Integrating inverter-based generators in power systems introduces several challenges to conventional protection relays. The fault characteristics of these generators depend on the inverters’ control strategy, which matters in the detection and classification of the fault. This paper presents a comprehensive machine-learning-based approach for detecting and classifying faults in transmission lines connected to inverter-based generators. A two-layer classification approach was considered: fault detection and fault type classification. The faults were comprised of different types at several line locations and variable fault impedance. The features from instantaneous three-phase current and voltages and calculated swing-center voltage (SCV) were extracted in time, frequency, and time–frequency domains. A photovoltaic (PV) and a Doubly-Fed Induction Generator (DFIG) wind farm plant were the considered renewable resources. The unbalanced data problem was investigated and mitigated using the synthetic minority class oversampling technique (SMOTE). The hyperparameters of the evaluated classifiers, namely decision trees (DT), Support Vector Machines (SVM), k-Nearest Neighbors (k-NN), and Ensemble trees, were optimized using the Bayesian optimization algorithm. The extracted features were reduced using several methods. The classification performance was evaluated in terms of the accuracy, specificity, sensitivity, and precision metrics. The results show that the data balancing improved the specificity of DT, SVM, and k-NN classifiers (DT: from 99.86% for unbalanced data to 100% for balanced data; SVM: from 99.28% for unbalanced data to 99.93% for balanced data; k-NN: from 99.64% for unbalanced data to 99.74% for balanced data). The forward feature selection combined with the Bag ensemble classifier achieved 100% accuracy, sensitivity, specificity, and precision for fault detection (binary classification), while the Adaboost ensemble classifier had the highest accuracy (99.4%), compared to the other classifiers when using the complete set of features. The classification models with the highest performance were further tested using a new dataset test case. They showed high detection and classification capabilities. The proposed approach was compared with the previous methodologies from the literature.
26 July 2010 in IEEE Transactions on Power Delivery
In Nordic countries, distribution networks are unearthed or compensated. Earth faults, particularly in compensated networks, provide fault currents that are low compared to the load currents. Identification of the faulty feeder is therefore difficult. A preliminary description of discrete wavelet transform (DWT)-Bayesian selectivity technique was introduced in reference 1 to identify the faulty feeder. It was dependent on a conditional probabilistic method applied to transient features extracted by using the DWT. However, a practical setting for this technique has not yet been presented. Furthermore, its sensitivity is limited to 1.5-k fault resistances, and is further reduced to 170 when considering current transformer and network noise. In this paper, the ratio between the absolute sums of the DWT detail level from each feeder is used as an input to the conditional probability approach, providing an enhanced selectivity decision. This input contributes to discriminating the faulty feeder during high resistance faults. The relay setting is introduced as a function of the number of feeders and their characteristic impedances, as the proposed algorithm is dependent on the discharge's initial transients. The performance is evaluated taking into account current-transformer and network noises. A digital implementation is experimentally verified by using two digital signal processing boards.
Adaptive protection schemes have been developed to address the problem of behavior-changing power systems integrated with inverter-based generation (IBG). This paper proposes a machine-learning-based fault detection and classification technique using a setting-group-based adaptation approach. Multigroup settings were designed depending on the types of power generation (synchronous generator, PV plant, and type-3 wind farm) connected to a transmission line in the 39-Bus New England System. For each system topology, an optimized pretrained ensemble tree classifier was used. The adaptation process has two phases: an offline learning phase to tune the classifiers and select the optimum subset of features, and an online phase where the circuit breaker (CB) status and the active output power of the generators are continuously monitored to identify the current system topology and to select the appropriate setting group. The proposed system achieved an average accuracy of 99.4%, a 99.5% average precision, a 99.9% average specificity, and a 99.4% average sensitivity of classification. The robustness analysis was conducted by applying several fault scenarios not considered during training, which include different transmission network configurations and different penetration levels of IBGs. The case of incorrect selection of the appropriate setting group resulting from selecting the wrong topology is also considered. It was noticed that the performance of developed classifiers deteriorates when the transmission network is reconfigured and the incorrect setting group is selected.
This article presents the development of a platform for real-time monitoring of multi-microgrids. A small-scale platform has been developed and implemented as a prototype, which takes data from various types of devices located at a distance from each other. The monitoring platform is interoperable, as it allows several protocols to coexist. While the developed prototype is tested on small-scale distributed energy resources (DERs), it is done in a way to extend the concept for monitoring several microgrids in real scales. Monitoring strategies were developed for DERs by making a customized two-way communication channel between the microgrids and the monitoring center using a long-range bridged wireless local area network (WLAN). In addition, an informative and easy-to-use software dashboard was developed. The dashboard shows real-time information and measurements from the DERs—providing the user with a holistic view of the status of the DERs. The proposed system is scalable, modular, facilitates the interoperability of various types of inverters, and communicates data over a secure communication channel. All these features along with its relatively low cost make the developed real-time monitoring platform very useful for online monitoring of smart microgrids.
1 June 2007 in 2007 IEEE Power Engineering Society General Meeting
Current traveling waves are utilized to localize faults on transmission lines using conventional current transformers. This paper investigates high frequency characterization of the current transformer. The current transformer parameters at high frequencies are experimentally determined. They are obtained using frequency response measurements where the current transformer transfer function is calculated from measured impulse signals. Therefore its performance under transient conditions can be studied using different programs such as EMTP and MATLAB. The current transformer modeling and its experimental verifications show that there are good replicas of high frequency traveling waves over a suitable range for fault locators. The test results insure the CT applicability to traveling wave fault location.
Many studies indicate the importance of distributed generation (DG) as one of the best alternatives to conventional energy sources. The growth in population is causing a significant increase in electricity demand all over the world, which creating a considerable need for further investment in the electric network infrastructure. However, integration of new renewable energy source in the distribution network (DN) could impact the stability and the reliability of the distribution network if successful integration of renewable and distributed generation into the distribution network not achieved. Successful integration of distributed energy resources requires optimized control and planning for the network configuration which relies basically on the decentralized energy source management and protection system coordination. The aim of this study is to investigate the impact of utilizing DG on the performance of DN, especially on the over-current protective devices. Also, introducing a proposed adaptive protection scheme to mitigate the influence of DG penetration on the protection coordination by adding group setting, which exists in nowadays numerical protective relays, for different operating conditions such as islanding mode, and full MVA capacity of a DG. The simplified power system simulated in NEPLAN and ETAP packages represents Benghazi medium voltage distribution network. The simulation and results show the effectiveness of the proposed approach on the reliability and stability of the distribution network.