DEVELOPMENT OF NOVEL PROTECTION SCHEMES FOR DC MICROGRIDS

Abstract

DC Microgrids (MGs) have gained considerable attention in the electricity industry due to the clean and green DC-based distributed generations (DGs) and the presence of DC loads. DC MGs are a better alternative than the AC ones because of several advantages, like the absence of a synchronization process, zero reactive power absorption, improved overall efficiency, more straightforward control strategies, low cost, etc. However, DC MG protection remains one of the challenging issues as there is a lagging concept related to the control strategies, standards, protection schemes, DC MGs grounding strategies, topologies, and operating voltage. Another unavoidable issue is the high fault current, which should be cleared quickly. However, AC protection devices are unsuitable in DC due to the absence of the zerocrossing. Some of the protection issues of DC MGs are fault detection, isolation, islanding detection, and coordination. This thesis aims for the development of islanding detection schemes and protection coordination strategies for DC MGs. Islanding detection is a critical issue for both AC and DC MGs. During islanding, the utility grid connection is absent. So, DGs are the main sources at this time. But, the power can reach the consumers without any interruption during islanding due to DGs. Here, employees can be affected by islanding as they may not notice that some part of the circuit is working, which may stop the reconnection of devices automatically. In order to maintain stability, it is essential to maintain a balance between load and generation in the islanded circuit. The above-stated reasons can be avoided by rapidly detecting the islanding so that it can be disconnected from the primary circuit as quickly as possible. This process is known as anti-islanding. If the detection technique fails, it may lead to instability in the system, affecting the load side devices and employees. Islanding detection in the DC grid is a relatively new concept as compared to AC grid islanding. In the DC grid, islanding detection majorly depends upon threshold settings with respect to under/over voltage limits (which fail during zero power mismatch (ZPM) conditions). Some active islanding techniques focus on disturbance injection, but the larger disturbance in those techniques causes unstable operations in the grid. In order to overcome the above-mentioned issues, hybrid detection schemes are implemented. This thesis aims for the development of different passive islanding detection schemes for DC MGs first to address the shortcomings of existing passive techniques. Then an impedance-based active islanding detection scheme is proposed. Here, the digital LockiIn Amplifier (D-LIA) and sensors in PV are used to accomplish the approach. The proposed islanding detection techniques are modeled in MATLAB/Simulink, with experimental validation on a DC Microgrids hardware test bed with the dSPACE (DSP-1104) controller. Next, to overcome the issues of active and passive schemes a two-stage hybrid islanding detection approach for DC microgrids is proposed here to address the significant non-detection zone (NDZ) and power quality degradation of prior detection schemes. The low inertia of DC MGs and converter behavior make the MGs sensitive to disturbance and faults. Disturbance in DC MGs is mainly due to the input power variation, fluctuation in load, and change in load-sharing proportions. The temporary faults and failures, delays in communications, and dynamics degrade the performance of the system. The voltage regulation in the DC bus is achieved with the help of a proper control strategy and communication protocol. Hence, the time taken to clear the fault depends on the line parameter affecting the stability of the system. Another problem in the protection of DC MGs is the fault discrimination strategy to discriminate the faults from other non-fault disturbances. For better protection system performance, it should categorize the disturbance and faults as temporary and permanent. Therefore, the design of a DC MG protection schemes are a challenging task due to these issues. For reliable and safe operation of electrical power, a robust protection system is needed. The absence of phasor measurements reduces the fault detection alternatives in DC MGs. Large capacitors in the converters and low cable impedance significantly increase the DC fault current. The withstanding capacity of power electronics converters is only up to twice the full load current. But, the DC fault current reaches approximately ten times the full load-rated current. Therefore, a fast and selective protection coordination approach is required to avoid load detriment and system disintegration to maintain the continuity of power supply and reliability of the DC MGs. This thesis proposes different protection coordination strategies for DC Microgrids. One is a localized backup fault detection method with assistance from extreme learning machines (ELMs) for medium voltage direct current microgrids (MVDC MGs). In this case, the system is simulated in MATLAB/Simulation platform to access the performance of the scheme. The performance of the scheme is also verified in real-time using a hardware test bench and LabVIEW environment. Another is an inverse time-current curve (ITCC) based optimal coordination strategies are proposed for both radial and non-radial (zonal DC Microgrid) DC systems. Here, the proposed scheme is simulated using the MATLAB/Simulink platform, and the experimental validation is assessed using a real-time digital simulator (RTDS) and dSPACE 1104 controller.