An intelligent protection scheme for DC networks using a machine learning-based multi-agent platform

Distance protection has been a challenging task due to the complexity of the power system. An effective protection scheme must perform fault detection, classification and location that help in quick restoration of the power supply. Although significant hybrid methods are available for fault detection, classification, and location, separately; there is always a need for an improved integrated approach performing fault detection, classification, and location. This chapter presents an integrated approach using autoregressive modeling and extreme learning machine (ELM). The performance of the scheme is investigated on 11 kV, 6 km medium voltage underground cable. The test system is simulated using PSCAD/EMTDC and the integrated protection scheme is implemented in MATLAB. The scheme uses half-cycle duration current samples. The data preprocessing is done using autoregressive (AR) modeling. The scheme detects the fault using the power of the critical pole. ELM binary classifier uses the AR coefficients to classify the type of fault. ELM in regression mode locates the fault distance. A total of 1260 data set is used to train the ELM fault classifier and ELM fault locator. The integrated protection scheme is tested with 10,800 test cases covering wide variation in operating conditions. The classification accuracy of the ELM fault classifier is 99.85% and the fault location error is less than 0.12% for the ELM fault locator, which is better than the other hybrid methods. Hence this integrated approach can be considered as quite suitable for protection of medium voltage underground cables.

ABSTRACTFaults in power transmission systems pose significant challenges due to the complexity and length of transmission lines. Effective fault detection, classification and location are essential for preventing further damage to the power grid. While travelling wave‐based algorithms are commonly used for fault location, they often focus on identifying the fault's location without classifying the fault type. Accurate classification is crucial for enabling efficient and timely responses from protection systems. This paper introduces an integrated model for fault detection, classification and location using voltage signals from a single terminal of a series‐compensated transmission line with a static synchronous series compensator (SSSC). The Gabor Transform (GT) is utilised for feature extraction, enabling both fault detection and classification. Travelling wave theory is then applied to identify the faulty segment and estimate the fault location. Additionally, a novel technique adaptively calculates the threshold value during the protection algorithm's execution. The proposed method is validated through a comprehensive analysis of various fault scenarios and sensitivity analysis. Numerical simulations in MATLAB/Simulink show that the model achieves 100% accuracy for fault detection, classification, and faulty segment identification, with 99.7925% accuracy for fault location estimation, demonstrating its effectiveness in fault management.

This article proposes a localized protection scheme for dc microgrids with radial configuration under the impact of constant power loads (CPLs) to determine the location of faults accurately. The proposed fault location scheme is primarily designed for fault location of CPLs in dc microgrids. First, a local protection relay for CPL is designed based on the transient behavior of the current and voltage in the main distribution line. Then, the estimation of the fault resistance is formulated based on the power sharing in the system to improve the accuracy of the protection system. To realize a robust protection scheme considering the variation of fault resistance, a fault resistance estimation procedure is employed to design a system that locates both low- and high-impedance faults. Finally, the effectiveness of the proposed strategy is evaluated based on offline digital time-domain simulations in Digsilent PowerFactory software environment and experimentally verified by implementing on a laboratory scale hardware setup. The obtained simulation and experimental test results, and comparison with other methods prove that the proposed scheme is immune against these disturbances and can efficiently and reliably estimate the location and resistance of faults with high accuracy and acceptable error margin.

A rapid diagnosis technology of short circuit fault in DC microgrid

In this study, discrete wavelet transform (DWT) and deep neural network (DNN) are utilised for fault location in a series‐compensated three‐terminal transmission line. The series compensation causes challenges in fault location schemes of the three‐terminal transmission lines. The presented fault location method has been extensively tested using the SIMULINK model of a three‐terminal transmission line. Features extracted from synchronous measurements of fault currents at the three terminals using DWT are fed to the DNN. Faulted section determination and fault distance calculation are carried out using a single intelligent network simultaneously. Faulted section is determined with 100% accuracy, and the efficiency of algorithm is validated for symmetrical and unsymmetrical faults, and different values of fault resistance, inception angle, and location. The accuracy of the algorithm is acceptable for large fault resistances (above 100 Ω) and fault inception angles near zero. Total mean error for test data is 0.0458% which is much improved with respect to other similar works.

A novel protection scheme for low voltage DC microgrid using inductance estimation

This paper proposes a pattern recognition based differential spectral energy protection scheme for ac microgrids using a Fourier kernel based fast sparse time-frequency representation (SST or simply the sparse S-Transform). The average and differential current components are passed through a change detection filter, which senses the instant of fault inception and registers a change detection point (CDP). Subsequently, if CDP is registered for one or more phases, then half cycle data samples of the average and differential currents on either side of the CDP are passed through the proposed SST technique, which generates their respective spectral energies and a simple comparison between them detects the occurrence and type of the fault. The SST technique is also used to provide voltage and current phasors and the frequency during faults which is further utilized to estimate the fault location. The proposed technique as compared to conventional differential current protection scheme is quicker in fault detection and classification, which is least effected from bias setting, has a faster relay trip response (less than one cycle from fault incipient) and a better accuracy in fault location. The significance and accuracy of the proposed scheme have been verified extensively for faults in a standard microgrid system, subjected to a large number of operating conditions and the outputs vindicate it to be a potential candidate for real time applications.

Fault location and distance protection in transmission lines are essential smart grid technologies ensuring reliability of the power system and achieve the continuity of service. The objective of this chapter is to presents an accurate algorithm for estimating fault location in Extra High Voltage (EHV) transmission lines using Artificial Neural Networks (ANNs) for high speed protection. The development of this algorithm is based on disturbed transmission line models. The proposed fault protection (fault detection/classification and location) uses only the three phase currents signals at the one end of the line. The proposed technique uses five ANNs networks and consists of two steps, including fault detection/classification and fault location. For fault detection/classification, one ANN network is used in order to identify the fault type; the fault detection/classification procedure uses the fundamental components of pre-fault and post-fault sequence samples of three phase currents and zero sequence current. For fault location, four ANNs networks are used in order to estimate the exact fault location in transmission line. Magnitudes of pre-fault and post-fault of three phase currents are used. The ANNs are trained with data under a wide variety of fault conditions and used for the fault classification and fault location on the transmission line. The proposed fault detection/classification and location approaches are tested under different fault conditions such as different fault locations, different fault resistances and different fault inception angles via digital simulation using MATLAB software in order to verify the performances of the proposed methods. The ANN-based fault classifier and locator gives high accuracy for all tests under different fault conditions. The simulations results show that the proposed scheme based on ANNs can be used for on-line fault protection in transmission line.

In this article, a new full-scheme distance protection for a series-compensated transmission line is proposed. The new combination of hyperbolic S-transform and learning machines is applied for fault detection, classification, and location, which are the three main aspects of distance relays. The hyperbolic S-transform is used for extracting useful features from the current and voltage signals sample of the power system from one terminal. The extracted features are employed for distance protection using support vector classification and support vector regression methods. The proposed method is evaluated by the different series capacitor location in the transmission line, where a metal oxide varistor protects the series capacitor against over-voltages. Extensive simulation experiments show that the proposed protection scheme provides accurate tripping logic with good accuracy of fault location. Also, the non-tripping decision for external faults that occurred on adjacent uncompensated transmission line is embedded in the proposed scheme. The various operating conditions of the series-compensated transmission line are simulated in the PSCAD/EMTDC (Winnipeg, Manitoba, Canada) environment. The obtained results of support vector classification and support vector regression confirm that the new extracted features from the time-frequency hyperbolic S-transform-matrix are very effective and reliable.

DC microgrid provides the horizontal infrastructures to integrate distributed generation (DG) and loads. Unlike traditional AC systems, DC systems cannot survive or sustain high magnitude fault currents. It makes locating faults very difficult. The conventional protection techniques completely de-energies the DC link in the DC microgrid. A new protection scheme for multi-terminal DC microgrid against line-to-line fault and the low resistance earth fault is presented in this study. The scheme isolates the faulted section from the DC microgrid. Healthy sections are operated without any disturbance and supply continuity is maintained in a ring main DC bus system. The current sensors are mounted at DC bus segments to monitor the entering and outgoing current at different nodes. Further, the current sensors are also mounted at both ends of service mains to monitor their current difference at both ends of the service mains. The controller detects this current difference and opens circuit breakers. To meet the requirement of fast interrupting time and high short-circuit current withstanding capability, insulated-gate bipolar transistors used as circuit breakers. The fault location scheme gives the fault location in various sections (service mains) and faults resistance in the microgrid. The proposed concepts have been verified by computer simulation.

Fault detection, classification, and location estimation are paramount in transmission lines to ensure uninterrupted power supply, enhance reliability, and safety. Faster and more accurate fault localization is critical for prompt repair and efficient power restoration, thus minimizing economic losses. This paper presents a novel Fault Detection Isolation and Localization (FDIL) scheme, a machine learning (ML)-based framework, to increase the accuracy of fault location (FL) prediction. The proposed framework utilizes the phasor voltage, current, and frequency data obtained from optimally placed Phasor Measurement Units (PMUs) for fault detection, isolation, and localization. This paper also proposes a hybrid deep learning-based approach using Bidirectional Long Short-Term Memory (Bi-LSTM) networks and an attention mechanism for FL prediction. Extensive simulations using the IEEE 9-bus system, with different types and locations of faults, are performed on the OPAL-RT simulator-based testbed to generate and train the ML models. The performance of the proposed scheme is evaluated by calculating the Mean Absolute Error (MAE), Mean Squared Error (MSE), and R-squared (<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">R</i><sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>) metric. The results demonstrate that the proposed approach achieves high accuracy in FL prediction, with an MAE of 0.0228 km, an MSE of 0.001, and an <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">R</i><sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> of 0.9894. Compared to the best-performing technique reported in the literature (CNN-LSTM, MAE = 0.211km), the proposed approach achieves an 89.2% reduction in localization error. The proposed scheme and model outperform the contemporary model, as reported in various literature, making it an effective solution for real-time fault monitoring in transmission systems.

This study aims to improve the performance of transmission line directional relaying, fault classification and fault location schemes using fuzzy system. Three separate fuzzy inference system are designed for complete protection scheme for transmission line. The proposed technique is able to accurately detect the fault (both forward and reverse), locate and also identify the faulty phase(s) involved in all ten types of shunt faults that may occur in a transmission line under different fault inception angle, fault resistances and fault location. The proposed method needs current and voltage measurements available at the relay location and can perform the fault detection and classification in about a half‐cycle time. The proposed fuzzy logic based relay has less computation complexity and is better than other AI based methods such as artificial neural network, support vector machine, and decision tree (DT) etc. which require training. The percentage error in fault location is within 1 km for most of the cases. Fault location scheme has been validated using χ2 test with 5% level of significance. Proposed scheme is a setting free method and is suitable for wide range of parameters, fault detection time is less than half cycle and relay does not show any reaching mal‐operation so it is reliable, accurate and secure.

In this paper, we propose an application of a developed fault location platform to the Taiwan transmission grid using transmitted synchronized IED (Intelligent Electronic Device) data as inputs to the platform via Ethernet. The proposed fault location platform can also be applied to utilities with IEDs around the world. IED measurements are synchronized with embedded phasor measurement unit (PMU) function in IEDs or with unsynchronized measurements correction method in the platform. The functions of the fault location platform consist of COMTRADE file converter, faulted waveform display function, digital filter, two-terminal and three-terminal transmission line fault location algorithms, unsynchronized measurements correction method, and graphical display of fault location. Due to the high accuracy and fast analysis features of the platform, utilities can reduce large amount of manpower cost and power interruption cost during a fault searching period. Field event results of the platform over the Taiwan power system are provided to demonstrate the claimed features.

Power system fault type classification and location prediction is critical in assessing the reliability of the power system, and later restoring it to a stable operating point followed by a fault. State of the art methods include sequence component, impedance measurement from the origin of the fault, and traveling wave based methods for fault type detection and classification problem. It is important to identify and classify the fault as quickly as possible for restoring power system stability to normal operation. Machine Learning and Deep Learning methods allow the analysis of large data of fault voltages and currents by using fast and efficient algorithms. These methods require large amount of data, however, with the recent advances in the field of power system, data acquisition using smart meters and Phasor Measurement Units (PMU), huge amount of system-wide data can be made available to analyze the problem of fault type classification and location prediction. This paper presents a comparative study of Stochastic Gradient Descent (SGD) based Deep Neural Network (DNN) and Machine Learning (ML) applied to power system fault type and location prediction problem. DNN architecture uses 10 hidden layers, each layer having 60 units with hyperbolic tangent as activation function, and a combination of Support Vector Machine (SVM) and Principal Component Analysis (PCA) method are considered. Comparative results in terms of time taken to run the algorithms, and accuracy of the results obtained are presented for a 3 machine 9 bus system. Results indicate that SVM method is an optimal choice for fault classification with high accuracy for location prediction and low computational requirements compared with DNN.

Fault location in DC microgrids (DCMGs) is a critical challenge due to the system’s inherent complexities and the demand for high reliability in modern power systems. This study proposes an explainable artificial intelligence (XAI)-based quantum deep neural network (QDNN) framework to address fault localization challenges in DCMGs. First, voltage signals from the DCMG are collected and analyzed using high-order synchrosqueezing transform to detect traveling waves (TWs) and extract critical fault parameters such as time of arrival, magnitude, and polarity of the first and second TWs. These features are fed into the proposed QDNN model that integrates advanced learning techniques for accurate fault localization. The cumulative distance from the fault point to the bus connecting the DCMG to the power network is considered the output vector. The model uses a combination of deep learning and quantum computing techniques to extract features and improve accuracy. To ensure transparency, an XAI technique called Shapley additive explanations (SHAP) is applied, enabling system operators to identify critical fault features. The SHAP-based explainability framework plays a critical role in translating the model’s predictions into actionable insights, ensuring that the proposed solution is not only accurate but also practically implementable in real-world scenarios. The results demonstrate the QDNN framework’s superior accuracy in fault localization even in noisy environments and with high-resistance faults, independent of voltage levels and DCMG configurations, making it a robust solution for modern power systems.