IEEE Transactions on Smart Grid Submission Guide

TSG covers integrated energy networks in which distributed resources, transportation, communications, prosumers, markets, data, and cyber-physical control interact. Define the system decision your work improves and the grid conditions that make the problem distinctive.

For machine learning or optimization papers, show why a generic method is insufficient, which physical or operational structure is used, and how violations, uncertainty, or distribution shift are handled. For control or market designs, show feasibility under realistic information, timing, network, and participant assumptions.

The current TSG information-for-authors document says Transactions papers are limited to 10 pages at submission and papers over that limit are returned without review. Revision rules and final overlength charges differ, so verify the latest Author's Kit.

Compress repetition, not evidence. Keep the grid problem, contribution boundary, essential model, primary validation, comparison, limitations, and operational interpretation in the main paper. Use supplementary material only where current IEEE policy allows and where reviewers can still evaluate the headline claim from the paper.

Do not hide constraint violations behind an average performance score. Report what happens in the difficult intervals and for which system sizes the method remains usable.

Use the current PES Author's Kit and template. Follow the online route linked from the current journal instructions, confirm ORCID details, and inspect the converted PDF for equations, fonts, figure labels, references, and author information.

Choose Traditional or open-access publication only after checking the current terms. Fees are volatile and should not be copied from an old guide. Confirm current overlength and color-print choices with IEEE before approval.

Generic AI on a grid dataset. Expose the power-system structure, operational constraint, and decision consequence.

Unfair baseline comparison. Re-run baselines with comparable data, forecasts, constraints, and tuning budgets.

Page-limit damage. Remove duplicated literature and exposition before moving decisive validation out of view.

Simulation-only certainty. State model boundaries and add realistic sensitivity, hardware, field, or historical evidence where the claim requires it.

In our pre-submission review work, the comparison most likely to change a TSG manuscript is the equal-information test. A proposed controller or predictor may appear superior because it receives a better forecast, sees future values, uses network parameters unavailable to the baseline, or is tuned on the evaluation system. Those differences can be scientifically valid, but they must be disclosed and separated from the contribution attributed to the method.

Create a comparison ledger listing data, forecasts, constraints, compute budget, tuning access, topology knowledge, and retraining frequency for every baseline. Re-run the central comparison under equal information where possible. When equalization is impossible, state the asymmetry and explain why the comparison remains useful.

This audit also catches data leakage. Random train-test splits can put adjacent time points, the same disturbance, or the same operating regime on both sides of the split. A deployment-shaped split should reflect the time, location, asset, or event transfer claimed in the paper.

Consider an algorithm that coordinates batteries and flexible loads and reports lower operating cost. The editorially meaningful question is not only whether the objective value improves. It is whether the method respects voltage, thermal, device, communication, privacy, and market constraints under realistic uncertainty.

A stronger evidence sequence begins with the grid decision and failure cost, defines what information each actor has, compares against current operational and research baselines, and reports constraint violations alongside cost. Stress tests vary forecast error, communications delay, device availability, topology, and participation. The paper then states whether the contribution is an optimization result, a control architecture, a market design, or an empirical system finding. That specificity prevents a technical paper from promising all four.

The current information-for-authors document imposes a 10-page initial limit for Transactions papers and says papers over the limit are returned without review. It also states the current ORCID requirement. These are hard checks, not editorial preferences.

The same document describes single-anonymous review by at least two independent reviewers and plagiarism screening. It also describes final overlength charges and a USD 2,495 open-access fee for the version reviewed, but fees and terms can change. Verify all monetary details in the live IEEE Author Center before selecting a publishing route.

Keep an archive of the submitted source, generated PDF, supplement, portal abstract, keywords, author order, funding, conflicts, and copyright or license choices. A later revision is easier when the team can reconstruct exactly what the editor and reviewers received.

Before final approval, give a coauthor who did not build the model one hour with the paper. Ask that reader to locate the operational decision, information available at decision time, baseline, binding constraint, principal failure case, and validity boundary. If those cannot be found quickly, the problem is likely argument visibility rather than a need for another algorithm. Repair the first-read path before adding more experiments.

Browse the journal directory, compare the IEEE Transactions on Power Systems submission guide and how to choose a journal for a paper, then use a free manuscript readiness scan to check whether the claim, validation, and limitations remain aligned. The owner decision should follow the system contribution and reader, not the IEEE brand alone.

Official and authoritative sources accessed 2026-08-17.

• IEEE Transactions on Smart Grid information for authors, IEEE.
• IEEE PES Author's Kit, IEEE Power & Energy Society.
• Submit an article for peer review, IEEE Author Center.