Search

Electricity load exhibits daily and weekly cyclical patterns as well as random characteristics. At present, prevailing deep learning models cannot learn electricity load cyclical and stochastic features adequately. This results in insufficient prediction accuracy and the scalability of current methods. To tackle these difficulties, this paper proposes a framework for electrical load prediction based on an Attention Mechanism Time Series Depthwise Separable Convolutional Neural Network (ELPF-ATDSCN). The framework starts by using the Maximum Information Coefficient for exogenous variable selection. It then incorporates a seasonal decomposition algorithm with manual feature engineering to extract the cyclical and stochastic features of the electrical load. Subsequently, the framework employs the ATDSCN to learn the cyclical and stochastic features of the electrical load. In addition, the Bayesian algorithm optimizes model hyperparameters for optimal model performance. Experimental results of point and interval load prediction on datasets from the US and Nordic power markets reveal that the ATDSCN model proposed in this paper enhances load prediction accuracy compared with other models. It can provide more reliable predictions for power system operation and dispatch. • A new attention mechanism is designed to improve the model memory capability. • The ATDSCN model is proposed for learning periodic and stochastic features of electricity loads. • A framework for electricity load prediction is proposed to improve the accuracy of electricity load forecasting. • The effectiveness of the proposed method is validated on the real world dataset. [ABSTRACT FROM AUTHOR]

Because the carbon load inside a diesel particulate filters (DPF) affects the DPF regeneration, and the carbon load recognition is significant for the particulate matter (PM) emission control. It is necessary to investigate an on-board DPF carbon load recognition method because the carbon load cannot be directly measured by sensors. Aiming to build a DPF carbon load prediction model adopting the deep learning method, this paper proposes a DPF carbon load identification model based on different experimental parameters using a layered one dimension convolutional neural network (1D-CNN) method. To improve data validity, this paper adopts two data-processing methods. The data pre-processing adopts data splicing method to complete the construction of the original sample set, and the data after-processing uses wavelet packet transform method to establish the feature sample sets. The model adopts the optimal feature dataset constructed by three input parameters, i.e., temperature difference, pressure difference, and exhaust mass flow, and has both high training accuracy and test accuracy above 90 %. The pressure difference is the most important influencing input parameter, and the three-parameter sample set (Δ T + Δ P + Q) has great recognition accuracy and good model stability with the high training accuracy and test accuracy as well as less iteration. • A heavy-duty diesel engine was tested under various conditions of carbon load inside DPF. • The data processing was done by two methods: data splicing and wavelet packet transform. • We proposed a DPF carbon load identification model based on a one-dimension CNN method. • The model adopts 3 parameters, and the 3-parameter sample set has optimal performance. • The 1D-CNN model has both high training accuracy and test accuracy above 90 %. [ABSTRACT FROM AUTHOR]

Obtaining the aerodynamic performance of the turbine blade by Computational Fluid Dynamics (CFD) methods is accurate. However, it consumes time and computational resources. This paper proposes an evaluation method based on Convolutional Neural Network (CNN) and Artificial Neural Network (ANN) to obtain the aerodynamic performance of the turbine blade accurately and quickly. Compared with the existing data-driven modeling methods, this method innovatively introduces the Residual Network (ResNet), employs a transfer learning strategy for network design, and realizes the automatic extraction of blade profile features and non-parametric input. In processing boundary conditions, the ANN is utilized to fuse the blade profile features with the boundary conditions to realize the mapping between blade profile and aerodynamic performance under different conditions. In addition, to minimize the prediction deviation caused by the severely uneven distribution of the data set, we combined ensemble learning with transfer learning and proposed a two-step prediction strategy. The numerical simulations results show that the ResNet-ANN model established in this paper has a prediction relative error of 5 % on turbine blade aerodynamic parameters under various working conditions. The error is reduced by more than 90 % under −40°-10° incidence angle of incoming flow compared with the empirical model. • The method for predicting turbine blade profile loss is non-parametric. • The scheme is effective on both design and off-design conditions. • A two-step strategy is proposed for precise prediction on unevenly distributed data. • The TSP strategy is based on transfer learning and ensemble learning. • The approach is of high accuracy with considerable generalization. [ABSTRACT FROM AUTHOR]

In the context of prominent energy crisis, photovoltaic power (PV) generation has received increasing attention, then accurate PV generation forecasting is crucial for ensuring the smooth operation of power stations. However, existing research is insufficient in comprehensively analyzing the impact of PV generation volatility. To fill the gaps and enhance the prediction accuracy, this paper proposes a new hybrid forecasting method. We first introduce the Locally Weighted Scatterplot Smoothing (LOWESS) method to process the data and enhance the data stability, and use Pearson correlation coefficient (PCC) and Random Forests (RF) for feature selection to improve the quality of input data. Then we use Attention mechanism and Convolutional Neural Network (CNN) layer to optimize Bi-directional Gate Recurrent Unit (BiGRU) model and form a new hybrid model. Finally, based on the Bagging algorithm, we use ensemble learning to further optimize the hybrid BiGRU model to enhance the depth and performance. The proposed method is validated through case analysis results from two different locations, Xuhui District in Shanghai, China and the DKASC area in Alice Springs, Australia. The results demonstrate that, compared with other models, the developed method exhibits exceptional prediction performance and effectively enhances the accuracy of PV generation forecasting. Keywords : photovoltaic power generation; Locally Weighted Scatterplot Smoothing; feature selection; ensemble learning; Bi-directional Gate Recurrent Unit. • LOWESS smoothing technology is used to reduce the variability of PV generation data. • Pearson correlation coefficient and Random Forest are used to for extract feature. • Attention mechanism and CNN layer are used to optimize BiGRU model. • Two BiGRU layers are stacked to improve the depth and prediction performance. • Ensemble learning is introduced to further optimize the hybrid BiGRU model. [ABSTRACT FROM AUTHOR]

Accurate wind speed forecasting is of great significance for the utilization of Arctic wind energy resources. The traditional single model is difficult to fully depict the nonlinearity of wind speed and its wide range of variations. In this paper, a hybrid model is proposed for multi-step wind speed forecasting, which combined complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN), convolutional neural network (CNN) and long-short term memory neural network (LSTM). The wind speed series is firstly decomposed into several intrinsic mode functions (IMF) by CEEMDAN to provide more stable data. Secondly, four-step-ahead forecasts are realized using well-tuned CNN-LSTM model for each IMF. Finally, the forecasted wind speed is obtained by reconstruction. The effectiveness and feasibility of the proposed method is validated based on thorough evaluation and step-by-step analysis. The RMSE of the proposed model is 0.4046 m/s, which are reduced by 58 % compared with 7 benchmark models. Furthermore, the average prediction interval of the proposed model is also reduced by 20 %, 16 % and 7 % compared to CEEMDAN-FCNN, CEEMDAN-CNN and CEEMDAN-LSTM respectively. The results prove that all three parts of the proposed model contribute to a better performance in wind speed forecasting. [Display omitted] • CEEMDAN can reduce the nonlinearity of wind speed data and unstable IMFs have a large impact on the wind speed forecasting. • Combining LSTM with CNN can enhance feature extraction and improve the accuracy of multi-step wind speed forecasting. • The proposed model achieves more accurate forecasting for catastrophe points which provides more critical information. [ABSTRACT FROM AUTHOR]

Power stations operating on PhotoVoltaic (PV) power generation are challenged by demand forecasting as PV power generation is random and intermittent. This paper presents an Efficient Shrinkage Temporal Convolutional Network (ESTCN) model, which combines the Temporal Convolutional Network (TCN) and an improved Deep Residual Shrinkage Network (DRSN) to forecast PV power output. First, the attention sub-network of the DRSN is improved, the fully connected layer of the sub-network is canceled, and feature extraction is done directly on the model features following global average pooling using a one-dimensional convolution to obtain cross-channel interaction and increased model efficiency. Next, an improved attention mechanism and adaptive soft thresholding are introduced into TCN to automatically determine the noise threshold to address the issue of information weight dispersion caused by redundant information in the input samples. By incorporating dilated causal convolution, attention module, soft thresholding, and residual connection, the ESTCN model is formed and is shown to enhance PV power output prediction with only a minimal increase in the number of parameters. The power station in Ningxia, China is adopted as a validation example with data taken from the year 2020. The fitting and prediction results of the ESTCN model are compared against the Convolutional Neural Network, Long Short-Term Memory, and TCN models. The ESTCN model yielded the following values of the evaluation metrics: Root Mean Square Error (RMSE) of 1.47 kW, Mean Absolute Error (MAE) of 0.79 kW, and coefficient of determination of 0.999. Compared to the TCN model, the prediction errors based on RMSE and MAE improved by 0.39 kW and 0.18 kW, respectively. Applying the ESTCN model to predict PV power generation of power stations in two regions in China, Ningxia and Xinjiang, shows the ESTCN model to be superior, scalable, and universal over other deep learning models. • Study the problem of photovoltaic power using a deep learning method. • A new Efficient Shrinkage Temporal Convolutional Network model is presented. • Attention sub-network of DRSN is improved by cancelling the fully connected layer. • Feature extraction is done on the model features using a one-dimensional convolution. • Improved attention mechanism and adaptive soft thresholding are introduced into TCN. [ABSTRACT FROM AUTHOR]

A precise and cost-effective prediction tool for fluid-structure interaction (FSI) analysis is crucial for optimizing the structural design of tidal turbine blades. However, the high computational costs associated with fluid dynamic analysis pose a significant challenge, as the current lack of efficient FSI prediction methods hinders the advancement of cutting-edge tidal turbine designs. To address this issue, this paper proposes a novel consolidated framework that integrates deep learning convolutional neural networks (CNN) with blade element momentum (BEM) theory and finite element method (FEM) to perform deformation analysis of turbine blade structures. The proposed CNN-BEM-FEM integrated framework efficiently identifies the geometric features and predicts the hydrodynamic parameters of turbine blades and thus, achieving accurate assessments of the structural behavior of tidal turbines. The study applies two-step verification procedures to validate the prediction accuracy of the CNN-BEM-FEM framework and the result demonstrates excellent agreement with experimental tests for hydrodynamic performance and blade deformation. When compared with the static one-way FSI calculated by Ansys Workbench software, the computational efficiency of CNN-BEM-FEM framework increases by more than 18 times, with discrepancies in blade deformation and equivalent stress calculations generally less than 5 %. By applying the proposed method to predict the FSI performance of tidal turbine blades with various shear web structures, the practical applicability for composite turbine blade design is successfully demonstrated. The results underscore the potential of the CNN-BEM-FEM framework as an efficient and accurate prediction tool for optimizing the structural design of tidal turbine blades. • A high-efficient framework that combines CNN, BEM and FEM for tidal turbine rotor structural analysis is proposed. • Convolutional neural network (CNN) is used for identifying hydrodynamic performance of turbine blade section. • Blade element momentum (BEM) is for turbine prediction, while finite element method (FEM) is for structural analysis. • It shows great agreement with experiments, ensuring precise prediction of turbine blade deformations and hydrodynamics. • It outperforms benchmark software by more than 18 times, demonstrating exceptional computational efficiency. [ABSTRACT FROM AUTHOR]

The accurate estimation of the state of charge (SOC) is crucial for the safe and reliable operation of Li-ion batteries. In this paper, novel dual-channel deep learning methods were proposed for SOC estimation during dynamic driving cycles. For dual-channel deep learning methods, each channel of the deep learning methods could be a single convolutional neural network (CNN) layer, single long short-term memory (LSTM) layer or sequentially combined structure of CNN and LSTM. After evaluation, it could be found that dual-channel deep learning methods could achieve higher accuracy than conventional single-channel method, however, the model size and floating-point operations (FLOPs) could be either reduced or increased. Among all the dual-channel deep learning methods, CNN + LSTM-CNN + LSTM method could achieve the highest accuracy and the shortest training time with sacrifices in computational complexity. CNN–CNN + LSTM method could achieve higher accuracy than the single-channel deep learning method with reduced computational complexity and model size. As the error of SOC estimation is highly correlated with voltage signals, conditional judgement was then integrated with dual-channel deep learning method to obtain higher accuracy. After integrated with conditional judgement, the averaged error of SOC estimation could be reduced by more than 45% without a substantial sacrifice in computational time. • Dual-channel deep learning methods are proposed for SOC estimation. • Two channels are responsible for processing of transient and moving averaged signals. • Model hyperparameters are optimized by Tree-structured Parzen estimator. • Cross validation method is applied to evaluate the proposed methods. • Conditional judgement is proposed to obtain higher accuracy. [ABSTRACT FROM AUTHOR]

Wind power has quickly developed in the world owing to the advantages of pure, inexpensive, and inexhaustible. However, strong volatility, unmanageable, and randomness make it difficult to achieve secure wind power generation. An excellent wind power prediction is effective for power system scheduling and safely stable operation. Vision Transformer (ViT) model is introduced for building a connection of the extracted characteristics and desired output. Long-short term memory (LSTM) is combined with ViT, and a new wind power forecasting model is proposed in this paper. For the proposed LSTM-ViT model, the temporal aspects of the weather data and correspondence properties are extracted based on LSTM. The link of the output and characteristic is established in view of the ViT, and the multi-headed self-attentiveness mechanisms in ViT fully exploit the relationship between the inputs. The validity and sophistication of the LSTM-ViT method are validated by the climate statistics and statistics of wind power. The results indicate that the wind power forecasting model is provided with higher prediction accuracy. The forecast results for the fourth quarter are used as analysis cases. The root mean square error of the method is reduced by 41.77%, 16.60%, 28.72%, 26.81%, and 16.25% compared to gate recurrent unit (GRU), LSTM, ViT, convolutional neural network (CNN)-ViT, and GRU-ViT respectively. The mean absolute error of the LSTM-ViT method in the first quarter is 0.327, with model comparison values reduction of 33.71%, 38.30%, 32.99%, 17.63% and 10.65% respectively. • A novel model called LSTM-ViT is proposed for ultra-short term wind power forecasting. • ViT model is introduced for building a connection of the extracted characteristics and desired output. • The multi-headed self-attentiveness mechanisms in ViT fully exploit the relationship between the inputs. • The complex nonlinear relationships are considered in the proposed model. [ABSTRACT FROM AUTHOR]

Accurately estimating the state of charge (SOC), state of energy (SOE), and state of health (SOH) online is a critical and urgent concern in the management of lithium-ion batteries for electric vehicle applications, particularly in terms of safety and reliability. This paper develops a hybrid neural network, abbreviated as CNN-SAM-LSTM model, which combines a convolutional neural network, self-attention mechanism, and long-short term memory neural network to jointly estimate the state parameters of lithium-ion batteries, including SOC, SOE, and SOH. Additionally, a joint loss function considering homoscedastic uncertainty is developed to optimize weight adjustments for the training losses associated with the three state parameters. Experimental data collected under UDDS, BBDST and CC discharge conditions are employed to showcase the effectiveness of the proposed CNN-SAM-LSTM model. The results demonstrate that the proposed model is capable of simultaneously and accurately estimating SOC, SOE, and SOH for lithium-ion batteries under different dynamical operating conditions. Moreover, when applied to randomly segmented data, the proposed model exhibits robustness, effectively handling deviations from random discharge segments. • A CNN-SAM-LSTM hybrid neural network is developed to estimate SOC, SOE and SOH. • The loss function of the hybrid network considers homoscedastic uncertainty. • The proposed method achieves high accuracy under dynamic operating conditions. • The proposed method is strongly robust against random discharge segments. [ABSTRACT FROM AUTHOR]

Solar radiation is an essential meteorological parameter for building energy efficiency analysis, and the quality of the data directly affects the analysis results. This paper investigates the estimation of hourly solar radiation based on the generation of the typical meteorological year(TMY) using various real meteorological parameters and limited solar radiation data. The focus of this paper is to use two types of neural network algorithms to improve the estimation accuracy and applicability, and to solve the problem of hourly solar radiation data acquisition in non-radiation areas. First, select two city station data and use three methods to generate TMY. Then, two neural network models, BP Neural Network (BP)，Convolutional Neural Network (CNN) are used to estimate the hourly solar radiation data and verify the results. Finally, by constructing a photovoltaic-integrated office building model, the accuracy of the hourly solar radiation estimation model is verified using energy consumption simulation and photovoltaic (PV) power generation simulation. The results show that this paper can well solve the problem of limited radiation data, which provides a new idea for the study of building energy efficiency in areas where radiation data is missing. • Multiple neural network is used to estimate the generation of solar radiation data using other meteorological parameters. • An advanced neural network is used to provide an effective TMY generation method in missing hourly solar radiation data. • The accuracy of the experimental results is verified by energy consumption simulation and photovoltaic power generation. [ABSTRACT FROM AUTHOR]

Accurate health estimation of massive battery packs and efficient optimization of data storage have become major technical challenges with the development of big data platforms. In this paper, multi-level energy indicators are defined to reflect the overall health state of the battery pack, and battery pack health assessment is achieved through energy estimation. A visual feature learning method is proposed to extract features from partial cell charging voltage profiles image and the relationship between features and energy indicators is constructed by the hybrid convolutional neural network. The effectiveness of the proposed method is verified on the dataset generated by the battery pack model considering cell inconsistency, and the mean absolute percentage error of each energy indicator estimation is less than 1%. Additionally, validations are carried out on simulated data with sampling noise and two cases of experimental data to verify the stability of the method. The proposed visual feature learning method provides a new idea for the data storage and health monitoring of massive battery packs on the big data platform. • Multi-level energy indicators are proposed to reflect the battery pack health. • Battery pack dataset generated based on orthogonal combination of inconsistency parameters. • Extract visual features from partial cell charging voltage profiles image. • Good anti-interference ability against sampling noise. [ABSTRACT FROM AUTHOR]

Many industrial processes belong to complex thermal system coupled by multiple heat sources, in which the energy distribution and transmission for each heat source is closely related to its adjacent neighbors, and the thermal dynamics have obvious nonlinear time-series characteristics, these factors limit the modeling and predict accuracy for this type of systems in actual practice. In this paper, a convolution and memory network-based spatiotemporal modeling approach is proposed to model the nonlinear spatial features and temporal dynamics of these heating systems. First, a deformation convolutional neural network (D-CNN) strategy is developed to extract the spatial features, it designs a deformable kernel matrix for convolution layer to obtain the spatial multi-scale context information of the DPS, and the feature order is reduced by pooling operation. Then, a cascade long-short time memory network (LSTM) model is constructed to model the temporal dynamics of the time-series data. In this model, the sliding window method is used to obtain the sequential dataset for several sampling time, which is taken as the input of each LSTM unit to reconstruct the nonlinear temporal dynamics in this type of systems. The effect of the proposed model for spatiotemporal dynamics is verified by the experiment of serial-connected lithium batteries (S-LIBs), and the result shows that the proposed model has smaller modeling errors for spatiotemporal dynamics than the other two methods. • A convolution and memory network-based modeling approach is proposed to model the spatiotemporal dynamics in thermal processes. • A deformation convolutional neural network strategy is developed to extract the spatial features using a deformable kernel matrix. • A cascade long-short time memory network model is constructed for the temporal dynamics of the time-series data. • The effect of the modeling approach is verified by the experiment of serial-connected lithium batteries (S-LIBs). [ABSTRACT FROM AUTHOR]

This paper presents a novel architecture, termed Fusion-Fission Optimisation (FuFi) based Convolutional Neural Network with Bi-Long Short Term Memory Network (FuFi-CNN-Bi-LSTM), to enhance state of charge (SoC) estimation performance. The proposed FuFi-CNN-Bi-LSTM model leverages the power of both Convolutional Neural Networks (CNN) and Bi-Long Short Term Memory Networks (Bi-LSTM) while utilizing FuFi optimization to effectively tune the hyperparameters of the network. This optimization technique facilitates efficient SoC estimation by finding the optimal configuration of the model. A comparative analysis is conducted against FuFi Algorithm-based models, including FuFi-CNN-LSTM, FuFi-Bi-LSTM, FuFi-LSTM, and FuFi-CNN. The comparison involves assessing performance on SoC estimation tasks and identifying the strengths and limitations of models. Furthermore, the proposed FuFi-CNN-Bi-LSTM model undergoes rigorous testing on various drive cycle tests, including HPPC, HWFET, UDDS, and US06, at different temperatures ranging from -20 to 25 degrees Celsius. The model's robustness and reliability are assessed under different real-world operating conditions using well-established evaluation indexes, including Relative Error (RE),Mean Absolute Error (MAE), R Square (R 2) , and Granger Causality Test. The results demonstrate that the proposed FuFi-CNN-Bi-LSTM model achieves efficient SoC estimation performance across a wide range of temperatures at higher and lower ranges. This finding signifies the model's efficacy in accurately estimating SoC in various operating conditions. • FuFi-CNN-Bi-LSTM combines CNN and Bi-LSTM with FuFi optimization for SoC estimation. • Optimal hyperparameter tuning in FuFi-CNN-Bi-LSTM enhances SoC estimation accuracy. • Comparison with FuFi-based models shows FuFi-CNN-Bi-LSTM's superior performance. • Tested across HPPC, HWFET, UDDS, US06 cycles, proving robustness in diverse conditions. • Achieves accurate SoC estimation across temperature ranges, demonstrating reliability. [ABSTRACT FROM AUTHOR]

To facilitate the ongoing transition to fourth-generation district heating systems, it is necessary to find resource-efficient methods to model the differential pressure in a district heating network accurately. This paper has developed, tested, and compared data-driven methods to create a soft sensor for a pressure transmitter in the district heating network. The sensor is used to control the district heating load and therefore, is the most critical spot of the sensor network. The data-based modelling approaches chosen were transfer functions and neural networks. The data set was collected from Hafslund Oslo Celsio's historical database for January–March 2021, when the heating demand is highest. The best convolutional neural network and a first-order transfer function give acceptable results in estimating the pressure transmitter signal. Both models have the simplest architectures within their model type, suggesting that the need for complex models in either approach is redundant. [Display omitted] • Modelling approaches in deriving a soft sensor for the differential pressure in a district heating system. • Full description of the process of deriving a purely data-driven model for a pressure transmitter soft sensor. • Presents a soft sensor for a pressure transmitter with an accuracy of 0.018 (Convolutional Neural Network) and 0.021 (System Identification). [ABSTRACT FROM AUTHOR]

Time-based pricing for residents is regarded as a critical component to facilitate integration of renewable energy, but it remains one of the significant barriers to target high potential customers. Randomly recruiting customers into time-based pricing may lead to customers' bad experiences and utility companies' economic losses. Focusing on this issue, this paper proposes a new customer selection framework for time-based pricing using the smart meter data. To utilize smart meter data effectively, a novel feature engineering procedure is designed, covering both of statistical characteristics and load patterns. Then, a Bayesian neural network (BNN) based prediction model is developed by integrating the Bayesian theory into neural networks for identifying high-potential customers. Finally, the prediction uncertainty is quantified to help utility companies to identify when to trust the prediction results. Case studies on an actual time-based pricing pilot demonstrate the effectiveness of the proposed framework. As for the comprehensive metric F1-score, BNN is in the lead, which is 9.0 % higher than neural network, 6.0 % higher than convolutional neural network (CNN), and 3.3 % higher than long short-term memory (LSTM). In addition, by leveraging the quantified uncertainties, the percentage of low-potential households among selected customers drops by 8.92 %, compared to blindly trusting the prediction results. Furthermore, it's found that load pattern features contribute more to model performance than statistical features, revealing that it's the shape, rather than the amplitude, of daily load curves that determines residential price responsiveness. © 2017 Elsevier Inc. All rights reserved. • A novel feature engineering approach is proposed to dig out information from smart meter data. • A Bayesian neural network-based prediction model is developed to identify customers with great price responsiveness. • The prediction uncertainty is quantified to determine when to trust the prediction results. • The effectiveness of the proposed customer selection framework is validated on an actual time-based pricing dataset. [ABSTRACT FROM AUTHOR]

The integrated energy system plays an important role in the energy conservation, emission reduction and the resource-efficient utilization. Accurate load forecasting is a significant basis for the optimal scheduling of the integrated energy system. The integrated energy system has coupling interaction between different energy sources in production, distribution and storage. However, the traditional method cannot effectively extract multi-scale features and utilize the coupling information between the multivariate energy sources. In this paper, a novel multi-scale fusion convolutional neural network integrating the bi-directional long short-term memory network and multi-domains hierarchical decoding is proposed to extract and analyze multivariate load data coupling in the integrated energy system data. The multi-scale fusion convolutional neural network is constructed by the multi-dimension convolution layer to obtain multi-scale feature of the integrated energy system data. Meanwhile, the bi-directional long short-term memory network is applied to extract the time dependencies of the integrated energy system data. Finally, the multi-domains hierarchical decoding extracts the coupling characteristics of different domains to predict multiple domains values. Compared with the backpropagation neural network, the support vector machine, the short and long-time memory network, the bi-directional long short-term memory network, and the convolutional neural network - bi-directional long short-term memory network, the proposed method achieves state-of-the-art results in terms of the mean absolute percentage error with 0.365 %, the explained variance score with 99.987 % and the R2-score with 99.984 %, which proves the effectiveness of the proposed model in the load prediction of the integrated energy system. In addition, the proposed method can provide operational guidance for energy production and storage. Through the operation guidance of the proposed method, the carbon emissions are reduced by 238791.58 kg every week. [Display omitted] • Novel multi-scale fusion convolutional neural network is proposed. • Load prediction of integrated energy systems for energy saving and carbon emission is built. • The proposed method achieves state-of-the-art results in terms of the MAPE, the EV and the R2-score. • The proposed method can provide operational guidance for energy production and storage. • The carbon emission is reduced by 238791.58 kg every week. [ABSTRACT FROM AUTHOR]

Capturing the lifespan trajectory of lithium-ion (Li-ion) batteries in the early stage is critical for the operation and maintenance of battery energy storage systems (BESSs). Recently, data driven model is a promising solution to implement this task, yet the battery's early cycling stage can only provide very limited information in the training phase and the scenarios of the BESS applications are not immutable. To alleviate the above issues, this paper proposes a multi-scenario transferable learning framework with few-shot to predict the Li-ion battery lifespan trajectory. An easily constructed model is chosen to generate various pseudo trajectories from only the full lifespan trajectory of a single cell. Then, a transferred deep learning method integrating the gate recurrent unit (GRU) and one-dimensional convolutional neural network (1D CNN) is proposed to utilize the pseudo curves, whose training ability can be adjusted to a new scenario using simply the first 100 cycling data and a pre-trained strategy. Finally, the proposed framework can accurately predict the whole Li-ion battery aging trajectory for multi-scenarios in two different datasets. • A multi-scenario transferable learning framework is proposed to predict the lifespan trajectory of lithium-ion battery. • Various pseudo trajectories are generated by an easily constructed model from only one cell. • A transferred deep learning method is proposed to utilize the pseudo curves. • The proposed framework can accurately predict the battery aging trajectory in two different datasets. [ABSTRACT FROM AUTHOR]

Photovoltaic (PV) power generation exhibits significant variability due to the unpredictable nature of solar energy and volatile weather conditions. This paper introduces a novel integrated model that combines parallel Bi-directional Long Short-Term Memory (BiLSTM) and Convolutional Neural Network (CNN), utilizing multimodal decomposition. The proposed model provides precise photovoltaic (PV) forecasts, essential for optimizing short-term dispatches and scheduling in PV power stations. Firstly, Pearson correlation coefficient is employed to assess the correlation between meteorological data and PV power. Variational mode decomposition (VMD), complementary ensemble empirical mode decomposition (CEEMD), and singular spectrum analysis (SSA) are utilized to decompose the highly correlated features including global radiation and radiation global title. Secondly, employing PV power as output, this study introduces sequences from decomposition methods, temperature, humidity, diffuse radiation, wind direction, and tilted diffuse radiation into the training of the Parallel BiLSTM-CNN (PBiLSTM-CNN) network. Finally, the feasibility of the proposed method is demonstrated by example verification and comparative analysis with alternative methodologies. By employing multiple decomposition methods to extract features, the PBiLSTM-CNN model achieves an average accuracy improvement of approximately 19 % and 37 % in different weather conditions and seasons. Moreover, the implementation of PBiLSTM-CNN results in an enhanced forecasting accuracy of about 48 % and 23 %. • Applying 3 decomposition methods to obtain more informative feature. • a parallel BiLSTM-CNN network is proposed with good capacity. • The mentioned method has good generalization and accuracy. [ABSTRACT FROM AUTHOR]