Because insulators provide electrical insulation and mechanical support for electric transmission lines, these components are of paramount importance to safe reliable operations power systems. However, often considered be prone different faults, e.g., bunch-drop, which demands a novel solution accurate fault detection location. Current research efforts have primarily focused on the bunch-drop glass insulators, study ceramic has not been reported date. To this end, paper proposes an...

Bolts are the most numerous fasteners in transmission lines and prone to losing their split pins. How realize automatic pin-missing defect detection for bolts so as achieve timely efficient troubleshooting is a difficult problem long-term research target of power systems. In this article, an model called visual shape clustering network (AVSCNet) constructed. First, unsupervised method shapes proposed applied construct that can learn difference shape. Next, three deep convolutional neural...

Transmission lines' status detection and fault diagnosis are based on the high-accuracy of typical fittings in aerial inspection photographs. Detecting tiny-size complicated sceneries is still an unresolved issue, nevertheless, because constraints small number feature pixels. Considering characteristics fitting connection structures, we proposed a novel method named context-based graph reasoning model (CGRM), which multiscale context, self-attention map, learning. This includes one...

Currently, reliable and accurate object detection in high-resolution remote sensing images still faces significant challenges, such as color, aspect ratio, complex background, scale variations. Even the results obtained based on latest convolutional neural network (CNN) methods are not satisfactory. To obtain more from large-scale images, we proposed a multiscale algorithm oriented to adaptively attentional feature fusion YOLOX algorithm. First, (MSAFF) structure was added increase...

During the last decade, neural networks have emerged as one of most powerful and accurate nonlinear models for load forecasting. However, using requires users to in-depth knowledge determine model structure parameters, which limits their wide application. To overcome this weakness, paper proposes an integrated approach combines a self-organizing fuzzy network (SOFNN) learning method with bilevel optimization method. SOFNNs can automatically both while selects best pre-training parameters...

Aiming at the problems of complex background, tiny-size objects, long-tailed distribution, and so on, a hybrid knowledge region-based convolutional neural network (HK R-CNN) is proposed to detect multiple fittings in aerial images transmission lines. First, structure combination rules line are studied, relationships co-occurrence connection spatial location between effectively extracted through data-driven way. Second, position-sensitive score map (PSSM) utilized express immobilized extract...

Aiming at the problems of complex background, diverse shapes, and object occlusion in aerial images, a cascade reasoning graph network (CRGN) is proposed for multi-fitting detection on transmission lines. First all, these three mentioned above, co-occurrence knowledge, semantic spatial knowledge were constructed to represent co-relation objects by analyzing characteristics line fittings. Next, Supervised Graph Learning (SGL), Attention (GAT), Convolutional Network (GCN) employed reason...

Accurate infrared thermal image instance segmentation of substation equipment is a prerequisite for intelligent analysis its temperature status. To address the issues low accuracy and false detection in existing methods, we propose visual feature reasoning-based method to compensate limitations deep learning methods improve accuracy. We utilizing distinctive features as priori knowledge three types construct two-branch model (FR-SOLOv2) based on power domain expertise reasoning. FR-SOLOv2...

An approach of a mean hourly wind speed forecasting in farm is proposed this paper. It applies support vector regression as well error estimation. Firstly, applied to the forecasting. Secondly, classifier trained estimate error. Finally, results can tailor themselves estimated error, and thus improve precision. To test approach, three-year data from given process, addition Experimental show that achieve higher quality forecasting; also it has lower square compared with traditional

Accurate and timely detection of insulator flashover on power transmission lines is paramount importance to utilities. Most available solutions mainly focus the exploitation mechanism or discharge area detection, rather than identification a damaged due flashovers using captured aerial images. To this end, paper proposes multi-saliency aggregation-based porcelain fault approach. The target determined Faster-Pixelwise Image Saliency by Aggregating (F-PISA) algorithm based color structural...

With the widespread application of digital images in various domains, accurate measurement image quality has become particularly crucial. This paper introduces a novel multi-branch multi-layer feature fusion network (MFFNet) to address inadequate expression multi-scale and semantic features local visual consideration existing no-reference assessment algorithms. MFFNet comprises primary sub-branch. Through convolutional neural extraction, main branch uses enhancement (MSFE) module capture...

During an automatic power transmission line inspection, a large number of images are collected by unmanned aerial vehicles (UAVs) to detect existing defects in components, especially insulators. However, with twin insulator strings the inspection images, when umbrella skirts rear string obstructed front string, defect detection becomes difficult. To solve this problem, we propose method self-shattering insulators based on spatial features contained images. Firstly, segmented according...

Abstract To better detect targets that may cause external damage to transmission lines, the authors present You Only Look Once‐Asymptotic Feature Pyramid Network (YOLO‐AFPN), a lightweight but efficient model. Firstly, adopt feature comparison strategy based on knowledge of line scenes, which facilitates increased attention target features during training. Secondly, YOLOv8 detection network is built, and backbone adds three layers simple parameter‐free module, extracts while maintaining...