Introduction: To address the optimal scheduling problem of renewable power distribution systems, this paper proposes an integrated framework combining deep learning-based load forecasting with particle swarm optimization. Objectives: The first objective is accurate multivariate load forecasting (cooling, heating, and electric loads). The second objective is minimizing network loss by optimally scheduling electric vehicle (EV) charging locations and times. Methods: For load forecasting, a hybrid RCNN-SVR model is constructed. The convolutional neural network (CNN) acts as a feature extractor to implicitly capture representative patterns from input data, while support vector regression (SVR) produces the final load predictions. Missing and outlier data are pre-processed. For optimal scheduling, a dual-layer particle swarm optimization (PSO) algorithm is developed. The inner layer enforces system constraints, and the outer layer minimizes network loss. EV charging load is simulated using the Monte Carlo method, and two cases (variable vs. fixed charging addresses) are optimized. Results: Experimental results demonstrate that the proposed RCNN-SVR model achieves high prediction accuracy, with mean absolute percentage error as low as 2.41% for winter electric loads. The dual-layer PSO reduces peak system load from 11.2×10³ kW to 10.5×10³ kW and decreases network loss by 2.13%, effectively smoothing grid fluctuations. Conclusion: The RCNN-SVR model significantly improves multivariate load prediction accuracy compared to separate forecasting methods. The dual-layer PSO successfully converts disorderly EV charging into orderly scheduling, reducing peak load and network loss. Together, they provide a practical solution for renewable distribution system scheduling.

With the widespread application of deep learning models across various domains, concerns regarding data privacy and security have become increasingly prominent. In particular, the risk of maliciously tampering with data to implant backdoors during the model training process poses a significant security threat. Although Transformers are widely used, research on detecting backdoors in these models remains insufficient. This paper proposes an attention map-based backdoor detection method for vision Transformers. By extracting the model’s attention maps and transforming them into histograms, we effectively analyze the differences between clean and poisoned data. Experimental results demonstrate that the proposed method successfully identifies anomalous data across multiple datasets. The findings confirm that leveraging attention map analysis substantially enhances the accuracy of backdoor detection, offering an effective solution to mitigate data contamination in deep learning models.

To solve the problems of high sensitivity to data quality and limited edge computing resources, a data operation scheduling method based on multi-agent collaboration and reinforcement learning is proposed. The distributed double-layer reinforcement learning framework is constructed innovatively. The upper layer optimizes the adjustable load scheduling in the time dimension with the minimum cost of multi-agent interaction. The lower level adjusts the task allocation strategy in real time by continuous reinforcement learning based on the complementarity of the agents. Experimental results show that the proposed method improves the power data fitting degree and makes it close to 1, and effectively solves the problem of data point offset.

It is difficult for traditional data processing methods to make full use of the potential of braiding driven data, unable to quickly collect and preprocess data, and difficult to ensure the accuracy of data. Rote learning (RL) is part of the research field of artificial intelligence, which aims to enable computers to learn autonomously, just like humans. This allows understanding of relationships and patterns between data and helps computers process information quickly. In order to solve the problems of poor data integrity, slow data processing efficiency and poor information sharing in traditional data processing, and further optimize the braiding driven data processing technology, this paper combined RL with braiding driven data. Through the method of mechanical learning, the potential of weaving driving data is fully exerted, so that it can better cope with nonlinear relations and high-dimensional features. It used the effective method provided by the RL to process the braiding drive data, collect the data, and preprocessed the collected data to ensure the correctness of the data. It extracted the features of the data, which was convenient to classify the data according to its attributes. At the same time, this paper verified it by the steps of feature extraction, model training and data analysis. In order to test whether braiding drive data processing by RL can effectively solve the problems existing in traditional drive data technology, this paper tested the performance of compiled drive data processing, and the analysis results were as follows. The data integrity rate of braiding drive data was as low as 81%, which was much higher than that of traditional drive data processing. The recognition ability of data acquisition and matching was much higher than that of traditional drive data processing. Compared with the traditional drive data processing, the information sharing has been greatly improved. In terms of data processing efficiency, it is also much higher than the traditional drive data processing. It can be seen that the method of braiding drive data processing through RL effectively improves the accuracy of data processing. It strengthens the identification ability of data collection and matching, improves the sharing of information, enables users to obtain data and analyze it faster, and also improves the processing efficiency of data.

Based on the analysis of big data, this paper studies the impact of user behavior response on the cost structure of the microgrid system in the power grid system. The article first conducts in-depth research on distributed power generation and energy storage systems, focusing on the principles and output characteristics of smart power distribution and utilization in power grids, researches smart power distribution and utilization systems in power grids, and makes a more comprehensive discussion of the current situation. Secondly, a big data analysis platform was built, and distributed storage and computing were studied. The platform was used to perform distributed storage and regulation of electricity consumption data, and the electricity consumption information data was divided into the important load, controllable load and transferable load, constructed a microgrid system model based on electricity consumption behavior response, and analyzed a calculation example. After that, a micro-grid system was simulated based on HOMER software, and the optimal capacity configuration of the system was performed. Under this configuration, the micro-grid system has the highest economic efficiency. At the same time, a demand-side load control system was built. Introduced distributed power as a controllable load, integrated new energy access and load control technology, coordinated the contradiction between the power grid and distributed power, and completed a cost-benefit analysis. Finally, for demand-side management electricity price response, peak-valley time-of-use price, the most important implementation method, according to the cost structure theory, the peak-valley period is divided according to the membership function in fuzzy mathematics, and the user's response model to peak-valley time-of-use price is established. The experiment uses the original data to simulate, find the user response model parameters based on the load transfer rate, and complete the comparative analysis of the effect under the peak-valley time-of-use electricity price, which is of great significance to the implementation and improvement of the peak-valley time-of-use electricity price project. The analysis results of the calculation examples show that the method constructed in this paper can effectively realize the power quality analysis of the distribution network in the big data environment. The research results provide technical support for the management of the rural grid voltage deviation of the power company, and lay the foundation for improving the safe operation and management of the power grid.

Currently, the data management of power enterprises faces the need to analyze data sources from multiple places. However, traditional multi-source data fabric systems have problems such as low analysis efficiency and high error rates, which brings great inconvenience to the data analysis of power enterprises. In order to improve the accuracy and efficiency of data analysis in data structure systems, the intelligent system architecture is applied to the construction of source data structure systems. The main modules are data collection, data matching, data integration, and data analysis. This article uses simulated annealing genetic algorithm to perform high-performance calculations on system timing data, thus achieving data matching. This article conducted data level data integration, feature level data integration, and decision level data integration. The access survey method was used to analyze the current data management problems faced by power companies. The evaluation and analysis of general multi-source data fabric systems and multi-source data fabric systems based on intelligent system architecture were conducted using the evaluation panel evaluation method. The analysis results showed that the operational convenience of the multi-source data fabric system based on intelligent system architecture could reach 60%–80%, which greatly improved compared to general multi-source data fabric systems; the information sharing of multi-source data fabric systems based on intelligent system architecture was greatly improved; the data processing efficiency of general multi-source data fabric systems was much lower than that of multi-source data fabric systems based on intelligent system architecture; however, the symmetry of data collection and matching in the multi-source data fabric system based on intelligent system architecture was slightly insufficient, and further improvement was still needed. In order to benefit more power companies through the intelligent system architecture based multi-source data fabric system, it was necessary to strengthen the management of data collection and matching symmetry.

Motivated by the progress in artificial intelligence such as deep learning and IoT networks, this paper presents an intelligent flink framework for real-time voltage computing systems in autonomous and controllable environments. The proposed framework employs machine learning algorithms to predict voltage values and adjust them in real-time to ensure the optimal performance of the power grid. The system is designed to be autonomous and controllable, enabling it to adapt to changing conditions and optimize its operation without human intervention. The paper also presents experimental results that demonstrate the effectiveness of the proposed framework in improving the accuracy and efficiency of voltage computing systems. Simulation results are provided to verify that the proposed intelligent flink framework can work well for real-time voltage computing systems in autonomous and controllable environments, compared with the conventional DRL and cross-entropy methods, in terms of convergence rate and estimation result. Overall, the intelligent flink framework presented in this paper has the potential to significantly improve the performance and reliability of power grids, leading to more efficient and sustainable energy systems.

原有的230 MHz射频技术在授权用户受到认知用户干扰时,通信过程误码率较高,为此,提出一种基于电力信息可视化采集的230 MHz射频技术.设计射频技术中接收灵敏度、动态范围、接收带外抑制等接收发射指标,优化基带对有用信号的调节,消除认知用户的干扰;引入可视化信息采集技术,由230 MHz无线专网承担通信工作,设计时序逻辑控制,采用过分配方式定义GTS和CFP的分配,保留通信过程中的处理时间,减小干扰.对比实验结果表明在光谱强度相同时,设计的技术误码率较低,具有有效性.

由于目前方法未能分析和挖掘电网用户行为,使用户的商品属性偏好与预计营销偏好存在差异,导致电网企业营销推荐结果不理想,为此提出基于用户行为数据的电网企业营销推荐系统.通过系统硬件和软件相互协作设计,从用户历史行为出发,优先分析处理用户的历史交互行为,对用户的行为喜好进行分类,挖掘用户的商品属性偏好,实现用户近期需求预测以及意向商品推荐.实验结果证明,所设计系统能够有效提升推荐速率和用户满意度,获取效果较好的推荐结果.

为了提高智能电网管理水平,提出基于熵权法的智能电网管理水平评价指标量化方法.采用多直流馈入方法构建智能电网管理水平评价指标体系,以输入电流、功率因素以及电压等参数为约束指标,建立智能电网管理水平评价指标参数模型;结合无功潮流耦合分析方法进行智能电网管理水平参数分析,通过恒功率、恒电流补偿方法进行智能电网管理的最优代价函数分析,建立智能电网管理的代价约束模型;通过电压电流变化约束分析的方法,进行智能电网管理水平评价的熵权指标参数分析,提取智能电网管理水平评价体系的熵权特征量;结合大数据挖掘和智能调度,实现智能电网管理水平评价指标的量化评价.仿真结果表明,采用该方法进行智能电网管理水平评价的量化分析能力较好,评价结果准确可靠,提高了智能电网管理水平和电力资源调度能力.

由于主数据驱动质量控制过程对于资产档案架构的适配性较差,导致数据传输能力较差,因此,设计基于IPv6流标签的企业资产档案主数据驱动质量控制方法.更新主数据IPv6流标签报头格式,将处理后的数据输入到驱动控制器中,通过设定约束条件,将控制器等价为动态线性化数据模型.通过实验结果对比可知,使用此方法后数据包的转发能力与传输能力得到明显的提升,在日后的资产档案主数据驱动管理中可使用此方法,为资产管理提供便利.

Modeling interactions among cyber-physical services is a crucial problem for analyzing behaviors of a composed system, where these cyber-physical services communicate with each other via synchronous and asynchronous messages. In this paper, we present an approach for modeling interactions of cyber-physical systems, which can be automated under the tool support. First, a service model is proposed to represent a cyber-physical service using a labeled transition system, where the order of executing transitions in a cyber-physical service is defined. Second, a synchronous composition is defined to compose a set of cyber-physical services communicating with synchronous messages into a cyber-physical system. Third, an asynchronous composition is defined to compose cyber-physical services interacting asynchronously through FIFO buffers, where three types of interactions are considered. Finally, experimental results show that our approach is automated and effective.

The quality of training sample is an important factor for wind speed prediction using data‐driven approaches, such as deep learning. This paper proposes a novel local predictor based on dynamic time wrapper (DTW) as training sample adaptation for wind speed prediction. After analyzing the similarity of wind speed time series using dynamic time wrapper, a local predictor is applied to improve the quality of training sample. An evaluation index is firstly proposed for estimating the quality of selected training sample. To verify the effectiveness of the proposed method, a random forest model with local predictor based on dynamic time warping (RF‐DTWLP) is applied to predict the wind speed. The simulation results demonstrate the prediction of wind speed by local predictor based on DTW has higher prediction accuracy. © 2021 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC.

本次研究详细介绍了配电网雷害风险评估的数据处理方案,提出了导线因雷击而产生过电压的计算方法.最后基于雷电灾害停电风险计算原理引入蒙特卡罗抽样调查法对配电网各个块区域的故障状态进行抽样检测,并阐述了抽样检测的具体流程.

针对输变配电网图模数据由不同系统先后维护,导致图模数据未建立有效的图模交互和共享机制,制约了输变电可靠性、站线变户关系等应用正常开展的问题,文章采用图模数据拓扑关系分析、Trie树算法与模糊匹配相结合的方式,提出了一种维护调度自动化系统的变电站图模数据和GIS系统的输配电图模数据模型拼接的方法.首先对数据进行拓扑追踪,记录拓扑关系和关键属性;其次根据拓扑关系和设备关键字段,采用Trie树算法进行相似性判断,探索并提出了一种新的基于过滤方法的Trie算法,对中文字符串进行模糊匹配,保证模糊匹配的效率和有效性,为EMS变电图模数据与GIS输配电线路的模型拼接提供了一个可行的方案,保证了图模数据在不同系统中的集成贯通,全面推进管理业务的融合和提升.

为提升电网运行的稳定性与经济性,提出基于CSMC模型的电网规划指标相关性计算方法.根据电网规划指标边缘分布函数、节点指标空间与时间的相关性,综合多元指标概率分布模型与CSMC模型,设计可同时分析电网规划指标空间与时间相关性的多元指标一阶CSMC模型,并利用状态转移核与状态转移密度描述该模型.选取核密度估计法确定电网规划指标核密度,对核密度实施积分处理得到电网规划指标的边缘分布函数,基于电网观测指标二元频率直方图选取Copula函数,并利用极大似然法估算Copula函数内的未知参数,根据得到的Copula函数求取相应的状态转移核或状态转移密度.实验结果表明,该方法能够准确分析电网规划指标的特征与规律,提升电网建设工程的稳定性与经济性.

In the three-dimensional scene,to improve the management efficiency of substation equipment and the accuracy of data display,this paper uses GIS technology to realize the management and data display of substation equipment.First,the laser scanner is used to obtain the information of substation equipment,and the specific location of the equipment is obtained according to GIS technology.Combined with the two kinds of information obtained above,the three-dimensional model of substation equipment is constructed.Then,the substation equipment data management database is established,and the spatial database is constructed by AreGIS Server software,which is imported into the substation equipment data management database to carry out comprehensive management.Based on this,in the graphic data display unit,the composite object technology of 2D GIS and 3D GIS is used to display the graphic data of 3D model.Experimental results show that the model has a higher efficiency of substation equipment management,can ensure a higher success rate of data display,and its imaging definition is higher.

目前,在电力设备的日常巡检和试验中,积累了大量关于设备故障情况的记录,缺乏相应的故障处理措施.传统集中式数据处理模型无法支持当前的强大电网系统.针对这一缺陷,文章提出了一种基于热点数据的电网边缘侧设备缺陷智能识别模型,模型分为云中心层、边缘层和现场层.在现场层,使用动态时间规整(Dynamic Time Warping,DTW)补充算法补充传感器数据;采用树突神经元模型(Dendritic Neuron Model,DNM)在边缘层进行故障初等分类,并将分类结果上传至云中心层;在云中心层利用数据之间的相关性实现故障分类.最后在公开数据集上进行设备缺陷识别模型验证,验证了模型的有效性和可行性.