What changed: the 2019 Template vs the 2026 Template
The Reliability Panel's Template for Compliance Programs replaces the 2019 Template for Generator Compliance Programs — the document most compliance programs in service today were built against.
It was published on 30 July 2026 and applies from 27 August 2026. Compliance programs must be consistent with it by 27 February 2027.
This is a rewrite, not a refresh. Here's what actually moved.
What changed structurally
2019 | 2026 | |
Title | Template for Generator Compliance Programs | Template for Compliance Programs |
Structure | 2 chapters + 1 table | 4 chapters + Appendices A, B, C |
Principles | 10 | 7 |
Rules version referenced | v129 | v251 |
Table columns | Standard / Methodology / Suggested frequency of testing / Notes / Basis | Method / Suggested assessment frequency / Basis |
Format | PDF plus a companion spreadsheet of the compliance tables |
Two of those matter more than they look.
The Notes column is gone. In 2019 it carried the applicability qualifiers — "applies to synchronous and conventional plant", "appropriate to use where high speed monitors are available". In 2026 that information sits in the method headings: Method 1 (continuous high-speed monitoring, with or without automated assessment tools), Method 2 (investigation of events where continuous high-speed monitoring is not practical). Your monitoring capability now determines which method you're in, and that choice is visible on the face of the table.
Supplementary methods are new. The 2019 table said plainly that where more than one method is provided, only one needs to be used. The 2026 tables introduce supplementary methods explicitly to be used in conjunction with other suggested methods. Method selection is no longer strictly either/or.
The companion spreadsheet, incidentally, is a genuine convenience and worth grabbing early — it makes re-mapping your existing clause table substantially faster.
What happened to the principles
Ten became seven, and the register changed. The 2019 principles read as design rationale, written from the Panel's point of view. The 2026 principles read as obligations. Four of the seven use "must".
Three are new: continuous monitoring as the preferred approach (Principle 2), assessment of plant changes including firmware (Principle 5), and the treatment of variations from the Template (Principle 7). Three 2019 principles have no direct successor — old #7 and old #8 are absorbed into the tables and event definitions, and old #9's statement that a Generator is accountable only for its own plant now appears as a note within the Table A.13 methods rather than as a general principle.
What's new in the guidance
Chapter 3 is almost entirely new material. Plant changes get a full section, with the clause 5.3.9(d) deemed list and active encouragement to use AEMO's scoping process. Model validation and updates expand from a few paragraphs to four sub-sections — including model overlays, an obligation to detect when models have gone stale, and the five-business-day update requirement under NER 5.3.8(f). Records now encourage retaining monitoring data for events demonstrating compliance, events indicating potential non-compliance, and reviewable operating incidents. Non-compliance becomes a full section covering AEMO's rectification period determination, the AER escalation mechanism, and what the AER tests in a technical compliance audit.
Two smaller additions worth knowing: section 4.6.1 covers monitoring away from the connection point, and section 4.9 addresses hardware-in-the-loop testing — the Panel's position being that HIL can complement other work but does not on its own provide reasonable assurance of ongoing compliance.
What changed in the tables
The 2019 table ran from S5.2.5.1 to S5.2.8. Appendix A now adds S5.2.5.5A (disturbance ride-through capability), S5.2.5.15 (short circuit ratio), S5.2.5.16 (voltage phase angle shift), and A.21, a pointer section directing participants to the MASS, the PFRR, the Power System Data Communications Standard, the Communication System Failure Guidelines, SO_OP_3705 and SO_OP_3708, and the Guide to VAR Dispatch.
Technology-specific guidance appears for the first time: Table A.13 directs testing across the conditions where voltage control behaves differently — low and high solar irradiance, reactive power on demand mode, low and high state of charge for BESS — and Table A.3 adds a method built around generating-unit testing for wind, solar and BESS.
The event definitions have been given teeth. Significant disturbance, major disturbance and major event existed in 2019 as short definitions. In 2026 each carries numbered worked examples with specific criteria. Section 4.8 adds that where no qualifying event occurs in the relevant period, you should still select and review the most relevant available event — and if nothing suitable occurred, document that.
The plant trip definition has broadened, too. 2019 tied the multi-unit case to systems of more than ten identical units. That threshold is gone; the 2026 definition applies to any production system of multiple units where a significant number of them, or critical ancillary plant, trips. Smaller multi-unit farms that sat outside the old wording are now inside it.
And frequencies are now framed as a floor. Section 4.7.3 states the recommended frequency should be considered the baseline requirement, with departure requiring documented reasoning under Principle 7. Section 4.7.1 makes the trade explicit: with continuous monitoring, testing focuses on edge cases; without it, more frequent testing may be necessary.
What this means in practice
Every generator compliance program in the NEM needs reviewing, and most will need amending — not because the underlying engineering changed, but because the Template yours was written against no longer exists in that form.
The documentation burden shifted from what you did to why you did it that way. You select a method and, where it varies from the Template's guidance or frequencies, record reasoning against seven named principles. That reasoning is what an AER technical compliance audit looks for.
"We don't monitor continuously" is now a position to be defended rather than a default.
Lead times will be non-zero. Data acquisition, IT/OT involvement, sometimes capital approval. Those don't compress.
Firmware and software updates are unambiguously in scope, and connect to a five-business-day model clock.
Model currency is a live obligation, not a commissioning milestone.
Event-triggered assessment now has defined thresholds — which means the trigger is testable by an auditor.
Records need to include the negative cases. Tests you couldn't perform and why. Periods with no qualifying event. Changes to the program and the reasons. These are precisely the records that don't get created by accident.
Storage and hybrid assets have the most new ground to cover.
The Template's scope has also broadened beyond generation to loads, distribution networks and HVDC links — the reason for the title change and the new appendices. It doesn't alter what Schedule 5.2 participants need to do.
We're walking asset owners through this in detail over the coming weeks — particularly the September continuous monitoring assessment, which is the decision most worth getting right early. If you'd like to work through what it means for a specific portfolio, we'd welcome the conversation.
