Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Energy Storage Arbitrage in Two-settlement Markets: A Transformer-Based Approach

Saud Alghumayjan, Jiajun Han, Ningkun Zheng, Ming Yi, Bolun Xu

TL;DR

This paper tackles maximizing energy storage profits in two-settlement electricity markets by integrating day-ahead bidding with real-time arbitrage. It combines a transformer-based predictor for real-time prices to inform day-ahead bids and a non-anticipatory LSTM-dynamic programming model for real-time bidding, under a price-taker framework that decouples DAM and RTM decisions. Evaluated on NYISO data across four zones, the approach yields up to ~29% profit gains over RTM-only strategies and reduces the number of negative-profit days, indicating improved profitability and risk management. The work advances practical two-stage bidding strategies for storage and demonstrates clear welfare and profitability benefits in a real-market setting.

Abstract

This paper presents an integrated model for bidding energy storage in day-ahead and real-time markets to maximize profits. We show that in integrated two-stage bidding, the real-time bids are independent of day-ahead settlements, while the day-ahead bids should be based on predicted real-time prices. We utilize a transformer-based model for real-time price prediction, which captures complex dynamical patterns of real-time prices, and use the result for day-ahead bidding design. For real-time bidding, we utilize a long short-term memory-dynamic programming hybrid real-time bidding model. We train and test our model with historical data from New York State, and our results showed that the integrated system achieved promising results of almost a 20\% increase in profit compared to only bidding in real-time markets, and at the same time reducing the risk in terms of the number of days with negative profits.

Energy Storage Arbitrage in Two-settlement Markets: A Transformer-Based Approach

TL;DR

This paper tackles maximizing energy storage profits in two-settlement electricity markets by integrating day-ahead bidding with real-time arbitrage. It combines a transformer-based predictor for real-time prices to inform day-ahead bids and a non-anticipatory LSTM-dynamic programming model for real-time bidding, under a price-taker framework that decouples DAM and RTM decisions. Evaluated on NYISO data across four zones, the approach yields up to ~29% profit gains over RTM-only strategies and reduces the number of negative-profit days, indicating improved profitability and risk management. The work advances practical two-stage bidding strategies for storage and demonstrates clear welfare and profitability benefits in a real-market setting.

Abstract

This paper presents an integrated model for bidding energy storage in day-ahead and real-time markets to maximize profits. We show that in integrated two-stage bidding, the real-time bids are independent of day-ahead settlements, while the day-ahead bids should be based on predicted real-time prices. We utilize a transformer-based model for real-time price prediction, which captures complex dynamical patterns of real-time prices, and use the result for day-ahead bidding design. For real-time bidding, we utilize a long short-term memory-dynamic programming hybrid real-time bidding model. We train and test our model with historical data from New York State, and our results showed that the integrated system achieved promising results of almost a 20\% increase in profit compared to only bidding in real-time markets, and at the same time reducing the risk in terms of the number of days with negative profits.
Paper Structure (16 sections, 2 theorems, 8 equations, 5 figures, 5 tables)

This paper contains 16 sections, 2 theorems, 8 equations, 5 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Motivation
  3. Contribution
  4. Related Work
  5. Energy Storage Arbitrage Framework
  6. Problem Statement
  7. Decomposition Reformulation
  8. Electricity Price Forecasting
  9. Real-Time Energy Storage Arbitrage Model
  10. Data Description
  11. Experiment Setup and Results
  12. Electricity Prediction Results
  13. Combined Profit Results
  14. Sensitivity Analysis
  15. Risk Analysis
  16. ...and 1 more sections

Key Result

Proposition 1

Real-time price expectation under price-taker assumptions. The DAM clearing results are statistically independent of real-time price realizations. This allows us to convert the whole-day summation of RTM intervals ($s\in \mathcal{S}$) into summation of hours ($t\in \mathcal{T}$) and summation of int

Figures (5)

  • Figure 1: Pipeline of the proposed solution.
  • Figure 2: Our real-time price forecasting model overview.
  • Figure 3: Density distribution of the day-ahead and real-time prices for NYC Zone price during 2021.
  • Figure 4: The results of the system for all NYISO four price zones during 2021. Note that the dashed lines are participating options using perfect real-time price forecast, and the solid lines are participating options using forecasted real-time price generated by the log-Transformer model.
  • Figure 5: Left: The cumulative profit of the trained models with randomly sampled hyperparameters. Right: the error of the models and their profits.

Theorems & Definitions (4)

  • Proposition 1
  • proof
  • Proposition 2
  • proof