Files in this item

FilesDescriptionFormat

application/pdf

application/pdfDimitra_Apostolopoulou.pdf (4MB)
(no description provided)PDF

Description

Title:Enhanced automatic generation control with uncertainty
Author(s):Apostolopoulou, Dimitra
Director of Research:Sauer, Peter W.
Doctoral Committee Chair(s):Sauer, Peter W.
Doctoral Committee Member(s):Domínguez-García, Alejandro D.; Overbye, Thomas J.; Voulgaris, Petros G.
Department / Program:Electrical & Computer Eng
Discipline:Electrical & Computer Engr
Degree Granting Institution:University of Illinois at Urbana-Champaign
Degree:Ph.D.
Genre:Dissertation
Subject(s):Automatic Generation Control
Area Control Error
Renewable-based Generation
Frequency Performance Criteria
Balancing Authority Area
Actual Frequency Response Characteristic
Abstract:Maintaining reliability is a key aspect in power system operations. One process that helps in achieving this goal is automatic generation control (AGC), which is responsible for restoring the system frequency to the nominal value, and the real power interchange between balancing authority (BA) areas to the scheduled values. In this dissertation, we present the limitations of current AGC system implementations, and propose modifications in their design in order to increase their efficiency. The AGC system goal has become more challenging due to the radical transformations occurring in the structure and functionality of power systems. These transformations are enabled by the integration of new technologies, such as advanced communication and power electronics devices, and the deepening penetration of renewable resources. For example, renewable-based generation is highly variable and intermittent, and might undermine the objective of AGC systems. A framework that may be used to quantify the effects of various uncertainty sources, such as load variations, renewable-based generation, and noise in communication channels, on the system characteristics is presented in this dissertation. To this end, we develop a method to analytically propagate the uncertainty from the aforementioned sources to the system frequency and area control error (ACE), and obtain expressions that approximate their probability distribution functions. We make use of the proposed framework and derive probabilistic expressions of the frequency performance criteria, developed by the North American Electric Reliability Corporation (NERC). Such expressions may be used to determine the limiting values of uncertainty that the system may withstand. Our studies show that some advances are necessary in AGC system implementations, due to changes in power systems, such as the deregulation of the power industry and the integration of new technologies. The basic concept of AGC systems that is used by the BA areas has not changed severely over the past years. We aim in proposing AGC system modifications that are realistic and implementable in real large-scale systems. The high complexity of power systems is an obstacle when performing several processes related to reliability. In order to overcome such issues, we propose a systematic reduction of the synchronous generator model with low computational effort. In addition, we use the derived reduced model to describe a BA area dynamic behavior by including only the BA area variables. We use the developed models to design adaptive AGC systems, with self-tuning gain techniques, that decrease the unnecessary regulation and reduce the associated costs, since they take into account the actual system conditions in the determination of the control gains. Furthermore, each BA area implements its own AGC system. However, if all the BA areas were operated as one single BA area, then the regulation amounts as well as the associated costs would be less. Operating separately and locally, individual BA areas are obliged to purchase more expensive ancillary services to accommodate the variability and uncertainty from high penetration of renewable-based resources. Thus, some level of coordination between BA areas is favorable for all entities. We propose a coordination scheme between BA areas that would decrease the regulation amounts and costs. Our approach is inspired from trying to mimic the AGC system, in the scenario where all areas are assumed to be one single BA area. To this end, we use the individual ACEs of each BA area to approximate the ACE in the scenario where all BA areas are assumed to be a single BA area. Then, we allocate the approximated ACE to the individual AGC systems proportionally to their size. Next, we mimic the AGC allocation for the entire area without the need for exchanging cost information between the BA areas. To this end, we develop a distributed algorithm that provides the same solution as the centralized AGC allocation, with the total mismatch of regulation being the only information exchanged between BA areas. Moreover, the AGC dispatch in many independent system operators (ISOs) is determined through a market mechanism, as mandated by the restructuring of power systems. However, we investigate the possibility of using the economic signals from the real-time markets (RTMs) instead of having AGC markets for the AGC dispatch. To do so, we start out by giving the formulation of the economic dispatch (ED) process, which is used to clear the RTM, and use it to obtain appropriate economic signals. We also discuss that the quality of the AGC service provided is affected by the ramping characteristics of the regulating units chosen to participate in AGC. We propose a systematic method for the AGC dispatch taking into account the economic signals from the ED process as well as the quality of the AGC service provided. The proposed ideas are illustrated through several test systems. We choose small systems to provide insights into the proposed methodologies and large-scale systems to demonstrate their scalability.
Issue Date:2015-01-21
URI:http://hdl.handle.net/2142/72755
Rights Information:Copyright 2014 Dimitra Apostolopoulou
Date Available in IDEALS:2015-01-21
Date Deposited:2014-12


This item appears in the following Collection(s)

Item Statistics