Non-Intrusive Load Monitoring is an effective method for disaggregating the power consumption of individual appliances from the aggregate load data of a building. The advent of smart meters, Internet of Things devices, and artificial intelligence technologies has significantly advanced the capabilities of non-intrusive load monitoring. However, challenges such as varying sampling frequencies and measurement sensitivities remain. This paper introduces an innovative model incorporating an Adaptive Multi-Scale Attention Integration Module (AMSAIM) to address these issues. The model leverages deep learning and attention mechanisms to improve the accuracy and real-time performance of non-intrusive load monitoring. Validated on the standard UK-DALE dataset, the model consistently demonstrated superior performance. In seen scenarios, our model achieved average F1-scores approximating 0.94 and notably reduced Mean Absolute Error (MAE) values. For washing machines, it achieved an F1-score of 0.99 and MAE of 41.64, outperforming the next best method's F1-score by 1 percentage point. In challenging unseen scenarios, the model showcased strong generalization, achieving an F1-score of 0.91 for washing machines and reducing MAE to 7.66. Furthermore, an ablation study rigorously confirmed the necessity of the AMSAIM module, showing that the synergistic integration of the efficient multi-scale attention (EMA) and the selective kernel (SK) adaptive receptive field unit is crucial for enhancing model robustness and generalization. Our results highlight the model's potential for enhancing energy efficiency and providing actionable insights for energy management across various conditions.

The industrial sector accounts for a significant share of global energy consumption and greenhouse gas emissions, making the optimal operation of industrial parks a key pathway for sustainable energy transition. This study proposes a day-ahead coordinated optimal scheduling framework for a multi-sector Industrial Park Energy Hub (IPEH) that integrates electricity, heating, and cooling systems with renewable generation and multi-energy storage. The model captures sectoral diversity across industrial, commercial, residential, and administrative sectors, enabling coordinated inter-sector operation through electricity and heating energy sharing. The scheduling problem minimizes total operating cost, including penalties for greenhouse gas (GHG) emissions and for power curtailment from photovoltaics (PV) and wind turbines (WT), while considering the physical constraints of the heating network and power tie-lines. The optimization problem is solved using the CPLEX solver in MATLAB. Results under three scenarios show that, compared with independent operation, electricity sharing alone reduces operating cost by 3.22% and renewable curtailment by 58.19%. Coordinated electricity and heat exchange further improves system performance, achieving a 6.95% reduction in operating cost, a 58.19% decrease in renewable energy curtailment, and emission reductions of 18.11% for CO2, 23.80% for SO2, and 38.42% for NOx.

As a large number of large-scale photovoltaic (PV) stations are integrated into the power grid, the penetration rate of PV power is growing higher and higher. The intermittency and volatility of PV power generation bring great pressure to the safe and stable operation of the distribution network. In order to realize scientific energy dispatching and optimization, the predicted output of large PV stations is the data basis and prerequisite. The output prediction method of large PV stations is studied in this paper, and a prediction method based on gradient-boosting decision trees is proposed. In the method, the original data are first collected, and the sample set is established through the steps of data interpolation, supplement, and integration, and then the sample set is pre-processed by data cleaning and normalization. The model training and PV output prediction during the test period are carried out based on the pre-processed data. Finally, the prediction results are imported into the error analysis module. The feasibility and accuracy of the proposed method are analyzed by comparing it with the traditional method. The results show that the normalized mean absolute error (nMAE) and normalized root mean square error (nRMSE) of the proposed method are 7.31% and 11.78%, respectively, while the nMAE and nRMSE of the traditional method are 11.67% and 20.39%, respectively. Thus, the prediction performance of the proposed method is superior to that of the traditional method.

As renewable energy gains broader adoption, the significance of photovoltaic storage microgrids in residential and commercial settings is on the rise. This study introduces an innovative model for decomposing load through local cross-channel interactions, coupled with an application strategy designed to enhance the operational efficiency of domestic photovoltaic storage microgrids. By integrating the Efficient Channel Attention (ECA) module into a non-intrusive load decomposition framework, we address the trade-off between the performance of channel attention mechanisms and model complexity, substantially enhancing decomposition accuracy. The ECA module minimizes the model's parameter count and computational demands through a strategy that avoids dimensionality reduction and leverages local cross-channel interactions, without compromising performance. Our experiments demonstrate high precision in decomposing power-hungry appliances, with F1 scores of 94% for washing machines and 97% for dishwashers. Building on the ECA module's application to home appliances, we propose a multidimensional feature-based load forecasting and scheduling strategy derived from the local cross-channel interaction model. This approach harnesses the decomposed appliance load features to enrich the forecasting models for photovoltaic energy storage systems. Additionally, exemplified by home microgrids, we introduce a dispatch control strategy for photovoltaic storage microgrids that ensures a reliable power supply for critical loads and aims to minimize energy costs.

Accurate photovoltaic (PV) power prediction is extremely critical for the safe and stable operation of power systems. Aiming at the problems of inaccurate cluster delineation and difficulty in improving the short-term prediction accuracy in the current PV power cluster prediction, this paper proposes a short-term power prediction method for PV power clusters that integrates the fuzzy C-mean (FCM) and the improved Transformer-Temporal Convolutional Network (TCN). Initially, subsets are clustered using the FCM clustering algorithm. Subsequently, the ITransformer-TCN model is employed to leverage its dual feature extraction capabilities for modeling and predicting each subset. A PV power station in Guangxi Province is applied to test this proposed method c, and comparative analysis demonstrates that this method achieves an average reduction of Root Mean Square Error by 8.9% compared to other methods.

Energy router (ER) achieves efficient management of electrical energy, optimizes distribution, and increases utilization of renewable energy sources. Due to the limited voltage gain range of bidirectional full-bridge resonant converter (CLLC) and the asymmetrical voltage range required during charging and discharging of energy storage batteries from the dc bus, the switching frequency of CLLC converter has to be varied over a wide range, which leads to inefficiencies. A bridge-arm multiplexed CLLC converter is proposed to improve the performance, voltage gain range, and applicability of CLLC converters in ERs. Moreover, the interleaved parallel technique can be applied to realize the high-power charging and discharging of batteries. To solve the problem of power imbalance caused by interleaved parallel operation, a gain compensation control strategy is proposed to equalize the power split of the converters, which leads to more uniform thermal distribution across the components and reduces the voltage and current ripples. A 2-kW experimental platform is built to validate the performance of the proposed converter. With the BMCLLC converter and the gain compensation control strategy proposed in this article, the efficiency can be maintained above 92% during high-power charging and discharging of the batteries.

The common cascade converter structure is a dual-active-bridge (DAB) converter in the front stage and an inverter with an LCL filter backstage. The feedback controller of the front-stage DAB converter typically has a small input impedance amplitude, leading to instability in the cascaded converter system. An impedance compensation controller based on DC bus feedback is proposed, which can enhance the system stability without additional hardware circuits; secondly, for the effect of weak grid impedance changes on the resonance peak of the LCL filter, an improved active damping method of notch filter based on grid impedance detection is proposed, which improves the adaptability of the system to weak grid impedance fluctuations. Finally, experimental verification on a 2KW cascaded converter platform proves that the optimized controller and the improved notch filter are effective in improving the stability of this cascaded system.

The powersphere is a spherical enclosed receiver composed of multiple photovoltaic cells. It serves as a replacement for traditional photovoltaic panels in laser wireless power transmission systems for optoelectronic conversion. The ideal powersphere aims to achieve a uniform distribution of light within the cavity through infinite reflections, reducing energy losses in the circuit. However, due to the high absorption rate of the photovoltaic cells, the direct irradiation area on the inner surface of the powersphere exhibits a significantly higher light intensity than the reflected area, resulting in a suboptimal level of light uniformity and certain circuit losses. To address the aforementioned issues, a method of intra-cavity beam splitting in the powersphere is proposed. This solution aims to increase the area of direct illumination and reduce the intensity difference between direct and reflected lights, thereby improving the light uniformity on the inner surface of the powersphere. Utilizing the transformation matrix of Gaussian beams, the q parameters for each optical path with beam splitting were calculated, and the equality of corresponding q values was demonstrated. Further, based on the q parameter expression for the electric field of Gaussian beams, the intensities for each optical path were calculated, and it was demonstrated that their values are equal. Additionally, an optical software was utilized to establish a model for intra-cavity beam splitting in the powersphere. Based on this model, a beam-splitting system was designed using a semi-transparent and semi-reflective lens as the core component. The light uniformity performance of the proposed system was analyzed through simulations. To further validate the effectiveness of the calculations, design, and simulations, multiple lenses were employed to construct the beam-splitting system. An experimental platform was set up, consisting of a semiconductor laser, monocrystalline silicon photovoltaic cells, beam expander, Fresnel lens, beam-splitting system, and powersphere. An experimental verification was conducted, and the results aligned with the theoretical calculations and simulated outcomes. The above theory, simulations, and experiments demonstrate that the intra-cavity beam-splitting method effectively enhances the optical uniformity within the powersphere.

The WT-PV integrated multi-energy industrial park can play a strategic role in reducing emissions and enhancing energy supply efficiency. The fluctuating electrical and thermal demand, the system's synergistic composition and uncertainties require optimal scheduling to address the alarming economic and environmental concerns. Based on the energy hub concept, as a practical approach to multi-energy system management, the structure of an industrial park energy hub (IPEH) is developed primarily as a basis for the optimization strategy. Then to meet the demand, a multi-objective day-ahead scheduling model is developed, yielding mixed-integer linear programming. The objective function aims to minimize the overall economic and environmental cost, gas emissions costs and the WT-PV power abandonment penalty cost. The scheduling strategy utilizes source-load day-ahead predicted curves of the IPEH as input to tackle renewables uncertainties and load fluctuations. The time-of-use electricity prices and natural gas prices realized electricity price peak shaving. Case studies are conducted on a typical Industrial Park for three scenarios taken on a typical winter and summer day. The results of different scenarios and the role of energy storage devices were discussed. Simulation results indicate that the scheduling model is effective across the energy-dispatching horizon with minimal overall cost.

This paper presents the MUSENILM model, a non-intrusive load decomposition model incorporating a parallel multi-scale attention mechanism to enhance energy monitoring and management in smart grids. The core innovation of the proposed model is its ability to extract multi-scale features, enhancing the model's understanding of time series data and achieving significant performance improvements on the UK-DALE and REDD public datasets. Specifically, when MUSENILM identifies the fridge electricity consumption pattern on the UK-DALE public dataset, compared to previous models, the accuracy improves from 88% to 91%, and the F1 score increases from 87% to 90%; on the load decomposition tasks of the remaining four appliances, the F1 scores are all improved, while the mean absolute error (MAE) and cumulative absolute error (SAE) for the five appliances are also reduced. Additionally, it shows better results compared to previous models on the REDD dataset. Moreover, when the MUSENILM model is transplanted to embedded devices and applied in smart grids, it can effectively identify illegal lithium battery charging events of electric bicycles in different scenarios, which is crucial for ensuring grid security and optimizing energy distribution. This research not only provides an efficient method for the field of NILM but also offers practical solutions for violation monitoring and management in smart grids.

The construction process of offshore wind farms involves multiple complexities, which is very complex to be scheduled manually, and the coordinating and optimized scheduling not only decreases project construction costs but also increases the construction speed. The impact of meteorological conditions on offshore wind power construction has been considered, and optimizing resource-allocation strategies under complex influencing factors has been analyzed. Then, a comprehensive strategy optimization index system is developed, which includes key indicators, such as the minimum working hours, resource-allocation-optimization rate, window period utilization rate, and cost–benefit ratio. Additionally, an offshore wind power resource-allocation-optimization model is formulated based on discrete event simulation (DES). A statistical analysis of each optimization index was performed using this model to assess the impact of resource-allocation strategies. The simulation results demonstrate that the model can not only simulate the construction process of offshore wind farms and monitor the state of wind turbines, personnel, and meteorological conditions in real time but also accurately calculate key indicators, such as the minimum working hours, resource-allocation-optimization rate, window period utilization rate, and cost–benefit ratio. This strategy effectively enhances resource-allocation efficiency during the wind farm installation phase and improves the overall construction process efficiency.

The air source heat pump (ASHP) using electric energy instead of traditional wood burning to dry cash crops has the advantages of high automation, energy saving and environmental protection, cost reduction, and efficiency. Aiming to evaluate the energy efficiency in the process of electric heat conversion and drying heating of air source heat pumps, two levels of energy efficiency indexes have been proposed. The primary energy efficiency index can measure the utilization efficiency of electric heat conversion and recovery, and the secondary index can directly measure the output efficiency of electric energy on output. Aiming to improve the automatic ability of the drying system, a remote monitoring system of air source heat pump drying system has been designed and developed based on the Internet of Things (IoT), which can monitor the operation process and status of air source heat pump drying system online in real-time in the cloud and handheld terminal, and control the process flow of drying operation, and also can calculate the two-level energy efficiency indexes. The remote monitoring system based on IoT plays the role of protection and electricity safety. The drying experiments of edible mushrooms show that the IoT air source heat pumping drying system has the advantages of high automation, high efficiency, and convenience, and can effectively evaluate the energy efficiency of the air source heat pump in the drying process of edible mushrooms.

The high randomness and complexity of arc fault make it difficult to be accurately identified.Aiming at the problem that the traditional arc recognition algorithm has low real-time performance and high hardware computing power,an error minimization extreme learning machine(EM-ELM)arc fault detection method suitable for edge computing,multi-load types and multi-feature combination is proposed.Through fast Fourier transform(FFT)and db4 wavelet decomposition,the period mean difference,pulse width percentage,inter-harmonic factor and wavelet high-frequency energy are extracted as the input characteristics of the arc fault detection algorithm on the edge side.On this basis,OS-EM-ELM combined with online sequence(OS)method is proposed,and the algo-rithm is improved by using field operation data to improve adaptability.The experimental results show that the pro-posed edge side arc fault detection method can effectively distinguish the normal and arc fault waveform,and it is suitable for the complex situation of working with a variety of loads at the same time.The calculation amount is small,the real-time performance is high,the adaptability is strong,and the application cost is low,which is more in line with the requirements of edge calculation of arc detection device.

针对航测作业过程中遇到云雾等恶劣气候条件时,会产生影像色彩失真、饱和度低、背景模糊不清等问题,本文进行了影像去雾与影像增强算法的研究,提出了一种改进 Retinex 算法,并以重庆某山区为例,与两种传统经典算法进行比较分析.结果表明,本文算法可对带雾影像进行有效处理,获得更真实准确的细节、更清晰真实的光感色彩;使用去雾与增强后的影像进行三维建模,从而生成高分辨率三维网格.分别从主观和客观开展对影像效果的评价,所得到的优化影像与三维模型都有良好的效果.

In order to better grasp the quantitative potential of provincial power substitution， a prediction model based on Lasso-XGBoost-Stacking is proposed. Five influencing factors， including economic development， environmental protection， energy price， policy support and technological progress， are quantified by cross-features to reduce the multicollinearity among the influencing factors. Besides， the weights of the quantified influencing factors are evaluated by using Lasso regression model. The predicted MAPE （mean absolute percentage error） of Zhejiang province is up to 12.22%， which can meet the requirements of quantitative potential analysis of power substitution in the province. The analysis of the multi-scenario power substitution scenarios in Zhejiang reveals that economic development has the greatest impact on power substitution potential， and the transportation industry has the greatest power substitution potential relative to industry， agriculture and other sectors.

With the increasing proportion of distributed photovoltaic (DPV) installations in county-level power grids, to improve the centralized operation and maintenance of the stations and to meet the needs of power grid dispatching, the output of the county-level regional DPV stations group needs to be predicted. In this paper, the weather prediction information is used to predict the output based on the model input average strategy. To eliminate the effect of the selected non-optimal training sample collection period on the prediction accuracy, an ensemble prediction method based on the minimum redundancy maximum relevance criterion and TabNet model is carried out. To reduce the influence of weather prediction errors on the power output prediction, a modified model based on error prediction is proposed. The ensemble prediction model is used to predict the day-ahead output, and a combination prediction model based on the proposed ensemble prediction model and the proposed modified model is established to predict the hour-ahead output. The experimental results verify the effectiveness of the proposed models. Compared with the corresponding reference models, the proposed ensemble prediction method reduces the normalized mean absolute errors (nMAEs) and the normalized root mean square errors (nRMSEs) of the day-ahead output prediction results by 2.86% and 5.51%, respectively. The combination prediction model reduces the nMAE and nRMSE of the hour-ahead output prediction results by 3.05% and 3.05%, respectively. Therefore, the prediction accuracy can be improved by the proposed models.

The LLC converter is a key component of the bidirectional power converter for mobile energy storage vehicles (MESV), it is difficult to obtain small gains at low power levels, so the power control in the pre-charging stage of the Li-ion battery cannot be achieved. In addition, the bus voltage may be lower than the peak grid voltage due to LLC reverse gain limitation. The low SoC of the lithium battery will result in the inverter not being able to connect directly to the grid. Aiming at these problems, a dual-mode switching full-bridge LLC control strategy is proposed to expand the range of pre-charge gain. In the reverse mode, the voltage gain is boosted by the gate signal on the asynchronous primary and secondary sides, and enables the inverter to be successfully connected to the grid. A 1.5 kW experimental platform was built to verify the strategy. Compared with the traditional single-mode pulse frequency modulation (PFM) control, the voltage control range was expanded from 58V-100V to 38V-100V, which met the control requirements in the pre-charging stage of lithium batteries. the inverter reverse gain is increased to 1.4. the problem is solved that the low SoC grid-connected total harmonic distortion (THD) does not meet the standard. Finally, the feasibility and effectiveness of this method are verified.