Independent Engineer IE

Whose engineer it is

Who pays and who relies are deliberately different parties. The borrower typically bears the cost under the mandate letter or an expenses undertaking, while the engagement letter names the lenders as the parties entitled to rely on the report — and reliance is often extended to hedge providers, tax-equity investors and later purchasers as well. Check the reliance list before you read the conclusions; it tells you who the document was written for.

The contrast with the owner's engineer is structural, not a matter of temperament. An owner's engineer is retained to argue the owner's position and is doing its job when it does. The IE's product is worthless the moment it advocates for anyone, which is why lenders care about prior engagement and why the report reads like an audit rather than a recommendation.

The IE does not seal the design. It reads the sealed set, the studies, the contracts and the model, and it forms an opinion for a lender. Legal responsibility for the design stays with the engineer of record who stamped it, and the IE picking up an error does not transfer that responsibility anywhere.

The report that gates financial close

The technical due diligence report is a condition precedent to financial close, and the first-drawdown conditions usually mirror it.

Its contents are predictable: technology and vendor review covering cell and integrator track record and the certification file; contract review across the EPC, battery supply, long-term service and offtake agreements looking for scope gaps and guarantees that are not back-to-back; capex and opex review against the budget and contingency; the energy production model; permits, interconnection status and schedule; and a risk register with the sponsor's mitigations tested one by one.

The certification file gets specific attention because it is easy to check and expensive to get wrong. The UL 9540 listing must cover the exact model and configuration being bought, the UL 9540A test data must match the hardware revision actually shipping, and the NFPA 855 compliance path — hazard mitigation analysis, spacing, explosion control — must be accepted by the AHJ that will issue the permit. A 9540A report run on a different cell revision than the one on the purchase order is a classic finding, and it is a finding precisely because the setbacks on the site plan depend on that data.

Findings do not simply get noted. An open item becomes a condition precedent, a funded reserve, a contract amendment, or a smaller loan. That is the mechanism by which an engineering observation becomes a commercial outcome.

Testing the curve

A thermal plant's technical case compresses to a heat rate and an availability number, and a week of testing settles both. A battery's case is a retention curve stretched across fifteen to twenty years, conditioned on a duty profile, a temperature window, an SOC window and a C-rate ceiling — while the revenue model dispatches against that curve every hour for the life of the debt. The IE's real question is whether the curve and the dispatch describe the same machine.

The checks are concrete. Benchmark the warranted table and the vendor's expected curve against the supplier's own cycle-test data and third-party degradation models. Confirm the model's energy in every year is usable energy at the POI — degraded, net of round-trip efficiency, auxiliary load and availability — rather than beginning-of-life nameplate.

Confirm the augmentation plan is funded, physically possible in the reserved footprint and DC bus headroom, and consistent with the warranty it is meant to satisfy. Confirm which availability metric the guarantee uses, since time-based and capacity-weighted are different measurements, and since excluded hours removed from the denominator and excluded hours deemed available in the numerator give different answers from the same outage log.

Each haircut lands in the same place. The loan is sized against the base case the IE signs off, not the one the sponsor submitted, so every assumption struck arrives at financial close as a smaller loan and a bigger cheque from equity. The sponsor argues its curve hard for exactly that reason.

Certifying the money

Once the loan is drawn the IE moves from opinion to certification. Construction milestones are not self-declared: the IE inspects, reports periodically to the lenders, and issues the certificate the agent bank needs before it releases each tranche. A sponsor that scheduled equipment payments against an optimistic milestone date discovers the gap here, not in the schedule narrative.

Completion is certified against a named document, and this is where precision matters most. Mechanical completion, substantial completion and the commercial operation date mean what the specific contract says they mean — and the EPC contract, the offtake and the loan agreement frequently define them differently, so a project can satisfy one and fail another on the same day. These are contractual definitions, not statutory ones. The IE certifies to the definition in the financing documents, whatever the EPC contractor has already certified to its own.

The capacity test is where the IE's earlier scepticism gets settled with a number. Because the measured energy at the POI becomes the beginning-of-life baseline for every degradation claim that follows, the IE reads the test conditions as closely as the result: ambient and cell temperature, resting SOC, auxiliary-load treatment, correction factors, and the exact measurement point.

After COD the role continues at lower intensity — periodic operating reports, annual or biennial capacity-test results plotted against the warranted curve, and lender sign-off where the loan requires it for augmentation capex or amendments to material contracts.

What independence actually requires

Independence is not a different letterhead. Lenders test for prior engagement on the same scope, because a firm that drafted the owner's technical requirements or advised the sponsor on vendor selection is reviewing its own work when it reappears as IE. The standard mitigations are disclosure, separate teams and ring-fenced deliverables; on a large financing the answer is often simply a different firm.

The pressure comes from supply. The population of firms holding real operating data across multiple BESS fleets is not large, so the same names turn up on both sides of a deal, especially on smaller projects where one advisory fee has to cover several roles. Treat it as a risk to manage and document rather than a rule to recite — the useful question for a lender is not whether the firm has ever worked for the sponsor, but whether this team has an interest in this project's assumptions surviving review.

One tell is worth remembering. An IE report that arrives with no findings has not tested anything: on an asset whose performance is a twenty-year curve built from vendor models, there is always something to strike out.