An Enhanced Differential Protection Scheme for LVDC Microgrid

In the low-voltage (LV) ac microgrids (MGs), with a relatively high R/X ratio, virtual impedance is usually adopted to improve the performance of droop control applied to distributed generators (DGs). At the same time, LV dc MG using virtual impedance as droop control is emerging without adequate power flow studies. In this paper, power flow analyses for both ac and dc MGs are formulated and implemented. The mathematical models for both types of MGs considering the concept of virtual impedance are used to be in conformity with the practical control of the DGs. As a result, calculation accuracy is improved for both ac and dc MG power flow analyses, comparing with previous methods without considering virtual impedance. Case studies are conducted to verify the proposed power flow analyses in terms of convergence and accuracy. Investigation of the impact to the system of internal control parameters adopted by DGs is also conducted by using proposed method.

As the grid-connected AC-DC converter is the key equipment that links the low voltage direct current (LVDC) microgrids and the utility grid, the safe and stable operation of the converter is of great significance to improve the reliability of the LVDC microgirds. Since the converter's reliability is affected by the converter's topological structure and modulation strategy, a civil LVDC microgrid case study is analyzed, and the power losses, thermal stresses, lifetimes and reliability of components in the AC-DC converters with different topologies and modulation strategies are evaluated in this paper. The influence of these two factors on the reliability and efficiency of the converter is discussed, and the normalized weighted total harmonic distortion (NWTHD) with different modulation strategies is analytically determined and compared. The results show that when the converter operates with unity power factor in this application scenario, among the two-level converter, flying capacitor converter, neutral point clamped (NPC) converter and T-type converter, the T-type converter has the highest reliability. The effects of the converter's topology and modulation strategy on the capacitor reliability are relatively small. In view of reliability and power quality, third harmonic injection pulse width modulation (THIPWM) or space vector pulse width modulation (SVPWM) is more suitable for the converter in this microgrid.

Low-voltage dc (LVdc) microgrids emerge as a viable alternative to ac microgrids. A large research interest is noted toward fast and selective protection of dc grids, typically focusing on hybrid or full solid-state solutions. In this article, the use of fuses as short-circuit protection in low-voltage dc microgrids is evaluated. The main advantage of fuses is that they are simple, cheap, standardized, and have low steady-state losses. A theoretical basis is formed to model dc short-circuit currents in grids with a limited short-circuit availability. The outcomes are applied to evaluate the possibilities of fuse protection in LVdc grids. It was found that fuses are an effective means of protection, although the required amount of capacitance at the output of the voltage balancing converter can be high, which impacts the total system cost. A fuse-based protection strategy is presented that highlights the need for additional capacitance to clear faults compared to the necessary capacitance for system stability. An experimental setup was built to validate the claims.

This paper addresses load current sharing and circulating current issues of parallel-connected dcdc converters in low-voltage dc microgrid. Droop control is the popular technique for load current sharing in dc microgrid. The main drawbacks of the conventional droop method are poor current sharing and drop in dc grid voltage due to the droop action. Circulating current issue will also arise due to mismatch in the converters output voltages. In this work, a figure of merit called droop index (DI) is introduced in order to improve the performance of dc microgrid, which is a function of normalized current sharing difference and losses in the output side of the converters. This proposed adaptive droop control method minimizes the circulating current and current sharing difference between the converters based on instantaneous virtual resistance <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\bm{R}_{{\bf droop}}$</tex-math> </inline-formula> . Using <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\bm{R}_{{\bf droop}}$</tex-math> </inline-formula> shifting, the proposed method also eliminates the tradeoff between current sharing difference and voltage regulation. The detailed analysis and design procedure are explained for two dcdc boost converters connected in parallel. The effectiveness of the proposed method is verified by detailed simulation and experimental studies.

A new differential protection scheme based on Synchro-squeezing transform applied to AC Micro-Grid

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

The article presents a data-mining model-based enhanced differential relaying scheme for fault detection and classification task in the microgrids. The proposed research work aims to provide a fixed setting, fast, and reliable scheme for accurately detecting and classifying faults in the microgrid. The presented scheme fetches the three-phase current signals from both ends of the distribution lines and computes the differential current phasor, including the distributed generation (DG) in the fault model. The deep neural network (DNN)-based data-mining models for fault detection and classification are built using differential current phasor datasets generated by wide variations in fault and operating conditions of the microgrid, including different types of faults with the change in fault location, fault resistance, DG penetration in grid-connected and islanding mode of operation with types of DG units. The developed DNN models are validated for their ability to detect the faults in a low-voltage CERTS microgrid and subsequently for a medium-voltage IEEE-34 bus microgrid on a MATLAB/SIMULINK platform. Furthermore, the proposed intelligent scheme is also tested on the Typhoon hardware-in-loop platform for real-time performance evaluation. The accuracy and speed in detecting and classifying faults indicate that the proposed DNN-based intelligent differential relaying scheme can be a potential candidate for providing reliable protection measures for the microgrid.

A differential protection scheme for the ring microgrids based on the power change rate of the bus is proposed. By setting the power change rate on both sides of the bus as the differential amount, the power change rate is maximized at the moment of short-circuit, which can immediately indicate whether a fault has occurred. Compared to the protection scheme that uses the current for fault identification, the proposed scheme reduces the time for the fault current to increase to the action threshold, saves the time required for fault detection. Moreover, it is proportional to the square of the current, and the higher orders increase the speed of the change rate, meanwhile, it is proportional to the short-circuit impedance which broadens the fault identification range and improves the protection sensitivity. Finally, the results of simulations prove that the proposed protection scheme has improved speed, sensitivity, and interference resistance while maintaining good selectivity for differential protection.

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.

Effectiveness of hybrid differential-adaptive overcurrent protection scheme for microgrid

MODWT-XGBoost based smart energy solution for fault detection and classification in a smart microgrid

This study provides designing and techno-economic analysis of a low voltage (48 V) DC microgrid for domestic load application. In present DC microgrid, photovoltaic (PV) is considered as main power generator to deliver the load demand while fuel cell (FC) and battery are considered as backup power sources. To establish a 48 V DC bus, suitable DC-DC converter having wide input voltage range (22–65 V) is designed to stabilize the variable output voltage of PV and FC generators. The converter efficiency was found to be ≈ 87%. A techno-economic study is also carried out to assess the levelized cost of electricity in microgrid using different energy storage systems, i.e. battery and hydrogen. Lead-acid and lithium-ion are considered as battery storages while high-pressure cylinder and metal hydride storages are considered for hydrogen storage. It is found that PV-battery is most economical configuration; however, PV-FC-battery is better choice for high reliability of overall system. PV-FC-battery configuration has lower cost of electricity as compared to PV-FC alone system. In battery storage system, lithium-ion battery shows faster charging/discharging and efficient operation as compared to lead-acid battery application. In hydrogen storage system, if high-pressure hydrogen is considered, the levelized cost of electricity is reduced significantly as compared to metal hydride storage. However, metal hydride is much better choice for hydrogen storage as compared to the high pressure considering the safety concern.

This article proposed a Gated-Recurrent Learning Deep Neural Network (GRL-DNN) to predict the gate driving sequence of SiC-MOSFET in a 380 V/3A Low Voltage DC microgrid (LVDC). It is devised that the proposed intelligent fuse Active Gate Driving (AGD) requires a current limitation of 3 A to avoid fault and excess power loss. During over-current conditions in microgrid the proposed data-driven Intelligent Fuse (iFuse) acts as a circuit breaker with current limitation. The proposed GRL-DNN switching model training dataset is obtained for various operating condition in MATLAB. Training features like gate triggering, temperature, over-current magnitude, and tripping time are obtained by the proposed circuitry sequence predictor model in LTspice. By obtaining large solution space, without any delay automatic AGD can cutoff over-current in LVDC microgrid system. By comparing with conventional gate driving solid state fuse by Angelov model, the proposed iFuse integrated merits like fast reaction, accurate tripping, and high reliability to any system for improving fault protection. Online data analysis was carried to validate the outcome of the proposed iFuse AGD in hardware-in-loop simulation. The conducted real-time scaled-down experiment under over-current conditions on 380 V/3A LVDC system shows the proposed GRL-DNN based iFuse as a valuable active gate driving component for protection.

In a microgrid, the overcurrent relays see significant variation in the short-circuit currents during transition between grid-connected and islanded modes of operation. Further, with the disconnection of one or more distributed generators (DGs), the changes in the short-circuit currents are observed by the relays. As a consequence, the existing protection scheme may either delay the relay operation or fail to operate. This paper proposes an optimal and single protection scheme suitable for all operating modes of microgrid along with every type of phase faults in the system, i.e., three-phase symmetrical ( $3\text{-}\phi$ ), line-to-line, and double line-to-ground faults. The optimal setting of the directional overcurrent relays is obtained using a metaheuristic-based optimization algorithm. Moreover, the proposed scheme is validated using GE Multilin, model-750/760, overcurrent relay in ETAP (industrial simulation software) on the modified IEEE 9-bus 4-DG microgrid.

Fault detection (FD) algorithms based on supervisory control and data acquisition (SCADA) data have been widely used in the operation and maintenance (O&M) of wind turbines (WTs). However, the performance of FD models will degrade due to the time-varying operating conditions (TVOC). Incremental learning (IL) methods can be employed to update the models online to adapt to TVOC, but the error accumulation caused by false negatives will lead to a continuous decrease in the FD rate (FDR). A novel IL method considering false negatives is proposed to improve the performance of WT FD models. First, a data buffer is built to cache some new normal data used in model updating. Second, a processing strategy for false negatives is proposed to block some high-risk data from being added to the buffer, thereby weakening the error accumulation. Third, when the buffer is full, update the FD model using the data in it. A real-world SCADA dataset with gearbox faults and four different FD algorithms with various model updating strategies are used in the experiments. The results demonstrate that the proposed method can lower the false alarm rate (FAR) of all four FD algorithms. After processing false negatives, for example, for the FD model based on Gaussian kernel regression, its FDR increased by about 11%, and for the FD model based on multivariate state estimation technique, its FDR increased by about 14%. The results of hyperparameter experiments and early stopping experiments show that the proposed method has good potential for practical applications.