DCCMAPI/CCET

Uma banca de QUALIFICAÇÃO de DOUTORADO foi cadastrada pelo programa.
DISCENTE: RAFAEL MARTINS BARROS
DATA: 29/03/2023
HORA: 09:00
LOCAL: Sala Google Meet
TÍTULO: Optimal Sequencing of Control Adjustments to Minimize Operational Losses in Transmission Systems by Means of Graph Shortest Path, Dynamic Programming, and Parallel Computing
PALAVRAS-CHAVES: dynamic programming, graph shortest path, load flow, optimal power flow, parallel computing, sequencing control adjustments
PÁGINAS: 50
GRANDE ÁREA: Ciências Exatas e da Terra
ÁREA: Ciência da Computação
RESUMO: Minimizing operational losses in electrical power transmission systems by the Optimal Reactive Dispatch (ORD) problem consists in a timely research topic that contributes to secure power system operation, the rational the use of scarce resources, the reduction of operational costs, and even the mitigation of greenhouse gas emissions by the power sector (especially in regulated electricity markets) by optimally adjusting its controllable variables, i.e., the terminal voltage magnitude settings of synchronous generators and condensers, the tap ratio settings of on-load tap-changing transformers, and the equivalent susceptances of switchable shunt capacitor banks and reactors, to attain the optimum. However, power system operators cannot inadvertedly or simultaneously realize such ORD control adjustments and have a limited time frame in which such controls must be adjusted prior to a significant change in the active power output of thermal, hydro, wind, and/or solar generation units in the short term. Moreover, most works in the literature do not account in the ORD problem formulation for the sequence in which such control adjustments must be realized, i.e., the path that effectively leads power system operation toward the optimum. Therefore, this work proposes the optimal sequencing of ORD control adjustments for n controllable variables that minimize operational losses in transmission systems by the statement of such a specific problem as a graph-shortest-path problem modeled as a large mixed-integer (combinatorial) nonlinear programming problem with n! possible solutions. The resulting graph-shortest-path problem is solved by means of a methodological approach based on parallel computing and dynamic programming; the asymptotic time complexity of the proposed methodological approach is also presented and discussed. As opposed to most works in the literature, which heuristically sequence ORD control adjustments, the methodological approach featured in this work guarantees that the obtained path toward minimal operational losses is optimal based on the ORD solution and from the graph-shortest-path problem perspective. Numerical results, average algorithm run times, and some parallel-application performance metrics considering transmission systems with up to 27 controllable variables, and a comparison of results between the proposed methodological approach and other techniques in the literature are presented to validate the effectiveness and to show the straightforward application of this proposal.
MEMBROS DA BANCA:
Presidente - 912.843.763-20 - RICARDO DE ANDRADE LIRA RABÊLO
Interno - 003.378.853-70 - JOAO CARLOS DE OLIVEIRA SOUZA
Externo à Instituição - JOSE CARLOS DE MELO VIEIRA JUNIOR - USP
Externo à Instituição - LEONARDO NEPOMUCENO - UNESP
Co-orientador externo à instituição - GUILHERME GUIMARÃES LAGE - UFSCAR