Search Results (5,173)

Accurate and stable wind power prediction is essential for effective wind farm capacity management and grid dispatching. Wind power generation is influenced not only by historical data, but also by turbine conditions and external environmental factors, such as weather. Although deep learning has made significant progress in the field of wind power forecasting, it often fails to account for two key characteristics of the data: dynamic variability and heterogeneity. Specifically, the influence of external variables on wind power changes over time, and due to the diverse nature of the information carried by different variables, simple weighted fusion approaches are insufficient to fully integrate heterogeneous data. To address these challenges, this paper introduces a dynamic heterogeneous graph attention recurrent network (DHGAR), which incorporates dynamic graphs, heterogeneous graph attention mechanisms, and gated recurrent units. Dynamic graphs capture real-time associations between wind power and external variables, while heterogeneous graph attention allows for more effective aggregation of diverse information. These two components are integrated into the gated recurrent units, replacing traditional fully connected layers to better capture temporal dependencies in the wind power time series. Experimental results on three real-world datasets demonstrate the superior performance and practical applicability of the proposed model. Full article

With the rapid development of the semiconductor industry, the demand for high-speed testing in the large-scale production of semiconductor devices and integrated circuit production lines continues to grow. As one of the key tools in semiconductor device performance testing and integrated circuit testing, source measure unit (SMU) plays a crucial role in high-precision transient response testing scenarios. In high-precision measurement scenarios, multiple measurements are often required and averaged to improve measurement accuracy, but this can slow down the measurement speed. This article proposes a measurement acceleration algorithm based on BA-Informer time series prediction to solve the problem of decreased measurement speed in high-precision measurement. On the one hand, this algorithm improves the encoder structure. Traditional time series prediction models may have limitations in handling long-term dependencies and trend extraction. BiRNN is an extended version of recurrent neural network (RNN), which consists of two directional RNN. One forward RNN processes data from the beginning to the end of the sequence, while the other reverse RNN processes data from the end to the beginning of the sequence. In the end, the outputs from both directions are merged at each time step. Compared to traditional one-way RNN, BiRNN can more effectively handle data with before and after dependencies. Based on its characteristics, this article integrates BiRNN into the encoder structure. This algorithm can simultaneously process input sequences from both positive and negative directions, effectively limiting the bidirectional contextual information of data and significantly enhancing the model’s ability to capture time series trends. In this paper, BiRNN is integrated into the encoder structure, and the algorithm can simultaneously process input sequences from both positive and negative directions, more effectively capturing the bidirectional contextual information of data and significantly enhancing the model’s ability to capture time series trends. This improvement enables the model to more accurately grasp the overall trend of data changes during prediction, thereby improving prediction accuracy. On the other hand, an attention discrete cosine transform (ADCT) module is introduced between the encoder and decoder to convert time-domain signals into frequency-domain representations. This not only reveals the spectral characteristics of the signal but also reduces data redundancy and improves the efficiency of subsequent processing by combining attention mechanisms. Finally, the algorithm performance is analyzed by analyzing the output characteristic curves of loads with different properties. The experiment shows that the prediction algorithm and the combination of measurement and prediction method proposed in this article save half of the measurement time by combining measurement and prediction while ensuring the same amount of data obtained, verifying the effectiveness of the proposed method. Full article

In the face of global climate change, accurately predicting carbon dioxide emissions has become an urgent requirement for environmental science and policy-making. This article provides a systematic review of the literature on carbon dioxide emission forecasting, categorizing existing research into four key aspects. Firstly, regarding model input variables, a thorough discussion is conducted on the pros and cons of univariate models versus multivariable models, balancing operational simplicity with high accuracy. Secondly, concerning model types, a detailed comparison is made between statistical methods and machine learning methods, with a particular emphasis on the outstanding performance of deep learning models in capturing complex relationships in carbon emissions. Thirdly, regarding model data, the discussion explores annual emissions and daily emissions, highlighting the practicality of annual predictions in policy-making and the importance of daily predictions in providing real-time support for environmental policies. Finally, regarding model quantity, the differences between single models and ensemble models are examined, emphasizing the potential advantages of considering multiple models in model selection. Based on the existing literature, future research will focus on the integration of multiscale data, optimizing the application of deep learning models, in-depth analysis of factors influencing carbon emissions, and real-time prediction, providing scientific support for a more comprehensive, real-time, and adaptive response to the challenges of climate change. This comprehensive research outlook aims to provide scientists and policymakers with reliable information on carbon emissions, promoting the achievement of environmental protection and sustainable development goals. Full article

Precise prediction of mooring tension is essential for the safety and operational efficiency of semi-submersible aquaculture platforms. Traditional numerical methods struggle with real-time performance due to the nonlinear and dynamic characteristics of environmental loads. This study proposes a novel neural network approach to enhance real-time forecasting of mooring line responses, combining Ensemble Empirical Mode Decomposition (EEMD), Temporal Convolutional Networks (TCNs), and a Self-Attention (SA) mechanism. The training dataset encompasses time-domain analysis results, including mooring tensions, motion responses, and total structural forces. Firstly, Pearson Correlation Analysis (PCA) is utilized to assess the linear relationships among the hydrodynamic variables. Subsequently, EEMD is applied to decompose the mooring tension data, which is then combined with the highly correlated variables to form the input dataset. Finally, the TCN model is trained to predict the time series, while an SA mechanism is integrated to weigh the significance of different moments within the sequence, thereby further enhancing prediction accuracy. The results demonstrate that the evaluation metrics of the EEMD-TCN-SA model outperform those of other neural network models, effectively predicting mooring tension for semi-submersible platforms and significantly reducing prediction errors. Full article

Background: Hospital-acquired pressure injuries (HAPIs) affect approximately 2.5 million patients annually in the United States, leading to increased morbidity and healthcare costs. Current rule-based screening tools, such as the Braden Scale, lack sensitivity, highlighting the need for improved risk prediction methods. Methods: We developed and externally validated a machine learning model to predict HAPI risk using longitudinal electronic medical record (EMR) data. This study included adult inpatients (2018–2023) across five hospitals within a large health system. An automated pipeline was built for EMR data curation, labeling, and integration. The model employed XGBoost with recursive feature elimination to identify 35 optimal clinical variables and utilized time-series analysis for dynamic risk prediction. Results: Internal validation and multi-center external validation on 5510 hospitalizations demonstrated AUROC values of 0.83–0.85. The model outperformed the Braden Scale in sensitivity and F1-score and showed superior performance compared to previous predictive models. Conclusions: This is the first externally validated, cross-institutional HAPI prediction model using longitudinal EMR data and automated pipelines. The model demonstrates strong generalizability, scalability, and real-time applicability, offering a novel bioengineering approach to improve HAPI prevention, patient care, and clinical operations. Full article

Despite many fuzzy time series forecasting (FTSF) models addressing complex temporal patterns and uncertainties in time series data, two limitations persist: they do not treat fuzzy and crisp time series as a unified whole for analyzing nonlinear relationships between different moments, and they fail to effectively capture how uncertainty in temporal patterns affects predictions. In this paper, we propose an FTSF model integrating Bayesian networks to overcome the limitations. Bayesian network (BN) structure learning is employed to extract fuzzy–crisp dependencies between historical fuzzified data and predicted crisp data alongside temporal crisp dependencies within crisp data. Integrating fuzzy logical relationship groups (FLRGs) and the two BNs representing the fuzzy–crisp and crisp relationships identifies temporal patterns efficiently. BN parameter learning models the occurrence uncertainties of dependencies through conditional probability distributions in BNs, while fuzzy empirical conditional probabilities quantify the occurrence uncertainties of the elements in FLRGs. The defuzzification stage infers the crisp predicted value using the fuzzy-empirical-probability weighted FLRGs and the two BN. We validate the forecasting performance of the proposed model on sixteen diverse time series. Experimental results demonstrate the competitive forecasting performance of the proposed model compared to state-of-the-art methods. Full article

The method based on convolution neural networks (CNNs) has been widely developed and applied to residual life prediction, and many excellent results have been achieved. However, CNN models can only learn feature information relative to size, and it is difficult to extract complex time series features from data of long time series. In addition, the existing models still have some problems, such as capturing the correlation of each time series and generating a large amount of redundant information. In order to alleviate the above problems, this study proposes a residual life prediction method of rolling bearings based on a multi-scale improved temporal convolutional network (MITCN) model. It is used to solve problems such as the low accuracy of bearing life prediction and the difficulty of the temporal convolutional network (TCN) model to capture the correlation of each time series. The model adopts the framework of a time convolution network and has good ability to extract time series information. By introducing a multi-scale expanded causal convolution residual structure, improved temporal convolutional network (ITCN) modules with different expansion factors capture information on different time scales and combine soft threshold functions and channel attention mechanisms to adaptively generate thresholds and eliminate redundant information. Finally, the carbon border adjustment mechanism (CBAM) is an attention mechanism used to enhance useful features and suppress useless features, so as to realize the effective fusion of multi-scale features. The IEEE PHM 2012 challenge data set is hereby used to verify the proposed method, which can effectively solve the problem of the low prediction accuracy of the remaining life of bearings. Full article

Accurate deformation prediction is crucial for ensuring the safety and longevity of bridges. However, the complex fluctuations of deformation pose a challenge to achieving this goal. To improve the prediction accuracy, a bridge deformation prediction method based on a bidirectional gated recurrent unit (BiGRU) neural network and error correction is proposed. Firstly, the BiGRU model is employed to predict deformation data, which aims to enhance the modeling capability of the GRU network for time-series data through its bidirectional structure. Then, to extract the valuable information concealed in the error, a transformer model is introduced to rectify the error sequence. Finally, the preliminary and error prediction results are integrated to yield high-precision deformation prediction results. Two deformation datasets collected from an actual bridge health monitoring system are utilized as examples to verify the effectiveness of the proposed method. The results show that the proposed method outperforms the comparison model in terms of prediction accuracy, robustness, and generalization ability, with the predicted deformation results being closer to the actual results. Notably, the error-corrected model exhibits significantly improved evaluation metrics compared to the single model. The research findings herein offer a scientific foundation for bridges’ early safety warning and health monitoring. Additionally, they hold significant relevance for developing time-series prediction models based on deep learning. Full article

The transition to electric vehicles (EVs) presents challenges and opportunities for the management of electrical networks. This paper focuses on developing and evaluating probabilistic forecasting algorithms to understand and predict EV charging behaviours, crucial for optimising grid operations and ensuring a balance between electricity demand and generation. Several forecasting approaches tailored to different time horizons are proposed across diverse model classes, including direct, bottom-up, and adaptive approaches. In all approaches, the target variable can be the load curve quantiles from 0.1 to 0.9 with 0.1 increments or prediction sets with a target coverage of 80%. Direct approaches learn from past load curves using GAMLSS or QGAM methods. Bottom-up approaches predict individual charging session characteristics (arrival time, charging duration, and energy demand) with mixture models before reconstructing the load curve. Adaptive approaches correct in real-time the prediction sets issued by direct or bottom-up approaches with conformal predictions. The experiments, conducted on real-world charging session data from Palo Alto, demonstrate the effectiveness of the proposed methods with regard to different metrics, including pinball loss, empirical coverage, and RPS. Overall, the results highlight the importance of quantifying uncertainty in load forecasts and the potential of probabilistic forecasting for EV load management. Full article

In July 2023, 19 continuous days of very hot days in Hong Kong brought inconvenience to citizens and disasters to society. This long-lasting heat wave event is closely linked to the atmospheric variability on the quasi-biweekly to intraseasonal timescales. While extreme weather has aroused the attention of scientists and society, limited studies focus on quasi-biweekly to intraseasonal extreme (QBIE) weather. Thus, to address this issue, this study aims at examining the climatology and long-term variability of these QBIE events in Hong Kong. This study serves as one of the very few fundamental works that construct a century-long record of QBIE temperature events, based on in situ observation in Hong Kong, and further examines the climatology, diversity, and variability of these QBIE temperature events. A total of 382 QBIE heat waves and 510 QBIE cold surges are identified from 1885 to 2022, exhibiting various characteristics in their occurring time and seasonality. Based on ARIMA model and time series analyses, we find that while apparent interannual variability exists in QBIE heat wave and cold surge activity, short-term climate prediction of QBIE temperature events based on past patterns or common climate indices is largely unfeasible. This research provides a valuable historical reference for understanding QBIE weather in the Guangdong–Hong Kong–Macau Greater Bay Area and highlights the need for further studies on the predictability of QBIE weather in the future. Full article

The importance of enhancing the accuracy of time-series forecasting using artificial intelligence tools is increasingly critical in light of the rapid advancements in modern technologies, particularly deep learning and neural networks. These approaches have already shown considerable advantages over traditional methods, especially due to their capacity to efficiently process large datasets and detect complex patterns. A crucial step in the forecasting process is the preprocessing of time-series data, which can greatly improve the training quality of neural networks and the precision of their predictions. This paper introduces a novel preprocessing technique that integrates information from both the time and frequency domains. To achieve this, the authors developed a feature extraction–extension scheme, where the extraction component focuses on obtaining the phase and amplitude of complex numbers through fast Fourier transform (FFT) and the extension component expands the time intervals by enriching them with the corresponding frequency characteristics of each individual time point. Building upon this preprocessing method, the FFT-LSTM forecasting model, which combines the strengths of FFT and Long Short-Term Memory (LSTM) recurrent neural networks, was enhanced. The simulation of the improved FFT-LSTM model was carried out on two time series with distinct characteristics. The results revealed a substantial improvement in forecasting accuracy compared to established methods in this domain, with about a 5% improvement in MAE and RMSE, thereby validating the effectiveness of the proposed approach for forecasting applications across various fields. Full article

Tourism is a critical sector for economic growth and cultural exchange, particularly for destinations like Turkey, which consistently attracts millions of visitors annually. This study investigates the dynamics of tourism demand in Turkey between 2008 and 2024, with a focus on seasonality, long-term trends, and predictive modeling accuracy. Time-series data were analyzed, and the impacts of economic indicators and digital search trends were evaluated using SARIMA and SARIMAX models. The results demonstrate that the SARIMA models outperformed the SARIMAX models, highlighting the dominance of intrinsic seasonal patterns over external regressors, such as exchange rates and inflation. The findings emphasize that geographic proximity and cultural similarities drive consistent tourist flows, while behavioral data like Google Trends provide supplementary insights into demand shifts. However, economic variables showed limited short-term predictive power. These results underscore the importance of prioritizing time-series structures in forecasting frameworks while complementing them with behavioral indicators for enhanced accuracy. This study contributes to the literature by addressing a critical gap in understanding how various factors influence tourism demand in Turkey and offers practical implications for policymakers and tourism planners to optimize strategic planning and resource allocation, ensuring sustainable tourism growth. Future research should explore hybrid models that incorporate sentiment-driven data and cultural factors for more robust forecasting. Full article

Climate change poses a significant global challenge that necessitates concerted efforts toward carbon neutrality. Circulating fluidized bed (CFB) boilers have gained prominence in various industries due to their adaptability and reduced emissions. However, many current control systems rely heavily on manual operator intervention and lack advanced automation, which constrains the operational efficiency. This study addressed the need for dynamic models capable of monitoring and optimizing NO_x emissions in CFB boilers, especially under fluctuating loads and strict regulatory standards. We introduced the TimesNet model, which utilizes fast Fourier transform (FFT) to extract key frequency components, transforming 1D time series data into 2D tensors for enhanced feature representation. The model employs Inception blocks for multi-scale feature extraction and incorporates residual connections with amplitude-weighted aggregation to mitigate catastrophic forgetting during training. The results indicated that TimesNet achieved R² values of 0.98, 0.97, and 0.95 across training, validation, and testing datasets, respectively, surpassing conventional models with a reduced MAE of 1.63 mg/m³ and RMSE of 3.35 mg/m³. Additionally, it excelled in multi-step predictions and effectively managed long-term dependencies. In conclusion, TimesNet provides an innovative solution for the precise monitoring of NO_x emissions in CFB boilers by enhancing predictive stability and robustness and addressing salient limitations in existing models to optimize combustion efficiency and regulatory compliance. Full article

Photovoltaic (PV) clusters in deserts such as the Gobi and other regions are frequently affected by sand and dust, which causes great deviation in power prediction and seriously threatens the safe operation of new power systems. For this reason, this paper proposes a short-term cluster PV power prediction method based on statistical upscaling, considering the effect of sand and dust. Firstly, the sand and dust events are identified, and then time series generative adversarial networks (TimeGANs) are used to solve the problem of small sample scarcity in sand and dust and construct a power correction model for sand and dust scenes. Secondly, for different weather scenes, a combination of conventional prediction and correction prediction is used to solve the problem of large differences in the predictability of a single model. Finally, a statistical upscaling method is utilized to calculate the cluster prediction power to solve the prediction difficulties of large-scale newly installed PV field stations. Through a case study and comparison with other models and methods, the cluster prediction method established in this paper effectively improves the prediction accuracy of the power of large-scale PV clusters affected by sand and dust, with the RMSE reduced by 8.28%. Full article

Traditional algorithms and single predictive models often face challenges such as limited prediction accuracy and insufficient modeling capabilities for complex time-series data in fault prediction tasks. To address these issues, this paper proposes a combined prediction model based on an improved temporal convolutional network (TCN) and bidirectional long short-term memory (BiLSTM), referred to as the TCN-BiLSTM model. This model aims to enhance the reliability and accuracy of time-series fault prediction. It is designed to handle continuous processes but can also be applied to batch and hybrid processes due to its flexible architecture. First, preprocessed industrial operation data are fed into the model, and hyperparameter optimization is conducted using the Optuna framework to improve training efficiency and generalization capability. Then, the model employs an improved TCN layer and a BiLSTM layer for feature extraction and learning. The TCN layer incorporates batch normalization, an optimized activation function (Leaky ReLU), and a dropout mechanism to enhance its ability to capture multi-scale temporal features. The BiLSTM layer further leverages its bidirectional learning mechanism to model the long-term dependencies in the data, enabling effective predictions of complex fault patterns. Finally, the model outputs the prediction results after iterative optimization. To evaluate the performance of the proposed model, simulation experiments were conducted to compare the TCN-BiLSTM model with mainstream prediction methods such as CNN, RNN, BiLSTM, and A-BiLSTM. The experimental results indicate that the TCN-BiLSTM model outperforms the comparison models in terms of prediction accuracy during both the modeling and forecasting stages, providing a feasible solution for time-series fault prediction. Full article