Time-Frequency Analysis based Approach to Islanding Detection in Micro-grid System

This paper offers a pattern identification-based scheme for discovering the islanding conditions in synchronous generator-based distributed generation (DG) systems. The chief purpose of the suggested scheme in this investigation is measuring the Changes of Magnetomotive Force (MF) in the airgap of DG (ΔMF) within 412 ms after events, Which is less than 2 seconds by IEEE std.1547. First, The MF signal sampled at the DG was inspected by Mathematical Morphology Filter (MMF) for noise reduction. Then, a k-Nearest Neighbor (k-NN) classifier was taught by ΔMF of DG to recognize the islanding conditions from the other typical operation conditions. The consequences of the inspections on the different islanding conditions incorporating system loading and network operation conditions of a typical system were showen to inquiry the capacity of the suggested scheme for islanding discovery. The consequences exhibit that the suggested algorithm firmly discovers and classifies the islanding and non-islanding conditions within 467 ms. The precision quantity of discovering the islanding and non-islanding conditions is 100% for clean and noisy (30 and 20 dB, respectively) test data under the mismatches ΔP=-5% and ΔQ=- 0.5 % for the typical system.

Islanding is a major problem that is related to a grid connected distribution generator. During islanding the distribution generator gets disconnected from the grid while still being energized. It should be detected as fast as possible in order to prevent the distribution generator from getting damaged or to ensure the safety of the personnel working. The islanding detection makes use of system parameter to detect islanding at the time of occurrence. In this paper a passive islanding detection method based on ripple content of voltage at the point of common coupling is being used to differentiate between islanding and non-islanding conditions. A threshold level is set upon the ripple content of voltage at the point of common coupling and islanding is detected whenever the threshold level is exceeded for some certain period of time. This method effectively distinguishes the islanding condition from other non-islanding condition and detects the islanding condition efficiently. The working of this method is being ensured by simulating a variety of islanding and non-islanding condition. No non-detection zone was shown by this method and islanding conditions were detected within 150ms of the allowed delay time. Theoretical analysis and simulation results are presented to validate the proposed method.

Islanding is a phenomenon that may prove dangerous for the utility workers who are unaware that the connection is powered. Thus control over the network is essential and its detection may prevent further fatality and automatic reconnection of the circuit when not required. The research work aims at developing a technique in islanding detection in a microgrid environment. In this paper, the rate of change of negative sequence voltage (ROCONSV) is considered for identify islanding and non-islanding condition. The negative sequence voltage at Vpcc is calculated at the point of common coupling (PCC) and compared with a set threshold to verify the technique. The proposed technique is tested for islanding and non-islanding event in MATLAB/Simulink environment. The proposed technique is verified in various system conditions such as 0% and 50% active power mismatch, DG disconnection, different fault cases, and capacitor switching. In all cases, the technique works properly. MATLAB R2019a software is used for the proposed islanding detection method (IDM).KeywordsIslandingPassive methodAlert signalDistributed generationCircuit breaker

The ever increasing demand for electricity leads to the advancement of distributed generation (DG). Almost all DG sources are renewable nature. One of the major complications with the high penetration of DG sources is islanding. The islanding may damage the clients and their equipment. As per the IEEE 1547 DG interconnection standards, the islanding will be identified in two seconds and the DG must be turned off. In this paper, an advanced islanding detection process stands on a deep learning technique with continuous wavelet transforms and convolution neural networks implemented. This approach transforms the time series information into scalogram images, and later, the images are used to train and test the islanding and non-islanding events. The outcomes are correlated with the artificial neural networks and fuzzy logic methods. The comparison shows that the proposed deep learning approach efficiently detects the islanding and non-islanding events.

Summary In a microgrid, the islanding condition occurs when the utility circuit breaker (CB) is opened. In this case, the microgrid is disconnected from the main grid. Therefore, the islanding condition can be detected by monitoring the variables that caused the utility CB to open. In this paper, the utility CB current is measured at the grid side, and the islanding condition is detected based on a feature extracted from the measured signal before the utility CB opening. Discrete wavelet transform is used to extract the features of the measured current, and then, the artificial neural network is trained in order to detect the islanding conditions based on the extracted features. Since the disturbances of distributed generation (DG) units or fault occurrence have no effect on the measured current, the proposed method is an effective one for islanding detection in distribution networks with high penetration of DGs. Copyright © 2014 John Wiley & Sons, Ltd.

This study present a new islanding detection technique based on relevance vector machine (RVM) containing various types of distributed generations (DGs). The proposed scheme is based on utilising negative sequence component of current (I 2), acquired at the terminal of the target DG. Various islanding and non-islanding events with variable real and reactive power, change in network topology and diverse X/R ratio have been generated by modelling IEEE 34 bus system in real time digital simulator (RTDS®/RSCAD) environment. The samples of I 2 for one cycle duration from the inception of islanding/non-islanding events are given as input to the proposed RVM classifier. The proposed classifier is able to discriminate between islanding and non-islanding events with an accuracy of 98.62% by utilising only 40% of the total dataset in training. Moreover, it is capable to classify islanding condition rapidly and correctly even with perfect power balance situation. Furthermore, it remains immune to change in network configuration/entirely different network and X/R ratio of various types of DGs. Comparative analysis of the proposed scheme in terms of accuracy, testing time and number of relevance vectors (RVs) with the existing schemes shows its superiority in discriminating islanding situation with non-islanding events.

In this paper, a new hybrid islanding detection method is presented for protecting the distributed system with unintentional islanding, with low-frequency current harmonic injection and over/under frequency relay for inverter-based multiple distributed generation (DG) systems. Low-frequency current harmonic component is injected into the system through the q-controller of the grid-side controller. The injected low-frequency current component causes the system frequency to deviate during islanding. This deviation separates islanding and non-islanding events. This paper evaluates the performance of this hybrid ROCOF relay when load and generation are matched and also at different active power imbalances. This paper will also clearly differentiate between islanding and non-islanding conditions for mixed types of DG’s connected to the grid. The test system results are carried out on the MATLAB/Simulink setting that shows the strength of this technique.

Islanding detection needs are becoming a pivotal constituent of the power system, since the penetration of distributed generators in the utility power system is continually increasing. Accurate threshold setting is an integral part of the island detection scheme since an inappropriate threshold might cause a hazardous situation. This study looked at the islanding conditions as well as two transient faults, such as a single line to ground fault and a three-phase balance fault, to assess the event distinguishing ability of the proposed method. Therefore, the goal of this research was to determine the threshold of the island if the distributed generator (DG) capacity is greater than the connected feeder load, which is the over-frequency island condition, and if the DG capacity is less than the connected feeder load, which is the under-frequency island condition. The significance of this research work is to propose a new island detection threshold setting method using the slip angle and acceleration angle that comes from phasor measurement unit (PMU) voltage angle data. The proposed threshold setting method was simulated in the PowerWorld simulator on a modified IEEE 30 bus system equipped with DG. There are three different interconnection scenarios in the test system and the performance of the proposed method shows that getting the island threshold for all the scenarios requires a single time step or 20 mile seconds after incepting an island into the network. In addition, it can distinguish between the real islanding threshold and the transient faults threshold.

The present study attempts to model the spatial variability of three groundwater qualitative parameters in Guilan Province, northern Iran, using artificial neural networks (ANNs) and support vector machines (SVMs). Data collected from 140 observation wells for the years 2002–2014 were used. Five variables, X and Y coordinates of the observation well, distance of the observation well from the shoreline, areal average 6-month rainfall depth, and groundwater level at the day of water quality sampling, were considered as primary input variables. In addition, nine qualitative variables were also considered as auxiliary input variables. Electrical conductivity (EC), sodium concentration (Na+), and sulfate concentration (SO4 2−) of the groundwater in the region were estimated using ANNs and SVMs with different input combinations. The results showed that both ANNs and SVMs work well when the only primary input variable is the well location. The ANN yielded an RMSE of 1.03 mEq/l for SO4 2−, 1.05 mEq/l for Na+, and 203.17 μS/cm for EC, using the X and Y coordinates of the observation wells in the study area. In the case of SVM, these values were, respectively, 0.87, 0.87, and 176.68. Considering the auxiliary input variables (pH, EC, and the concentrations of Na+, K+, Ca2+, Mg2+, Cl−, SO4 2−, and HCO3 −) resulted in a significant decrease in the RMSE of both ANNs (0.22, 0.30, and 33.04) and SVMs (0.26, 0.34, and 36.23). Comparing these RMSE values with those of cokriging interpolation technique (0.59, 0.98, and 177.59) indicated that ANNs and SVMs produced more accurate estimates of the three qualitative parameters. The relative importance of auxiliary input variables was also determined using Gamma test. The output uncertainty of ANNs and SVMs were determined using p-factor and d-factor. The results showed that SVMs have less uncertainty than ANNs.

The fault detection process is very difficult in transmission lines with a fixed series capacitor because of the nonlinear behavior of protection device and series-parallel resonance. This paper proposes a new method based on S-transform (ST) and support vector machines (SVMs) for fault classification and identification of a faulty section in a transmission line with a fixed series capacitor placed at the middle of the line. In the proposed method, the fault detection process is carried out by using distinctive features of 3-line signals (line voltages and currents) and zero sequence current. The relevant features of these signals are obtained by using the ST. The obtained features are then used as input to multiple SVM classifiers and their outputs are combined for classifying the fault type and identifying the faulty section. Training and testing samples for the proposed method have been generated with different types of short-circuit faults and different combinations of system parameters in the MATLAB environment. The performance of the proposed method is investigated according to the accuracy of fault classification and faulty section identification. To evaluate the validity of this study, the proposed method is also compared to both ST--neural network and previous studies. The proposed method not only provides a good classification performance for all types of faults, but also detects the faulty section at a high accuracy.

The development of Phasor Measurement Units (PMUs) in power system improves the possibilities of analyzing and monitoring the system dynamics. PMU data with its high sampling rate contains crucial information about the state of power system. Hence effective PMU data analysis can provide solution for most of the power system problems. This work proposes a data analysis method on PMU data. This combines the simplicity of passive islanding detection techniques and the robustness of artificial intelligence methods. PMU measurements are analyzed and classified to detect power system islanding. Artificial Neural Networks (ANN) and Support Vector Machines (SVM) algorithms are used to classify islanding and non-islanding events. Both the models classified the events correctly with good accuracy. The test results also show a better performance for SVM classifier compared to ANN classifier.

The paper deals with an optimization problem in an electricity distribution network with different types of distributed generation and a battery storage system in terms of a smart grid concept. The optimization problem considers two objectives, namely, the annual energy losses and the exchange of energy with the higher-level power grid. The decision variables of the problem are the allocation of the different distributed generation units and the battery storage system, the annual power profiles of the controllable distributed generation and the battery storage system, and the power factor profiles of the controllable and noncontrollable distributed generation. All decision variables are solved simultaneously in a single optimization problem. The variable load shapes of the grid consumers and the profiles of the photovoltaic and wind power systems are considered in the study. All data are observed at the annual level with hourly resolution. The problem solving method uses computational intelligence techniques, namely, metaheuristic optimization methods and artificial neural networks. The study proposes a framework for optimizing the decision variables in the planning phase of distributed generation and battery storage, and for controlling the variable power and power factor profiles based on an artificial neural network in the implementation phase. The optimization problem is solved with a power system simulation program and a metaheuristic optimizer in cosimulation synergy. The three cases of distributed generation and battery storage are considered simultaneously. The proposed method is applied to the test grid operator IEEE with 37 buses, and reductions in annual energy losses and energy exchange are obtained in the ranges 34–86% and 41–99%, respectively.

Distributed Generation (DG) sources predominantly renewable energy-based like solar photovoltaic and wind energy sources are considerably penetrated in power system networks. This high penetration of DG sources may lead to the formation of unintentional islands which is hazardous to both the equipment and personnel, if sustained. In this paper, 46 passive parameters are considered to find out which candidate(s) are promising for detecting sustained unintentional islands and avoid false DG trips. These 46 parameters are derived from point of common coupling (PCC) voltage, DG output current, frequency, power factor angle, active and reactive powers. All these parameters are tested on standard IEEE 13 and 34 bus distribution networks integrated with inverter-interfaced DGs at different locations for different penetration levels in MATLAB/SIMULINK environment. The parameters are tested for different islanding events (comprising of various local loads such as conventional parallel RLC resonant load, R, RC, and RL loads) and for different non-islanding events. The parameters are then ranked accordingly by a suitable averaging approach based performance ranking technique. From this analysis, the best candidates are obtained for detecting inverter DG islands at single-point (or single PCC) location and multi-point (or multiple PCCs) locations. Multiple best passive candidates are obtained for different island scenarios from which a set of promising islanding detection indicators are proposed.

With increase of distributed generators (DG) in power system, there arise the potential for operation of power system during islanding mode as well. Along with its advantages, there come challenges with protection system due to change in short circuit current (which is the primary factor for overcurrent relay setting). In this paper, investigation of impact of Distributed Generators (DG) in protection system of the distribution network has been carried out. Simulation based study regarding protection issue during islanding condition has been carried out for distribution network in mid Himmerland (Denmark). Use of definite time overcurrent relay along with directional overcurrent relays has been proposed for the reliable protection scheme. Simulation in ‘DIgSILENT Power Factory’ shows the proper selectivity and coordination of overcurrent relays for 3 phase faults during various cases, such as islanding or shutdown of some DG.

This study aims to develop a real-time model for detecting and classifying various fault types in power distribution networks to enhance reliability and operational efficiency. A hybrid DWT–SVM approach is essential for accurately detecting and classifying faults in modern power distribution systems, especially under non-stationary and dynamic conditions. The proposed method addresses a range of fault scenarios, including single-phase-to-ground, line-to-line, double-line-to-ground, and three-phase faults, within a 33 kV distribution network. A hybrid approach combining Discrete Wavelet Transform (DWT) and Support Vector Machine (SVM) is introduced. Using the Debauchies-4 (Db4) wavelet, DWT effectively decomposes transient fault currents at the source terminal, capturing critical time-frequency domain features. Fault classification is performed using an SVM optimized with a Radial Basis Function (RBF) kernel and Particle Swarm Optimization (PSO), enabling precise mapping of data into higher-dimensional spaces for optimal separation. Validation conducted with MATLAB R2023b demonstrates a detection accuracy of 100% and a classification accuracy of 99%. Comparative analyses against models such as PSO-based Support Vector Machine (PSO-SVM), DWT-DNN and wavelet transform with artificial neural networks (WT-ANN) on the IEEE 13-bus system highlight the proposed method's superior performance. This innovative approach proves to be robust, adaptable, and highly effective across diverse fault scenarios, offering significant improvements in accuracy and reliability for fault detection and classification in power distribution networks. The method supports real-time operation by using high-speed simulation platforms MATLAB/Simulink to process signals and classify faults within milliseconds using a pre-trained SVM model.