07 · Essays

The limits of remote audits

Last updated 2026-04-21
Essay (opinion)
Long-form opinion piece. Reflects Opnor's engineering positioning, not industry consensus.

Opnor offers a free 5-minute online diagnostic and a paid remote EMD engagement that requires no site visit. We will also tell you, in the same conversation, that we won't sign off on a deep-dive audit without walking your floor. The two positions aren't in tension — they're a function of what each tier is actually defensible as.

What can be remote

A surprising amount, given the right inputs. Remote diagnostics can:

Reconcile utility data against a partial asset list. Twelve months of utility bills + a CSV of major equipment is enough to flag the most obvious mismatches: a plant that's using 30% more electricity than its asset list explains, a thermal carrier missing from the spreadsheet, a sub-meter that's reading nonsense.

Identify directional ECM categories. If we know your sector, your annual spend, and your carrier mix, we can tell you the likely top-3 ECM categories without ever seeing the plant. Sawmills get compressed-air leaks and kiln VFDs. Mining gets ventilation. Foundries get furnace heat recovery. The categories are sector-stable.

Estimate program eligibility. Hydro-Québec SEA eligibility is a function of province, sector, electrical share, and consumption magnitude. All readable from utility bills.

Build a baseline regression. 12 months of meter data plus weather and production figures gives a defensible R²-grade baseline. That's the foundation of post-audit M&V (IPMVP Option C).

What needs to be on-site

The line shifts when you cross from identifying opportunities to sizing them defensibly:

Load factor verification. A 75 kW pump on the asset list could be running at 40% load or 90% load. The savings difference between those two scenarios is a factor of 4 (cube law). You can't determine load factor from a utility bill or a spreadsheet — you need clip-on ammeters, ideally over a representative production cycle. Field work.

Asset-list completeness. Plants always have unregistered load. The boiler feed pump that was added five years ago and never made it onto the maintenance database. The compressed-air dryer that someone's old-school maintenance manager installed without telling engineering. You only find these by walking the floor with a clipboard.

Operational condition assessment. A nameplate doesn't tell you the kiln fan is running on a soft-foot misalignment that's consuming 8% extra power. The boiler stack temperature isn't in any spreadsheet. The compressed-air system isn't actually pressure-banded the way the controls schematic says it is. Site walks find these.

ECM specification. Sizing a VFD retrofit requires knowing the cable run, the panel space, the existing harmonics, and the local access constraints (confined-space lockout-tagout, off-shift work windows). All site-specific. All affecting capex and labour estimates by 20%+.

The load-factor case in particular
We get asked occasionally: "why can't your AI just predict load factors from the asset spec?" We can — to about ±15%, which on a 75 kW pump translates to a ±20,000 kWh/yr swing in the savings number. For directional analysis, fine. For a deliverable that's going into a funding application or a capital request, not fine.

Why this matters for the buyer

The reason this line matters isn't process purity — it's defensibility. ISO-50002 Level-2 conformance requires sampled site measurement. Hydro-Québec SEA reviewers reject applications that lack on-site verification of major loads. A P.Eng. won't sign an audit report for an asset they haven't physically observed in operation.

Vendors selling fully-remote audits are doing one of three things: (1) delivering Level-1 walkthroughs marketed as Level-2, (2) signing reports without P.Eng. accountability and hoping no one challenges, or (3) operating outside Quebec where SEA conformance isn't the bar. None of those work for an industrial plant looking to actually fund and implement ECMs.

How we structure the tiers honestly

Three tiers, each defensible at its claimed depth:

Tier 0 — Free Energy Makeover Report. Self-serve, 5 minutes, web-based. Gives you a directional savings band, an ECM category list, and HQ eligibility. No commitment, no claim that it's a substitute for an audit.

Tier 1 — EMD (Energy Makeover Diagnostic). Paid, remote. Analyst review of utility bills + asset list + sub-meter data you upload. Output: focused report identifying the top-3 ECMs and whether a 2-week audit is justified. Defensible at the "screening" level. Not SEA-eligible on its own.

Tier 2 — 2-Week Deep-Dive Audit. Paid, on-site. Includes one day of field walk-through, sampling of major loads, ISO-50002 Level-2 conformance, P.Eng. sign-off. SEA-eligible. The deliverable a plant can hand to a funding reviewer or a capital committee.

We will tell you in scoping which tier matches your need. If you ask for a 2-week audit and your situation only justifies an EMD, we'll say so and quote you the smaller engagement. If you need a 2-week audit but you're hoping the EMD is enough, we'll explain why it isn't.

🚧 Author backlog (Opnor team to fill)
  • Add a comparison table: what each tier verifies vs what it doesn't
  • Founder review — does this match how Opnor explains tier selection in scoping calls?
  • Cross-link to the SEA application page once it goes deeper on what HQ requires