Search

In the smart grid, residential inverter air conditioners (AC) with significant demand response (DR) potential due to their load flexibility and as the major contributors to peak electricity, need to be grid-responsive to relieve power supply-demand imbalance and ensure thermal comfort. Model predictive control (MPC) has strong capabilities for unlocking the flexibility of residential buildings to realize DR by responding to electricity prices. However, the high computational requirements and complex control system integration processes make the application of MPC a significant challenge. A hierarchical nonlinear MPC (HNLMPC) is developed to realize grid-responsive control for residential inverter ACs by responding to real-time electricity price signals. The controller consists of three parts: the upper-level supervisor MPC, the lower-level optimal PID controller, and the signal converter. The indoor air temperature is selected as the optimized setpoint sequence passed from the upper level to the lower level. A nonlinear prediction model is developed considering the dynamic performances of the inverter AC and the coupled thermal response of an air-conditioned room. A test platform is constructed using Simulink and Simscape to access the DR performance of HNLMPC by comparing it with different rule-based control methods, hierarchical linear MPC, and centralized MPC. The control results show that HNLMPC can achieve peak load shifting and peak shaving without sacrificing thermal comfort by adjusting the room temperature to charge and discharge cooling for the building's thermal mass. Additionally, it enables plug-and-play capability for practical applications, reducing the dependency on local computing power and the need for accurate models. Compared to basic rule-based control, HNLMPC reduces peak-hour energy consumption by 31.6% and total electricity costs by 14.3% over the entire cooling season. • A hierarchical nonlinear MPC is developed for inverter air conditioners. • The prediction model considers the coupled response of an air-conditioned room. • The HNLMPC realizes plug-and-play and has strong robustness. • The total peak-hour energy saving is 31.6% for the cooling season. [ABSTRACT FROM AUTHOR]

Factors such as high penetration of renewable energy, load, geographic location, and interactions between price zones make accurate electricity price forecasting (EPF) very challenging, especially day-ahead electricity price forecasting (DAEPF). To address the issue, A spatio-temporal graph neural network prediction model based on multi-view fusion is proposed in this paper, which learns and analyzes distance relationships, price correlations, and similarities in price distributions across multiple regions, four kinds of graph matrix are constructed to represent the complex spatio-temporal interaction in electricity market. To realize information aggregation between multiple perspectives, a novel multi-view fusion module (MVF) is proposed, which actively mines and utilizes the correlation between nodes within the graph and nodes across the graph through spatial attention and graph attention mechanism, and a temporal embedding module is proposed. The temporal information between nodes is represented by multi-head temporal attention mechanism and the time dependence of multiple receptive fields is obtained by multi-scale gated convolution. Massive experiments are conducted on multiple price zones in the European power market with a high proportion of new energy sources. The results show that MVF can effectively integrate multiple scenario information and improve the prediction accuracy of the network, and the proposed combined network has significant advantages over other models involved in this study. • A novel spatio-temporal graph model based on multi-view fusion is proposed. • Multiple adjacency matrices are constructed from multiple perspectives. • Spatial-graph attention is used to focus on each node and graph feature. • A novel spatio-temporal prediction model is proposed by combining MVF module with temporal series embedding module. • The proposed method is effective in multi-region electricity price prediction. [ABSTRACT FROM AUTHOR]

The present work introduces a general and novel quality-of-experience-aware energy management system. The said system is designed to be responsible for supervising the operation of a smart house, where it accounts for both economic performance and demand management (DM) actions taking into account user comfort, and adopting a quality of experience (QoE) metric. Considering the availability of distributed generation (DG), smart houses are taken as hybrid nano-grids (NGs), and the Energy Management System (EMS) works as a Nano-Grid-Central Controller. Part of the energy storage in this NG is done using renewable hydrogen, which results in a reduction of pollutant emissions. A Model Predictive Control (MPC) algorithm is the foundation for the proposed smart-house EMS, and its formulation as a mixed-integer quadratic programming (MIQP) optimization problem is given, which avoids the use of nonlinear optimization tools. Validated by simulation, the system achieves the required standards: runs the smart house for a year with a 21% electricity bill reduction and 77% reduction in user discomfort. • A quality-of-experience-aware model predictive control energy management system-EMS. • The EMS accounts for economic performance, demand management and user comfort. • Part of the energy storage is done using renewable hydrogen, reducing emissions. [ABSTRACT FROM AUTHOR]

Heating Ventilation and Air Conditioning (HVAC) are energy-intensive systems that greatly contribute to peak demand, which can cause stability and reliability issues in the grid. The use of adaptive smart temperature controllers combined with demand response programs play an important role in addressing this challenge. However, deploying such controllers is difficult, mainly due to the need for detailed models of buildings which can be expensive to acquire. This work proposes a general purpose approach to alleviate this problem through the use of continuous-time neural differential equations models to predict the temperature and HVAC power usage. A fully data-driven approach is used to train the models, thus requiring no prior knowledge about buildings apart from metering data. Therefore, this approach is easily adaptable to different buildings. Furthermore, we show that we can adapt such models to incorporate high-level prior knowledge about building physics to further enhance their efficiency and interpretability. A planning algorithm embedded in an MPC framework is designed to control the temperature setpoint to limit the HVAC power consumption and increase eligibility to a demand response program. Extensive empirical tests are conducted on simulation and real data to assess each model's performance. The experiments show that continuous-time models are more sample-efficient and robust to missing and irregular observations. Our experiments also reveal that for the same level of planning accuracy, continuous-time models require fewer samples than their discrete-time counterparts when adjusting temperature setpoints to lower power consumption during demand response events. • We develop continuous-time models based on Neural Differential Equations to forecast the indoor temperature and HVAC power consumption in buildings. • Models are fully data-driven and include a physics-informed structure. • An empirical comparison shows that continuous-time models are more sample efficient than discrete-time models. • Models are embedded in a model predictive control algorithm to efficiently lower the HVAC power consumption during demand response events. [ABSTRACT FROM AUTHOR]

The transition to renewable energy sources, particularly sources like wind and solar induces a dependency on weather in the supply side of electrical grids. At the same time, the move to electric mobility with uncontrolled charging induces extra peak loads on these grids. These developments can cause grid congestion or an imbalance between the renewable power supply and the demand. Locally balancing the power supply and demand in grid-connected microgrids can alleviate such issues on the main grid. This paper presents a model based control strategy to address the challenge of locally balancing the power supply and demand in a grid-connected microgrid to avoid reaching the threshold rated power output set for large buildings. The microgrid under consideration consists of photovoltaic power sources and a large fleet of electric vehicle chargers (> 150). A model predictive controller is developed that treats the daily vehicle charging as a batch process. Given vehicle charge objectives, the controller utilizes vehicle charger occupancy and photovoltaic power generation forecasting services to distribute power optimally over a fixed period of time. The optimization problem is formulated as a quadratic programming problem and is implemented utilizing open-source Python libraries. The controller was integrated into the control system of a microgrid situated at a corporate office in the Netherlands. The control system oversaw the operation of 174 vehicle chargers. The effectiveness of the model predictive control technology was demonstrated over a three-week period and led to an average daily grid peak power reduction of 59%. • A model-based control strategy is presented in a grid-connected microgrid. • The microgrid under consideration consists of PV and 174 vehicle chargers. • The controller's strategy minimizes power grid peaks and maximizes EV charge energy provided by the PV cells. • The controller was integrated and implemented in the microgrid system. • An average daily grid peak power reduction of 59% was achieved over a three week period. [ABSTRACT FROM AUTHOR]

This paper introduces an innovative approach for achieving optimal energy management (OEM) in renewable-based hybrid microgrids (HMGs). The HMGs exhibit enhanced efficiency, minimizing power loss while maximizing generated power, making them advantageous over traditional MGs. This new approach consists of a modified model predictive control based improved Firefly1to3 algorithm (MMPC-IFA1to3). Here, MMPC adeptly regulates converter switching phenomena, ensuring stability during uncertain conditions in HMG. On the other hand, FA1to3 states that for each member movement of a firefly population, occurs towards three members. Here, the improved nature of FA1to3 is achieved by using the chaotic function to fine-tune the regulation parameter. This unified method generates the best power references for both active and reactive powers, making switching operations in voltage source converters simpler. This paper mainly emphasizes active and reactive power management through objective function minimization. The proposed IFA1to3 approach effectively incorporates constraints to minimize costs, ensure power availability, and mitigate voltage deviations in renewable-based HMGs. Moreover, charging/discharging of energy storage devices and power exchange between the utility grid and DC MG are carried out through the MMPC-IFA1to3 algorithm by monitoring active and reactive loads simultaneously. The corresponding converters are modeled based on the bidirectional power flow to cope with active and reactive power management. Further, the robustness of the proposed approach is verified under different operating conditions, and outcomes are compared with benchmark techniques in a MATLAB/Simulink environment to evidence its superior EM in HMG. Finally, the proposed MMPC-IFA1to3 approach is validated in a real-time environment through the dSPACE DS 1104 embedded processor to evidence its industrial applications through its robustness and OEM during various case studies. • In the hybrid MG, MPC regulates converter of the connected sources, while MMPC controls the VSC converter. • A novel FA, combining FA1to3 with chaotic logistic function, optimizes power references generation in this study. • Proposed MMPC-IFA1to3 approach ensures voltage stability and effectively manages ARP during uncertainties in MG. • The proposed IFA1to3 is more effective for multi-objective functions optimization. • Real-time validation of the proposed approach in dSPACE DS1104 evidences its practical applicability. [ABSTRACT FROM AUTHOR]

Lithium plating in lithium-ion batteries for electric vehicles, occurring due to low-temperature or high-rate charging, is a significant factor impacting safety and service life. To address this issue, a novel adaptive charging approach is proposed, combining ultrasound-assisted diagnosis and model predictive control (MPC). In the method, a discrete state-space electrochemical model is used to describe the dynamic characteristics of the battery, and a model predictive controller (MPC) is utilized to optimize the charging current to avoid lithium plating. Considering that factors such as battery performance degradation and variable working temperature affect the battery model's judgment of lithium plating, an ultrasound-assisted diagnosis method is used to determine the critical point of lithium plating. The effectiveness of the method is validated through low-temperature charging and cycle aging experiments. The results indicate that without complex model parameter calibration in different temperatures, the new charging method not only has a higher charging speed than constant current charging, but also can effectively suppress the occurrence of lithium plating on the negative electrode of the battery. The method is expected to be applied in electrochemical energy storage systems to enhance safety and service life. • Simplified battery electrochemical model is designed for charging control. • Model predictive controller is utilized to optimize the charging current. • Ultrasound-assisted diagnosis method is used to diagnose lithium plating. • Model parameters do not need to be recalibrated at different temperatures. [ABSTRACT FROM AUTHOR]

Currently, building energy consumption prediction models are usually based on a large amount of historical operational data in high demands of building operating hours and monitoring systems. However, many buildings may lack operational data due to relatively limited monitoring systems, causing the failure to use data-driven methods to characterize the energy profile. In this context, transfer learning is a promising method to establish the knowledge transfer between many high-quality building operation datasets and a small amount of target building data, and to help predict energy consumption in the target building. This paper studies the possibility of employing transfer learning to achieve both short and long-term building energy consumption prediction. Firstly, a similarity analysis method, based on variable modal decomposition and dynamic time warping, is proposed for identifying the source buildings with the most similar energy features to the target building. Then, transfer learning for long-term prediction air-conditioning energy consumption is developed with weather parameters generated by the Morphing method as inputs. For the short-term single-step prediction, the proposed model CV-RMSE improves 81.3% (AEC) and 77.4% (EEC), respectively, compared to the prediction model that does not implement the transfer learning strategy and directly uses the target BEC data. As for the short-term multi-step prediction, the proposed model CV-RMSE improves 62.0% (AEC) and 65.5% (EEC), respectively. For the long-term prediction, the average CV-RMSE for the whole year is 6.62% and 11.15% for the proposed and directly target domain-based model, respectively. The proposed method explores the practicality of transfer learning in building energy forecasting, contributing to the use of existing building operation data for energy management at different timespan. • A similarity analysis method for building energy consumption data based on VMD-DTW is proposed. • The Seq2Seq-Transformer model is applied to short-term multi-step prediction of building energy consumption. • Transfer learning idea is first introduced to the long-term building energy consumption prediction. [ABSTRACT FROM AUTHOR]

In the current studies on energy management strategy (EMS) for vehicle-following scenarios, the accuracy of vehicle state predictions based on mechanistic models is influenced by the time-varying conditions, affecting the optimization control performance. To address this issue, a data-driven Koopman model predictive control for hybrid energy storage system (HESS) of electric vehicles (EVs) in vehicle-following scenarios is proposed, combining the safety speed planning and energy management strategy. Firstly, a data-driven Koopman vehicle state prediction model is constructed in the upper layer for estimating parameters, such as road surface smoothness and slope. This model is then integrated into Model Predictive Control (MPC) to optimize the speed of the following vehicle. Subsequently, in the lower layer, utilizing the output from the upper layer and predicting load power, the load power is further allocated within the HESS. Simulation results demonstrate that in scenarios considering factors like slope, the hierarchical MPC with the Koopman model reduces energy consumption by 5.55% compared to the hierarchical MPC with mechanistic model. • Data-driven Koopman MPC for EVs under vehicle-following scenarios is proposed. • A Koopman prediction model is developed considering parameters such as slope. • The proposed method reduces energy consumption by 5.55%. [ABSTRACT FROM AUTHOR]

Accurate crude oil price forecasting is crucial for maintaining energy economy stability and enhancing investment and operational decision-making. Nowadays, the link prediction model for crude oil prices based on a complex network model has become an emerging and promising approach. However, existing link prediction methods are unable to extract complex information about interval-valued crude oil prices, and fail to consider the impact of multi-source information. Therefore, this paper proposes the link prediction model for interval-valued crude oil prices using complex network and multi-source information, which converts crude oil price interval series into complex networks. Through link prediction, it takes into account both the impact of past interval-crude oil prices on future interval time series and the interference of external real-time information on interval-valued prices. First, news headlines and geopolitical risk indices related to crude oil prices are captured, and search index data and news sentiment scores associated with interval data are analyzed. Subsequently, the interval sequence data and related influencing factors are decomposed and reconstructed respectively. The reconstructed interval center and radius sequence are then converted into a directed network, and node similarity in the network is calculated to generate similar nodes of the interval subsequence as feature values. Finally, multiple machine learning models are employed to acquire the optimal weighted combination prediction with interval prediction values. According to the experimental data, this approach outperforms other benchmark methods in terms of prediction accuracy. • Predictive modeling to transform interval-valued into complex networks is proposed. • A link prediction model based on multi-source information is constructed. • Combining AI, link prediction, and complex networks forms a multidisciplinary mothed. • The proposed model is well validated in predicting interval WTI crude oil prices. • The constructed model can effectively extract multi-scale data features. [ABSTRACT FROM AUTHOR]

Accurate crude oil price forecasting is crucial for maintaining energy economy stability and enhancing investment and operational decision-making. Nowadays, the link prediction model for crude oil prices based on a complex network model has become an emerging and promising approach. However, existing link prediction methods are unable to extract complex information about interval-valued crude oil prices, and fail to consider the impact of multi-source information. Therefore, this paper proposes the link prediction model for interval-valued crude oil prices using complex network and multi-source information, which converts crude oil price interval series into complex networks. Through link prediction, it takes into account both the impact of past interval-crude oil prices on future interval time series and the interference of external real-time information on interval-valued prices. First, news headlines and geopolitical risk indices related to crude oil prices are captured, and search index data and news sentiment scores associated with interval data are analyzed. Subsequently, the interval sequence data and related influencing factors are decomposed and reconstructed respectively. The reconstructed interval center and radius sequence are then converted into a directed network, and node similarity in the network is calculated to generate similar nodes of the interval subsequence as feature values. Finally, multiple machine learning models are employed to acquire the optimal weighted combination prediction with interval prediction values. According to the experimental data, this approach outperforms other benchmark methods in terms of prediction accuracy. • Predictive modeling to transform interval-valued into complex networks is proposed. • A link prediction model based on multi-source information is constructed. • Combining AI, link prediction, and complex networks forms a multidisciplinary mothed. • The proposed model is well validated in predicting interval WTI crude oil prices. • The constructed model can effectively extract multi-scale data features. [ABSTRACT FROM AUTHOR]

Major nations across the globe are increasingly concerned about the rising trends in carbon dioxide (CO 2) emissions, particularly in societies of varying scales. Against this backdrop, precise prediction of carbon emissions becomes critically important, especially for the formulation and adjustment of near-term carbon reduction policies. However, the non-linear, non-stationary, and complex nature of daily carbon emission data poses a great challenge for daily-level forecasting especially in the big data context. To address this issue, we propose a novel composite forecasting approach named DCEF dedicated to the estimation of daily carbon emissions. In concrete, our approach employs the Empirical Mode Decomposition (EMD) for data stabilization and the Auto-regressive Integrated Moving Average (ARIMA) model for forecasting, while integrating the Truncated singular value decomposition(TSVD) technique for data compression and mitigating noise. Finally, DCEF is empirically validated with real daily carbon emission datasets collected from 6 sectors in 13 countries of varying sizes. Experimental results demonstrate the advantages of our approach compared to other baseline models in terms of prediction accuracy and efficiency. • Establish a model that is capable of daily CO 2 emissions forecasting in 6 sectors. • Improves prediction accuracy and efficiency with the proposed ensemble model. • Model performance consistently outshines competitors in 13 country datasets. [ABSTRACT FROM AUTHOR]

Accurate machine learning (ML) predictions for the early stages of the building design are crucial to construct energy-efficient buildings utilizing limited resources. Several studies have employed ML methods for energy consumption (EC) prediction without considering the utmost crucial PCM-integrated building design parameters. In addition, reducing the dataset by considering only the most significant building design parameters before applying the ML-based method would be beneficial for reducing the computational power and memory usage of the system as well as utilizing less time in the modelling process. To this end, this research presents a novel framework to establish the most robust and reliable ML-based prediction model with less complexity, considering only the most influential PCM-integrated building design parameters. These parameters were identified for future scenarios of hot semi-arid (BSh) climate zones using multi-stage sensitivity analysis. Afterward, a reduced EC database based on the most significant building's early-design-stage parameters (EDSPs) was utilized to formulate several multi-expression programming (MEP) and support vector machines (SVM)-based forecasting models, considering the variations in their hyperparameter values. Formulated prediction models have shown less time utilization through the training and testing phases for the EC evaluations of selected PCM-integrated building compared to the physical-modelling process. Several statistical parameters were used to test and validate the performance of the formulated prediction models. The acquired model evaluation and validation results demonstrated that the MEP-based prediction model (MEP 15) exhibited the highest level of reliability and accuracy, showing an R2 value of >95% for both the training and testing phases. The model's interpretability showed that, throughout the parametric analysis, the developed prediction model adhered to the system's physical boundary constraints. Also, the best-performing prediction model showed energy savings of up to 12% for building integrated with PCM having a melting temperature of 28 °C. Conclusively, this research demonstrated that the developed MEP-based prediction model could be employed to precisely forecast the EC for selected PCM-incorporated building in the whole BSh climate, considering important EDSPs while utilizing minimal resources. • Proposed a novel machine learning based framework using MSA. • MSA identified influential design parameters and reduced dataset substantially. • Formulated several MEP and SVM-based prediction models on reduced dataset. • Performed parametric and energy savings analysis for best prediction model. • Best prediction model (MEP 15) with R2 > 97% showed energy savings of up to 12%. [ABSTRACT FROM AUTHOR]

The accuracy of offshore wind power forecasting (OWPF) is the basis for guaranteeing the safe dispatch and economic operation of power systems, which can reduce the technical and economic risks faced by power market participants. Combining the theoretical approaches of data space and deep learning, this study proposes a hybrid short-term OWPF framework based on the integrated consideration of spatial and temporal correlations of regional field clusters in the data space and the 3DTCN-CBAM-LSTM model. Firstly, other offshore wind farms with spatio-temporal correlation with the target offshore wind farm are screened out using the K-means spatial clustering algorithm. The historical power sequences determined to be similar to the target offshore wind farm and feature data of target offshore wind farms are used as inputs to the prediction model to improve the efficiency of the framework's power prediction. Secondly, the short-term power prediction framework of offshore wind power based on the 3DTCN-CBAM-LSTM model is constructed, using the 3DTCN model's ability to capture the spatio-temporal dynamic characteristics of offshore wind farm data to improve the accuracy of the prediction model. The temporal characteristics of the data are strengthened using the LSTM model to improve the stability of the prediction model. Finally, the CBAM mechanism is introduced to combine channel attention and spatial attention to eliminate irrelevant information and capture the spatio-temporal characteristics of the data in a more targeted way, which further improves the accuracy of OWPF. To verify the practicality and reliability of the wind power prediction framework described in this study, the differences between the prediction results of the framework and those of the benchmark model are compared using single-step prediction and multi-step prediction (4, 6, 8, and 12 steps). Compared with the benchmark model (BPNN), the single-step prediction results show that the MAE, MAPE, and MSE metrics of the 3DTCN-CBAM-LSTM model are reduced by 20.004 MW, 50.13%, 55.61%, and 80.29%, respectively, and the R2 is improved by 3.39%, which is a substantial improvement in prediction performance. The multi-step prediction results show that, with the progression of the prediction steps, the 3DTCN-CBAM-LSTM model has the highest prediction accuracy and model fit at each step. Therefore, the proposed 3DTCN-CBAM-LSTM model presents obvious advantages and reliability in the prediction of offshore wind power multi-step power. • A data-driven model considering the spatio-temporal correlation is adopted for OWPF. • K-means spatial clustering algorithm is used to screen offshore wind farms. • A 3DTCN is used to capture spatio-temporal dynamic properties of offshore wind farm data. • The CBAM is introduced to improve the model's ability to capture the spatio-temporal features of offshore wind power data. • The superiority of the proposed prediction method is demonstrated using single-step and multi-step predictions. [ABSTRACT FROM AUTHOR]

To improve the granularity of load decomposition within integrated energy system (IES), a novel multiple load forecasting approach is proposed. This method leverages an aggregated hybrid modal decomposition (AHMD) strategy and a composite model approach to thoroughly investigate the inherent characteristics of load profiles. Improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (ICEEMDAN) is utilized for the primary decomposition of electric, cooling, and heating loads, yielding a series of subsequences. To further refine the analysis, Fuzzy Dispersion Entropy (FDE), and Over-zero Rate (OZR) are employed to aggregate the decomposed subsequences. Subsequently, Successive Variational Mode Decomposition (SVMD) is applied to break down the high-frequency components in a secondary manner. The adoption of the rolling decomposition strategy effectively circumvents the information leakage problem caused by traditional modal decomposition-prediction methods. In concert with environmental features, the Multiple Linear Regression (MLR) model forecasts the low-frequency components. Meanwhile, the Temporal Convolutional Network-Bidirectional Gated Recurrent Unit (TCN-BiGRU) fusion neural network is deployed for predicting mid-frequency and high-frequency components. Additionally, a Multi-Head Attention (MHA) mechanism is introduced to effectively assign weights to the features. The predictions of each component are superimposed to obtain the final forecast. An example of an integrated campus energy system is used to compare the predictive effectiveness of different models. The results show that the proposed method exhibits superior prediction accuracy across multiple load types. This approach significantly enhances energy efficiency and fosters the advancement of sustainable energy practices. • Quadratic decomposition is used to reduce data complexity. • Rolling decomposition strategy avoids information leakage. • Aggregation strategy balances prediction time and accuracy. • Different prediction models are applied for load component characteristics. [ABSTRACT FROM AUTHOR]

Predictive solar-desalination models are becoming more widespread using ML and AI. However, forecasting solar still water productivity based on numerous designs still needs to be improved. Herein, we used transfer learning to create precise supervised predictive ANN regression models for water productivity (L/m2.day) predictions based on literature findings. Such observation datasets from single-basin solar stills were utilized to build the random initialization ANN model. The transfer learning method was applied to the latter model by taking the learned network (weights) for fine-tuning the hyperparameters from the earlier developed novel hybrid solar still known as the source (pre-trained) ANN model, to predict the target ANN model. Based on most minor statistical errors, the pre-trained model with 5–64–64-1 architecture and ReLU activation function was the most appropriate for water productivity prediction. All created ANN models were compared to the MLR model. The results revealed that the generated target ANN model outperformed the ANN RI and MLR with OI values of 0.872, 0.834, and 0.803, respectively, in all modeling stages. The target ANN model's accuracy and generalization were sufficient. The target ANN model had residuals of forecasted distillate values of around 1%. This work discusses the significance of transfer learning to generate accurate target ANN models for predicting freshwater outputs in single-slope solar stills, which can be integrated with established theoretically tuned parameters to enhance performance and maximize distillate water yields. • Leveraging Transfer Learning: This study enhances distillate forecasts in solar desalination through predictive ANN regression models, utilizing transfer learning. • Optimal Model Architecture: The pre-trained ANN model (5–64–64-1,ReLU)proves optimal for water prediction, showcasing transfer learning efficacy. • Superior Performance: Comparative analyses highlight the target ANN model over RI ANN and MLR models, with higher OI values and lower residuals. • Robust Generalization: The target ANN model exhibits exceptional accuracy and robust generalization for forecasting freshwater outputs. • Practical Implications: This research aids expense optimization, enhances water production, and impacts solar desalination research and industry. [ABSTRACT FROM AUTHOR]

The fast-increasing applications of renewable energy and electric vehicles (EVs) pose significant challenges to the safe and economic operation of the power grid. However, through Virtual Power Plant (VPP) technology, EV fleets can be aggregated into large-scale EV generalized energy storage systems, providing scarce dispatchable flexible resources for the future power system. Rapid and accurate prediction of the Aggregated Schedulable Capacity of EV (EVASC) fleets is a crucial technology for VPP with Electric Vehicles (EV-VPP) to participate in various ancillary services of the power system. In this paper, orienting for various dispatch scenarios in the power system, a multi-time scale prediction model of the EVASC for EV-VPPs is presented based on the enhanced Prophet algorithm combining a light gradient boosting machine (LGBM) algorithm. The enhanced Prophet-LGBM algorithm is initially proposed here by incorporating the LGBM algorithm into the traditional Prophet algorithm in series and adding extra sequential features in its input, such as temperature, weather, and alike. In this way, the problem of overfitting and low capability in addressing additional input features in the traditional Prophet is overcome. The proposed enhanced Prophet-LGBM-based prediction method of the EVASC for EV-VPPs is validated using 1.8 million actual charging records of over 4000 charging piles for half a year at the provincial level. The 30-day simulation results of the day ahead and intraday predictions of the EVASC for EV-VPPs show that much higher performances in the prediction accuracy can be achieved, specifically in holidays. • An electric vehicle schedulable capacity forecasting model is presented. • An enhanced Prophet plus a light gradient boosting machine is proposed for the model. • The model can forecast the multi-time scale electric vehicle schedulable capacity. • The model is validated using 1.8 million charging records of 4000 piles • The prediction simulations show much higher accuracy, specifically in holidays. [ABSTRACT FROM AUTHOR]

With the rapid development of proton exchange membrane fuel cell (PEMFC) commercialization, a comprehensive knowledge of multi-physics fields in large-scale PEMFC stacks has become ever more critical. Although conventional three-dimensional computational fluid dynamic (CFD) models have achieved great success, the application in the commercial-size stack-scale simulation remains inapplicable due to enormous computational resource requirements. Herein, based on the latest 3D CFD model, multi-physics digital twin (DT) technology and 3D stack flow distribution prediction model, a novel multi-scale upscaling prediction model is proposed. The voltage, water and thermal management characteristics of a 164-cell PEMFC stack with an active electrode area of 292.5 cm2 are studied and analyzed in details. For the analysis of commercial-size PEMFC stacks, the most comprehensive multi-physics fields are covered in this paper to date. And the results suggest that by introducing the DT technology, the time requirement of the multi-physics field prediction for unit scale prediction can be reduced by hundreds of thousands of times with a maximum global relative deviation of 1% under 10 groups of random test conditions, giving a solution from the cell scale to stack scale performance prediction, design, heat and thermal management in the PEMFC research and application. • A novel 3D + Digital twin+3D upscaling strategy for PEMFC stack is proposed • The multi-physics fields in a commercial-size PEMFC stack are analyzed • The upscaling strategy can save plenty of computational resource [ABSTRACT FROM AUTHOR]

Cloud computing is an integral component of modern IT infrastructure and addressed as an energy consumer due to its increasing utilisation. Cloud workload prediction is regarded as an effective method that can assist cloud service providers with more appropriate resource scheduling, thereby increasing the overall resource utilisation with reduced energy wastage. Herein, accurate prediction models are pivotal to the success of prediction driven energy efficient datacentre management. Despite existing prediction models for cloud datacentres, their accuracy and dependability under limited computational resources still remains a concern due to the resource intensive nature of large prediction models. This study aims to propose different cloud workload statistical methods tailored to the computational resource limitations of various analysts and develop a more comprehensive and accurate cloud computing prediction model based on advanced time-series prediction models to help cloud service providers optimise resource utilisation and reduce energy consumption. Accordingly, we first propose two designs to statistically analyse the cloud workloads: one that is more accurate but consumes more computational resources and the other that simplifies the computation process to require fewer resources while maintaining a certain degree of accuracy. Second, we develop the Mixed Channel Time-series Dense Encoder (MC-TiDE) to efficiently learn information between different time-series, based on the Time-series Dense Encoder (TiDE) model. Experiments conducted on real-world cloud trace logs (including the Alibaba 2018 and Google 2019 datasets) show that our proposed MC-TiDE model outperforms other notable prediction models, demonstrating its prediction accuracy while ensuring efficient training and inference processes. • Propose two statistical methods for cloud workload analysis tailored to resource limits. • Develop a Mixed Channel Time-series Dense Encoder (MC-TiDE) model for accurate cloud predictions. • MC-TiDE efficiently learns information between different time-series datasets. • Experiments show MC-TiDE outperforms other models in prediction accuracy and efficiency. [ABSTRACT FROM AUTHOR]

The large-scale penetration of intermittent sources brings serious stability problems to power systems with multiple energy complementary characteristics. The regulation performance of pumped storage plant (PSP) in suppressing frequency fluctuations induced by intermittent energy sources is essential to a multi-energy system. This paper aims to design an optimal control strategy for the pump-turbine governing system (PTGS) for a faster and smoother regulation process. To precisely describe the complicated hydraulic-mechanical transients in PTGS, a novel complete state-space model (CSSM) is proposed based on the electrical analogy method. The mathematical expressions of the subsystems in PSP are developed, and the simulation algorithm are detailed. The accuracy of the proposed model under various conditions is revealed by comparison with the method of characteristics (MOC) and measured data. Furthermore, the temporal-spatial constraint of CSSM when dealing with pipe segmentation is discussed. Subsequently, a linear quadratic regulator with anti-disturbance optimal tracking (LQR-ADOT) capability is designed based on the proposed nonlinear model. The integral of rotation speed deviation and the augmented state vector are employed to eliminate the steady-state error caused by the disturbance term and solve the trajectory tracking problem, respectively. Comparative experiments of LQR-ADOT with proportional-integral-differential (PID), fractional order PID (FOPID) and model predictive control (MPC) reveal that the proposed controller can effectively attenuate rotational speed oscillation and improve overall performance under various conditions. Notably, in frequency regulation, the settling time is decreased from 28.83 s to 16.82 s and the overshoot is reduced from 9.45% to 0.31% compared to traditional PID. • A novel state-space model of the PTGS, addressing pipelines and pump-turbine nonlinearities, is proposed. • A LQR with anti-disturbance and optimal target-tracking capability is designed. • The regulation quality of PTGS is analyzed comparatively in different scenarios. • The accuracy of the CSSM is validated by comparison with measured data and the MOC. • The computation complexity and efficiency of the CSSM is discussed. [ABSTRACT FROM AUTHOR]

With the increasing penetration of photovoltaic (PV) power into the grid, the impact of PV power on the stable operation of the grid is becoming increasingly significant. Our aim is to design a multi-step PV power forecasting method to reduce the uncertainty associated with grid-connected PV operation based on the time and accuracy requirements of real-world forecasting scenarios. A novel data augmentation method is introduced and a new model PTFNet (Parallel Temporal Feature Information extraction Network) is designed to achieve more accurate forecasting, and the prediction results are evaluated to match the real scenarios. Firstly, noting that past work has rarely considered the particularities of PV data, which are reflected in the physical modeling of PV, we incorporate site-specific and future temporal information by adding several physical modeling intermediate variables to the raw dataset. The input data consists of past measured data and numerical weather prediction (NWP) data. Then, to better establish the relationship between these two types of data and PV power, the proposed model utilizes a parallel structure to extract the temporal dependency and inter-feature dependency of these two types of data and PV power, respectively, and combines these two types of information for forecasting. Our model can produce 24-, 36- and 40-h forecasts with 15-min resolution, which means that these forecasts meet the forecast horizon requirements of most current day-ahead PV power forecasting scenarios. Transformer architecture-based models have been increasingly used in the field of time series forecasting in recent years, and we selected 10 typical long-term forecasting models for comparison, the first time that all of these models have been applied to a PV dataset. We conducted experiments on two public datasets, and the results show that PTFNet achieves the overall most competitive forecasting ability. Finally, to match the actual forecasting scenarios, we intercepted the last 24 h of the 40-h forecasts as the actual forecasts submitted to the grid operator and post-process the forecasts. The proposed model achieved the best results on all evaluation metrics of the post-processed data and the predictions of the model meet the accuracy requirement of the grid operator in most cases. • Data augmentation considering PV physical modeling improves prediction accuracy. • A Transformer-based day-ahead photovoltaic power prediction model is established. • Explicitly extract temporal and inter-feature dependencies of the data for prediction. • Post-process the predictions to obtain output consistent with application scenarios. [ABSTRACT FROM AUTHOR]

Photovoltaic (PV) distributed generators (DGs) are inherently stochastic and have low inertia owing to their weather dependence and connection to an inverter. Frequency regulation presents a significant challenge for the high penetration of PV-DGs to microgrids. Recently, a virtual synchronous generator (VSG) has been proposed to enhance the inertia of microgrids because it mimics the behavior of synchronous generators. However, few studies have utilized VSG-based DGs for secondary frequency regulation owing to their stochastic power output. In this study, a model predictive control scheme integrated with two-layer moving-horizon estimation (TL-MHE) observer is proposed and applied to the secondary frequency control of a PV high-penetration microgrid. The local observer estimates the power disturbances in each PV-DGs, whereas the centralized observer estimates the system states and load disturbances. The two-layer observer design decouples multiple observations and reduce the order of the observation equations. Compared with conventional centralized observer-based controllers, the proposed scheme can obtain better observation results. The model prediction controller optimizes the restorative power from each DG in real time to minimize frequency fluctuations, thus affording better control performance. Simulation results show that the proposed TL-MHE method can accelerate the system frequency recovery, reduce frequency- fluctuations peaks, and reduce frequency deviation by more than 80% compared with the conventional method. • A model predictive control scheme integrated with two-layer moving-horizon estimation observer is proposed. • The two-layer observer design decouples multiple observations and reduce the order of the observation equations. • We use deloading virtual synchronous generator control to make the renewable energy participating in the secondary frequency regulation. • The proposed method overcomes the stochasticity of each generating unit and load. [ABSTRACT FROM AUTHOR]

Rare or missing data pose significant challenges in the prediction of wind power (WP) and photovoltaic power (PV). Many methods address the data scarcity issue solely through augmentation techniques, often neglecting the impact of missing data on the augmentation process. When data augmentation is performed on missing datasets, the prediction accuracy cannot be further improved, and this is called as an extreme data scarcity problem. To solve this problem, we introduce two methods, called Information Maximizing Collaborative Adversarial Variational Bayes (InfoCAVB) and MemoryFormer, to achieve data augmentation based on missing data reconstruction. In this paper, the augmentation and reconstruction process are performed at the same time. When the reconstruction process is performed, InfoCAVB utilizes adversarial training to construct variational bayes that approximates the posterior distribution of real data to recover missing data. Meanwhile, in the augmentation process, InfoCAVB maximizes the mutual information between real data and reconstruction data to establish correspondences between specific dimensions of augmentation data and the features of real and reconstruction data, respectively. The advantage of InfoCAVB lies in incorporating data augmentation into data reconstruction through the information maximizing principle. Finally, we propose a MemoryFormer adapted for InfoCAVB to predict WP and PV, and the benefit of MemoryFormer lies in excavating the potential temporal correlations between reconstructed and augmented data. MemoryFormer embeds the distribution of real data into the generated data through memory units, ensuring consistent distribution during the training process of discrete reconstructed and augmented data. Experimental results indicate that the proposed InfoCAVB-MemoryFormer reduces the RMSE averages by 15.7%, 14.1%, and 5.92% for WP prediction and 28.23%, 22.67%, and 12.21% for PV prediction compared to other state-of-the-art model models, demonstrating the effectiveness of the proposed approach in extreme data scarcity scenarios. • Propose a new data augmentation method InfoCAVB based on missing data reconstruction. • InfoCAVB uses adversarial training to construct variational bayes that approximates the posterior distribution to recover missing data. • InfoCAVB maximizes the mutual information between real data and reconstruction data to establish correspondences. • Propose a prediction model MemoryFormer, which can excavate the potential temporal correlations between reconstructed and augmented data. [ABSTRACT FROM AUTHOR]

The tremendous energy usage from buildings leads to research studies on their improvement, among which advanced building control plays an important role. In advanced building controls, data-driven predictive control (DDPC), differentiable predictive control (DPC), and reinforcement learning (RL) have shown advantages, but their comparison often lacks in existing studies. The simulation-based prior comparison studies have inconsistent results due to different assumptions and simplifications. Therefore, to comprehensively compare the three advanced strategies for real-time building HVAC controls, we implemented DDPC, specifically, hierarchical DDPC (HDDPC), DPC, and RL in a real building testbed for more than 5 months. The results show that all three advanced controls maintained the indoor environmental quality (IEQ) cost-effectively. Overall, HDDPC outperformed the baseline control with more than 50% energy savings, followed by RL with 48%, and DPC with 30.6%. Most control failures were related to API communication issues. Besides, the information gaps between room and system level controllers and non-optimal control decisions will degrade HDDPC's performance. Such degradation did not happen in DPC and RL, which led to better performance of agent-based control over HDDPC. Moreover, HDDPC needs minutes to make control decisions whereas DPC and RL need milliseconds, indicating higher online computing resources required by HDDPC. For agent training, DPC is faster than RL, as DPC training needs minutes and RL needs hours, but its performance is not as good as RL. This study provides a comprehensive understanding and assessment of the pros and cons of advanced building controls and sheds light on future research on building controls. • Real-time implementation of advanced controls in the real building. • Long-term experiment for advanced building controls. • A comprehensive analysis and comparison of different advanced building controls. [ABSTRACT FROM AUTHOR]

Mechanism-data-integrated methods are promising technologies for safe and flexible operation of power stations, which play an important role in compensating for the renewable energy intermittency and fluctuation. As an attempt in this direction, this work is devoted to the tube overheating problem in the boiler with the aim of developing effective predictive methods. First, we obtain insights from the real incidents with excessive metal temperatures. With data collected over a six-month period from a 350-MW cogeneration unit operating in a flexible mode, we quantified 230 overheat events using steam-pressure-dependent permissible limits. Power law distributions with tails are revealed for the severity of overheating, and three types of events can be classified. Among them, the 'moderate-type' overheating can be accompanied with simultaneous overheating of multiple neighboring tubes, and is thus recognized as the primary target for the real-time prediction. Our data-driven model rests on the long-short-term memory neutral network and gives satisfactory outputs under normal operating conditions with a mean absolute error of 3.40 °C. However, the original model fails to reach the extreme values whilst tube overheating due to severe dataset imbalance as only 0.012% of the data correspond to overheating. Finally, we devise an additional strategy making use of the change in the predictive capability of the model. The integrated method successfully predicts all overheat events of a tube over two minutes in advance, and the false alert is kept at a minimum level. • Simultaneous overheating of multiple adjacent tubes is found from operating data. • The severity of tube overheating exhibits power law distributions in frequency. • Data driven model fails to predict tube overheat due to a lack of training samples. • A new strategy is proposed that combines the rarity nature of tube overheating. • Our method predicts tube overheat two minutes in advance with a minimum false rate. [ABSTRACT FROM AUTHOR]

To make full use of the flexible adjustment capability of DC tie-lines in multiarea AC/DC systems and to coordinate the generation resources and load demand of multiarea AC/DC systems, this paper presents an intraday two-stage hierarchical optimal scheduling model for multiarea AC/DC systems based on analytical target cascading (ATC). To avoid repeated adjustment and overadjustment of DC tie-lines after wind power integration, a two-stage rolling coordinated scheduling model for the area subsystem based on model predictive control (MPC) is proposed. The two-stage rolling coordinated scheduling method takes into consideration the influence of the predicted value in the future finite time domain and the latest measured DC tie-line power on the current scheduling state. Based on the ATC and the area decomposition criterion of the AC/DC grid, an optimal scheduling model for the upper-level system is proposed that takes into consideration the DC tie-line adjustment constraints and the area coupling constraints. The optimal scheduling model for the upper-level system formulates the two-stage DC tie-line plan for the multiarea AC/DC system, and the two-stage rolling coordinated scheduling model of the area subsystem solves the subproblems of the generation plan for each area subsystem in a parallel manner considering the area coupling constraints. The proposed method can achieve intraday two-stage decoupling scheduling of multiarea AC/DC systems and promote the cross-area consumption of large-scale wind power through flexible adjustment of the DC tie line. This approach also reduces the communication burden between the area subsystems and ensures the efficiency of the solution algorithm. • Based on analytical target cascading technology (ATC), we integrate the hierarchical control structure of decomposition-coordination into the model predictive control (MPC) method to achieve decentralized dispatch for multiarea AC/DC systems. • A two-stage rolling coordinated scheduling model for the area subsystem considering the area coupling constraints based on the MPC is proposed. • Based on the ATC and the area decomposition criterion of the AC/DC grid, an optimal scheduling model for the upper-level system is proposed. • Flexible adjustment of the DC tie-line and decoupling scheduling of multiarea AC/DC systems are achieved. • The hierarchical optimal scheduling model reduces the communication burden between the subsystems and ensures the efficiency of the solution algorithm. [ABSTRACT FROM AUTHOR]

Recently, battery fast charging strategies have gained increasing interest as range anxiety and long charging time have been the main obstacles to the wider application of electric vehicles. While simply increasing the current can reduce charging time, it might also tend to accelerate the irreversible capacity degradation and power fade. To solve the dilemma between charging speed and battery lifetime, in this work, we proposed a life-extending optimal charging method that considers the charging time and the aging-related effects within the battery. A multi-physics battery model coupled with thermal and electrochemical degradation dynamics is developed and integrated into a model predictive control framework to manipulate the optimal charging current considering the constraints of safe requirements and the rate of internal aging side reactions. The design method was extensively validated by in-situ and ex-situ experiments. Results show that by reducing the rates of side reactions and minimizing detrimental morphological changes in the anode material, the proposed charging method can prolong the battery lifetime by at least 48.6%, compared with the commonly used constant current and constant voltage charging method without obviously sacrificing charging speed. [Display omitted] • An electrochemical-thermal-aging battery model is developed and widely validated. • Battery degradation mechanisms are analyzed at various operating conditions. • Optimal charging profile is designed by considering multiple physical constraints. • The impact of algorithm hyper-parameters is discussed. • Capacity loss is reduced by 48.6% without sacrificing the charging speed. [ABSTRACT FROM AUTHOR]

Dual fluidized bed (DFB) gasification is a promising method for producing valuable gaseous energy carriers from biogenic feedstocks as a substitute for fossil fuels. State-of-the-art DFB gasification plants mainly rely on manual operation or single-input single-output control loops, and scientific contributions only exist for controlling individual process variables. This leaves a research gap in terms of comprehensive control strategies for DFB gasification. To address this gap, we propose a multivariate control strategy that focuses on crucial process variables, such as product gas quantity, gasification temperature, and bed material circulation rate. Our approach utilizes model predictive control (MPC), which enables effective process control while explicitly considering process constraints. A simulation study is given demonstrating how different MPC parametrizations influence the behavior of the closed-loop system. Experimental results from a 100 kW pilot plant at TU Wien demonstrate the successful control achieved by the proposed control algorithm. [Display omitted] • Model predictive control for product gas quantity and gasification temperature. • Consideration of process constraints such as remaining oxygen in the flue gas. • Control structure composed of a high-level MPC and a bed material circulation MPC. • A simulation study compares and evaluates different controller parametrization. • Experimental results are shown for a 100 kW pilot plant at TU Wien. [ABSTRACT FROM AUTHOR]

This paper presents a refined quantum Harrow Hassidim Lloyd (HHL) algorithm for microgrid control. The first novelty of the developed method is that a mapping shift function enables the original HHL algorithm to handle general linear equations with non-singular and indefinite matrix. Second, a method of Matrix Extension for Amplifying Sampling Probabilities of Intended Solution (ME-ASPI) is proposed to design the reformulated linear algebraic equations, allowing for improved sampling efficiency of the quantum tomography in the refined HHL algorithm. Then, we applied the method to solve the model predictive control (MPC) problem in nonlinear dynamical microgrids. Specifically, with the ME-ASPI method, the refined HHL algorithm can effectively obtain the intended partial optimal control inputs for MPC. The optimization of quadratic programming problem in each time step of MPC is transformed into a linear system problem, which is addressed by the proposed quantum solver through using only partial information, with the time complexity improved from O (N 2. 37286) classically to O (N 2 log N × p log p) in quantum. Numerical examples have validated the effectiveness of the refined HHL algorithm with the proposed mapping function and the ME-ASPI method. By leveraging quantum properties, the proposed method provides a hybrid quantum–classical framework for microgrid control. This generic method can also potentially tackle many other challenges in analyzing and controlling general complex engineered systems. • Mapping shift function for the implementation of Harrow-Hassidim-Lloyd algorithm. • Enhanced Sampling for control input of interest in Model Predictive Control. • A bound for an important parameter in Harrow-Hassidim-Lloyd algorithm. [ABSTRACT FROM AUTHOR]

The use of model predictive control (MPC) to optimally control district heating (DH) networks can support the transition to a carbon-neutral heating sector. DH systems are inherently subject to non-linear physics and integer controls, which results in a mixed-integer non-linear program (MINLP). In this work, an MINLP-based MPC strategy is developed for the optimal control of a DH network, building upon an existing decomposition approach (combinatorial integral approximation). The main novelty of this work is the application of an integrated MINLP-based MPC to a DH network and its comparison to a non-linear program (NLP)-based MPC. To successfully develop this MINLP-based approach, the pycombina tool is efficiently integrated in the existing NLP-based MPC toolchain (TACO) and the concept of an augmented time horizon is introduced to manage dwell time constraints. The MINLP-based MPC is applied to two use cases: a relatively simple nine-zone terraced house and a more complex fourth generation DH network. The simulation study shows that the MINLP-based MPC yields a comparable control performance to that of a previously developed NLP-based MPC, but the CPU time is approximately eight times higher. However, the absence of a post-processing step (which requires ample engineering practice and time) and the improved match between the MPC's controller model and the actual system show promise for the MINLP-based MPC to control complex DH networks. • An MINLP-based MPC methodology for district heating systems is developed. • The MINLP-based MPC methodology can cope with dwell time constraints. • NLP- and MINLP-based MPC reach similar performance for the envisaged use cases. • NLP-based MPC has a computational advantage. • MINLP-based MPC avoids parameter tuning and better matches the actual system. [ABSTRACT FROM AUTHOR]

In this paper an alkaline electrolysis water system with 50Nm3H 2 /h is established, and the electrolyzer is equipped with PPS (Polyphenylene sulfide) diaphragm. The characteristics of oxygen purity are investigated in this study. The oxygen purity is characterized by HTO (Hydrogen to Oxygen). Three different types of experiments are carried out to determine the stability, safe operation boundary, and dynamic response characteristics of oxygen purity. Experimental study reveals that the lowest load limit is just 40% rated load under maximum pressure. The change of HTO has the 3 min delay and dynamic change time of HTO from minute to hour levels after the step of operating conditions. A completed oxygen purity prediction model for this industrial-scale alkaline water electrolysis system is proposed. The steady-state, time-delay, and inertia of HTO are considered in the model, which effectively predicts the steady-state HTO of all operating points and the variation of HTO after the step of operating conditions with minimal error. The model is used to quantitatively analyze the hydrogen crossover driven by concentration difference, which exceeds 90% of the total hydrogen crossover. And the model reveals the reason for the poor gas purity at low current. • An industrial-scale alkaline water electrolysis system is established. • Three oxygen purity experiments are conducted. • A comprehensive oxygen purity prediction model considering steady-state, time delay, and inertia has been proposed. • The oxygen purity characteristics of the electrolyzer equipped with PPS and Zirfion diaphragm is compared. [ABSTRACT FROM AUTHOR]

Accurately predicting the remaining useful life (RUL) of lithium-ion batteries (LiBs) is crucial for improving battery management system design and ensuring device safety. However, achieving accurate long-term predictions of aging trajectories is challenging due to error accumulation in multi-step ahead forecasts. This study shows that considering future internal resistance (R), which is related to the aging process, and the capacity regeneration phenomenon (CRP) that occurs during aging can help reduce error accumulation. Specifically, we propose a hybrid method that incorporates future R and CRP to predict the aging trajectories and RULs of LiBs. Experiment results demonstrate: (1) for the same charging/discharging policies and battery types, the proposed method can accurately predict the aging trajectory and RUL using only the first 20 cycles' data (approximately 5% of the complete data); (2) for different charging/discharging policies and battery types, with transfer learning, the proposed method can predict the aging trajectory and RUL using the first 40 cycles' data. These results demonstrate that the proposed model is both accurate in long-term prediction and robust for estimating the aging trajectory and RUL across various datasets. • Our method can make accurate long-term prediction in aging trajectory. • Internal resistance and capacity regeneration are considered in the proposed method. • Only the first 5% of the complete data are used in RUL prediction. • Transfer learning adapts the proposed model to different dataset. [ABSTRACT FROM AUTHOR]

The exploration of sewage heat as a clean energy source is an innovative and promising concept in the realm of sustainable energy development. By recovering heat from sewage through regional pipelines, there is enormous potential for maximizing its use. However, this field remains underexplored. This study proposes an evaluation method to enhance the sewage heat utilization strategies using Urban Sewage State Prediction Model (USSPM). With a specific focus on building heat demand and drainage features, the relationship among spatial distribution, building types, and sewage heat utilization potential was investigated through multiple case studies involving both real-world buildings and hypothetical scenarios. This study first demonstrates a decision-making process for identifying priority buildings for sewage heat system installation through a scenario in an actual area based on a comparison of their energy consumption. Moreover, the results of hypothetical scenarios provide recommendations for building locations and the characteristics of different types in sewage heat utilization. For instance, we suggest that residential buildings with high drainage temperatures, stable flow rates, and relatively low heat demand can predominate as ideal contributors to the overall heat supply and should be placed upstream. Hospitals and hotels, with larger heat demand, are advisable to position downstream. Through the proposed method and its application, this study provides recommendations for sewage heat utilization. It is expected to better leverage sewage heat in the future and achieve the goal of energy conservation through the utilization of clean energy. • Evaluation approach for sewage heat utilization potential on urban scale. • Capturing the heat demand and drainage features of different building types. • Decision-making approach for selecting priority buildings to utilize sewage heat. • Explore the impacts of various building types and spatial distribution. • Discuss sewage heat utilization strategies in real-world and hypothetical scenarios. [ABSTRACT FROM AUTHOR]

Direct internal reforming solid oxide fuel cells (DIR-SOFCs) are economically viable devices for power generation, while their reliability requires further improvement. Advanced process control can effectively facilitate the long-term operation of SOFCs with high efficiency and safety. DIR-SOFCs exhibit complex coupling effect and a high degree of nonlinearity, making it worthwhile to embed high-quality dynamic models including spatial distributions into nonlinear model predictive control (NMPC). A two-layer control architecture aiming for rapid load tracking and thermal management with high economic efficiency is designed for planar DIR-SOFCs, containing a set point optimizer and an NMPC controller based on the distributed parameter model expressed by partial differential–algebraic equations (PDAE). PDAE-constrained dynamic optimization problems are solved in the NMPC controller to seek optimal control strategies, considering safety constraints about the maximum temperature gradient and physical constraints of actuators. The high-fidelity one-dimensional PDAE model is validated through steady and dynamic simulations, and the optimal steady-state distributions are analyzed under three power demands. Behaviors of the closed-loop system under power demand step-up and step-down scenarios are investigated to demonstrate the effectiveness of the designed control scheme. The proposed NMPC controller can efficiently realize rapid load tracking, maintain good thermal management, and reduce the transition time, with minor steady-state errors and acceptable real-time performance. • Two-layer control architecture for rapid load tracking and thermal management. • High-fidelity one-dimensional PDAE model for DIR-SOFCs. • NMPC controller design based on PDAE-constrained dynamic optimization. • Dynamic behavior analysis and interpretation of the closed-loop system. • Effective multivariable optimal control and acceptable real-time performance. [ABSTRACT FROM AUTHOR]

Model Predictive Control (MPC) emerges as a promising solution to address the substantial greenhouse gas emissions from the building sector. By employing advanced control strategies, such as MPC, for peak energy shifting, there is a significant potential to enhance energy efficiency and reduce emissions through effective demand response within a smart grid. Although MPC has been the subject of extensive research, the practical implementation of cost-effective and easily deployable solutions in buildings remains limited. This study proposes a cost-effective MPC approach, employing the Internet of Things (IoT) and dynamic pricing. The control strategy, developed in MATLAB, is locally deployed on Raspberry Pi hardware via WiFi. The proposed MPC was tested in a controlled environment at the University of Nottingham, UK, where it regulated a radiator heating device to maintain indoor comfort in response to dynamic hourly electricity prices, using real-time indoor temperature feedback. The results confirmed the proposed MPC's accuracy in predicting indoor temperature responses and controlling indoor temperature within setpoints over a typical winter week. Performance analysis further revealed that the proposed MPC strategy resulted in a 20% electricity cost reduction compared to a conventional control strategy. Additionally, alongside the proposed MPC strategy, a Phase Change Material (PCM) wallboard system was integrated into a co-simulation platform. The developed PCM wallboard model underwent a verification and validation process, utilizing both numerical simulations and experimental data. The results demonstrate that the proposed MPC-controlled PCM wallboard system saved 35% on electricity cost compared with the original case study room. This study provides valuable insights into the development of intelligent localized demand response control for the built environment, offering a range of choices for IoT equipment. [Display omitted] • Proposed low-cost, IoT-based model predictive control (MPC) strategy for buildings. • MPC strategy implemented in real-world settings using Raspberry Pi hardware. • Integrated PCM wallboard system, achieving 35% electricity cost savings. • Co-simulation platform for interactive modelling of MPC and PCM-integrated buildings. • Peak energy shifting using MPC in PCM-integrated buildings. [ABSTRACT FROM AUTHOR]

Controlling flexible hybrid microgrids (MGs) is difficult due to the system's complexity, which includes multiple energy sources, storage devices, and loads. Although adding new components to the MG system through the plug-and-play (PnP) feature enables operating of the system in different modes, it adds to the system's complexity, hence necessitates careful control system design. The most challenging aspect of designing the control system is ensuring that it can control the MG optimally in its various modes of operation. Previous methods based on logical control allow for synthesizing a controller capable of controlling the MG in its various operational modes. However, the resultant controller does not optimally operate the MG. Classical model predictive control allows optimal control of the MG only in specific operating modes. On the other hand, switched model predictive control (S-MPC) can optimally control the MG in its various modes. However, the design of S-MPC is complex, particularly for MGs with many operating modes or complex switching logic. Multiple factors contribute to the complexity, including model development, mode detection, and switching logic. This paper presents a hybrid method based on ɛ -variables and classical MPC for constructing the S-MPC for flexible hybrid MG with PnP capabilities. Our results show that the proposed controller synthesis approach provides an effective solution for optimally controlling flexible hybrid MGs with PnP capabilities as the proposed method enables: (i) an increase in the amount of energy export to the utility grid by 50.77% and (ii) a significant decrease in the amount of energy import from the grid by 46.7%. [Display omitted] • Managing complex control systems for plug-and-play flexible hybrid MGs. • Addresses complexities in mode detection, switching logic, and model development. • Use ɛ -variables and MPC to create S-MPC for plug-and-play-ready MGs. • Increase in exported energy (+50.77%) and a decrease in imported energy (−46.9%). [ABSTRACT FROM AUTHOR]

Accurate cluster photovoltaic power prediction (CPPP) is crucial for the operation and control of renewable energy grid-connected power systems. The traditional modeling strategies for CPPP such as direct aggregation (DA) and statistical upscaling (SU) have limitations such as error accumulation and upscaling factor uncertainty. To address these issues, this paper proposed a novel hybrid approach for CPPP by combining machine learning models with an improved SU technique. Firstly, a robust broad learning system (BLS) model, in which the Generalized Maximum Correntropy Criterion (GMCC) is used to replace the original mean square error (MSE) loss in BLS, is proposed to solve the problem of multiple outliers affecting the prediction accuracy of regional cluster stations, and it is called GBLS. Then, the Relevance Vector Machine (RVM) as an effective nonlinear regression model is further utilized to compensate for the prediction errors obtained by the GBLS to form the hybrid prediction model, namely GBLS-RVM. Moreover, to mitigate the uncertainty associated with scaling factors in traditional SU strategy, a new SU strategy is developed to refine the relationship between the reference station and the cluster sub-region, enabling direct modeling for regional power prediction. Finally, data from two PV clusters in different regions of China are used to validate the effectiveness of the proposed model, and the results show that under the improved SU strategy, the GBLS-RVM model, reduced RMSE by approximately 33.7% compared to the traditional BLS model, and the RMSE decreased by 12.89% and 30.2% when compared to traditional DA and traditional SU strategies. • A novel hybrid method for CPPP is developed by combining ML and improved SU technique. • A BLS with GMCC (GMCC-BLS) as a prediction model is developed to handle the outliers in measurement data. • An innovative error correction strategy based on RVM is proposed to improve prediction reliability. • A new SU technique is used to establish direct relationships between reference stations and subregions. [ABSTRACT FROM AUTHOR]

Accurate predictions of photovoltaic power generation (PV power) are essential for the integration of renewable energy into grids, markets, and building energy management systems. PV power is highly susceptible to weather conditions. Therefore, as weather forecast accuracy improves, it has become increasingly important issue to effectively utilize weather forecast data to enhance prediction accuracy. In this study, an improved model that combines Long Short-Term Memory (LSTM) and self-attention mechanisms is proposed. Proposed model captures the time features through the LSTM network and the correlations among multivariate time series through the self-attention mechanism. Additionally, methods to efficiently integrate historical and forecast data into various time-series forecasting models are also proposed. To verify the effectiveness of the proposed method and the performance of the proposed model, an actual PV power data of a building in Japan is used for various types of experiments. The results demonstrate that the proposed method effectively leverages weather forecast data and enhances the prediction performance of all models, the coefficient of determination (R2) are improved 15.8% for LSTM model, and 26.4% for proposed model. Whether for short-term or long-term predictions, proposed model consistently provides superior accuracy, practicality, and adaptability across all output sequence lengths. Compared to the basic LSTM model, R2 on short-term and long-term forecasting increased by 3.9% and 22.5%, respectively. • A novel hybrid model using LSTM and a self-attention mechanism was proposed. • A method which can use weather forecast data for time-series forecasting models was proposed. • The proposed model and method are validated using actual data from a building in Japan. • The impact of weather forecast data on model prediction performance was studied. • Studied the effect of input and output sequence lengths on different models. [ABSTRACT FROM AUTHOR]

Deep learning models are increasingly applied in the field of solar radiation prediction. However, the substantial demand for labeled data limits their rapid application in newly established systems. Traditional transfer learning employs pre-training and fine-tuning methods to reduce the use of data in the target system. However, it still necessitates a small amount of labeled data for fine-tuning. This results in extensive time and cost for data collection, delaying the deployment of prediction models and optimization algorithms and leading to energy wastage. In this study, we employed the Adversarial Discriminative Domain Adaptation (ADDA) approach to achieve transfer learning under zero-label conditions in the target system, enabling new systems to harness the knowledge from other systems to create predictive models. Using the measured solar radiation data from Tokyo and Okinawa, two sets of experiments were designed with interchanged source and target domains to validate the efficacy and robustness of the proposed model. The results indicate that compared with the method of directly using the source domain model, transfer learning can enhance the predictive accuracy of the test set by at least 14% in both experiments, exhibiting more stable predictive performance and reduced prediction outliers. • We first employed the ADDA algorithm to address prediction with zero observed solar radiation data. • 14 % prediction accuracy boost over reused models in zero-label scenarios achieved. • The algorithm demonstrates better outlier prediction, and holds significant practical implications. • All the code will be open sourced in https://github.com/daxiang415. [ABSTRACT FROM AUTHOR]

This paper introduces a continuous 24/7 power purchase agreement (PPA) designed for contracting photovoltaic (PV) generation within sustainable plus energy neighbourhoods (SPENs) or local energy communities, aiming to ensure a stable economic revenue stream for community stakeholders. The PPA involves the sale of solar PV generation, auctioned at a fixed strike price, to an external off-taker. Employing statistical prediction tools such as long short-term memory and auto-regressive modelling, the proposed framework allows hour-to-hour power delivery commitments between seller and buyer, accurately estimating the agreed-upon volume of renewable energy to be exchanged. A fixed PPA price is negotiated utilizing Nash Bargaining Theory, aiming to optimize revenue for the SPEN while minimizing procurement costs for the buyer, thus achieving an economic equilibrium that mitigates the long-term price risk prevalent in wholesale energy markets. Additionally, the proposed methodology includes utilization of a battery energy storage system (BESS) to store excess power or address supply–demand contractual disparities during periods of low PV generation. Simulation results obtained under varying climatic conditions and energy market dynamics across different countries, demonstrate that the proposed PPA framework, by combining risk assessment strategies and statistical learning methods, can effectively reduce associated financial risks while maximizing payoff for the community. • A novel VaR and RNN-based approach for determining energy trading volumes in long-term forward contracts is proposed. • A cooperative bargaining solution for optimal pricing in long-term PPAs is outlined. • BESS flexibility in mitigating PV uncertainties for hourly base-load PPAs across diverse European conditions is studied. [ABSTRACT FROM AUTHOR]

Stochastic optimization is a promising approach to support a reliable and cost-efficient transition to a renewable generation-dominated power grid. However, utilization of this approach requires accurate and computationally affordable probabilistic forecasts that rigorously quantify the uncertainty in and the correlation between stochastic grid variables (wind/solar generation and load demand). We investigate and compare two types of probabilistic forecasting model architectures for jointly predicting stochastic variables in a power grid: sequence-to-sequence models and recursive one step ahead prediction models. A long short-term memory (LSTM) neural network model is chosen to represent sequence-to-sequence deep learning models and a periodic vector autoregressive (PVAR) model is chosen to represent the one step ahead recursive models. For each model architecture, we consider three forecasting approaches: predicting each time series variable separately, jointly predicting all time series variables of a given type (e.g., wind generation or load), or jointly predicting all time series variables. We quantify the predictive capability of these forecasting models with respect to accuracy degradation with increasing prediction horizon, ability to characterize uncertainty, and ability to capture cross-variable dependence structure. We train and test the forecasting models using publicly available zonal power supply and demand time series data for the French power grid. We find that different modeling architectures/approaches show superior forecasting capability for different prediction horizons. Our results show that in general, one step-ahead recursive (PVAR) models have high prediction accuracy for short forecast horizons, and sequence-to-sequence (LSTM) models have slower performance degradation for longer forecast horizons. • Developed probabilistic models for predicting solar/wind generation and load. • Investigated various modeling approaches to learn spatio-temporal correlations. • Developed probabilistic models to support stochastic UC optimization. • Compared recursive to sequence-to-sequence forecasting models. [ABSTRACT FROM AUTHOR]

The solar-air source heat pump hybrid heating system exhibits suboptimal performance in optimizing the control of indoor air temperature, enhancing thermal comfort, and reducing system energy consumption. Advanced Model Predictive Control (MPC) has found extensive application in heating ventilation air conditioning systems(HVAC). Leveraging a physical platform, this study has designed a mathematical simulation model and meticulously validated simulation results. In order to reduce computational costs, a Modal Analysis-based Model Predictive Control (MA-MPC) is introduced, achieving a 55-fold reduction in processing time for a single time step while ensuring precise system control. Comparative simulations were conducted to assess the performance of ON-OFF control and MPC in terms of pump frequency, flow rate, stratified tank temperatures, component operation, and system energy consumption. Under the MPC strategy, indoor temperature fluctuations, pump frequency response speed, and energy consumption outperform ON-OFF control, resulting in a reduction of temperature fluctuations by approximately 2.4 °C. In contrast to MPC control, MA-MPC exhibits a mere 0.255 °C deviation in indoor temperature. This advancement not only significantly shortens computation time but also ensures enhanced control precision and reduced system energy consumption. Relative to MPC, MA-MPC achieves remarkable energy savings of up to 16.32% for the entire system. • Employing model predictive control in composite heating systems featuring solar-air source heat pumps. • Fusing modal analysis with MPC to achieve streamlined model order reduction. • Assessing contribution thresholds in modal analysis: trading off computational time and control precision. • Empirical validation of the proposed methodology in an authentic heating system with contrast MPC. • MA-MPC attains a noteworthy 55-fold reduction in processing time compared to conventional MPC. [ABSTRACT FROM AUTHOR]

Building photovoltaic (PV) power generation is intermittent, volatile, random, and uncontrollable; thus, the use of solar grid-connected power generation can lead to a series of problems, such as grid voltage fluctuations and power imbalance. As a typical flexible load, the scheduling and regulation of the heating, ventilation, and air conditioning (HVAC) system load can help a grid-connected solar grid achieve balanced and flexible operation. This study proposes a PV capacity dynamic tracking model predictive control strategy for air-conditioning systems with flexible loads. This strategy aims to effectively address the issue of unstable grid-connected output of solar PV systems. In particular, a solar radiation prediction model based on a long short-term memory neural network optimized by the grey wolf optimization algorithm was established, which effectively improved the model prediction accuracy. The cost function of HVAC flexible load control considers the law of solar PV output and the human thermal comfort model. A genetic algorithm (GA) is used to obtain the optimal control parameters for dynamic regulation. The GA was employed to efficiently and effectively optimize the control of flexible HVAC loads, considering the variability of the solar PV output and ensuring human thermal comfort. A dynamic regulation model was developed for a typical office building with a grid-connected solar PV power generation system. TRNSYS was utilized for verification based on the actual measured data and relevant meteorological parameters of the case building. A new evaluation metric, volatility, was proposed to evaluate net load fluctuation. The simulation platform based on the physical building yield results indicated a strong resemblance between the energy consumption curve under the predicted control conditions using the flexible model and the PV generation curve. Additionally, the net load volatility was measured to be 2.63. Compared with the predicted control condition without the flexible model, the net load volatility of the grid was reduced by 47.08%, and the energy-saving rate reached 10.89% in the summer. [Display omitted] • Tracking MPC strategy is proposed for fine-tuning control of HVAC system. • The law of photovoltaic output is considered in the proposed strategy. • Grey Wolf Optimization algorithm is used to tune LSTM neural network. • Control strategies comparisons exhibit the effectiveness of the proposed strategy. • A new metric-"volatility" is proposed to evaluate load fluctuations. [ABSTRACT FROM AUTHOR]

Power generation and consumption predictions are fundamental for smart grid operations, addressing challenges posed by renewable energy variability and irregular consumer demand. This study introduces the Dual Sequence Predictive Model (DSPM) based on Spatiotemporal CNN (STCNN) architecture, offering a holistic solution for forecasting Electricity Generation (EG) and Electricity Consumption (EC) collectively. Leveraging STCNN, the DSPM efficiently extracts spatial and temporal information, improving prediction accuracy and processing speed. A 1D spatial attention module is added to capture crucial spatial dependencies in historical data. Incorporating shared historical weather information streamlines learning process and reduces model complexity. Extensive benchmark evaluations demonstrate the DSPM superior performance, achieving the lowest error rates compared to baseline models. The DSPM also employs the Kullback–Leibler Divergence (KLD) algorithm for power generation and consumption matching, ensuring efficient power distribution within smart and microgrids. Furthermore, the DSPM is evaluated alongside Single Sequence Prediction (SSP) models, providing a comprehensive analysis of its capabilities and comparisons with state-of-the-art models. The SSP model achieved an average RMSE reduction of 17.3% for solar EG data and 15.86% for IHEPC residential EC data, while DSPM reduced RMSE by 7.57% over PowerGrid dataset compared to baseline models. • Developed a DSPM for collective prediction of power generation and consumption. • Unified architecture for spatiotemporal feature learning. • Model evaluation for load and demand prediction individually. • Integration of the KLD algorithm for efficient power matching. • DSPM outperforms baseline models and reducing operational complexity by up to 20%. [ABSTRACT FROM AUTHOR]

In this paper, a dynamic prediction model for electrolyte capacity fade in vanadium redox flow batteries (VRFBs) is proposed. The capacity fade characteristics of VRFBs were analyzed quantitatively from both microscopic ions crossover and macroscopic electrolyte volumetric change perspectives. The dynamic behavior of vanadium ions (V2+, V3+, VO2+, and VO+ 2), protons, and water molecules in VRFB half-cells and electrolyte tanks has been thoroughly explored. The parameters related to the capacity fade were obtained from a self-discharge test and formation charge processes. The accuracy of the proposed model was experimentally validated using a unit cell (25 cm2) in a long-term operation. The results reveal that after 200 continuous charge/discharge cycles (duration of 739,257 s) with a current density of 80 mA cm2, the model can accurately predict changes in electrolyte volume and capacity fade with errors of 0.632 mL and 0.0295 Ah, respectively. The electrolyte capacity fade under various current densities for long-term operations has also been predicted by the model and experimentally validated. Under all current densities applied to the VRFB, the discrepancies between simulation and experimental results for electrolyte volume change and capacity fade were <8.46% and 4.99%, respectively. The proposed model significantly improves the accessibility and reliability of accurate capacity fade prediction for VRFBs, enhancing the competitiveness of VRFBs in energy storage applications. • Dynamic model of VRFB which is capable of predicting capacity fade was developed. • Electrolyte volume change of VRFB was evaluated by considering mass transfer, migration, and self-discharge phenomena. • Proposed model was validated with the various long-term experiments, which confirmed the high fidelity of the model. [ABSTRACT FROM AUTHOR]

In the realm of economic scheduling and optimal operation within integrated energy systems, integrated energy load aggregation plays a pivotal role. Accurate prediction of integrated energy load aggregation enables the refined management of energy resources, thereby enhancing energy utilization efficiency. This paper merges the advantages of clustering algorithms, prediction algorithms and ensemble learning to propose a short-term prediction approach for integrated energy load aggregation. Firstly, the intrinsic distribution of energy consumption data is analyzed through Affinity Propagation (AP) clustering with multi-dimensional similarity. This analysis yields the division of all samples into clusters based on data characteristics. Subsequently, the Transfer entropy is harnessed to conduct correlation analysis within each cluster. For constructing predictive models for each individual cluster, a prediction algorithm based on the Feature-Temporal Attention-enhanced Bi-directional Simple Recurrent Unit (FTA-Bi-SRU) is employed. The Dynamic Weight Average (DWA)-based multi-task loss equilibrium and Multi-Term Adam (MTAdam) loss function are combined to enhance the efficiency of the forecasting model. Finally, a Light Gradient Boosting Machine (LightGBM) model with Bayesian optimization is employed to integrate the predictive outcomes of each individual cluster to generate the overall predictive outcomes for integrated energy load aggregation. The proposed methodology is experimented on real-world energy consumption data to ascertain the efficacy of the prediction method. The experimental results substantiate the superiority of the suggested forecasting approach. • This marks the inaugural discussion of load forecasting for integrated energy load aggregation is discussed. • A multi-dimensional similarity Affinity Propagation (AP) clustering algorithm is introduced to group multi-energy users. • A method based on transfer entropy for correlation analysis is presented. • An Feature-Temporal Attention based Bi-directional Simple Recurrent Unit (FTA-Bi-SRU) prediction model is introduced. • A Bayes-optimized Light Gradient Boosting Machine (BO-LightGBM) ensemble framework is proposed to enhance overall prediction. [ABSTRACT FROM AUTHOR]

Deterministic wave elevation prediction is crucial for improving the power generation efficiency of offshore energy structures (OESs). Although phase-resolved wave models may predict deterministic wave information, a comprehensive understanding of wave theory and mathematics is necessary to guarantee their accuracy. Inspired by the capability of physics-informed neural networks (PINNs) in solving physical equations, we propose an irregular long-crested wave prediction model named LWT-PINN that combines the linear wave theory (LWT) with PINNs. To the best of our knowledge, this is the first time that PINNs are used to predict waves. Five sets of wave elevation time series with steepness ranging from 0.0174 to 0.0349 are generated in a wave basin to optimize and evaluate the LWT-PINN. The computing time for LWT-PINN is 0.13 s, close to real-time. When predicting downstream wave elevation, the LWT-PINN significantly outperforms the linear wave prediction (LWP) model in terms of accuracy and prediction length. Specifically, the LWT-PINN achieves satisfactory accuracy with only 100 s upstream wave elevation, whereas the LWP requires 175 s; LWT-PINN's high-precision prediction duration is 13.7 s, which is 2.5 times longer than that of LWP (5.5 s). Moreover, LWT-PINN can provide wave autoregressive predictions of around 5 s without the aid of statistical methods, which offers a new direction for developing wave autoregressive prediction methods. The excellent performance of the proposed model can substantially assist the stable operation of OESs. • A physics-informed neural network with linear wave theory for wave prediction. • The proposed model accurately predicts the downstream wave for more than 10 s. • The proposed model achieves autoregressive prediction without using statistical methods. • It offers phase-resolved wave models a new direction for obtaining wave components. [ABSTRACT FROM AUTHOR]

The ultra-short-term forecasting of wind speed is of great significance to the stable power supply of the power system. Current wind speed forecasting methods aim to improve forecasting precision while disregarding model training speed and model deployment complexity. This research proposes a lightweight hybrid model named SLF-EMD-MGM-NS for wind speed forecasting. EMD-MGM is designed as the network's fundamental structure for reducing the hybrid model's training time and ensuring the hybrid model has high forecasting precision. The study presents the switching loss function (SLF) mechanism. When the quantile is 0.5, an MSE-based loss function is employed for training all subsequences. When the quantile is 0.5 and 0.95, first use the wind speed fluctuation threshold to select primary subsequence, and then use the Log-Cosh-based loss function for training primary subsequences. The SLF mechanism can increase point prediction precision and interval prediction boundary stability. Moreover, a novel subsequence superposition (NS) mechanism is proposed for getting high confidence level and narrow-width interval prediction results. The NS mechanism superimposes the interval prediction results of the fluctuation subsequence with the point prediction results of the model to generate the final interval prediction results. According to the experimental results, the SLF-EMD-MGM-NS model has a high confidence level, acceptable prediction results, a narrow-width interval prediction result, and a significantly shorter training time than the other hybrid models. • Proposed EMD-MGM as the base network architecture to do wind speed forecasting. • Three point regression loss functions were extended to wind speed interval regression. • A new mechanism named SLF was proposed to do quantile regression. • A new mechanism named NS was proposed to superpose subsequences' forecasting results. [ABSTRACT FROM AUTHOR]

Grid-interactive communities are an emerging solution for optimal energy dispatching, improving grid stability, and enhancing the usage of on-site renewable energy by leveraging community flexibilities. This work introduces a novel Physically Consistent Deep Learning (PCDL)-based optimization framework for day-ahead energy dispatching and thermal comfort control within grid-interactive communities. Specifically, the PCDL model is developed to capture the thermal dynamics within the community while ensuring strict adherence to physics consistency. This consistency is defined with the aim of generating physically viable solutions in response to modified system inputs. Subsequently, the PCDL model is utilized for load estimation and indoor climate prediction within a proposed scheduling and control strategy: (1) an optimal energy dispatching plan is calculated based on the day-ahead grid market; (2) a real-time model predictive control (MPC) method is applied to minimize the utilization error and indoor comfort constraint violations caused by following the scheduled energy dispatching plan. A simulation case of a residential hall community at Cornell University with on-site renewable energy resources is implemented to demonstrate the effectiveness of the proposed approach. Comparative simulations, which include the Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and Gaussian Process (GP) regression models, confirm the performance advantage of capturing community thermal dynamic and control-oriented generalization ability when using the PCDL model. These results leads to notable improvements in indoor thermal comfort control, ranging between 65.4 and 68.7%, 63.6–79.3%, and 60.4–62.2% for each comparative model, respectively. Moreover, by harnessing community demand flexibility, we achieve energy cost savings between 29.5 and 39.7% compared to the baseline controller. • A physically consistent deep learning (PCDL) model for thermal dynamic prediction. • A PCDL-based community energy dispatching and indoor comfort control framework. • Novel case study results compared to alternative data-driven models. [ABSTRACT FROM AUTHOR]

Shale gas production prediction is of great significance for shale gas exploration and development, as it can optimize exploration strategies and guide adjustments to production parameters for both new and existing wells. However, the dynamic production characteristics of shale gas wells under the influence of multiple factors such as reservoirs, engineering, and production, exhibit complex nonlinear and non-stationary features, leading to low accuracy in predicting shale gas production. To address this issue, a novel masked convolutional neural network (M-CNN) based on masked autoencoders (MAE) is proposed for shale gas production prediction. First, high-dimensional shale gas production data are transformed into images with unknown information using an encoding structure, thereby converting the regression task into images generation task. Then, convolutional neural network is used for image restoration prediction, and the corresponding numerical values at the image positions are extracted as shale gas production prediction results. Specifically, dilated convolution and multi-scale residual structure (MSRS) are developed to improve the feature representation capability of the network. Meanwhile, convolutional block attention module (CBAM) is adopted to enhance the feature extraction ability of the M-CNN. The performance of our method is validated experimentally on shale gas production data of Changning (CN) block in China. The average RMSE, MRE, and R 2 on the test sets are 0.211 (1 0 4 m 3 / d), 10.9%, and 0.906, respectively, which is much lower than the traditional time series models. Experimental results demonstrate the effectiveness and superiority of the proposed M-CNN method for shale gas production prediction. • Propose a novel masked convolutional neural network for shale gas production prediction. • Masking mechanism transforms high-dimensional shale gas time-series data into 2D images. • Multi-scale residual structure is developed to enhance the feature extraction ability of the network. • Dilated convolution is designed to extract global features. [ABSTRACT FROM AUTHOR]