Standards context

Standards AgenticGrid Proworks within

Four documents shape how the workbench organises evidence and governs its AI. Referencing them is context for your reviewers, not a claim of conformance.

Why this page exists

The documents your reviewers will ask about

When a transformer decision is challenged, the challenge is usually framed in the language of a standard: was the gas interpretation consistent with recognised guidance, was the loading assessment consistent with the loading guide, was the AI governed under a recognised risk framework. AgenticGrid Pro is built so those questions have short answers.

What follows is a plain statement of what each document covers and precisely how the workbench relates to it. Where the relationship is "we organise evidence in these terms", we say that. Where it would be a stretch, we say that too.

The four documents

What each covers, and how the workbench relates

IEEE C57.104 — DGA interpretation
IEEE guidance for interpreting dissolved gases in mineral-oil-immersed transformers. AgenticGrid Pro screens multi-gas trends and generation rates against C57.104 context and shows the measured values, units and sampling dates behind every screening result. It does not present itself as an implementation of the standard, and it issues no final DGA conclusion of its own — interpretation remains subject to qualified engineering review and your own procedures.
IEC 60599 — gas and ratio interpretation
IEC guidance on interpreting dissolved and free gases in oil-filled electrical equipment, including the basic gas ratios. AgenticGrid Pro computes those ratios and displays Duval Triangle placement alongside the source values, so a reviewer sees the ratio and the report it came from together rather than a derived verdict on its own.
IEEE C57.91 — loading and thermal aging
The IEEE guide for loading mineral-oil-immersed transformers and step-voltage regulators, including the thermal aging model. AgenticGrid Pro places hot-spot and top-oil estimates in the context of that guide and carries the loading profile and assumptions with the estimate. It is not a substitute for a loading study.
NIST AI RMF — AI governance posture
The NIST AI Risk Management Framework, which organises AI risk work as govern, map, measure and manage. AgenticGrid Pro's AI posture is framed in those terms: bounded agent tasks, visible evidence, an explicit human decision point, and an append-only audit trail. A deployment review under this framing plugs into your existing OT cyber-security programme; it does not replace it.

Thermal aging

Why the loading guide is load-bearing here

Per the IEEE C57.91 transformer loading guide's Arrhenius aging model, winding insulation aging rate approximately doubles for every 7 °C increase in hot-spot temperature above the 110 °C reference for 65 °C-rise thermally upgraded insulation. That relationship is the engineering basis for treating thermal and loading condition monitoring as insulation-life protection rather than a comfort metric.

It is also why AgenticGrid Pro refuses to show a hot-spot number without the load and ambient conditions that produced it. A temperature without its operating context cannot be read against the aging model at all.

The boundary

Context, not certification

GridAPM makes no claim of certification, accreditation, conformance assessment or endorsement by IEEE, IEC, NIST or any other body. None of these organisations has reviewed, approved or endorsed AgenticGrid Pro. Naming a standard on this site describes the framework our evidence model and terminology are organised around — nothing more.

Standards are also not interchangeable with your own procedures. Where your utility's practice is stricter or differently structured, your practice governs; the workbench is designed to record the assumptions actually used so that the difference is visible in the audit trail rather than buried.

Nothing here is engineering advice for a specific asset. Every output of AgenticGrid Pro is draft material until a qualified engineer in your organisation approves it.

Put your own evidence through the workflow

A bounded evaluation starts with the records you already have and ends with an evidence pack your reviewers can inspect line by line. Your units, your engineers at the gate.

Type to search research, platform pages, and tools.