Search

Insulators are critical components of transmission lines but are prone to failures that can jeopardize the safe operation of electrical power systems. Accurate detection of insulator defects is essential for timely maintenance. With advancements in object detection algorithm and artificial intelligence, insulator defect detection has garnered significant attention. However, detection accuracy remains an issue. To address this, we propose an improved transfer learning model. Our approach incorporates the Mish activation function and a global context network module to enhance the model's performance. The improved YOLOv9 model is trained and tested using two public datasets: the insulator defect image dataset and the China power line insulator dataset. Experimental results demonstrate that our model achieves optimal detection precision and recall rates of 99.84 and 99.92%, respectively—improvements of 1.06 and 1.09% over the actual YOLOv9. Additionally, our model outperforms other algorithms, such as RTDETR and SSD, particularly in adapting to complex backgrounds and detecting small targets. [ABSTRACT FROM AUTHOR]

Transient processes that occur during any damage to electrical power facilities carry information about the location and nature of the damage. Specific information carriers are not only the readings of observed currents and voltages but also the spectral components of these quantities. Attention is paid to the harmonics of the fundamental frequency. However, their information content cannot be sufficient to identify the damage. Moreover, there are modes in which they are compensated or absent altogether because of the disconnection of sources. The article focuses on spectral components of the free process, which is an integral part of the transient process that accompanies any damage. Being the natural frequencies of the observed network, the complex values of the frequencies of the spectral components are the same for all observation locations of a given object. The equations of the linear model of the object are compiled for these locations on the basis of the complex method, however, based only on the own complex frequency value. The spectral analysis procedures that isolate spectral components with their complex values of frequencies and amplitudes are considered. It is demonstrated that the mode of each individual spectral component can be divided into two parts, normal and local. The ratio of local complexes of the same complex frequency value on various sides of the line is uniquely related to the location of its damage. [ABSTRACT FROM AUTHOR]

The demand for real-time efficient scene comprehension has been increasing rapidly in the drone-based automatic inspection of power transmission lines (PTL). The extensive application of semantic segmentation in urban scenes proves that it can meet the requirements for scene understanding. However, existing methods have difficulty adapting to changes in the scene, which leads to problems of performance degradation and fuzzy contours of segmented objects. To overcome the existing problems, a class-aware edge-assisted lightweight semantic segmentation network is proposed in this paper. Class-aware edge detection is introduced as an auxiliary task, and a two-branch network is designed to locate instances and refine contours. Specifically, hybrid graph learning uses task-specific graph-based structures to reason attention information of region and edge features. Based on the complementary characteristic of region and edge features, cascaded shared decoders adopt specific interaction functions to enhance the ability of region features to locate targets and the ability of edge features to improve contour details. In addition, to verify the effectiveness of the proposed method, we construct two datasets named the transmission tower component recognition dataset (TTCRD) and the transmission line regional classification dataset (TLRCD). Comprehensive experiments on TTCRD and TLRCD prove that the proposed method can accurately refine the contour of objects and overcome the challenges in the two datasets. Comparison experiments and ablation experiments also demonstrate the superior performance of the proposed method and the effectiveness of each component in our architecture. [ABSTRACT FROM AUTHOR]

The questions of the determination of the fault location (DFL) of a power transmission line by the instantaneous values of currents and voltages observed on two sides were examined. There is no universal criterion of the DFL based on instantaneous values. The task is to search for a general approach to the construction of criteria that conform to the substance of the observed processes. Such an approach is the division of the observed values into two components – local and normal. The local mode is created in the transmission line model by an unknown source acting at the point of the short circuit. Significantly, the model inputs are shunted. There is the possibility of training the indicator of the determination of the fault location in local mode and obtaining its priori characteristics, activating the model in various places of probable fault, and recording local currents at the shunt inputs. In addition to the training stage, to which the main attention is paid, the fault location algorithm itself is also examined. This includes the determination of the normal mode of currents as the reaction of the model of an undamaged line to observed voltages, and determination of local currents as the differences between the observed and normal currents. [ABSTRACT FROM AUTHOR]

The commonly used fault diagnosis method of the power transmission line (PTL) is the traveling wave method which uses the traditional mode components α , β and o to realize the fault type recognizing and the fault location in distribution network. However, the traditional mode components of the traveling wave method may take the value of zero that cannot provide the effective fault feature for the A-phase ground fault, or the A and B double phases short-circuit fault. Moreover, a single traditional mode component cannot recognize all the types of the faults, and the fault recognition has to use the multiple mode components that increases the complexity. To address the aforementioned issues, this paper proposes an edge computing scheme based on a novel mode component γ for PTL fault diagnosis in distribution network. Firstly, the novel mode component γ is fused according to the traditional mode components α and β , then the propagation property of the traveling wave- γ is derived. Secondly, the edge computing scheme and the PTL fault diagnosis method are designed for distribution network by using the only novel mode component γ . Thirdly, the relationship table of the fault type and the boundary conditions as well as the component γ expressions is derived for the fault recognition using the only traveling wave- γ . In addition, the fault distance computing is explored by using the traveling wave- γ . The computing is composed with the wave- γ head identification using the optimal Wavelet vanishing moment parameter, and the arrival time determination, and the fault location method. To evaluate the efficiency of the proposed edge computing scheme, the simulation experiment and the laboratory tests are conducted. The simulation experiment results show that, the novel mode component γ not only provides the effective fault features for all the ten types of the power transmission line faults, but also increases fault locating accuracy by the average 2.08 % compared with the methods using the traditional mode component α , β and o. Meanwhile, the laboratory tests show that the edge computing method is practical. [ABSTRACT FROM AUTHOR]

The tasks of phase selection (recognition of damaged phases) and location (determination of the fault location) are solved by a universal algorithm for monitoring an arbitrary site of the presumed fault location. The currents and voltages observed on one side are converted into estimates of the electrical quantities at that location, which are determined on the assumption that the line is not damaged before that position. The converter is an algorithmic model of the initial section of the line with variable length. The algorithmic model is created based on a priori information about the line parameters. Information about the structure and parameters of the rest of the network is required to determine a priori coordinate functions that express the currents in the damage branches through the components of the output voltages or currents of the algorithmic model. The type of components is determined by the completeness of the observer's information base. The most valuable emergency components in terms of information are available, when a short circuit (SC) occurs against the background of the previous mode. However, in the situation when voltage is applied to an already damaged line, then one must be satisfied with the components of the zero and negative sequences. The null value of the reactive power of all damaged phases serve as the criterion for determination of the fault location (DFL), and the boundary conditions for single- and two-phase short circuits as the criterion for phase selection. The tasks of selection and location are closely intertwined. Generally speaking, a complete information base provides an opportunity to perform the detection and ranging without regard to the type of short circuit. But this will be a preliminary estimate of the location of the short circuit. Recognition of damaged phases will result in its refinement. But with an incomplete information base, phase selection certainly precedes location. [ABSTRACT FROM AUTHOR]

A new approbation of the previously presented method for recognizing a failed branch in one-sided observation of a power transmission line has been carried out. A real line of short length with four branches, two of which are almost identical, is considered. The line connects a high-power source with a low-power one. Each branch is characterized by an a priori hodograph, and the observed mode is displayed by a coordinate hodograph, the expected intersection of which with the a priori hodograph of the failed branch should occur when the coordinates of the branch and the hodograph coincide. An ambiguity of the solution was found when one of the identical branches was failed, if the observation was carried out from the sending side of the line. However, when the observation was carried out from the receiving side, the failure was unambiguously recognized. It is shown that unambiguous recognition can also be provided from the sending side of the line by resorting to cascade disconnection of the line. [ABSTRACT FROM AUTHOR]

In this paper, a system is proposed for visualizing and analyzing icing monitoring data of power transmission lines. The distributions of temperature and humidity are visualized by two-dimensional maps with customizable map layers. The multi-dimensional monitoring data are visualized as parallel coordinates. Moreover, a prediction algorithm that is based on a hybrid deep belief network is integrated into the system for predicting the icing thickness. If the icing thickness of a certain location exceeds the threshold value, the historical meteorological data of the location can be visualized as line graphs, which helps to choose the appropriate de-icing measures. According to the experimental results, our system is capable of reflecting the statistical features of icing monitoring data with high accuracy of icing thickness prediction. [ABSTRACT FROM AUTHOR]

The results of using a mathematical model to study advanced algorithms for the automatic load-frequency control (ALFC) system of 800 MWPower Unit 1 of the Perm TPP islanded as part of the Perm-Zakamsky energy hub from the UES of Russia are presented. To prevent long-term oscillations of power and frequency of the power unit during power surges due to the reduced capacity of the power lines, advanced ALFC algorithms are proposed and analyzed. A criterion for the activation of these algorithms is proposed, and the optimal one is chosen. [ABSTRACT FROM AUTHOR]

This article considers the effects of wire breakage in networks with an insulated neutral, as well as issues related to detecting the location of such damage. The need to use means and techniques for rapidly localizing line sections with wire breakage is revealed and justified. The analysis conducted allows proposing a technique for detecting line sections with phase wire breakage in branched networks with an insulated neutral. This technique is based on analyzing voltage readings from the meters installed in the low-voltage part of transformer substations connected to the damaged network. The implementation of this technique requires using software that is integrable with available electronic supervisory control resources without installing additional equipment because it is possible to use electric counters with data transmitters installed everywhere at transformer substations as voltage meters. [ABSTRACT FROM AUTHOR]

Since the power transmission line (PTL) passes through the high mountain and heavy snowfall region, it is necessary to keep the stability of the PTL. In this study, PTL is modeled as the mass-spring-damper system by using multi-body dynamics analysis program. In order to analyze the dynamic behavior of PTL, a damping coefficient for a multi-body model is derived by using the free vibration test and Rayleigh damping theory. The icing cross section of the transmission line is considered by ellipse and triangle shape. The aerodynamic coefficient for each cross sections are derived by using the commercial CFD program, ANSYS Fluent. The fluctuating wind velocity is regenerated with time history by using Kaimal spectrum. Galloping simulations are performed for the elliptical and triangular iced sections by using the generated fluctuating wind velocity. There is an attack angle which showed the maximum vertical displacement according to the icing section. As a results, the triangular icing shape on the fluctuating wind velocity is more dangerous for the galloping phenomenon. This results can be available to the tower design for the power transmission line. [ABSTRACT FROM AUTHOR]

When the power transmission line (PTL) passes through a region of high mountains and heavy snowfall, maintaining the stability of the PTL and avoiding sleet jump and galloping are necessary. In this study, PTL is modeled as the mass-spring-damper system by using the multi-body dynamics analysis program, RecurDyn. The lumped mass model is compared with the finite element model based on deflection. To analyze the dynamic behavior of the PTL under icing and wind conditions, we obtained a damping coefficient for a multibody model from the free vibration test and Rayleigh damping theory. The icing cross-section of the transmission line is assumed to have ellipse and triangle shapes. The aerodynamic coefficients for each cross-section are derived by using the commercial CFD program, ANSYS Fluent. The occurrence of galloping is simulated for each shape according to the attack angle. Results indicate that the dynamic behavior of the PTL and the galloping conditions such as the icing shape, thickness, and attack angle, can be analyzed. Furthermore, the effects of each factor are evaluated. In the elliptical icing section, the effect of icing thickness is high; whereas in the triangular icing section, the wind velocity highly affects galloping. [ABSTRACT FROM AUTHOR]

Due to the merits of high inspection speed and long detecting distance, Ultrasonic Guided Wave(UGW) method has been commonly applied to the on-line maintenance of power transmission line. However, the guided wave propagation in this structure is very complicated, leading to the unfavorable defect localization accuracy. Aiming at this situation, a high precision UGW technique for inspection of local surface defect in power transmission line is proposed. The technique is realized by adopting a novel segmental piezoelectric ring transducer and transducer mounting scheme, combining with the comprehensive characterization of wave propagation and circumferential defect positioning with multiple piezoelectric elements. Firstly, the propagation path of guided waves in the multi-wires of transmission line under the proposed technique condition is investigated experimentally. Next, the wave velocities are calculated by dispersion curves and experiment test respectively, and from comparing of the two results, the guided wave mode propagated in transmission line is confirmed to be F(1,1) mode. Finally, the axial and circumferential positioning of local defective wires in transmission line are both achieved, by using multiple piezoelectric elements to surround the stands and send elastic waves into every single wire. The proposed research can play a role of guiding the development of highly effective UGW method and detecting system for multi-wire transmission line. [ABSTRACT FROM AUTHOR]

The article contains the analysis of the influence of the phase shifter (PS) on the energy processes in the power transmission line in terms of the two-unit model of the electric network. The approach to synthesis of the models regulated by the phase shifter providing for both calculation of the steady operation modes of the electric networks with the phase shifters and research of the electromagnetic processes and designing of the device itself is offered. [ABSTRACT FROM AUTHOR]

In this paper, 50 Ohm coaxial line transitions required for sensing dielectric (or ice) coatings on power transmission lines (Goubau lines) at microwave frequencies (below 10 GHz) are proposed, designed and simulated. Two different types of transitions are needed: uni- and bi-directional. Computer simulations are performed using finite element methods (FEM). Good performance results using both types of microwave transitions are obtained. Design and simulation results show that Goubau line type transitions can be realized for various applications including detection of ice layers and microwave communications on electric power transmission lines. [ABSTRACT FROM AUTHOR]

In this paper, measurement of ice and ice/water layer thickness on power transmission lines using Goubau modes at 1, 2.54 and 10 GHz is analyzed, simulated and discussed. Ice layer thicknesses change the phase and attenuation of Goubau mode propagation over a given line length. It is found feasible to determine layer thickness by measuring the phase and attenuation of the Goubau mode at a few GHz. In addition to the thickness, it is also possible, to determine the composition of the layer by measuring the different dielectric properties of ice, water and their mixture over a wide bandwidth to develop a database for comparison with measured data. The sensitivity of suggested instrumentation is sufficient for short line lengths (up to 1 m) or long lengths (several km). The implementation of instrumentation on long lines is still to be done but no major obstacles are foreseen. The decision to de-ice transmission line using any given and established techniques is facilitated if average ice thickness on transmission line between supporting transmission line towers can be measured. [ABSTRACT FROM AUTHOR]

High resistance fault poses an enormous challenge to the existing algorithms of fault detection and fault classification. In this paper, the standard deviation and accumulation method are employed to perform the fault detection and classification. It is primarily built in two stages. Firstly, the standard deviations for the measured current's signals of the local and remote terminals is computed to extract the fault feature. Secondly, the cumulative approach is used to enlarge the fault feature to perform the high resistance fault. The proposed scheme is known as Standard Deviation Index (SDI), and it is obtained for the three phases and zero sequence. The proposed algorithm has been tested through different fault circumstances such as multiple faults locations, fault resistances, and fault inception time. Moreover, far-end faults with high-resistance, faults happened nearby the terminal, faults considering variable loading angle, sudden load change, different sampling frequency, bad signaling and a fault occurred in the presence of series compensation are also discussed. The results show that the proposed scheme performed remarkably well regarding the fault with resistance up to 1.5kΩ and can be detected within a millisecond after the fault inception. Additionally, the computational simplicity that characterizes the processes makes it more efficient and suitable for domain applications. [ABSTRACT FROM AUTHOR]