Pith. sign in

REVIEW 3 major objections 4 minor 41 references

Emission-Forecasting-Based Spatial-Temporal Carbon Response: A Multi-Agent Attention-Enhanced Deep Learning Framework

T0 review · 3 major / 4 minor · reviewed 2026-08-01 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Day-ahead forecasts of nodal carbon intensity, paired with spatial-temporal flexible loads, cut simulated emissions by more than 30%.

desk verdict A genuine forecasting-dispatch integration let down by an unsupported headline: the only 30% emission cut reported measures spatial-temporal flexibility, not the claimed one-hour latency reduction. read the letter →

arxiv 2607.26560 v1 pith:JHPDVV6C submitted 2026-07-29 eess.SY cs.SYecon.GNq-fin.EC

Feiyu Cai , Jing Qiu , Yi Yang , Chenxi Zhang , Xinlei Wang , Baichuan Liu , Junhua Zhao This is my paper · ORCID
classification eess.SYcs.SYecon.GNq-fin.EC
keywords Carbonintensityforecastingspatial-temporalresponsehierarchicaldesigndual-stageattentionmechanismmulti-agentcooperationmobileenergystoragedistributeddatacentersdemand
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Carbon-oriented demand response has relied on nodal carbon intensity (NCI) computed after the fact, so flexible loads learn about clean-energy windows too late. This paper argues that the same NCI can be predicted a day ahead with enough accuracy to drive proactive dispatch, and that the prediction matters operationally: in simulations, giving mobile battery-storage systems and distributed data centers day-ahead NCI forecasts instead of ex-post calculations, and letting them shift across both time and network location, reduces system emissions by more than 30% when one hour of scheduling latency is removed. The forecasting engine is a hierarchical network that first predicts renewable output and then feeds those predictions, together with historical NCI, through a CNN-GRU model with dual-stage attention and large-language-model agents that clean input data and refine forecasts. The scheduling layer treats the forecasts as a carbon price map and moves GDLs to the cleanest nodes at the cleanest hours, subject to travel time and workload-balance constraints. If the central claim is right, carbon accounting in power systems shifts from a retrospective scoreboard to an actionable prediction, and flexible assets can be dispatched toward decarbonization without waiting for the meter.

What carries the argument

The key machinery is a two-tier forecasting pipeline plus a spatial-temporal scheduling model. Tier 1 is a small ANN per renewable source that predicts next-day wind and solar output from historical production and weather. Tier 2 is a CNN-GRU hybrid with a dual-stage attention mechanism: a feature-attention module (FAM) weights input features (including the Tier-1 renewable forecasts) at each time step, and a temporal-attention module (TAM) uses Spearman rank correlation to weight past hidden states for the decoder. A large-language-model data-preprocessing agent cleans and normalizes inputs, and an evaluation agent analyzes forecast errors and suggests parameter adjustments that are applied

What would settle it

A direct simulation comparing Case 1 (ex-post NCI, spatial-temporal GDLs) and Case 3 (ex-ante NCI, spatial-temporal GDLs) under identical GDL flexibility, with the one-hour decision lag in Case 1 explicitly modeled, would settle the attribution; if that comparison shows less than 30% emission reduction, the paper's headline claim fails.

Watch

Extended reading notes

Core claim

The paper's central claim is that ex-ante NCI forecasting can replace ex-post carbon-flow calculation as the operating signal for low-carbon demand response. It reports that on a modified 33-bus distribution system with Australian market and weather data, the proposed framework achieves over 30% emission reduction under a one-hour reduction in scheduling latency. The supporting internal comparison shows 33.86% emission reduction when spatial-temporal GDL dispatch (mobile storage and data centers moving across buses) is added to temporal-only ex-ante scheduling. The paper also claims that the hierarchical ANN-CNN-GRU-DSAM (ACGD) forecasting model, with LLM agents at the data-input and output-

Load-bearing premise

The >30% reduction is credited to removing one hour of scheduling latency, but the paper reports no emission number for the comparison that isolates that latency (Case 1 vs Case 3) and provides no mathematical model of dispatch latency, so if the ex-post baseline or the latency scenario is modeled differently the claimed reduction could shrink or vanish.

Editorial extensions

If this is right

  • If correct, the framework turns NCI from an ex-post accounting metric into a day-ahead operational signal, so carbon-aware demand response can be proactive rather than reactive.
  • Mobile storage and data-center load can be dispatched both in time and across network locations, giving emission reduction even when renewable forecasts are imperfect.
  • The hierarchical two-tier design separates renewable forecast error from downstream NCI prediction, so any improvement in renewable forecasting should automatically improve carbon-signal quality.
  • The reported 33.86% emission reduction between temporal-only and spatial-temporal ex-ante scheduling is a direct incentive to deploy geographically movable flexible loads in distribution networks.
  • The attention weighting of features and temporal states, rather than simply deeper networks, appears to carry the accuracy gain over recurrent and CNN-hybrid baselines.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Inference: Since the paper reports 33.86% for Case 2 vs Case 3 (both ex-ante, differing in spatial flexibility) and does not report a Case 1-vs-Case 3 number, the attribution of >30% reduction to one hour of latency removal is not directly demonstrated; a cleaner test would isolate latency with spatial flexibility held fixed.
  • Inference: The LLM agents are heuristic workflow helpers, not trainable layers; a natural extension is to test whether their fine-tuning suggestions generalize to unseen network topologies or whether they overfit the regional dataset used here.
  • Inference: There is likely a threshold forecast-error level below which spatial-temporal GDL dispatch stops beating temporal-only dispatch; mapping that threshold would clarify how much forecasting accuracy the operational benefit requires.
  • Inference: The paper simulates at hourly resolution with one-hour dispatch; real-world communication and travel delays would need to be modeled before deployment, which the authors explicitly leave to future work.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper proposes an ex-ante spatial-temporal carbon response framework for power systems. A hierarchical ANN-CNN-GRU-DSAM model with LLM-based preprocessing/evaluation agents is used to forecast day-ahead nodal carbon intensity (NCI). The forecasts drive a scheduling model for geographically dispatchable loads (mobile energy storage systems and distributed data centers) on a modified IEEE 33-bus system with New South Wales data. The abstract and Section 5 claim that reducing carbon-scheduling latency by one hour achieves over 30% emission reduction; the only reported 30%+ number is a 33.86% reduction in Case 3 versus Case 2, both of which use ex-ante signals and differ in whether spatial-temporal GDL dispatch is allowed.

Significance. If the claimed result were valid, the framework would be a useful contribution: it couples NCI forecasting with operationally meaningful carbon-aware dispatch, provides a hierarchical renewable/NCI forecasting architecture, and demonstrates a concrete emission-reduction pathway for MESS/DDC flexibility. The paper also ships code and reports noise-robustness experiments, forecasting ablations, and a reproducible test environment, which are strengths. However, the central quantitative claim about latency reduction is not supported by the reported experiments, because the only reported 30%+ result does not isolate latency. The ground-truth NCI is also defined within the authors' own MESS-modified CEF accounting model, so independent verification is missing.

major comments (3)
  1. [Abstract, §4.2, §5] The headline claim that "under a one-hour reduction in carbon scheduling latency, the proposed model can achieve over 30% emission reduction" is not supported by any reported comparison. The only numerical 30%+ result is "Compared with Case 2, Case 3 reduces emissions by 33.86%" in §4.2. Cases 2 and 3 both use ex-ante NCI signals; Case 2 forbids spatial-temporal GDL dispatch and Case 3 permits it (§4.1). Thus 33.86% measures the value of spatial-temporal flexibility, not the removal of dispatch latency. The latency-isolating comparison, Case 1 versus Case 3, is never reported numerically, so the abstract's causal claim is unsupported.
  2. [§3.4, Eq. (42), §6] The optimization model contains no dispatch-latency variable. Equation (42) constrains MESS travel time TD, but that is physical movement time, not the latency between carbon-signal calculation and dispatch. Case 1 is defined only as not considering ex-ante scheduling, with no equation mapping the asserted one-hour delay into the decision problem. The conclusion in §6 explicitly states "Future work should develop mathematical models of dispatch latency," confirming that the mechanism behind the headline result is absent. Without such a model, the claimed one-hour-latency reduction cannot be computed or attributed.
  3. [§3.3, Eq. (23)–(25), §4.1] The "observed" NCI used as both the training target and the evaluation ground truth is computed from the MESS-modified CEF model of Ref. [34], whose authors overlap with this manuscript. This makes the reported emission reductions self-referential to the authors' own carbon-accounting conventions. The paper should validate the ground-truth NCI against an independent accounting method (e.g., standard CEF or a published emission-intensity dataset) and report the sensitivity of the 33.86% reduction to accounting assumptions. Without this, the emission numbers are not independently verifiable.
minor comments (4)
  1. [Eq. (25)] The loss expression is typeset unclearly: the fraction involving T and the weighting coefficients is hard to parse, and the equation mixes forecast error terms without explicit summation limits over features and buses. Please rewrite it with clear indices and dimensions.
  2. [Fig. 9] The labels 'P' and 'Q' in the figure are not adequately described in the caption or text. It would help to state explicitly which time periods correspond to early-morning and noon renewable peaks.
  3. [Fig. 11] The multi-criteria trade-off radar chart lacks numerical axes and does not report the actual values used for each criterion. Without numeric support, the claimed trade-off advantage of Agents+ACGD over iTransformer is not quantitatively verifiable.
  4. [Nomenclature and notation] Several symbols are corrupted or inconsistently rendered (e.g., indices in Ω, superscripts such as Eᵉˡᵉ, and the rank sequences in Eq. (11)). The notation should be normalized so that each symbol is defined once and used consistently.

Circularity Check

2 steps flagged · score 6.0 of 10

The >30% 'latency reduction' claim is the Case 2-vs-Case 3 spatial-dispatch result renamed; ground-truth NCI is the authors' own CEF model.

  1. renaming known result [Abstract; §4.2 Case definitions and 33.86% result; §5]
    "Compared with Case 2, Case 3 reduces emissions by 33.86%. ... The results demonstrate that under a one-hour reduction in carbon scheduling latency, the proposed model and methodology can achieve over 30% emission reduction."

    By the paper's own case definitions, Case 2 ('The system considers ex-ante scheduling, but GDLs cannot be spatial-temporally dispatched') and Case 3 ('The system considers ex-ante scheduling, and GDLs can be spatial-temporally dispatched') differ only in spatial-temporal GDL dispatch, both using ex-ante signals. The reported 33.86% therefore measures spatial-temporal GDL flexibility, not removal of one-hour latency. The abstract and §5 rename this number as the latency-reduction result, while the latency-isolating Case 1-vs-Case 3 comparison is never reported. §6 concedes no latency model exists ('Future work should develop mathematical models of dispatch latency'), so the headline prediction is the Case 2/3 statistic repurposed for a different variable.

  2. self citation load bearing [§3.3 Eq. (23)-(24); §4.2 observed-NCI definition]
    "The mathematical model in Eq. (23) and (24) is the MESS-based modified CEF model [34], which provides the calculated value used for comparison with the forecasted value when evaluating forecasting performance. ... the observed NCI is defined as the theoretical carbon intensity calculated via the CEF model using actual historical generation and load data."

    The paper's training target, forecast-validation target, and operational emission accounting all use the NCI defined by Eq. (23)-(24), a MESS-modified CEF model cited to [34], whose authors overlap with the present authors. The 'observed' NCI used as ground truth is produced inside the authors' own accounting framework. The ACGD forecaster is fit to reproduce this calculated NCI, and the scheduling objective (Eq. 26) then uses that forecast; the reported emission reductions are thus evaluated in units of the authors' own carbon model rather than against an independent external carbon benchmark. This self-citation is load-bearing for the quantitative emission claims, even though the forecasting architecture itself is independently testable.

full rationale

The ACGD forecasting architecture, ablation studies, and noise-robustness tests are largely self-contained and compare against independent baselines (LSTM, GRU, CNN-LSTM, CNN-GRU, iTransformer); those parts are not circular. The circularity is concentrated in the headline quantitative claim. The only reported >30% figure (33.86%) is the Case 2-vs-Case 3 comparison, which by the paper's own case definitions varies spatial-temporal GDL dispatch while holding ex-ante signals fixed; it is then relabeled in the abstract, §5, and conclusion as a one-hour latency-reduction result. The latency-isolating Case 1-vs-Case 3 comparison is never numerically reported, and §6 explicitly states that future work must develop mathematical models of dispatch latency, confirming that no causal latency mechanism is modeled. A secondary self-citation issue is that the 'observed' NCI ground truth comes from the authors' own MESS-modified CEF model [34], making the carbon accounting by which reductions are measured internal to the authors' prior work. These issues affect the central >30% claim but do not undermine the independent benchmark comparisons of the forecaster.

Assumptions & free parameters 7 free parameters · 6 assumptions · 2 invented entities

The central claim rests on five externally unvalidated pillars: (1) ground-truth NCI defined by the authors' own CEF extension [34]; (2) an asserted one-hour latency scenario with no mathematical model (§6); (3) a single IEEE 33-bus / NSW configuration; (4) GDL physical models imported from [35]–[37]; (5) LLM agents whose mechanism of benefit is unexplained. Free parameters are mostly unreported validation-selected hyperparameters plus the unmodeled latency assumption the headline depends on.

free parameters (7)
  • Loss-balance weights ω_RES, ω_NCI (Eq. 25) = not reported
    Weights balancing Tier-1 RES-output and Tier-2 NCI loss terms; the LLM evaluation agent recommends adjustments during daily fine-tuning (Algorithm 1 steps 18–19). No fitted values or sensitivity analysis reported.
  • Temporal attention window size μ (Eqs. 13–15) = not reported
    Window over historical hidden states in the TAM; selected by validation-set performance; value not stated.
  • Network/training hyperparameters (GRU units 128/128/64, batch 64, lr 1e-3, early-stopping patience 20) = Table II
    Validation-selected; standard, but part of the forecast-performance claims in Table III.
  • LLM configuration (GPT-4-0613, temperature 0.2) = 0.2 / gpt-4-0613
    Chosen for deterministic responses; shapes the agent behavior behind the 9.483% MAPE and noise-resilience claims.
  • Ex-post scheduling latency in Case 1 = 1 hour (asserted)
    The scenario parameter behind the headline '>30% emission reduction under one-hour latency reduction'; asserted in prose, never modeled mathematically (§6).
  • MESS travel time TD (Eq. 42) = not reported
    Derived from physical road distance with a 'conservative average speed'; scenario parameter affecting spatial dispatch feasibility.
  • DDC power-model constants (e1_i, e2_i, e3_i, φ1_i, φ2_i, EP_i, EI_i, etc.) = from [35], [36]
    Imported empirical/physical constants for server, cooling, and network power; domain parameters not fitted here.
assumptions (6)
  • domain assumption CEF model, including the MESS-modified CEF from [34], yields correct ground-truth NCI for both training and evaluation (Eqs. 23–24)
    The forecasting target and the emission accounting both use the authors' own extension of carbon emission flow; no measured carbon data validates the accounting.
  • domain assumption A modified IEEE 33-bus system with parameters from [40] and one year of NSW AEMO/BOM data is a representative test bed
    All quantitative claims (MAPE, 33.86%) are conditional on this single configuration; §5 concedes scalability and generalization are unvalidated.
  • domain assumption Gaussian input noise is a valid proxy for renewable-generation uncertainty in evaluating agent robustness
    The Fig. 10 experiment; the connection to physical RES uncertainty is asserted, not derived.
  • domain assumption LLM agent outputs (JSON preprocessing rules and parameter recommendations) improve, and never degrade, the forecasting pipeline
    Agents+ACGD's gain over ACGD and the noise-resilience gain are ascribed to the agents; the mechanism is not analyzed.
  • domain assumption GDL physical models (DDC power Eqs. 31–34; MESS movement Eqs. 35–42) from [35]–[37] apply unchanged
    Imported wholesale from prior literature; no validation of these models in the paper's setting is provided.
  • standard math Standard deep-learning math: Xavier initialization, GRU gating (Eqs. 18–21), softmax attention (Eqs. 7–13), Spearman rank correlation (Eqs. 11–13)
    Uncontroversial background; no proof required.
invented entities (2)
  • MESS 'internal carbon intensity' e_m,t^MESS
    purpose: Assigns a carbon intensity to stored energy inside mobile storage so charging and discharging emissions can be netted in the CEF accounting (Eqs. 23, 26)
    A bookkeeping convention introduced in the authors' prior work [34]; no external measurement validates it, and it directly shapes the emission objective from which the headline claim is computed.
  • LLM-based preprocessing and evaluation agents
    purpose: Automate data cleaning/normalization at the input stage and error analysis/parameter recommendation at the output stage of the forecast pipeline (Alg. 1, Eqs. 22)
    Software components, not physical entities; the only evidence for their benefit is the paper's internal ablation (Table III) and noise test (Fig. 10), both without error bars or external validation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Emission-Forecasting-Based Spatial-Temporal Carbon Response: A Multi-Agent Attention-Enhanced Deep Learning Framework." pith.science (2026). https://pith.science/paper/JHPDVV6C

@misc{pith2026260726560,
  author       = {Pith},
  title        = {Pith review of: Emission-Forecasting-Based Spatial-Temporal Carbon Response: A Multi-Agent Attention-Enhanced Deep Learning Framework},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JHPDVV6C}},
  note         = {Machine review of arXiv:2607.26560}
}
read the original abstract

As a major contributor to carbon emissions, the decarbonization of power systems has garnered significant societal attention. Nodal carbon intensity (NCI), a critical factor in carbon-oriented demand response, has traditionally been determined through ex-post calculations. However, this ex-post approach introduces latency in low-carbon dispatch. To address this, this paper presents a proactive ex-ante spatial-temporal carbon response framework. At its core, we develop a novel deep learning-based hierarchical design, enhanced by a dual-stage attention mechanism and a large language model (LLM)-based multi-agent cooperation system, to accurately forecast day-ahead NCI. This design effectively mitigates the impact of renewable energy uncertainty and enhances predictive resilience. On the demand side, the framework proposes a spatial-temporal carbon scheduling model that integrates geographically dispatchable loads (GDLs), including mobile energy storage systems (MESSs) and distributed data centers (DDCs). Leveraging high-accuracy day-ahead NCI predictions, the framework can effectively reduce system emissions by quickly responding to carbon intensity fluctuations. The proposed framework is tested on the modified IEEE 33-bus system. According to the simulation results, the impacts of proposed framework on dispatching latency and emission outcomes are analyzed. The results demonstrate that under a one-hour reduction in carbon scheduling latency, the proposed model and methodology can achieve over 30% emission reduction. This research breaks through the limitations of passive carbon accounting, advancing toward proactive carbon management. It offers an intelligent solution that accelerates the transition to cleaner power systems while directly supporting sustainable production goals.

Figures

Figures reproduced from arXiv: 2607.26560 by the authors.

Figure 1
Figure 1. Overview of the proposed ACGD-based multi-agent spatial-temporal carbon response framework. The first model is a hierarchical multi-agent collaborative day-ahead hourly NCI forecasting model for microgrids based on ACGD. This hierarchical architecture consists of a first-tier ANN and a second-tier hybrid neural network that integrates CNN and GRU modules enhanced by DSAM. In the first tier, renewable generation in t… view at source ↗
Figure 2
Figure 2. Flow diagram of the hierarchical ANN-CNN-GRU-DSAM hybrid model. The FAM is proposed to serve as an encoder between the output layer of CNN and the input layer of the GRU. FAM learns the correlation between each input feature and the actual NCI to be forecast, and adaptively processes the input features to strengthen the influence of relevant features while weakening the influence of irrelevant ones. Given the 𝑘 –th … view at source ↗
Figure 3
Figure 3. Flow diagram of the DSAM-enhanced GRU cell. The improved structure of the DSAM-enhanced GRU cell is illustrated in [PITH_FULL_IMAGE:figures/full_fig_p013_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: Example prompt designs for LLM-based forecasting-performance enhancement agents. Example prompt designs for this LLM-based multi-agent cooperation mechanism are presented in [PITH_FULL_IMAGE:figures/full_fig_p014_4.png]
Figure 5
Figure 5. Figure 5: Topology of the tested IEEE 33-bus distribution network and MESS movement paths. The topology of the IEEE 33-bus distribution network is illustrated in [PITH_FULL_IMAGE:figures/full_fig_p021_5.png]
Figure 6
Figure 6. Figure 6: Spatial-temporal carbon response of MESS2 in Cases 1 and 3. The bars in the figure represent the charging and discharging power of MESS2. The bar colors indicate the bus to which MESS2 moves for charging or discharging according to its movement path. The different-colo…
Figure 7
Figure 7. Figure 7: Spatial-temporal carbon response of DDCs in Case 3 [PITH_FULL_IMAGE:figures/full_fig_p023_7.png]
Figure 8
Figure 8. Figure 8: System carbon-emission distribution with and without spatial carbon response over 24 h. Compared with the movement mechanism of MESSs, DDCs achieve spatial-temporal [PITH_FULL_IMAGE:figures/full_fig_p023_8.png]
Figure 9
Figure 9. Figure 9: Day-ahead carbon-intensity forecasting performance in Cases 4-a and 4-b. Within Case 4, Case 4-a is used as the baseline prediction model. In Case 4-a, only a CNN￾GRU-DSAM hybrid neural network is used to complete the day-ahead NCI forecasting task. The forecasting res…
Figure 10
Figure 10. Figure 10: Noise-resilience comparison between baseline ACGD and agent-assisted ACGD. Because the objective of this work is not only standalone forecasting accuracy but also forecasting-driven carbon response, a multi-criteria trade-off analysis is further conducted in [PITH_FU…
Figure 11
Figure 11. Figure 11: Multi-criteria trade-off comparison among forecasting models [PITH_FULL_IMAGE:figures/full_fig_p026_11.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

41 extracted references · 5 linked inside Pith

  1. [34]

    Distribution Network Planning Towards a Low-Carbon Transition: A Spatial-Temporal Carbon Response Method,

    Y. Yang, J. Qiu and C. Zhang, “Distribution Network Planning Towards a Low-Carbon Transition: A Spatial-Temporal Carbon Response Method,” IEEE Trans. Sustain. Energy, vol. 15, no. 1, pp. 429–442, Jan. 2024

    2024

  2. [1]

    Energy Requirements and Carbon Emissions for a Low-Carbon Energy Transition,

    A. Slameršak, G. Kallis, and D. W. O’Neill, “Energy Requirements and Carbon Emissions for a Low-Carbon Energy Transition,” Nat. Commun., vol. 13, no. 1, p. 6932, Nov. 2022

    2022

  3. [2]

    U. N. E. Programme, We are All in this Together-Annual Report 2024, Feb. 2025. [Online]. Available: https://wedocs.unep.org/20.500.11822/47082

    2024

  4. [3]

    DCCEEW, National Greenhouse Gas Inventory Quarterly Update: March 2022, Aug. 2022. [Online]. Available: https://www.dcceew.gov.au/climatechange/publications/national– greenhouse–gas–inventory–quarterly–update–march–2022

    2022

  5. [4]

    Emissions of Carbon Dioxide in the Electric Power Sector,

    Congressional Budget Office, “Emissions of Carbon Dioxide in the Electric Power Sector,” Congressional Budget Office, D.C., USA, Dec. 2022

    2022

  6. [5]

    Resilience of renewable power systems under climate risks,

    L. Xu, et al., “Resilience of renewable power systems under climate risks,” Nature Reviews Electr. Eng. 1, 53–66, 2024

    2024

  7. [6]

    Carbon-Oriented Operational Planning in Coupled Electricity and Emission Trading Markets,

    Y. Wang, J. Qiu, Y. Tao, and J. Zhao, “Carbon-Oriented Operational Planning in Coupled Electricity and Emission Trading Markets,” IEEE Trans. Power Syst., vol. 35, no. 4, pp. 3145– 3157, Jul. 2020

    2020

  8. [7]

    Calculating Probabilistic Carbon Emission Flow: An Adaptive Regression- Based Framework,

    M. Ma, et al ., “Calculating Probabilistic Carbon Emission Flow: An Adaptive Regression- Based Framework,” IEEE Trans. Sustain. Energy, vol. 15, no. 3, pp. 1576–1588, Jul. 2024. 30

    2024

Show all 41 references
  1. [8]

    A Market-Clearing-Based Sensitivity Model for Locational Marginal and Average Carbon Emission,

    Z. Lu, L. Yan, J. Wang, C. Kang, M. Shahidehpour, and Z. Li, “A Market-Clearing-Based Sensitivity Model for Locational Marginal and Average Carbon Emission,” IEEE Trans. Energy Markets, Policy Regulat., vol. 2, no. 4, pp. 579–582, Dec. 2024

    2024

  2. [9]

    Generation Expansion Planning Considering the Rehabilitation of Aging Generating Units,

    M. Farhoumandi, F. Aminifar, and M. Shahidehpour, “Generation Expansion Planning Considering the Rehabilitation of Aging Generating Units,” IEEE Trans. Smart Grid, vol. 11, no. 4, pp. 3384–3393, Jul. 2020

    2020

  3. [10]

    Reliability Evaluation for Integrated Power- Gas Systems with Power-to-Gas and Gas Storages,

    Z. Zeng, T. Ding, Y. Xu, Y. Yang, and Z. Dong, “Reliability Evaluation for Integrated Power- Gas Systems with Power-to-Gas and Gas Storages,” IEEE Trans. Power Syst., vol. 35, no. 1, pp. 571–583, Jan. 2020

    2020

  4. [11]

    Bi-Level Carbon Trading Model on Demand Side for Integrated Electricity- Gas System,

    Z. Yan, et al., “Bi-Level Carbon Trading Model on Demand Side for Integrated Electricity- Gas System,” IEEE Trans. Smart Grid, vol. 14, no. 4, pp. 2681–2696, Jul. 2023

    2023

  5. [12]

    Coordinated Low-Carbon Dispatching on Source-Demand Side for Integrated Electricity-Gas System Based on Integrated Demand Response Exchange,

    C. Li, Z. Yan, Y. Yao, Y. Deng, C. Shao, and Q. Zhang, “Coordinated Low-Carbon Dispatching on Source-Demand Side for Integrated Electricity-Gas System Based on Integrated Demand Response Exchange,” IEEE Trans. Power Syst., vol. 39, no. 1, pp. 1287– 1303, Jan. 2024

    2024

  6. [13]

    Carbon Emission Flow from Generation to Demand: A Network-Based Model,

    C. Kang, et al ., “Carbon Emission Flow from Generation to Demand: A Network-Based Model,” IEEE Trans. Smart Grid, vol. 6, no. 5, pp. 2386–2394, Sep. 2015

    2015

  7. [14]

    Key Scientific Problems and Resear ch Framework for Carbon Perspective Research of New Power Systems,

    C. Kang, et al., “Key Scientific Problems and Resear ch Framework for Carbon Perspective Research of New Power Systems,” Power Syst. Technol., vol. 46, no. 3, pp. 821–833, 2022

    2022

  8. [15]

    Carbon-Aware Comput ing for Datacenters,

    A. Radovanovic, et al. , “Carbon-Aware Comput ing for Datacenters,” arXiv preprint arXiv:2106.11750, 2021

    2021 arXiv

  9. [16]

    Automated Emissions Reduction (AER) methodology,

    WattTime, “Automated Emissions Reduction (AER) methodology,” WattTime, Oakland, CA, USA, Tech. Rep., 2022. [Online]. Available: https://www.watttime.org

    2022

  10. [17]

    Optimal Power Scheduling Using Data-Driven Carbon Emission Flow Modelling for Carbon Intensity Control,

    Y. Wang, J. Qiu, and Y. Tao, “Optimal Power Scheduling Using Data-Driven Carbon Emission Flow Modelling for Carbon Intensity Control,” IEEE Trans. Power Syst ., vol. 37, no. 4, pp. 2894–2905, Jul. 2022

    2022

  11. [18]

    Energy Forecasting: A Review and Outlook,

    T. Hong, P. Pinson, Y. Wang, R. Weron, D. Yang, and H. Zareipour, “Energy Forecasting: A Review and Outlook,” IEEE Open Access J. Power Energy, vol. 7, pp. 376–388, 2020

    2020

  12. [19]

    Intelligent Wireless Tool Wear Monitoring System Based on Chucked Tool Condition Monitoring Ring and Deep Learning,

    N. Chen, et al., “Intelligent Wireless Tool Wear Monitoring System Based on Chucked Tool Condition Monitoring Ring and Deep Learning,” Adv. Eng. Inform., 65, 103176, May. 2025

    2025

  13. [20]

    A Novel Algorithm for Tool Wear Monitoring Utilizing Model and Knowledge-Guided Multi-Expert Weighted Adversarial Deep Transfer Learning,

    Z. Gao, N. Chen, and L. Li, “A Novel Algorithm for Tool Wear Monitoring Utilizing Model and Knowledge-Guided Multi-Expert Weighted Adversarial Deep Transfer Learning,” Mech. Syst. Signal Process., 228, 112456, Apr. 2025

    2025

  14. [21]

    Carboncast: Multi-Day Forecasting of Grid Carbon Intensity,

    D. Maji, P. Shenoy and R. K. Sitaraman, “Carboncast: Multi-Day Forecasting of Grid Carbon Intensity,” in Proc. 9th ACM Int. Conf. Syst. Energy-Efficient Buildings, Cities, Transportation (BuildSys), pp. 198–207, 2022

    2022

  15. [22]

    Innovative approach to daily carbon dioxide emission forecast based on ensemble of quantile regression and attention BILSTM,

    Z. Zhou, L. Yu, Y. Wang, Y. Tian, and X. Li, “Innovative approach to daily carbon dioxide emission forecast based on ensemble of quantile regression and attention BILSTM,” J. Clean. Prod., vol. 460, 142605, 2024

    2024

  16. [23]

    Carbon emissions forecasting based on a new hybrid model under carbon reduction target: a case study of eastern region in China,

    B. Zhao and W. Yang, “Carbon emissions forecasting based on a new hybrid model under carbon reduction target: a case study of eastern region in China,” J. Clean. Prod. , vol. 494, 144957, 2025

    2025

  17. [24]

    Gpt-4 Technical Report,

    J. Achiam, et al., “Gpt-4 Technical Report,” arXiv preprint arXiv:2303.08774, 2023. 31

    2023 arXiv

  18. [25]

    Llama: Open and Efficient Foundation Language Models,

    H. Touvron, et al., “Llama: Open and Efficient Foundation Language Models,” arXiv preprint arXiv:2302.13971, 2023

    2023 arXiv

  19. [26]

    Large Language Model for Low- Carbon Energy Transition: Roles and Challenges,

    Y. Cheng, X. Zhou, H. Zhao, J. Gu, X. Wang, and J. Zhao, “Large Language Model for Low- Carbon Energy Transition: Roles and Challenges,” in 2024 4th Power System and Green Energy Conference (PSGEC), Shanghai, China, pp. 810–816, 2024

    2024

  20. [27]

    CarbonGPT: meta causal graph-enhanced large language models for carbon emission forecasting in large-scale power distribution networks,

    S. Li, J. Lai, H. Li, W. Tang, Y. Xue, and H. Jiang, “CarbonGPT: meta causal graph-enhanced large language models for carbon emission forecasting in large-scale power distribution networks,” IEEE Trans. Smart Grid. Early access: Feb. 03, 2026, doi: 10.1109/TSG.2026.3660905

    2026

  21. [28]

    Large Language Model-Based Bidding Behavior Agent and Market Sentiment Agen t-Assisted Electricity Price Prediction,

    X. Lu, J. Qiu, Y. Yang, C. Zhang, J. Lin, and S. An, “Large Language Model-Based Bidding Behavior Agent and Market Sentiment Agen t-Assisted Electricity Price Prediction,” IEEE Trans. Energy Markets, Policy Regulat., vol. 3, no. 2, pp. 223–235, Jun. 2025

    2025

  22. [29]

    Machine Learning: Trends, Perspectives, and Prospects,

    M. I. Jordan and T. M. Mitchell, “Machine Learning: Trends, Perspectives, and Prospects,” Science, vol. 349, no. 6245, pp. 255–260, 2015

    2015

  23. [30]

    Yolov4: optimal speed and accuracy of object detection,

    A. Bochkovskiy, C.-Y. Wang, and H.-Y. M. Liao, “Yolov4: optimal speed and accuracy of object detection,” arXiv Preprint arXiv:2004.10934, 2020

    2004 arXiv

  24. [31]

    Deep learning for cybersecurity in sm art grids: review and perspectives,

    J. Ruan, et al ., “Deep learning for cybersecurity in sm art grids: review and perspectives,” Energy Convers. Econ., vol. 4, no. 4, pp. 233–251, Aug. 2023

    2023

  25. [32]

    Application of deep learning image recognition for lithium battery state of health assessment,

    Y. Liu, Y. Su, S. Zhang, V. Terzija, and Z. Cheng, “Application of deep learning image recognition for lithium battery state of health assessment,” Energy Convers. Econ., vol. 6, no. 4, pp. 246–255, Jul. 2025

    2025

  26. [33]

    A dual-stage attention-based recurrent neural network for time series prediction,

    Y. Qin, D. Song, H. Chen, W. Cheng, G. Jiang, and G. Cottrell, “A dual-stage attention-based recurrent neural network for time series prediction,” arXiv preprint arXiv:1704.02971, 2017

    2017 arXiv

  27. [35]

    Aggregated Model of Data Network for the Provision of Demand Response in Generation and Transmission Expansion Planning,

    M. Chen, et al., “Aggregated Model of Data Network for the Provision of Demand Response in Generation and Transmission Expansion Planning,” IEEE Trans. Smart Grid, vol. 12, no. 1, pp. 512–523, Jan. 2021

    2021

  28. [36]

    Incentive-Compatible Demand Response for Spatially Coupled Internet Data Centers in Electricity Markets,

    M. Chen, C. Gao, M. Shahidehpour and Z. Li, “Incentive-Compatible Demand Response for Spatially Coupled Internet Data Centers in Electricity Markets,” IEEE Trans. Smart Grid, vol. 12, no. 4, pp. 3056–3069, Jul. 2021

    2021

  29. [37]

    Utility-Scale Portable Energy Storage Systems,

    G. He, J. Michalek, S. Kar, Q. Chen, D. Zhang, and J. F. Whitacre, “Utility-Scale Portable Energy Storage Systems,” Joule, vol. 5, no. 2, pp. 379–392, 2021

    2021

  30. [38]

    Available: http://www.aemo.com.au/

    Australian Energy Market Operator (AEMO) [Online]. Available: http://www.aemo.com.au/

  31. [39]

    Available: http://www.bom.gov.au/

    Australian Government Bureau of Meteorology [Online]. Available: http://www.bom.gov.au/

  32. [40]

    Flexible Integrated Network Planning Considering Echelon Utilization of Second Life of Used Electric Vehicle Batteries,

    Y. Yang, J. Qiu, C. Zhang, J. Zhao and G. Wang, “Flexible Integrated Network Planning Considering Echelon Utilization of Second Life of Used Electric Vehicle Batteries,” IEEE Trans. Transp. Electrif., vol. 8, no. 1, pp. 263–276, Mar. 2022

    2022

  33. [41]

    itransformer: Inverted transformers are effective for time series forecasting,

    Y. Liu, et al., “itransformer: Inverted transformers are effective for time series forecasting,” in Proc. 12th Int. Conf. Learn, Represent. (ICLR), Vienna, Austria, May 7–11, 2024

    2024

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.