Optimizing Bidding Curves for Renewable Energy in Two-Settlement Electricity Markets Journal: European Journal of Operational Research

Dongwei Zhao, Stefanos Delikaraoglou, Vladimir Dvorkin, Alberto J. Lamadrid L, Audun Botterud · 2025

Read Optimizing Bidding Curves for Renewable Energy in Two-Settlement Electricity Markets Journal: European Journal of Operational Research by Dongwei Zhao, Stefanos Delikaraoglou, Vladimir Dvorkin, Alberto J. Lamadrid L, Audun Botterud (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-18732.txt.

Introduction

Zhao and co-authors present a 2025 study on how wind and solar farms should bid into power markets that settle twice, once a day ahead and once in real time. Poor coordination between the two stages raises costs. The authors propose benchmark bidding curves for renewable power. A two-level model tunes the curves. It foresees how the market clears under each weather outcome and aims for the lowest total expected cost.

They prove that a simple one-step curve with a zero price can reach the best total cost when renewable power costs nothing to produce. Tests on a model of the New York grid with 1,576 nodes beat the standard practice of bidding the plain forecast. The method came close to the ideal outcome that a fully uncertain market design would give.

Catalog record

CREATORDongwei Zhao, Stefanos Delikaraoglou, Vladimir Dvorkin, Alberto J. Lamadrid L, Audun Botterud
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE42 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-18732

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Optimizing Bidding Curves for Renewable Energy in Two-Settlement Electricity Markets Journal: European Journal of Operational Research Dongwei Zhao, Stefanos Delikaraoglou, Vladimir Dvorkin, Alberto J. Lamadrid L, Audun Botterud Abstract Coordination of day-ahead and real-time electricity markets is imperative for cost-effective electricity supply and also to provide efficient incentives for the energy transition. Although stochastic market designs feature the least-cost coordination, they are incompatible with current deterministic markets. This paper proposes a new approach for compatible coordination in two-settlement markets based on benchmark bidding curves for variable renewable energy. These curves are optimized based on a bilevel optimization problem, anticipating per-scenario responses of deterministic market-clearing problems and ultimately minimizing the expected cost across day-ahead and real-time markets. Although the general bilevel model is challenging to solve, we theoretically prove that a single-segment bidding curve with a zero bidding price is sufficient to achieve system optimality if the marginal cost of variable renewable energy is zero, thus addressing the computational challenge. In practice, variable renewable energy producers can be allowed to bid multi-segment curves with non-zero prices. We test the bilevel framework for both single- and multiple-segment bidding curves under the assumption of fixed bidding prices. We leverage duality theory and McCormick envelopes to derive the linear programming approximation of the bilevel problem, which scales to practical systems such as a 1576-bus NYISO system. We benchmark the proposed coordination and find absolute dominance over the baseline solution, which assumes that renewables agnostically bid their expected forecasts. We also demonstrate that our proposed scheme provides a good approximation of the least-cost, yet unattainable in practice, stochastic market outcome. \newpageafter author Nomenclature Sets and Indexes \Lambda : Set of transmission lines, indexed by (n,m) , where n/m is sending/receiving end of a transmission line \Omega : Set of VRE generation scenarios, indexed by \omega \mathcal{I} : Set of conventional units, indexed by i or j \mathcal{I}^{\text{FS}} : Subset of \mathcal{I} with fast-startup generation units \mathcal{I}^{\text{SL}} : Subset of \mathcal{I} with slow-startup generation units \mathcal{K} : Set of VRE generation units, indexed by k \mathcal{N} : Set of grid buses, indexed by n \mathcal{T} : Set of market operation time slots, indexed by t \mathcal{S} : Set of VRE bidding curve segments, indexed by s (\cdot)_{n} : mapping of (\cdot) into the set of buses Parameters C_{i} : Variable generation cost of conventional units C_{i}^{0} : No-load cost of conventional units C_{i}^{\text{SU}} : Start-up cost of conventional units C^{\text{sh}} : Cost of load shedding (value of lost load) C_{i}^{\text{U/D}} : Real-time up-/downward re-dispatch costs \overline{F}_{n,m} : Transmission line capacities L_{n,t} : Day-ahead load forecast L_{n,t,\omega} : Real-time demand realization \overline{P}_{i} : Maximum generation capacity of conventional units \underline{P}_{i} : Minimum generation capacity of conventional units {R}_{i}^{\text{U/D}} : Upward/downward ramping capacity of conventional units \widetilde{W}_{k,t,\omega} : Real-time VRE power realization \overline{W}_{k} : VRE power capacity x_{n,m} : Transmission line reactance Decision Variables \delta_{n,t}^{\text{DA}} : Day-ahead voltage angle c_{i,t}^{\text{DA}} : Day-ahead start-up cost of conventional units p_{i,t}^{\text{C}} : Day-ahead generation of conventional units p_{k,t,s}^{\text{W}} : Day-ahead generation of VRE units u_{i,t}^{\text{DA}} : Day-ahead commitment status of conventional units C_{k,t,s}^{\text{W}} : Day-ahead bidding cost of VRE units over bidding curve W_{k,t,s} : Day-ahead bidding quantity of VRE units over bidding curve \delta_{n,t,\omega}^{\text{RT}} : Real-time voltage angle c_{i,t,\omega}^{\text{RT}} : Real-time start-up cost of conventional units l_{n,t,\omega}^{\text{sh}} : Real-time shedding of electrical loads p_{k,t,\omega}^{\text{W,cr}} : Real-time VRE power curtailment r_{i,t,\omega}^{\text{U/D}} : Real-time up-/downward re-dispatch of conventional units u_{i,t,\omega}^{\text{RT}} : Real-time commitment status of conventional units 1 Introduction 1.1 Background and Motivation The transition to zero-carbon power systems is driving the large-scale integration of variable renewable energy sources (VRE) in restructured electricity markets. In the United States, for example, the New York Independent System Operator (NYISO) and the California Independent System Operator (CAISO) are implementing measures to align with state policies mandating that 100% of electricity come from zero-carbon sources by 2040 ( NYISO, 2022 ) and 2045 ( CAISO, 2022 ) , respectively. Despite these efforts, current electricity market designs remain rooted in the technical and economic frameworks developed for dispatchable fossil-fueled generators. Short-term electricity markets typically operate in two settlements or stages ( Kirschen and Strbac, 2018 ) : a day-ahead market (DAM), cleared before actual operations, and a real-time market (RTM), which addresses imbalances closer to delivery. Dispatchable fossil-fueled generators are generally capable of meeting DAM commitments with minimal RTM adjustments. However, the increasing penetration of VRE introduces greater risks to electricity supply due to its inherent variability and uncertainty, posing significant challenges for efficiently managing RTM imbalances ( Zhou et al., 2022 ) . Since the DAM scheduling will determine RTM redispatch, efficiently scheduling VRE in the DAM is of vital importance to achieving market efficiency. Notably, VRE basically has zero marginal costs with uncertain outputs. If VRE producers bid with the true marginal cost, their bids will be dispatched in priority in the DAM based on the merit order. However, although VRE has zero marginal cost in the DAM, their uncertain output may lead to significant RTM re-dispatch costs. To reflect the RTM re-dispatch cost, some electricity markets allow VRE producers to bid prices higher than zero. For example, VRE producers can submit multi-segment bidding curves to indicate price and quantity preferences in NYISO ( NYISO, 2023 ) . These bidding curves are usually private decisions of VRE producers. However, if these curves are not properly constructed to reflect the true cost in DAM and RTM, they may jeopardize market efficiency. The above considerations motivate this work to study the key question: What are optimal DAM bidding curves for VRE producers in two-settlement electricity markets to achieve the expected least system cost? 1.2 Main Contributions To answer the above question, we formulate and analyze a bilevel optimization framework that jointly optimizes bidding prices and quantities of multi-segment bidding curves. We explore the solution method to solve the problem for large-scale systems. Specifically, we make the following four contributions: 1. Bilevel optimization framework for bidding curves: To the best of our knowledge, this is the first work to optimize the bidding curves for VRE in the two-settlement electricity market, which aims to achieve the minimum system cost. We present a bilevel optimization framework to optimize DAM bidding curves, which anticipates the real-time redispatch cost induced due to VRE uncertainty and minimizes the expected system cost of day-ahead and real-time dispatch. 2. Optimality guaranteed by single-segment bidding curves: The bilevel problem that jointly optimizes prices and quantities of bidding curves is challenging to solve. However, we prove theoretically that from the system perspective, a single-segment bidding curve with a zero bidding price and optimal quantity is sufficient to achieve optimality. 3. Simulations on large-scale systems: VRE producers in practice can be allowed to bid multi-segment curves with non-zero prices. Given bidding prices, we test the bilevel framework for single-segment and multi-segment bidding curves, respectively. We adopt our previously proposed method of strong duality and McCormick envelopes ( Zhao et al., 2024 ) to relax the bilevel optimization problem into a linear program (LP), which can solve at scale, e.g., a 1576-bus NYISO system. 4. Practical insights: Although VRE producers can bid with different prices under single- or multi-segment bidding curves in the DAM, we demonstrate that the bilevel

Read the complete text