The Hidden Cost of Late Engineering in Utility-Scale Solar: What the Pro Forma Doesn't Capture

Every discipline has its own version of "measure twice, cut once." Utility-scale solar's version comes with a number attached, and the number is ten.

Manufacturing engineers have tracked this relationship for decades: an error caught at the drawing-board stage costs a rounding error to fix. The same error, caught after the concrete has cured or the racking is torqued down, costs roughly ten times more and the multiplier climbs sharply again if it surfaces after commissioning. The principle is not folklore. It shows up in Six Sigma cost-of-quality studies, in NASA systems engineering literature, and in every EPC change-order log that has ever been audited. Utility-scale solar, with its combination of thin margins, compressed interconnection timelines, and distributed construction sites, feels this multiplier more acutely than almost any other asset class being financed today.

The pro forma does not see this coming. It has a line item for engineering - typically a fixed percentage of total installed cost, scoped once at financial model creation and rarely revisited. It does not have a line item for what happens when that engineering turns out to be wrong after the plant is half-built. This article maps where utility-scale solar projects most commonly absorb that cost, why the exposure is structurally invisible to early-stage financial models, and what an engineering scope that front-loads risk discovery requires.

The Rule of Ten: Why It Applies More Acutely to Solar Than Most Asset Classes

The cost-of-correction curve is not unique to solar. Every capital-intensive industry with a design-then-build sequence experiences some version of it: an error is cheapest to fix the moment it exists only on paper, and it becomes progressively more expensive the further downstream it travels before discovery. What makes utility-scale solar unusually exposed is the combination of four structural features that most other asset classes don't share simultaneously.

First, the design cycle is compressed relative to the physical footprint. A 150 MWac project can move from notice-to-proceed to mechanical completion in nine to fourteen months, which means design, procurement, and construction overlap far more than they would on a comparable-scale industrial facility with a multi-year build schedule. There is less slack in the schedule to absorb a discovered error without touching the critical path.

Second, the site is distributed, not contained. A structural or electrical error on a single-building facility is expensive to fix in one location. The same category of error on a utility-scale solar site - a torque spec, a grounding detail, a combiner box clearance, is frequently repeated across thousands of identical points before anyone notices, because the construction crew is executing the same detail at scale before an inspector or a commissioning engineer catches the pattern.

Third, the commercial structure front-loads commitment. EPC contracts, module and inverter procurement, and often the interconnection agreement is locked well before final design is complete. Correcting a design error after equipment has been ordered doesn't just cost engineering hours, it can mean change orders against fixed-price contracts, restocking fees, or procurement delays that ripple into the interconnection timeline.

Fourth, and most specific to this asset class: the financial model was built on the design that existed at financial close, not the design that gets built. Lenders size debt against a technical package. When the as-built diverges from that package even for defensible engineering reasons it can trigger a re-review, a covenant question, or at minimum an uncomfortable conversation with the independent engineer during a construction draw.

The pro forma has a line item for engineering. It does not have a line item for redesign which is what engineering errors in construction become.

The Five Categories of Late-Stage Engineering Discovery in Utility Solar

Across post-mortem reviews of utility-scale solar construction, the same five categories of late-discovered engineering issues surface repeatedly. None of them are exotic. All of them are preventable at the design stage, and all of them get expensive once construction has started.

Category	Typical Discovery Point	Median Cost Impact	Prevention Scope
Power system study gaps	Equipment submittal review / commissioning	Relay hardware change, interconnection delay	Preliminary study pre-procurement + re-run on as-procured data
Geotechnical / structural	Pile driving / foundation install	Pile redesign, remobilization	Site-specific boring data before final racking design
Protection / interconnection mismatch	Facilities Study reconciliation	Equipment re-order, IA amendment	SLD reconciliation against final Facilities Study before IFC
Constructability conflicts	Field installation	Change orders, schedule slip	Formal constructability review pre-IFC issuance
O&M / design misalignment	Post-COD operations	Elevated O&M cost, downtime	O&M as design-phase stakeholder review

Power system study gaps. Short-circuit, arc-flash, and protection coordination studies scoped too late or scoped against preliminary equipment data that changes before procurement closes routinely surface relay setting conflicts or equipment rating mismatches after switchgear has already been ordered.

Geotechnical and structural surprises. Racking and pile designs based on regional soil assumptions rather than site-specific boring data run into refusal, corrosion, or pull-out capacity issues once piles go into the ground usually discovered by the civil crew, not the engineer of record.

Protection and interconnection mismatches. Single-line diagrams that don't reflect the final utility Facilities Study requirements a different point-of-interconnection transformer impedance, a dual-breaker scheme the design didn't anticipate force a redesign of equipment that may already be on order.

Constructability conflicts. Drawings that are electrically and structurally correct but don't account for how the site is built conduit runs that collide with racking foundations, combiner box placement that ignores maintenance vehicle access generate field modifications that never make it back into the design record.

O&M and design misalignment. Monitoring architecture, spare-parts logic, and access provisions that get value-engineered out of the design package resurface as operational cost the moment the plant is in commercial operation, and the O&M contractor discovers the design didn't plan for their scope.

Each of these categories shares a common signature: the issue existed as a knowable risk at the design stage, and the cost of resolving it grew with every stage it passed through undetected.

Power System Studies: Why Running Them Late Is More Expensive Than Running Them Early

Power system studies, short-circuit, protection coordination, arc-flash, and load flow, are frequently scoped as a late-stage deliverable, run once equipment submittals are in hand, and the single-line diagram is considered final. That sequencing is backwards for utility-scale solar, and it is the single most common source of late-stage engineering cost in this asset class.

The reason is structural, not procedural. Inverter-based resources behave differently under fault conditions than the synchronous generation these protection philosophies were originally built around fault current is limited rather than proportional to equipment rating, which means relay settings calibrated against generic assumptions frequently don't coordinate correctly once real equipment data and real utility fault current contributions are plugged in. When that mismatch is caught during design, it's a settings change. When it's caught during commissioning or worse, during a utility witness test it can mean a hardware swap, a delayed interconnection, or a missed commercial operation date with contractual consequences under the PPA.

Running the preliminary protection coordination study against manufacturer datasheets before equipment is ordered and re-running it against as-procured equipment data before construction of the electrical infrastructure begins, costs a fraction of what a post-installation relay hardware change costs both in dollars and in schedule. The study itself is not expensive. The consequence of skipping the early pass is.

Field Modification Logs: What They Reveal About Design-Phase Gaps

Every utility-scale solar construction site generates a field modification log, the running record of every deviation from issued-for-construction drawings that the field crew had to make to keep work moving. Most of these logs are treated as a closeout formality, filed away once the project reaches substantial completion. Read as a diagnostic instead of an archive, they are one of the most useful documents available for understanding where a design process failed.

Patterns repeat across projects and across EPCs. Conduit routing conflicts with racking foundations because the electrical and structural design packages were developed on parallel tracks without a coordination pass. Combiner box and inverter pad clearances that satisfy code minimums on paper turn out to be unworkable once a maintenance vehicle or a crew with tooling is standing at the equipment. Grounding details that are technically compliant get field-modified because the specified conductor routing wasn't buildable given the as-graded site topography.

None of these are protection or structural failures in the strict sense the design usually would have worked, eventually, on paper. What the field mod log reveals is a coordination gap: disciplines that were designed in isolation and reconciled for the first time by the construction crew, in the field, under schedule pressure. That reconciliation is the constructability review that most engineering scopes never explicitly buy.

The Constructability Review That Isn't in Most Engineering Scopes

A constructability review is a discrete engineering deliverable: a structured pass across the completed design package, before it is issued for construction, specifically to identify where the electrical, structural, and civil disciplines conflict in the field not where any single discipline is wrong on its own terms.

It is frequently absent from engineering scopes for a straightforward reason: it doesn't map cleanly onto any single discipline's deliverable, so it's easy to assume someone else is covering it, or that the EPC's own field engineering team will catch anything material once construction starts. In practice, EPC field engineering catches these conflicts too but after mobilization, when the fix costs a change order instead of a drawing revision.

A constructability review scoped as an explicit deliverable typically two to three weeks of coordinated review across disciplines before IFC issuance costs a small fraction of the engineering budget. Skipping it doesn't eliminate the coordination problem. It just relocates the discovery point from the design office to the construction site and moves the cost from the engineering budget to the change-order log.

How to Structure Engineering Scope to Front-Load Risk Discovery

The traditional utility-scale solar engineering sequence treats risk discovery as something that happens to the project issues surface during construction, get logged, and get resolved under schedule pressure. Front-loading risk discovery means deliberately structuring the engineering scope so that the categories above are addressed at the point where correction is cheapest, not at the point where they happen to surface.

In practice, that means four adjustments to a conventional engineering scope. Run preliminary power system studies against manufacturer datasheets before equipment procurement closes and re-run them against as-procured data before electrical construction begins not after. Require site-specific geotechnical data, not regional soil maps, before finalizing pile and racking design. Scope a formal constructability review as a discrete deliverable between IFC-ready design and IFC issuance, with sign-off from structural, electrical, and civil disciplines jointly, not sequentially. And treat O&M input as a design-phase stakeholder, not a post-COD afterthought monitoring architecture, spare parts logic, and equipment access all belong in the design review, because the cost of omitting them shows up as a support contract, not a change order, but it is real cost, nonetheless.

None of these adjustments require a fundamentally different engineering approach. They require sequencing the existing scope differently and being willing to pay for review work that doesn't produce a stamped drawing but does produce a project that doesn't need one redrawn.

The rule of ten is not a scare tactic. It is a cost curve that every EPC has already lived through, whether it was ever named. The developers who build engineering scope to front-load discovery aren't spending more on engineering they're spending it earlier, where it's cheaper, and where the financial model can still absorb what it finds.

A change order is a design decision made under schedule pressure, by whoever happens to be on site the day the conflict shows up.

To zero re-runs & crazy puns!