The current ship target detection algorithm is not adaptable to changes in size and position, its accuracy speed are difficult balance the process of improvement. In particular, there little research on key parts ships. this paper, an improved YOLOv11 location proposed. It relationship between efficiency model complexity using algorithms. processing performance deforming according algorithms strong. Moreover, algorithms’ occluding overlapping targets small-size high. view practical problems...

Realizing accurate control of ship target information in complex marine environments is great significance for maintaining environment security and safeguarding maritime sovereignty. With the rapid development material technology manufacturing industry, types styles ships are increasing, distribution multi-type on sea widespread. How to realize detection identification dynamic targets an important difficult problem that needs be solved urgently current detection. In this paper, improved...

Improving decision-making in the autonomous maneuvering of unmanned aerial vehicles (UAVs) is great significance to improving flight safety, mission execution rate, and environmental adaptability. The method deep reinforcement learning makes decision UAVs possible. However, current algorithm prone low training efficiency poor performance when dealing with complex continuous problems. In order further improve level explore safe efficient methods environments, a based on hierarchical ‌Proximal...

To achieve the optimal planning of grid resource storage for a large-scale distribution network (DN), cluster partition-based zonal method DN, considering reliability, is proposed. Firstly, comprehensive clustering partition index proposed, which includes modularity index, power balance and node affiliation index. A hybrid genetic–simulated annealing algorithm employed to perform partition. Secondly, three-layer joint expansion model based on partitioning At upper level, route established...

The integration of distributed generation (DG) into distribution networks has significantly increased the strong coupling between power supply capacity and renewable energy acceptance capacity. Addressing this while enhancing both capacities presents a critical challenge in modern network development. This study introduces an innovative joint planning reconstruction strategy for storage, designed to simultaneously enhance proposed approach employs bi-level optimization model: upper level...

Considering the possible overvoltage caused by high proportion of photovoltaic systems (PVs) accessing distribution networks in future, traditional centralized control methods will be too complex to satisfy response time demands. To solve this problem paper presents a two-level voltage method. At day-ahead level, based on PV-output and load-demand forecast, community detection algorithm using an improved modularity index is introduced divide network into clusters; optimal scheduling drawn up...

This article proposes a two-layer optimal scheduling strategy to handle the overvoltage problem in high photovoltaic (PV) power-penetrated distribution networks. The voltage regulators can be classified as power utility and PV owners, which are referred stakeholders. proposed includes autonomous optimization layer coordination layer. In layer, min-max robust game model mixed-integer second-order cone programming-based respectively minimize operating costs of stakeholders stakeholder. A...

In order to realize the goal of constructing a "three-type, two-network" world-class energy internet enterprise and fully utilize value surplus resources in substations, integration substation, data center station, storage charging battery-replacing photovoltaic (PV) station 5G has become an inevitable trend. However multi-station is not suitable for all substations. this background, planning strategy substations power systems proposed. Firstly, requirements available substation are firstly...

The penetration of photovoltaic (PV) outputs brings great challenges to optimal operation active distribution networks (ADNs), especially leading more serious overvoltage problems. This study proposes a zonal voltage control scheme based on multiple spatiotemporal characteristics for highly penetrated PVs in ADNs. In the spatial domain, community detection algorithm using reactive/ power quality function was introduced partition an ADN into sub-networks. time short-term scheduling (SZS) with...

With the high penetrated photovoltaics (PVs) accessing distribution networks (DNs) in future, overvoltage of network will be more serious, and voltage control complex. To solve this problem, a cluster partition-based zonal method for DN is proposed study. First, partition using comprehensive performance index proposed: terms structure, global density quality function introduced to measure coupling degree nodes clusters. In function, inconsistency coefficient presented reflect PV output...

The low-voltage AC series arc fault brings huge hidden dangers to the safety of civil electricity, and it is difficult find a unified criterion. For such issues, neural network algorithm based on wavelet analysis feature value decomposition proposed. Firstly, DB5 base used decompose collected data into 5 layers wavelet. Then coefficients first layer are extracted construct Hankel matrix. matrix decomposed by Eigenvalue Decomposition (EVD) eigenvector σ. mean d1, square root d2 standard...