Search

A foreseeable relatively high wind power penetration of global electricity demand will be reached in the near future. This research aims to improve aggregation modeling techniques for large wind farms considering the effect of local windspeed variations. A new structurally complete dynamic model of a doubly-fed induction generator (DFIG) consisting of an induction machine, a back to back converter, and a step-up transformer are developed in a compact state-space form. A structure-preserving aggregation method to select the model’s parameters so that the major collective dynamic characteristics of a group of DFIGs is closely approximated. The proposed aggregation method maintains the same steady-state results compared with the original unreduced model. It also preserves the major collective dynamic responses of the clustered DFIGs. A particular three-winding transformer model with current magnifying functions is developed to realize the DFIGs aggregation for Electromagnetic Transient (EMT) simulations. Furthermore, the impact of the wake effect and varying wind speed within a wind farm on the aggregated model is investigated. Although complex models can estimate and simulate wind interaction with high accuracy, a simpler model is needed for system level simulations. First, Jensen’s model is used to estimate the local windspeeds at each generator location from the ambient windspeed. Next, the aggregation algorithm generates a single representative windspeed for a cluster of generators for use in the DFIG dynamic model. The mechanism responsible for loss of synchronization following ac faults for a DFIG is identified in the thesis. A dq frame theoretical model for the DFIG and its phase locked loop (PLL) is developed which identifies positive and negative feedback mechanisms involved in synchronization. This is combined with a large signal plot of the synchronizing signal (proportional to measured q axis voltage Vmq) and a signal dependent on the PLL’s output phase ang

This thesis proposes a new circuit topology for a capacitive ladder-based electronic voltage regulator (CL-EVR) which is a substitute for an electromechanical tap changer in power transformers. The proposed CL-EVR employs a combination of capacitors and TRIACs to generate voltage steps that can be added or subtracted from the grid voltage to achieve voltage regulation. In comparison to traditional on-load tap changers (OLTC), the CL-EVR is much faster and also eliminates arcing while increasing the speed of the voltage regulation, thereby potentially increasing the operating lifetime of transformers and reducing failures. Other advantages over conventional OLTCs are the elimination of di/dt limiting reactors and numerous tap changer contact leads to a potentially lower cost. The thesis begins with an examination of traditional approaches to voltage regulation such as electromechanical and solid-state tap changing transformers. Improved models for the change-over switch are introduced and used to design a better change-over switch. The focus then shifts to the pièce de résistance which is an entirely new topology. A converter-based on-load tap changer (COLTC) is proposed to be used as a substitute for the traditional tap changers. The proposed COLTC improves the transient behaviour, arcing, and response time as compared to the traditional tap changers. The proposed design includes a new failure detection method that detects a failed TRIAC switch and isolates it from the circuit not only to prevent further damage to the other CL-EVR elements but also to continue the voltage regulation, albeit with a slightly reduced capacity. To ascertain its performance and stability over the operating range, the thesis conducts a comprehensive theoretical analysis encompassing the steady state and transient performance of a simplified system consisting of the CL-EVR connected to resistive or inductive loads. A series of Electromagnetic Transient (EMT) simulations are conducted in or

High Voltage Direct Current (HVdc) transmission is seeing an increasing use as an efficient method for delivering renewable energy from remote areas to the load center(s). Maintaining the integrity of the HVdc transmission line is paramount to insuring the reliability and the sustainability of the supply. An important component, to this reliability and sustainability, is the rapid detection and repair of faults on the HVdc transmission line. Modern point-to-point HVdc transmission systems utilize a neutral conductor in addition to the power carrying conductors for HVdc power transmission. The voltage of the neutral conductor is normally zero volts since the neutral conductor is grounded at one end. Online dc Line Fault Locator (dcLFL) systems, such as the dcLFL developed by Manitoba Hydro International (MHI), can reliably and rapidly detect the fault location on the main power carrying conductors of an HVdc transmission line. This is accomplished by detecting the arrival time of the traveling wave from the voltage at the station. However, the low voltage on the neutral conductor prevents the dcLFL from detecting fault locations on the neutral conductor. The objective of this research is to design and implement a suitable method for reliably and rapidly detecting and locating faults on the neutral conductor. An online Neutral Line Fault Locator (nLFL) system which uses HVdc system measurements with regression analysis to derive the transmission line impedance to determine an accurate fault location is proposed. To surmount the low-voltage problem. The proposed nLFL system was developed and tested, with a highly detailed Electromagnetic Transients (EMT) simulation model. The nLFL system was constructed in hardware and tested in real time using a real-time EMT simulator. After detailed EMT studies of various scenarios, the nLFL system was field tested on the East Alberta Transmission Line (EATL) HVdc system.

This thesis deals with overvoltages caused due to a ground fault on pole conductor in an overhead bipolar Line Commutated Converter (LCC) HVdc system with a Dedicated Metallic Return (DMR). In HVdc systems with earth return, when a fault occurs, one pole may be lost causing the system to operate in monopolar mode with the ground return. During the fault, overvoltages are induced on the healthy pole resulting in its voltage to increase beyond its rated value. It is already known that these overvoltages reach up to 1.8 pu on the healthy conductor for an HVdc system with ground return due to a ground fault. Due to large ground currents flow during monopolar operation with ground return, nearby buried metallic structures are subjected to corrosion and these currents might interfere with buried metallic structures. To mitigate the effects caused by a large ground current flowing over a long period, a third Dedicated Metallic Return (DMR) conductor connecting both neutrals of converters, is laid along the line through which the pole current flows. The presence of a third conductor which is grounded only at one terminal and floating on other terminal completely changes the fault mechanics during a ground fault. In this thesis, the overvoltages on the healthy pole and neutral conductors of a dc overhead line with a DMR caused due to a monopolar ground fault are investigated. Ground faults are applied on a pole conductor at different locations on the line and overvoltages produced along the line are studied. Theoretic explanation of overvoltages is explained based on travelling waves theory and using lattice diagrams. Results showed that overvoltages of up to 2 pu on the healthy pole and 1.2 pu on the dedicated metallic return conductor are observed near the terminal close to ungrounded neutral. It is also observed that the highest overvoltage on the pole conductor occurs due to a fault closer to the ungrounded terminal unlike the case with ground return where maximum overvo

Frequency scanning is a numerical method for extracting the frequency response of a power electronic system (PES). This technique is performed via injecting a small-amplitude wide-band signal in the time-domain simulation of a PES at steady state. This thesis presents the stability analysis of the interactions between a PES and its ac grid using frequency scanning techniques. Frequency scanning techniques in different domains are used to extract the model of PESs and the ac grids independently. These results are then compared and the advantages and disadvantages of each of these techniques are presented. Using the obtained models of the PESs and the ac grids, two different approaches i.e. the Generalized Nyquist Criterion (GNC) and the Eigenlocus Stability Analysis are applied to access the stability of the combined system. A comparison is given on the use of these methods for stability analysis. The results obtained from the frequency scanning can be used for other applications such as finding the stability margin of the multiple input multiple output (MIMO) systems using the results of the eigenlocus approach. The adverse interactions can then be stabilized by increasing the strength of the ac grid. The approximate stability screening technique based on the positive and negative sequence impedance/admittance can also be used for stability screening in grid-connected PESs. Although this technique is widely used in industry and generally gives correct indications of stability, it can give erroneous results due to the small but existent cross-coupling of frequencies. As a result, this thesis recommends using the DQ-based frequency scanning for stability screening purposes instead of using the positive/negative sequence-based approximate stability screening technique.

This thesis presents a demonstration of Real-Time High-Speed controller hardware-in-the-loop (CHIL) testing for a digital signal processor (DSP) -based controller of a shunt active power filter (APF) for medium voltage (MV) grid power quality improvements. A recently developed Real-Time and high-speed simulator (RT Box) and a DSP are used for the CHIL simulation to validate the control algorithms which are implemented in a commercial low-cost DSP. The filter reference current is extracted based on synchronous reference frame theory. Hysteresis current control is employed to generate PWM switching signals for controlling the APF to inject compensating currents into the grid. The direct current (DC) side capacitor voltage level of the APF is maintained by a PI controller. The results show the APF associated with the implemented controller can mitigate variety power quality problems, such as harmonic elimination, power factor correction, and load balancing. The whole CHIL simulation system work stable with high switching frequency.