PROTECTION OF DISTRIBUTION SYSTEM INTEGRATED WITH RENEWABLE ENERGY SOURCES

Abstract

The proliferation of distributed energy resources (DERs) integration into distribution system (DS) is the commendable strides made to change the proclivity of the power system towards microgrid. A microgrid is a self-reliant small-scale power grid formed by the local source of supply, which can operate independently or in addition to the utility grid, giving local communities access to more robust and sustainable energy sources. A microgrid may contain several renewable energy sources based distributed generation (DG), energy storage systems, and other power-generating units to produce power for end consumers. This empowerment of DG elevates power generation and act as an auxiliary source to assist system performance. Besides the aim of improving system performance and supporting the primary generation, DGs are essentially employed to avoid power disruptions and perform as resources for fast system recovery. Substantial injection of real and reactive power by DGs into the utility grid decisively increases the grid resilience, reduces loss, and enhances power quality by mitigating voltage sag during the fault. Environment-friendly generation and sustainability, enhance the security and reliability of the system. The thriving benefits acquired from the integration of DGs in the emerging microgrid are undermined due to the substantial challenges associated with the protection. Due to the penetration of DERs, the fault level in the microgrid changes significantly. Conventional protection philosophies are highly affected by the nature of the DGs (inverter-tied or rotational machine based), their penetration level, location, and operation state (grid-connected or islanded) of the DGs, which degrades microgrid resiliency. Change in operation state from grid-connected to autonomous mode and vice versa or network reconfiguration, changes the fault level dynamically. This dynamic fault behaviour necessitates the protection devices to modify their operation settings for accommodating the system parameter changes. Adaptable relay characteristics profoundly help to deal with the dynamic fault current. However, high-speed reliable communication infrastructure is an indispensable requirement for this kind of protection scheme, which embarks some inevitable issues such as high cost, communication failure issues, and requirement of expedite maintenance. In order to avoid such complexities, the modern sense of protection has propounded the signal processing techniques as the distortion in signal is an obvious phenomenon during transients. Signal processing techniques realize the fault inception by using frequency information extracted from transient components of the signal. Addressing these aforementioned issues, the works presented in this thesis propose innovative solutions for microgrid considering different microgrid topologies, their modes of operation, and types of connected DGs into the microgrid in order to deal with the dynamic fault current of the microgrid. Primarily, for setting the energy of a signal as a legitimate feature for fault identification, this presented research work has implemented Teager-Keiser Energy Operator (TKEO), which estimates signal energy for diagnosing the fault event. Since the computation of signal energy using TKEO involves only three consecutive sample data and three arithmetic operations, TKEO is a distinctly suitable energy operator for processing zero-crossing signal applications where diagnosing real-time information is required. In this work, the suggested topology has extracted the energy of the current signal, retrieved from both ends of the line using TKEO. For fault detection, the energy difference of the current signals is used. In order to validate the effectiveness of this proposed scheme, modified IEEE 14 and 33 bus test microgrid systems operating in both grid-connected and autonomous modes with radial and looped topologies are modelled in PSCAD-EMTDC simulation software for several case studies. The programming for signal processing is developed and tested in MATLAB environment. This method principally relies on the energy difference of current signals; therefore, it can subdue the difficulties associated with dynamic fault current. Moreover, it does not suffer from computational complexity; thus, inherently hastening the fault detection process. The signal processing techniques have been elevated, when continuous wavelet transform (CWT) and discrete wavelet transform (DWT) have become an integral part of the protection algorithm. Where Fourier analysis uses the sine wave of specific frequency as a basis to decompose a signal, wavelet transform uses a wave-like oscillation called wavelets to decompose the signal. By translating and dilating the mother wavelet, a signal is decomposed into detail and approximation coefficients in different levels of decomposition. More precisely to say, wavelet transform results in a finer-scale representation of a signal. Henceforth, DWT-multiresolution analysis (DWT-MRA) has become more prevalent as it reduces the redundancy of CWT, and also emphasises more projection into detail spaces surrounded by the wavelets. However, transient signals generated in the power system are non-stationary and nonlinearly deterministic; therefore, to analyse such signals choosing the basis functions designed independently of the processed signal is not an adequate option; therefore, in this work empirical wavelet transform (EWT) based algorithm has been established. EWT develops the basis directly based on the information contained in the signal. It extracts different modes of a signal using adaptive wavelets designed by proper wavelet filter bank. This scheme estimates the multiresolution coefficient of symmetrical components of fault current signal using EWT, and then the adaptive threshold setting for detection and classification of fault is obtained by TKEO method for both gridconnected and autonomous mode of operation. This scheme is adaptive in nature as it can produce the threshold from the estimated Teager Kaiser Energy contained in the given signal.The fault current magnitude reduces, when the microgrid consists of inverter-tied DGs and operates in autonomous mode. Since inverters lack a rotational mass component, they do not acquire the necessary electromagnetic inertia to carry sufficient fault current and dynamically behave differently from synchronous or induction machines. For inverter-dominated microgrids, the curtailment of converter output current yields insufficient contribution to fault, which undermines the accuracy and viability of conventional overcurrent protection schemes. To address that in this another work, statistical measurement based algorithm in conjugation with EWT is established for effectively detection of system faults. This scheme estimates the multiresolution coefficient of fault current signal using EWT, and then the adequate threshold setting for detection of fault is obtained using the Fano factor based ratio test. Since, the generated signals due to power system transients are inherently non-linear and non-stationary by nature, in order to analyse those signals, adaptive decomposition methods such as empirical mode decomposition (EMD), variational mode decomposition (VMD) in conjugation with Hilbert Transform have been accomplished in this another work. These signal decomposition methods essentially rely on the local characteristics time scale of data; therefore, these are highly efficient for the application of non-stationary and nonlinear signals. The reconfiguration of microgrid from grid-connected mode operation (GCMO) to autonomous mode operation (AMO) and vice-versa further elevates this dynamic fault current phenomenon. This becomes more severe, when high impedance fault (HIF) is encountered in the system. Though HIF is not a common event; however, the occurrence of HIF is observed more in the distribution system of voltage levels below 15 kV. In HIF, the energized broken wire of the overhead power line falls on a low conductive surface (sandy soil, trees, grass), which imposes very high resistance for the fault current, resulting in low fault current magnitude with associated arcs. These arcs create low and high frequencies and offer asymmetry in the fault current waveforms. Moreover, the fault current signature does not produce significant current variation so as to be detected as a fault, leading to produce fatal damage, and can also be detrimental to human safety due to the downed conductor on the consumer side. In order to address this, another approach based on EMD in conjugation with TKEO is presented for HIF detection. The incidence of a fault is affirmed by measuring the differential energy signal and estimation of Shannon entropy. The proposed technique is tested for grid-tied and autonomous microgrid operation modes with several HIFs and non-fault cases.