A quantitative analysis of the effect of flexible loads on reserve markets

Sebastien Mathieu, Quentin Louveaux, Damien Ernst, Bertrand CornélussePscc 2014 · DOI

TL;DR

This paper introduces a novel agent-based model to analyze how flexible loads, when allowed to participate in electricity reserve markets, impact market prices and costs. The key finding is that enabling flexible loads to offer "modulation bids" significantly reduces reserve procurement costs for system operators, though it can increase reliance on more expensive non-contracted reserves.

What Problem Does This Paper Address?

The increasing integration of renewable energy sources introduces greater variability into electricity grids, necessitating robust reserve markets to ensure grid stability. Traditionally, these reserve markets have been dominated by generation units. However, there's a growing interest in leveraging "load flexibility" – the ability to shift electricity consumption in time – as a resource for ancillary services. This paper addresses the challenge of quantitatively analyzing the economic impacts of allowing flexible loads to participate in day-ahead reserve markets, specifically focusing on how this participation affects energy prices, reserve procurement costs, and the behavior of market actors (producers and retailers). A key gap identified is the lack of models that capture the "gaming" possibilities of individual actors optimizing their profits across energy, reserve, and imbalance markets, especially when demand-side resources offer new bidding strategies.

Methods

The authors employ an agent-based modeling (ABM) approach, simulating the interactions of various market participants: producers, retailers, and the system operator (SO). Each agent optimizes its decisions (bidding in energy and reserve markets, managing imbalances) using linear programming based on forecasts of future market prices and tariffs.

A core innovation is the introduction of "modulation bids" for flexible loads in the reserve market. Unlike traditional reserve bids that offer a fixed capacity, modulation bids represent a flexibility margin around a baseline consumption over a set of periods. This acknowledges that shifting consumption in one period necessitates adjustments in others to maintain overall energy balance and consumer utility. The model also incorporates:

  • Day-ahead energy and reserve markets: These are cleared sequentially.
  • Imbalance settlement: Actors pay or receive fees based on deviations from their declared positions.
  • Price forecasting: Agents use a weighted average of past prices.
  • Retailer optimization: Retailers manage both inelastic and flexible loads, optimizing consumption to minimize costs and potentially profit from reserve offers.
  • Producer optimization: Producers bid energy and reserve capacity, also managing imbalances.

Datasets

Not discussed in the paper. The paper uses simulated data generated by the agent-based model itself.

Key Insights and Ideas

  • Flexible loads can significantly reduce reserve costs: Allowing flexible loads to participate in reserve markets, particularly through modulation bids, drastically cuts down the SO's procurement costs.
  • Modulation bids are a more realistic representation of demand-side flexibility: They capture the inherent trade-offs in shifting consumption, acknowledging that a demand reduction in one period might require an increase later.
  • Retailers can act as aggregators: Retailers can leverage their control over flexible loads to participate in reserve markets, creating new revenue streams and optimizing their overall portfolio.
  • Gaming is possible: The ABM approach allows for the analysis of how individual actors strategically optimize their bids and positions to maximize profit, which can influence market outcomes.
  • Trade-offs exist: While modulation bids reduce procurement costs, they can lead to an increased reliance on more expensive non-contracted reserves, especially when the same sign of imbalance is required for extended periods. This suggests a need for careful calibration of reserve procurement levels.

Key Results

  • Energy Market Price: The mean energy market price remained largely unchanged (around 49.81 €/MWh) even with increasing flexibility.
  • SO Reserve Procurement Cost: Opening the reserve market to modulation bids reduced SO procurement costs significantly, to approximately 11% of the initial costs (from 42,683 € to a much lower value, though a specific final number isn't explicitly stated for the modulation bid scenario, the relative reduction is emphasized).
  • Total Imbalance: The total imbalance of the system decreased slightly with the introduction of flexible loads.
  • Non-Contracted Reserves: The volume of non-contracted reserves used increased with the rate of flexible consumption, particularly due to the limitations of modulation bids in sustaining imbalances of the same sign over their entire duration. For example, with 10% flexible consumption, the volume of non-contracted reserves used was around 200 MWh, compared to near zero in the baseline scenario.

Limitations and Future Work

The authors acknowledge several limitations and suggest future research directions:

  • Simplified Forecasting: Agents use a simple exponential average for price forecasting. More sophisticated forecasting methods could be explored.
  • Deterministic Model: The current model is deterministic, assuming perfect knowledge of future prices and imbalances for optimization. Introducing stochasticity would provide a more realistic representation.
  • Market Design Variants: The model could be extended to assess other market design variants, such as:
  • Energy-constrained bids: Incorporating bids that have limits on the total energy they can provide.
  • Market-based capacity prices: Allowing capacity prices for reserves to be determined by the market rather than being regulated. This could lead to gaming.
  • Congestion and over-voltage services: Extending the model to analyze the provision of services for distribution network management.
  • Contracting More Reserves: The paper suggests that contracting for more reserves could mitigate the issue of increased non-contracted reserve usage when using modulation bids. This is a practical consideration for system operators.

Summary updated

BibTeX
@inproceedings{mathieu2014quantitative,
  author = {Sebastien Mathieu and Quentin Louveaux and Damien Ernst and Bertrand Cornélusse},
  title = {{A quantitative analysis of the effect of flexible loads on reserve markets}},
  booktitle = {Pscc},
  year = {2014},
  doi = {10.1109/pscc.2014.7038428},
}
Figure