Graduation Semester and Year
2020
Language
English
Document Type
Dissertation
Degree Name
Doctor of Philosophy in Electrical Engineering
Department
Electrical Engineering
First Advisor
Ali Davoudi
Second Advisor
Ramtin Madani
Abstract
This dissertation investigates the applications of high performance optimization techniques for emerging power systems with augmented power electronics devices. One of the main sources of complexity in the analysis of power systems is rooted in the power flow equations modeling steady-state relationship between power injections and voltages. Hence, the present work is in-part focused on addressing the complexity of power flow equations in the presence of power electronic devices. In contrast to the classic optimal power flow (OPF) solution techniques, we employ convex optimization methods to reliably find globally optimal solutions in polynomial time. The first chapter investigates the optimization of droop control set-points, i.e., voltage and power levels at each bus, and the switching status of transmission lines, in multi-terminal direct current (MTDC) grids. Additional constraints, that ensure a safe operation in response to power fluctuation while updating droop set-points, are integrated into the problem formulation. This problem is expressed as a mixed-integer nonlinear program with three sources of computational difficulty: i) Non-convex power balance and flow equations, (ii) Non-convex converter loss equations, and iii) Binary variables standing for the on/off status of transmission lines. Second-order cone programming relaxation tackles the non-convexity of converter loss, power flows and power balance equations, and branch-and-bound search determines the optimal switching status of transmission lines. CIGRE B4 DC grid benchmark is emulated in a real-time hardware-in-the-loop environment to corroborate the proposed method. This dissertation next copes with the long-standing state estimation and topology identification problems in direct current (DC) networks. This problem is challenging due to binary decisions and nonlinear relations between sensor measurements and state variables. We introduce a non-convex nuclear norm estimator whose nonconvexity is addressed by incorporating two inertia terms. In the presence of noise, penalty terms are integrated into the objective function to estimate unknown noise values. Numerical results for the modified IEEE 9-bus, 14-bus, and 30-bus systems corroborate the merits of the proposed technique. Furthermore, this technique is experimentally validated for a converter-augmented 14-bus system in a real-time hardware-in-the-loop platform. Lastly, this work introduces an enhanced modeling for generator response in the security-constrained optimal power flow (SCOPF) problem, in which every contingency scenario corresponds to the outage of an arbitrary set of generators and lines. Integrating active and reactive power contingency response into SCOPF problem is a major computational challenge. We introduce a family of surrogate models for common-practice power system contingency response decisions such as PV/PQ switching and active power redistribution. The proposed models prevent physical and operational violations by means of optimally, allocating active power imbalances among available generators and determining the most efficient PV/PQ switching decisions. The efficacy and scalability of the proposed method is numerically validated on an IEEE benchmark system.
Keywords
Convex optimization, DC network, network configuration, optimal switching, state estimation, topology identification
Disciplines
Electrical and Computer Engineering | Engineering
License
This work is licensed under a Creative Commons Attribution-NonCommercial-Share Alike 4.0 International License.
Recommended Citation
Altun, Tuncay, "HIGH-PERFORMANCE OPTIMIZATION METHODS FOR EMERGING POWER SYSTEMS" (2020). Electrical Engineering Dissertations. 396.
https://mavmatrix.uta.edu/electricaleng_dissertations/396
Comments
Degree granted by The University of Texas at Arlington