Impact of Braess's Paradox and Simultaneous Imposition of Non-Coincidental Transmission Outages on FTR Auctions
MetadataShow full item record
This thesis identifies and resolves an issue caused by Braess's paradox in Financial Transmission Right (FTR) auctions. Braess's paradox in power systems is the situation where adding a new transmission line can reduce the transmission system capacity, and vice-versa. FTRs are auctioned by Regional Transmission Organizations (RTOs) to market parties who wish to hedge uncertain transmission costs. The issue can cause the RTO to over-allocate FTRs and become revenue inadequate which leaves the RTO the dilemma of how to recover the deficit. An auction process called the simultaneous feasibility test (SFT) limits the FTR awarded to ensure that sufficient congestion rents are collected by the RTO to pay the FTR holders. The problem stems from an SFT approximation coined in this thesis the Simultaneous Imposition of Non-coincidental Transmission Outages (SINTO) that models planned transmission outages concurrently rather than as scheduled. When Braess's paradox applies to FTR auctions, the SFT approximation defies the intuitive assumption that removing transmission lines will reduce transmission system capacity. Thus, two methods are proposed to mitigate the effects of Braess's paradox in FTR auctions. The first is the Chronological Imposition of Planned Transmission Outages (CHIMPO), which ideally models the transmission outages as scheduled but also considerably increases the auction's computational cost. The second method, called the Normally-Operated – SINTO (NO-SINTO), is a robust and computationally inexpensive approximation that adds a single set of transmission constraints to the SINTO model. The five contributions of this thesis are described through simple examples and case study simulation using actual historical FTR auction data. The first establishes, using the SINTO SFT approximation, that the existence of Braess's paradox can lead to revenue inadequacy in FTR auctions. The second demonstrates that modeling SINTO in FTR auctions may aggravate the impact of the paradox. The third offers two alternative FTR auction models (CHIMPO, NO-SINTO) to reduce the risk of revenue inadequacy from Braess's paradox. The fourth demonstrates that the ideal CHIMPO allocation of FTRs is better approximated by the NO-SINTO model than the SINTO model. The fifth indicates that RTOs may practically implement the NO-SINTO approximation on a realistically sized power networks.