Given that the current microgrid incorporates highly connected distributed energy sources, conventional model control methods do not suffice to support complex and ever-changing operating scenarios. This paper proposes a deep learning-based optimization method for management in new power system article constructs cloud edge collaboration architecture, which collects interactive network status data through terminal devices sides. A is constructed based on Bi-LSTM attention cloud. And sunk...

Energy conservation, emission reduction and vigorous development of new energy are inevitable trends in the power industry, but factors such as storage loss, solar loss line real situations have led problem to a complex direction. To address these intricacies, we use more precise modeling approach propose collaborative optimization method integrating Deep-Q-Network (DQN) algorithm with multi-head attention mechanism. This calculates weighted features system’s states compute Q-values...

In this paper, we propose an effective and intelligent prediction model that can well distinguish complex acute appendicitis from uncomplicated appendicitis. study, 358 patients admitted to the First Hospital of Jilin University in Changchun for past 5 years were included, data panel was constructed based on 32 factors collected. The framework comprised mainly a Principal Component Analysis (PCA) algorithm Support Vector Machine(SVM) with new kernel function named Mercer Machine(MSVM)...

By studying the classification of anomaly patterns in integrated energy systems, a deeper understanding their operational status can be gained, leading to improved reliability and efficiency. This ultimately result reduced consumption carbon emissions, contributing sustainability efforts. paper proposes method that employs conditional variational autoencoder attention mechanism for deep clustering identify distinguish between normal datasets. The proposed model effectively addresses issue...

To address the problem that non-intrusive load monitoring algorithm has limitations and low recognition rate, identification method based on convolution neural network is proposed. Firstly, obtains start-stop state by an incident detection algorithm, which includes steps of zero-crossing, curve similarity, threshold judgment, current waveform acquisition. Secondly, converted into a grayscale graph, varying performed according to database each convolutional network. Then, type during...

Abstract With the development of smart grid and energy Internet, more sensing devices are installed used in power system, thus forming advanced metering infrastructure (AMI). It makes system generate massive data all time, which may come from meters, digital protection so on. Behavior evaluation is to filter out selected tags that meet type user portrait through known partial tags, these similar together determine user’s behavior. How make good use collected big an important research topic...

To overcome the limitations of Fe( ii )-activated percarbonate process for ACT removal, this study introduced cysteine as a complexing agent into )/SPC system and enhanced degradation efficiency ACT.

Abstract A PID‐inspired accelerated distributed optimal control algorithm is proposed for the economic dispatch problem of a multi‐bus DC microgrid, which contains both conventional generators (CGs) and renewable (RGs). Firstly, constrained optimization with aim minimizing power generation cost microgrid established. To solve problem, an in discrete‐time domain proposed. The convergence speed significantly improved compared to existing algorithms without acceleration terms. More importantly,...

The rapid growth of renewable energy and electric vehicles (EVs) presents new development opportunities for power systems storage devices. This paper a novel integrated Green Building Energy System (GBES) by integrating photovoltaic-energy vehicle charging station (PV-ES EVCS) adjacent buildings into unified system. In this system, the building load is treated as an uncontrollable primarily utilized to facilitate consumption surplus photovoltaic (PV) generated EVCS. First, strategy...

“Flexible load” refers to the load whose power consumption varies within a specified interval or transfers between different periods, including adjustable transferable loads with demand elasticity. Establishing flexible resource control platform can carry out consultative and planned dissipation according actual supply situation regulation ability of industrial commercial users, while ensuring that their rigid is not affected by dissipation, supplemented economic compensation measures,...

Redundant complexities and inadequate representation of spatiotemporal features are included in the electricity price data. To address complex data redundancy features, a gated channel mechanism (GCM) combined with high-order pooling feature enhanced convolutional LSTM network (GCHCon-LSTM) prediction model is proposed. On standardized processed dataset, vertical correlation information expanded using dual neural (GDCNN) integrated adjustment features. filtered GCM. Temporal spatial...

For the strong stochastic fluctuation of EV charging load, feature decomposition is an effective technical means to improve accuracy load prediction. However, how effectively characterize and learn these different modes after difficulty further applicability this tool in field To solve problem, ultra-short-term prediction model based on sample entropy modal classification proposed. Firstly, historical decomposed into several functions adaptively, then reflecting characteristics extracted for...

Aiming at the problem that capacity detection error of an uninterruptible special transformer is large, which affects operation effect transformer, a algorithm for based on artificial immunity and multi-object optimization proposed. The combines immune to decompose vibration signal transformer. According decomposition results, winding deformation model constructed, possible problems in can be found time, so corresponding measures taken maintenance repair, thus realizing research experimental...

In order to summarize the development achievements of demand response (DR) in domestic recent years and promote sustainable business, implementation modes DR between China American are compared based on comparative analysis relevant policies, documents, work reports demonstration projects. The around types, time scales America, as well start-up conditions, advanced notification at home. scale America is by analyzing number participant users, effects such load reduction, subsidies or...