Paper: arXiv 2509.21337

Authors: Luis van Sandbergen

Abstract

The provision of renewable electricity is the foundation for a sustainable future. To achieve the goal of sustainable renewable energy, Battery Energy Storage Systems (BESS) could play a key role to counteract the intermittency of solar and wind generation power. In order to aid the system, the BESS can simply charge at low wholesale prices and discharge during high prices, which is also called energy arbitrage. However, the real-time execution of energy arbitrage is not straightforward for many companies due to the fundamentally different behavior of storages compared to conventional power plants. In this work, the optimized operation of standalone BESS in the cross-market energy arbitrage business is addressed by describing a generic framework for trading integrated BESS operation, the development of a suitable backtest engine and a specific optimization-based strategy formulation for cross-market optimized BESS operation. In addition, this strategy is tested in a case study with a sensitivity analysis to investigate the influence of forecast uncertainty. The results show that the proposed strategy allows an increment in revenues by taking advantage of the increasing market volatility. Furthermore, the sensitivity analysis shows the robustness of the proposed strategy, as only a moderate portion of revenues will be lost if real forecasts are adopted.

Complexity vs Empirical Score

  • Math Complexity: 7.5/10
  • Empirical Rigor: 6.0/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper uses advanced optimization techniques like Mixed-Integer Linear Programming (MILP) and Rolling Horizon frameworks, indicating high mathematical complexity. It also includes a dedicated backtest engine, sensitivity analysis, and case study with real-world data applications, showing substantial empirical rigor.

Research Flowchart

  flowchart TD
  A["Research Goal: Optimize<br>Standalone BESS Operation<br>for Cross-Market Arbitrage"] --> B["Input Data:<br>Historical Wholesale<br>Electricity Prices"]
  B --> C["Methodology:<br>Development of Optimization<br>Strategy & Backtest Engine"]
  C --> D{"Computational Process:<br>Forecast Price & Optimize<br>Charge/Discharge Scheduling"}
  D --> E["Case Study Execution<br>with Sensitivity Analysis"]
  E --> F["Outcome: Increased Revenue<br>via Market Volatility Exploitation"]
  E --> G["Outcome: High Robustness<br>Minimal Revenue Loss with Real Forecasts"]

Evidence Card

What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card

Proposes

The paper proposes a general trading‑integrated BESS operation framework, a backtest engine, and an optimization‑based bidding strategy for standalone battery arbitrage.

Source sentence
for optimally operating a standalone BESS for energy arbitrage. The paper’s primary contributions are: 1) the description of a general trading integrated BESS operation framework, 2) the development of a backtest engine for specific operation strategies, 3) the formulation of an optimization-based strategy for BESS bid

Assumes 6

  • The index used for intraday continuous trading is assumed tradable without liquidity constraints. market
    Source sentence
    Therefore, it was assumed that this index is tradable without liquidity constraints.
  • Bids are always accepted and can be executed over a 24‑hour tradable period. execution
    Source sentence
    Accordingly, it was also assumed that the bids were always accepted, with a tradable time period of 24 hours on IDC.
  • Transaction costs of the EPEX market are neglected. market
    Source sentence
    Transaction costs of the EPEX were also neglected.
  • The trading period from (d‑1) 16 to (d‑1) 24 o’clock is omitted, assumed to have subordinate influence. market
    Source sentence
    this approach for IDC trading neglects the trading period from (d-1) 16 to (d-1) 24 o’clock, which probably has a subordinate influence.
Show 2 more
  • Perfect forecasts are used in a benchmark scenario. model
    Source sentence
    Using perfect forecasts, the BESS generated a
  • Forecast errors for the ID1 index are simulated by applying volatility levels of 10, 20, 50 and 100 %. statistical
    Source sentence
    a volatility of 10, 20, 50 and 100 % was manually applied to the ID1 index to simulate forecast errors.

Data

Backtest sample: 2019 – 2023 · German electricity markets (EPEX Spot, DAA, IDA, IDC) · other · intraday

  • ICIS day‑ahead (DAA) electricity price forecast · ICIS market-prices commercial
    04/05/2024 to 04/06/2024
    Source sentence
    The Mean Absolute Error (MAE) for the ICIS DAA forecast in this period was 10.65 and the Root Mean Square Error (RMSE) 15.48.
  • ICIS intraday auction (IDA) electricity price forecast · ICIS market-prices commercial
    04/05/2024 to 04/06/2024
    Source sentence
    A MAE value of 14.94 and a RMSE of 20.91 were determined for the ICIS IDA forecast.
  • EPEX spot market price series · EPEX market-prices public
    Source sentence
    The input data mainly includes price forecasts and asset data, such as the current battery state of charge (SOC).
  • ID1 index (used for intraday continuous market) market-prices unknown
    Source sentence
    a volatility of 10, 20, 50 and 100 % was manually applied to the ID1 index to simulate forecast errors.

Costs

Not modeled A trading or execution result is reported without costs.

Source sentence
Transaction costs of the EPEX were also neglected.

Finds 3

  • The proposed strategy increases revenues by exploiting higher market volatility. performance
    Source sentence
    The results show that the proposed strategy allows an increment in revenues by taking advantage of the increasing market volatility.
  • Sensitivity analysis indicates that real day‑ahead forecasts have little impact, while intraday auction forecasts affect results more strongly. performance
    Source sentence
    the real prediction of the DAA did not significantly influence the result. In contrast, the real prediction for the IDA had a stronger influence.
  • The intraday continuous market (IDC) remains robust to volatility increases up to 50 %, with larger declines only beyond that level. performance
    Source sentence
    the IDC at least proved to be resistant to increased volatility, so that stronger declines were only recorded from a 50% increase in volatility.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 20 related papers). Coverage is partial, so this is neither replication nor refutation.

Code

No repository by the authors found in the paper.

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.

Cited References (45)

  1. Optimized operation of large scale battery systems (2022) — link
  2. Optimized O peration of L arge S cale B attery S ystems - E xperiences, N ew O pportunities and B ig D ata (2018)
  3. Betriebserfahrung und O ptimierung von G roßbatteriesystemen: E ine S imulationstudie zur O ptimierung des SoC -Managements (2020)
  4. Gurobi Optimizer Reference Manual (2023) — link
  5. JuMP: A Modeling Language for Mathematical Optimization (2017) — link
  6. Elektrochemische Speicher: Superkondensatoren, Batterien, Elektrolyse-Wasserstoff, Rechtliche Rahmenbedingungen — link
  7. Speicher: Für eine funktionierende Energiewende — link
  8. Speicher: Flexibilität für eine erfolgreiche Energiewende — link
and 37 more
  1. A R eview of M odeling, M anagement, and A pplications of G rid- C onnected L i- I on B attery S torage S ystems (2022) — link
  2. Modelling and optimal energy management for battery energy storage systems in renewable energy systems: A review (2022) — link
  3. Energy A rbitrage O ptimization W ith B attery S torage: 3D-MILP for E lectro- T hermal P erformance and S emi- E mpirical A ging M odels (2020) — link
  4. Energy S torage A rbitrage U nder D ay- A head and R eal- T ime P rice U ncertainty (2018) — link
  5. Optimal O peration of E nergy S torage S ystems C onsidering Forecasts and B attery D egradation (2018) — link
  6. Increasing the lifetime profitability of battery energy storage systems through aging aware operation (2023) — link
  7. Real- T ime S tochastic O ptimization of E nergy S torage M anagement U sing D eep L earning- B ased F orecasts for R esidential PV A pplications (2020) — link
  8. Robust MPC -based bidding strategy for wind storage systems in real-time energy and regulation markets (2021) — link
  9. Economic model predictive control of L i-ion battery cyclic aging via online rainflow-analysis (2021) — link
  10. Use of battery storage systems for price arbitrage operations in the 15- and 60-min G erman intraday markets (2018) — link
  11. Optimal S torage A rbitrage under N et M etering using L inear P rogramming (2019) — link
  12. Integration of sizing and energy management based on economic predictive control for standalone hybrid renewable energy systems (2019) — link
  13. Deep R einforcement L earning B ased R eal- T ime R enewable E nergy B idding W ith B attery C ontrol (2023) — link
  14. Deep R einforcement L earning- B ased E nergy S torage A rbitrage W ith A ccurate L ithium- I on B attery D egradation M odel (2020) — link
  15. An A mp-hour based electricity cost model for economic dispatch of batteries in microgrids (2014)
  16. Sim SES : A holistic simulation framework for modeling and analyzing stationary energy storage systems (2022) — link
  17. aging-aware- MPC (2023) — link
  18. Energy arbitrage through smart scheduling of battery energy storage considering battery degradation and electricity price forecasts (2016) — link
  19. YALMIP : A T oolbox for M odeling and O ptimization in MATLAB (2004)
  20. Alternative B attery T echnologies R oadmap 2030+ (2023) — link
  21. Factoring the C ycle A ging C ost of B atteries P articipating in E lectricity M arkets (2018) — link
  22. Convex optimization (2004)
  23. IBM CPLEX O ptimizer — link
  24. Model Predictive Control (2004)
  25. Decomposed Stochastic Model Predictive Control for Optimal Dispatch of Storage and Generation (2014) — link
  26. Co-Optimizing Battery Storage for the Frequency Regulation and Energy Arbitrage Using Multi-Scale Dynamic Programming (2018) — link
  27. Pyomo: modeling and solving mathematical programs in Python (2011)
  28. GLPK (GNU Linear Programming Kit) (2012) — link
  29. Optimal bidding in hourly and quarter-hourly electricity price auctions: Trading large volumes of power with market impact and transaction costs (2022) — link
  30. FlexIndex — link
  31. bess-optimizer — link
  32. pandas-dev/pandas: Pandas (2020) — link
  33. EU Power short-term forecast — link
  34. EAO — A Framework for Optimizing Decentralized Portfolios and Green Supply (2021) — link
  35. EUPHEMIA Public Description: Single Price Coupling Algorithm — link
  36. Optimized Sizing of Battery Energy Storage Systems in the Energy Arbitrage Business for Industrial Customers (2024)
  37. Array programming with NumPy (2020) — link

0 of 45 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.