes and another one featuring 69. These configurations allow evaluation of the aforementioned optimization strategies under different energy conditions while incorporating the power generation and demand variability corresponding to a specific region of Colombia. The analysis covers two-time horizons (a representative day of operation and a full week) in order to capture both short-term and weekly fluctuations. The variability is modeled via an artificial neural network to forecast renewable generation and demand. Each optimization method undergoes a statistical evaluation based on multiple independent executions, allowing for a comprehensive assessment of its effectiveness in terms of solution quality, average performance, repeatability, and computation time. The proposed methodology exhibits the best performance for the three objectives, with excellent repeatability and computational efficiency across varying microgrid sizes and energy behavior scenarios. Full article

The earth’s surface has an uneven solar energy density that is sufficient to stimulate solar photovoltaic (PV) production. This causes variations in a solar plant’s output, which are impacted by geometrical elements and atmospheric conditions that prevent it from passing. Motivated by the focus on encouraging increased PV production efficiency, the goal was to use machine learning models (MLM) to map the distribution of solar energy in Mozambique in a regressive literal and geospatial–temporal manner on a short measurement scale. The clear-sky index $\left(K_t^{*}\right)$ theoretical approach was applied in conjunction with MLM that emphasized random forest (RF) and artificial neural networks (ANNs). Solar energy mapping was the result of the methodology, which involved statistically calculating $K_t^{*}$ for the analysis of solar energy in correlational and causal terms of the space-time distribution. Utilizing data from PVGIS, NOAA, NASA, and Meteonorm, a sample of solar energy was gathered at 11 measurement stations in Mozambique over a period of 1 to 10 min between 2012 and 2014 as part of the FUNAE and INAM measurement programs. The statistical findings show a high degree of solar energy incidence, with increments $\left(∆K_t^{*}\right)$ in the average order of −0.05 and $K_t^{*}$ mostly ranging between 0.4 and 0.9. In 2012 and 2014, $K_t^{*}$ was 0.8956 and 0.6986, respectively, because clear days had a higher incident flux and intermediate days have a higher frequency of ${∆K}_t^{*}$ on clear days and a higher occurrence density. There are more cloudy days now 0.5214 as opposed to 0.3569. Clear days are found to be influenced by atmospheric transmittance because of their high incident flux, whereas intermediate days exhibit significant variations in the region’s solar energy. Full article

The exploitation of renewable energy is essential for mitigating climate change and reducing fossil fuel emissions. Wind energy, the most mature technology, is highly dependent on wind speed, and the accurate prediction of the latter substantially supports wind power generation. In this work, various artificial neural networks (ANNs) were developed and evaluated for their wind speed prediction ability using the ERA5 historical reanalysis data for four potential Offshore Wind Farm Organized Development Areas in Greece, selected as suitable for floating wind installations. The training period for all the ANNs was 80% of the time series length and the remaining 20% of the dataset was the testing period. Of all the ANNs examined, the hybrid model combining Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) networks demonstrated superior forecasting performance compared to the individual models, as evaluated by standard statistical metrics, while it also exhibited a very good performance at high wind speeds, i.e., greater than 15 m/s. The hybrid model achieved the lowest root mean square errors across all the sites—0.52 m/s (Crete), 0.59 m/s (Gyaros), 0.49 m/s (Patras), 0.58 m/s (Pilot 1A), and 0.55 m/s (Pilot 1B)—and an average coefficient of determination (R²) of 97%. Its enhanced accuracy is attributed to the integration of the LSTM and GRU components strengths, enabling it to better capture the temporal patterns in the wind speed data. These findings underscore the potential of hybrid neural networks for improving wind speed forecasting accuracy and reliability, contributing to the more effective integration of wind energy into the power grid and the better planning of offshore wind farm energy generation. Full article

Based on the Clouds and the Earth’s Radiant Energy System (CERES) satellite data from 2001 to 2023 and the climate indices from the National Oceanic and Atmospheric Administration (NOAA), this study analyzes the solar irradiance over mainland China and the impacts of clouds and aerosols on it and constructs monthly forecasting models to analyze the influence of climate indices on irradiance forecasts. The irradiance over mainland China shows a spatial distribution of being higher in the west and lower in the east. The influence of clouds on irradiance decreases from south to north, and the influence of aerosols is prominent in the east. The average explained variance of clouds on irradiance is 86.72%, which is much higher than that of aerosols on irradiance, 15.62%. Singular Value Decomposition (SVD) analysis shows a high correlation between the respective time series of irradiance and cloud influence, with the two fields having similar spatial patterns of opposite signs. The variation in solar irradiance can be attributed mainly to the influence of clouds. Empirical Orthogonal Function (EOF) analysis indicates that the variation in the first mode of irradiance is consistent in most parts of China, and its time coefficient is selected for monthly forecasting. Both the traditional multiple linear regression method and the Long Short-Term Memory (LSTM) network are used to construct monthly forecast models, with the preceding time coefficient of the first EOF mode and different climate indices as input. Compared with the multiple linear regression method, LSTM has a better forecasting skill. When the input length increases, the forecasting skill decreases. The inclusion of climate indices, such as the Indian Ocean Basin (IOB), El Nino–Southern Oscillation (ENSO), and Indian Ocean Dipole (IOD), can enhance the forecasting skill. Among these three indices, IOB has the most significant improvement effect. The research provides a basis for accurate forecasting of solar irradiance over China on monthly time scale. Full article

The short-term performance of ground heat exchangers (GHEs) is crucial for the optimal design of ground-source heat pumps (GSHPs), enhancing their contribution to sustainable energy solutions. Local short-time thermal response functions, or short-time g-functions (STGFs) derived from thermal response tests (TRTs), are of great interest for predicting the heat exchange due to their fast and simple applicability. The aim of this work is to perform a sensitivity analysis to assess the impact of thermal parameter variability and TRT operating conditions on the accuracy of the average fluid temperature (T_f) predictions obtained through a local STGF. First, the uncertainties associated with the borehole thermal resistance (R_b), transmitted from the soil volumetric heat capacity (C_S) or some models dependent on GHE characteristics, such as the Zeng model, were found to have a low impact in T_f resulting in long-term deviations of ±0.2 K. Second, several TRTs were carried out on the same borehole, changing input parameters such as the volumetric flow rate and heat injection rate, in order to obtain their corresponding STGF. Validation results showed that each T_f profile consistently aligned well with experimental data when applying intermittent heat rate pulses (being the most unfavorable scenario), implying deviations of ±0.2 K, despite the variabilities in soil conductivity (λ_S), soil volumetric heat capacity (C_S), and borehole thermal resistance (R_b). Full article

This paper presents a novel small-scale system for drinking water treatment from surface waters, designed to rely on gravity as the only source of energy driving the treatment process. The pilot-scale setup, designed for a flow rate of 0.5 L/s, was tested at the Cornell University Water Filtration Plant (CWFP) for a total period of 5 months of operation. The experiments evaluated the influence of selected process parameters on system performance. The identified best operation practices were used to complete a comparative study against CWFP’s full-scale treatment process and to conduct a performance assessment in the context of various legislative landscapes. The objective of the work was to determine both the advantages and disadvantages of the proposed technology over established solutions. Over the study period, the average turbidity of the produced water was equal to 0.54 NTU. The pilot complied with the United States Environmental Protection Agency (US EPA) turbidity standard of <0.3 NTU 47.1% of the time and <1 NTU for 89.9% of the time, thus falling short of the standard of <0.3 NTU 95% of the time and <1 NTU 100% of the time. For 99.5% of the time, it complied with the World Health Organization turbidity guideline of <5 NTU for chlorination treatment. The benchmark conventional system outperformed the tested prototype, complying with the US EPA standards for the entire duration of the study. The tested process also generated a waste stream, which accounted on average for more than 10% of the total raw water volume. Full article

Wind power, as a clean and renewable energy source, plays an increasingly important role in the global transition to low-carbon energy systems. However, its inherent volatility and unpredictability pose challenges for accurate short-term prediction. This study proposes an ultra-short-term wind power prediction framework that integrates multiple technical indicators with the extreme gradient boosting (XGBoost) algorithm. Inspired by financial time series analysis, the model incorporates K-line representations, power fluctuation features, and classical technical indicators, including the moving average convergence divergence (MACD), Bollinger bands (BOLL), and average true range (ATR), to enhance sensitivity to short-term variations. The proposed method is validated on two real-world wind power datasets from Inner Mongolia, China, and Germany, sourced from the European network of transmission system operators for electricity (ENTSO-E). The experimental results show that the model achieves strong performance on both datasets, demonstrating good generalization ability. For instance, on the Inner Mongolia dataset, the proposed model reduces the mean squared error (MSE) by approximately 11.4% compared to the long short-term memory (LSTM) model, significantly improving prediction accuracy. Full article

In the realm of renewable energy, harnessing wind power efficiently is crucial for establishing a low-carbon power system. However, the intermittent and uncertain nature of wind speed poses significant challenges for accurate prediction, which is essential for effective grid integration and dispatch management. To address this challenge, this paper introduces a novel hybrid model, NRBO-TXAD, which integrates a Newton–Raphson-based optimizer (NRBO) with a Transformer and XGBoost, further enhanced by adaptive denoising techniques. The interquartile range–adaptive moving average filter (IQR-AMAF) method is employed to preprocess the data by removing outliers and smoothing the data, thereby improving the quality of the input. The NRBO efficiently optimizes the hyperparameters of the Transformer, thereby enhancing its learning performance. Meanwhile, XGBoost is utilized to compensate for any residual prediction errors. The effectiveness of the proposed model was validated using two real-world wind speed datasets. Among eight models, including LSTM, Informer, and hybrid baselines, NRBO-TXAD demonstrated superior performance. Specifically, for Case 1, NRBO-TXAD achieved a mean absolute percentage error (MAPE) of 11.24% and a root mean square error (RMSE) of 0.2551. For Case 2, the MAPE was 4.90%, and the RMSE was 0.2976. Under single-step forecasting, the MAPE for Case 2 was as low as 2.32%. Moreover, the model exhibited remarkable robustness across multiple time steps. These results confirm the model’s effectiveness in capturing wind speed fluctuations and long-range dependencies, making it a reliable solution for short-term wind forecasting. This research not only contributes to the field of signal analysis and machine learning but also highlights the potential of hybrid models in addressing complex prediction tasks within the context of artificial intelligence. Full article

Exercise metrics are critical for assessing health, but real-time heart rate and respiration measurements remain challenging. We propose a physiological monitoring system that uses an in-ear microphone to extract heart rate and respiration from faint ear canal signals. An improved non-negative matrix factorization (NMF) algorithm combines with a short-time Fourier transform (STFT) to separate physiological components, while an inverse Fourier transform (IFT) reconstructs the signal. The earplug effect enhances the low-frequency components, thereby improving the signal quality and noise immunity. Heart rate is derived from short-term energy and zero-crossing rate, while a BiLSTM-based model can refine the breathing phases and calculate indicators such as respiratory rate. Experiments have shown that the average accuracy can reach 91% under various conditions, exceeding 90% in different environments and under different weights, thus ensuring the system’s robustness. Full article

Accurate building energy consumption prediction is essential for efficient energy management and energy optimization. This study utilizes bidirectional long short-term memory (BiLSTM) to automatically extract deep time series features. The nonlinear fitting and high-precision prediction capabilities of Random Forest (RF) and XGBoost models are then utilized to develop a BiLSTM-RF-XGBoost stacked hybrid model. To enhance model generalization and reduce overfitting, a Bayesian algorithm with an early stopping mechanism is utilized to fine-tune hyperparameters, and strict K-fold time series cross-validation (TSCV) is implemented for performance evaluation. The hybrid model achieves a high TSCV average R² value of 0.989 during cross-validation. When evaluated on an independent test set, it yields a mean square error (MSE) of 0.00003, a root mean square error (RMSE) of 0.00548, a mean absolute error (MAE) of 0.00130, and a mean absolute percentage error (MAPE) of 0.26%. These values are significantly lower than those of comparison models, indicating a significant improvement in predictive performance. The study offers insights into the internal decision-making of the model through SHAP (SHapley Additive exPlanations) feature significance analysis, revealing the key roles of temperature and power lag features, and validating that the stacked model effectively utilizes the outputs of base models as meta-features. This study makes contributions by proposing a novel hybrid model trained with Bayesian optimization, analyzing the influence of various feature factors, and providing innovative technological solutions for building energy consumption prediction. It also provides theoretical value and guidance for low-carbon building energy management and application. Full article