Search

This paper describes an algorithm for generating synthetic datasets of lightning flashovers on medium voltage overhead distribution lines (OHL). Datasets are heterogeneous, hierarchical and class--imbalanced. They are designed specifically for training machine learning (ML) models (i.e. binary classifiers) for predicting lightning flashovers on OHLs, and they are built from both direct and indirect lightning strikes. Algorithm for creating datasets employs the Monte Carlo method and makes use of the several underlying mathematical models of lightning interaction with the OHL, which introduce ``noise'' and make training of the ML models more challenging. Algorithm is versatile enough to account for the different statistical properties of lightning interactions with distribution lines, starting from the bivariate probability distribution of lightning--current parameters. Example of the synthetic dataset will be presented, using a typical medium voltage OHL geometry with a horizontal conductors arrangement., Preprint of the peer-reviewed and accepted version of the paper for the IEEE EUROCON 2023 international conference.

The paper deals with hardware-in-the-loop (HIL) laboratory setup consisting of the arbitrary waveform generator (AWG) and the device used for grid voltage fundamental harmonic amplitude and frequency real-time (RT) estimation. The device under test (DUT) is realized as desktop target PC with Phar Lap RT operating system using NI PCIe 6251 DAQ card. On the other hand, the AWG serves as a three-phase or single-phase grid disturbance simulator. The algorithm based on the linear Kalman filter and the frequency-locked loop has been used as robust and accurate estimation technique. Both AWG and DUT have been driven from the host PC. Furthermore, LabViewallows user to configure and perform remote online monitoring of the DUT, i.e., to track amplitude and frequency changes point-by-point on the host PC over Ethernet TCP/IP protocol. The AWG allows various real-time case study analysis to be performed in accordance with up to date power quality standards. For a selected case studies, the DUT is verified by comparing experimental results against IEEE and IEC standards for power quality analysis. An example of grid voltage amplitude and frequency online monitoring is also presented. Moreover, proposed HIL laboratory setup can be used for comparison of various estimation techniques for control and protection purposes in microgrids.

Dataset has been generated from the Monte Carlo simulations of lightning flashovers on medium voltage (MV) distribution lines. It is suitable for training machine learning models for classifying lightning flashovers on distribution lines, as well as for line insulation coordination studies. Following five different types of lightning interaction with the MV distribution line have been simulated: (1) direct strike to phase conductor (when there is no shield wire present on the line), (2) direct strike to phase conductor with shield wire(s) present on the line (i.e. shielding failure), (3) direct strike to shield wire with backflashover event, (4) indirect near-by lightning strike to ground where shield wire is not present, and (5) indirect near-by lightning strike to ground where shield wire is present on the line.

This synthetic dataset was generated from Monte Carlo simulations of lightning flashovers on medium voltage (MV) distribution lines. It is suitable for training machine learning models for classifying lightning flashovers on distribution lines. The dataset is hierarchical and class imbalanced. Following five different types of lightning interaction with the MV distribution line have been simulated: (1) direct strike to phase conductor (when there is no shield wire present on the line), (2) direct strike to phase conductor with shield wire(s) present on the line (i.e. shielding failure), (3) direct strike to shield wire with backflashover event, (4) indirect near-by lightning strike to ground where shield wire is not present, and (5) indirect near-by lightning strike to ground where shield wire is present on the line. Last two types of lightning interactions induce overvoltage on the phase conductors by radiating EM fields from the strike channel that are coupled to the line conductors. Three different methods of indirect strike analysis have been implemented, as follows: Rusck's model, Chowdhuri-Gross model and Liew-Mar model. Shield wire(s) provide shielding effects to direct, as well as screening effects to indirect, lightning strikes. Dataset consists of two independent distribution lines, with heights of 12 m and 15 m, each with a flat configuration of phase conductors. Twin shield wires, if present, are 1.5 m above the phase conductors and 3 m apart [2]. CFO level of the 12 m distribution line is 150 kV and that of the 15 m distribution line is 160 kV. Dataset consists of 10000 simulations for each of the distribution lines.

This dataset contains phasor measurements (PMU-type) signals from the IEEE New England 39-bus power system test case network, which are generated from a large corpus of systematic MATLAB®/Simulink electro-mechanical transients simulations. It was prepared to serve as a convenient and open database for experimenting with different types of machine learning (including deep learning) techniques for transient stability assessment (TSA) of electrical power systems. A dataset contains time-domain signals from 9360 simulations. Different load and generation levels of the New England 39-bus benchmark power system are systematically covered, as well as all three major types of short-circuit events (three-phase, two-phase and single-phase faults) in all parts of the network. The consumed power of the network was set to 80%, 90%, 100%, 110% and 120% of the basic system load levels (for different system load levels, both generation and loads are scaled by the same ratio). The short-circuits are located on the busbar or on the transmission line (TL). When they are located on a TL, it was assumed that they can occur at 20%, 40%, 60%, and 80% of the line length. Timing of the fault occurrences takes into the consideration a moment on the instantaneous sinusoidal reference voltage. The observation period of each simulation was set at 3 seconds and signals are sampled at 1/60 s resolution. Many different machine electrical and mechanical (rotor and stator quantities), as well as network (three-phase currents and voltages), time-domain signals are obtained from simulations with a PMU-type resolution.

Transformer health index (HI) is very useful for representing the overall health of a power transformer, due to the fact that it quantifies equipment condition based on different criteria that are related to the long-term degradation factors that cumulatively lead to the asset's end-of-life. Moreover, transformer HI is a very powerful tool for managing assets and identifying future investment needs, as well as prioritizing present capital allocations and asset maintenance programs. This paper presents a novel approach to the transformer asset management, which leverages a combination of Bayesian statistical model for transformer HI analysis and influence diagrams. Influence diagrams, as extensions of the Bayesian networks, offer transparent decision-making process under uncertainty, while accounting for transformer HI test reliability and maintenance costs. Asset-management using the proposed model will be demonstrated on a concrete transformer data., This is the pre-print of a paper submitted to the IEEE ICIAES 2019 conference: http://iciaes.com/

Health index (HI) is a very useful tool for rep- resenting the overall health of a complex asset, such as the power transformer, due to the fact that it quantifies equipment condition based on different criteria that are related to the long- term degradation factors that cumulatively lead to the asset’s end-of-life. The main concern with HI computation is with the practical management of the numerous criteria that are combined in different ways (with proprietary information and associated weighting factors) to produce a HI value. Hence, several authors have proposed different approaches to the HI calculation, e.g., analytical expressions, logistic regression, fuzzy logic, support vector machines, and artificial neural networks. This paper proposes using Bayesian multinomial logistic regression for the HI calculation. This approach offers high flexibility with multiple metric and/or nominal predictors, including correlation and interaction between predictors, and acknowledges the fact that the transformer HI is described with three to five categories. It further offers high model interpretability and benefits from the Bayesian ability to quantize uncertainty in model parameters.

This paper describes usage of several third-party Python (programming language) libraries, which could be seen as very useful for easy and efficient manipulation and analysis of data produced by the lightning detection networks (LDNs). Python is a high level interpreted language with extensive library of third-party packages and APIs. Several of its libraries, popular within the data science community, could be seen as very useful for efficient analysis of data produced by LDNs, such as: geographical and probability density distribution of lightning strikes, lightning flash-cell identification, its movement tracking, and nowcasting.

This paper introduces application of genetic algorithm (GA) in designing air-termination system of the high-voltage open-air substation external lightning protection system (LPS). The air-termination system, as a part of the external LPS of high-voltage substation, is known to be cumbersome to design using traditional techniques, where optimization is often relegated to the trial-and-error approach. In addition, traditional methods are often associated with geometric constraints and various implementation difficulties. This paper proposes a novel approach which utilizes a combination of statistical LPS efficiency and GA in designing techno-economically optimal external LPS of open-air substations. Proposed approach offers a LPS designer unique and valuable assistance in optimally arranging elements of air-termination system for obtaining maximum lightning shielding effects with minimum total investments.

Transmission line (TL) backflashover (BF) performance is traditionally ascertained using a single number, the backflashover rate (BFR), which is measured in number of BF events per 100 km-years. This paper aims at presenting a novel indicator for measuring performance of high- voltage TL tower's ability to tolerate direct lightning strikes without provoking a BF event. This indicator is also a single number, which can be computed for any TL tower (by means of EMTP simulations) and measures, in a novel way, its tolerance against BF events. It is given in terms of the risk of the BF occurrence, which means it is statistical in nature and depends on the total sum of conditions governing the BF events. The BF risk, as an indicator, is obtained from the probability density function of the shield wire(s) incident lightning currents and the cumulative distribution function of the BF currents statistical distribution. Hence, it merges complete probabilities of obtaining lightning currents striking a tower with probabilities of those currents provoking a BF events on that tower. This novel risk-based indicator can be related to the price of that risk and the associated investment costs, enabling the cost-effective optimisation solutions to the problems of TL arrester applications and station insulation coordination design.

This paper presents a method for estimating the backflashover rate (BFOR) on high voltage (HV) transmission lines. The proposed method employs a state-of-the-art model of HV transmission line (TL) for backflashover analysis, assembled within the EMTP-ATP program―with EMTP running in batch mode―to construct the curve of limiting parameters (CLP). The probability of backflashover occurrence at a particular TL tower is obtained by computing the volume under the surface of the probability density function of the bivariate log-normal statistical distribution (of lightning current parameters), bounded by the CLP in the coordinate system of lightning-current amplitude and wave-front duration. This probability, in combination with the estimated number of direct lightning strikes to TL towers, obtained from the application of the electrogeometric model, provides the associated BFOR, assuming it is further normalized on the basis of 100 km-years. The usage of the proposed method is demonstrated on a typical HV transmission line. A sensitivity study for estimating the BFOR with the proposed method is provided as well.

One of the major difficulties introduced with wind power penetration is the inherent uncertainty in production originating from uncertain wind conditions. This uncertainty impacts many different aspects of power system operation, especially the balancing power requirements. For this reason, in power system development planing, it is necessary to evaluate the potential uncertainty in future wind power generation. For this purpose, simulation models are required, reproducing the performance of wind power forecasts. This paper presents a wind power forecast error simulation models which are based on the stochastic process simulation. Proposed models capture the most important statistical parameters recognized in wind power forecast error time series. Furthermore, two distinct models are presented based on data availability. First model uses wind speed measurements on potential or existing wind power plant locations, while the seconds model uses statistical distribution of wind speeds., {"references":["Y. Zhang and K. W. Chan, \"The impact of wind forecasting in power\nsystem reliability,\" in Third International Conference on Electric Utility\nDeregulation and Restructuring and Power Technologies, April 2008,\npp. 2781–2785.","L. Soder, \"Simulation of wind speed forecast errors for operation\nplanning of multiarea power systems,\" in International Conference\non Probabilistic Methods Applied to Power Systems, Sept 2004, pp.\n723–728.","P. Haessig et al., \"Energy storage sizing for wind power: impact of the\nautocorrelation of day-ahead forecast errors,\" Wind Energy, pp. 1 – 18,\n2013.","X. Wang, P. Guo, and X. Huang, \"A review of wind power forecasting\nmodels,\" Energy Procedia, vol. 12, no. 0, pp. 770 – 778, 2011, the\nProceedings of International Conference on Smart Grid and Clean\nEnergy Technologies (ICSGCE 2011.","F. Marzbani, A. Osman, M. Hassan, and T. Landolsi, \"Short-term wind\npower forecast for economic dispatch,\" in 5th International Conference\non Modeling, Simulation and Applied Optimization (ICMSAO), April\n2013, pp. 1–6.","L. Landberg, \"Short-term prediction of the power production from\nwind farms,\" Journal of Wind Engineering and Industrial Aerodynamics,\nvol. 80, no. 1, pp. 207 – 220, 1999.","U. Focken et al., \"Short-term prediction of the aggregated power output\nof wind farmsa statistical analysis of the reduction of the prediction\nerror by spatial smoothing effects,\" Journal of Wind Engineering and\nIndustrial Aerodynamics, vol. 90, no. 3, pp. 231 – 246, 2002.","Y. Han and L. Chang, \"A study of the reduction of the regional\naggregated wind power forecast error by spatial smoothing effects in\nthe maritime canada,\" in 2nd IEEE International Symposium on Power\nElectronics for Distributed Generation Systems (PEDG), June 2010, pp.\n942–947.","M. Lange, \"On the uncertainty of wind power predictions - analysis of\nthe forecast accuracy and statistical distribution of errors,\" Journal of\nSolar Energy Engineering, vol. 127, no. 2, pp. 177 – 184, 2005.\n[10] X. Y. Ma, Y. Z. Sun, and H. L. Fang, \"Scenario generation of wind power\nbased on statistical uncertainty and variability,\" IEEE Transactions on\nSustainable Energy, vol. 4, no. 4, pp. 894–904, Oct 2013.\n[11] H. Bludszuweit, J. A. Dominguez-Navarro, and A. Llombart, \"Statistical\nanalysis of wind power forecast error,\" IEEE Transactions on Power\nSystems, vol. 23, no. 3, pp. 983–991, Aug 2008.\n[12] C. Cullen, Matrices and Linear Transformations: Second Edition, ser.\nDover Books on Mathematics. Dover Publications, 2012.\n[13] S. Kachigan, Multivariate Statistical Analysis: A Conceptual\nIntroduction. Radius Press, 1991."]}