Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Cherry-Picking Approach to Large Load Shaping for More Effective Carbon Reduction

Bokan Chen, Raiden Hasegawa, Adriaan Hilbers, Ross Koningstein, Ana Radovanović, Utkarsh Shah, Gabriela Volpato, Mohamed Ahmed, Tim Cary, Rod Frowd

TL;DR

The paper addresses how large flexible loads, such as data centers, can reduce grid CO2 emissions and electricity costs through day-ahead load shaping. It employs a calibrated ERCOT day-ahead DC-OPF framework to compare multiple shaping signals against an optimization-based upper bound, and introduces a practical cherry-picking rule that selects the next day's strategy based on grid net demand (GND) and daily minimum LMP ($min(LMP)$), informed by ML insights. Key findings show that while LMP-based shaping is strong among practical signals, cherry-picking across signals achieves CO2 reductions up to roughly $2$–$3\times$ the single-signal performance and can reach about $66\%$ of the optimization upper bound, with only minimal cost impact (≤$0.6$ per MWh). The results also reveal substantial additional CO2 savings from spatial shifting across two data-center locations, underscoring the importance of grid-aware, location-specific strategies for carbon-aware computing in data centers, DERs, and VPPs.

Abstract

Shaping multi-megawatt loads, such as data centers, impacts generator dispatch on the electric grid, which in turn affects system CO2 emissions and energy cost. Substantiating the effectiveness of prevalent load shaping strategies, such as those based on grid-level average carbon intensity, locational marginal price, or marginal emissions, is challenging due to the lack of detailed counterfactual data required for accurate attribution. This study uses a series of calibrated granular ERCOT day-ahead direct current optimal power flow (DC-OPF) simulations for counterfactual analysis of a broad set of load shaping strategies on grid CO2 emissions and cost of electricity. In terms of annual grid level CO2 emissions reductions, LMP-based shaping outperforms other common strategies, but can be significantly improved upon. Examining the performance of practicable strategies under different grid conditions motivates a more effective load shaping approach: one that "cherry-picks" a daily strategy based on observable grid signals and historical data. The cherry-picking approach to power load shaping is applicable to any large flexible consumer on the electricity grid, such as data centers, distributed energy resources and Virtual Power Plants (VPPs).

A Cherry-Picking Approach to Large Load Shaping for More Effective Carbon Reduction

TL;DR

The paper addresses how large flexible loads, such as data centers, can reduce grid CO2 emissions and electricity costs through day-ahead load shaping. It employs a calibrated ERCOT day-ahead DC-OPF framework to compare multiple shaping signals against an optimization-based upper bound, and introduces a practical cherry-picking rule that selects the next day's strategy based on grid net demand (GND) and daily minimum LMP (), informed by ML insights. Key findings show that while LMP-based shaping is strong among practical signals, cherry-picking across signals achieves CO2 reductions up to roughly – the single-signal performance and can reach about of the optimization upper bound, with only minimal cost impact (≤ per MWh). The results also reveal substantial additional CO2 savings from spatial shifting across two data-center locations, underscoring the importance of grid-aware, location-specific strategies for carbon-aware computing in data centers, DERs, and VPPs.

Abstract

Shaping multi-megawatt loads, such as data centers, impacts generator dispatch on the electric grid, which in turn affects system CO2 emissions and energy cost. Substantiating the effectiveness of prevalent load shaping strategies, such as those based on grid-level average carbon intensity, locational marginal price, or marginal emissions, is challenging due to the lack of detailed counterfactual data required for accurate attribution. This study uses a series of calibrated granular ERCOT day-ahead direct current optimal power flow (DC-OPF) simulations for counterfactual analysis of a broad set of load shaping strategies on grid CO2 emissions and cost of electricity. In terms of annual grid level CO2 emissions reductions, LMP-based shaping outperforms other common strategies, but can be significantly improved upon. Examining the performance of practicable strategies under different grid conditions motivates a more effective load shaping approach: one that "cherry-picks" a daily strategy based on observable grid signals and historical data. The cherry-picking approach to power load shaping is applicable to any large flexible consumer on the electricity grid, such as data centers, distributed energy resources and Virtual Power Plants (VPPs).
Paper Structure (15 sections, 10 figures, 3 tables, 1 algorithm)

This paper contains 15 sections, 10 figures, 3 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Background on carbon accounting methodologies
  3. Methodology
  4. Daily optimization-based benchmark
  5. Comparative evaluation of load shaping strategies
  6. Impact over the course of a year
  7. Key factors influencing strategy effectiveness
  8. Seasonal variations
  9. Insights from using ML to select a load shaping strategy for a day
  10. Impact of next day’s minimum LMP on strategy effectiveness
  11. Cherry-picking for increased CO2 reduction from load shaping
  12. Relating the rule-based cherry-pick policy to historically observed grid dynamics
  13. The impact of price-driven load shaping strategies on generation dispatch
  14. Temporal and spatial shifting of energy demand between data centers
  15. Conclusions and future work

Figures (10)

  • Figure 1: How a signal (in this case hourly LMP) translates into data center demand changes.
  • Figure 2: Generation mix in North and East ERCOT weather zones in 2023.
  • Figure 3: Reduced yearly carbon emissions and average electricity price from following each load shaping signal.
  • Figure 4: Daily CO2 emission savings across the most effective strategies.
  • Figure 5: Carbon impact for low and high grid net demand (GND) regimes.
  • ...and 5 more figures