What a Structural Assessment for Rooftop Solar Should Contain

Engineering & Design   6 min read

Of the commercial rooftop projects I have seen stall after contract signature, structural findings are the most common cause. Not because the roofs were unsound, but because nobody checked properly until the design was already priced.

Why it is not obvious

Solar adds a relatively light distributed load – 8-25 kg/m² depending on mounting. That sounds trivial against typical design loads. The complication is that it is permanent load added to a structure whose spare capacity was long ago allocated to snow, wind, services, and whatever has accumulated over four decades of occupancy.

Industrial buildings also change. Extra ductwork, additional plant, a mezzanine somebody added in 1998 without updating the drawings. The as-built capacity is frequently not the as-designed capacity.

What the assessment needs to cover

Existing structure verification. Not just drawings – physical confirmation. Member sizes, spans, connection details, and the condition of what is actually there. On older buildings drawings are missing or wrong often enough that measurement is essential.

Current load state. What is already on the roof: HVAC units, ductwork, cable trays, walkways, previous alterations. Each needs locating and weighing, because the new array shares the structure with them.

Combined load cases. The critical check is not solar alone. It is solar plus snow plus wind, in the combinations required by the governing code. In snow regions this is decisive, and it introduces a subtlety many assessments miss.

Snow drifting around the array. Rows of tilted modules obstruct wind-driven snow, which then accumulates in drifts against and between them. Local loads in those drift zones can substantially exceed the uniform design snow load. Both ASCE 7 and EN 1991-1-3 address drifting, but applying them to a solar array requires deliberate thought rather than a uniform load check.

Wind uplift by zone. Corners and edges see far higher suction than the roof field. The assessment must map these zones and confirm both the array attachment and the underlying structure handle them.

Point loads at supports. Ballast blocks and mounting feet concentrate load. A deck adequate for a distributed load may not be adequate for concentrated loads landing between purlins. Support positions frequently need to align with structural members, which constrains the array layout – and therefore the yield.

Findings that commonly change the project

  • Purlin capacity exceeded near the perimeter. Usually solvable by pulling the array in from the edges, at the cost of some capacity.
  • Deck cannot take concentrated ballast loads. Leads to load-spreading trays, a switch to penetrating fixings aligned with purlins, or localised strengthening.
  • Snow drift zones overloaded. Often resolved by increasing row spacing or reducing tilt – again, less capacity.
  • Corrosion at connections. Common on older buildings, particularly in humid or chemically aggressive process environments. Remedial work, sometimes extensive.
  • Existing overload discovered. Uncomfortable but valuable – the building was already carrying more than it should, and solar simply prompted someone to check.

Sequencing it correctly

The assessment should happen before the array layout is finalised, because its conclusions constrain the layout. The frequent pattern – design the array, price it, then commission a structural report – means either redesigning or discovering a cost nobody budgeted.

A reasonable order:

  1. Structural survey and desk study of available drawings.
  2. Preliminary capacity assessment, producing an allowable load map by roof zone.
  3. Array layout designed within those constraints.
  4. Final verification against the specific mounting system and ballast plan.

Who should do it

A structural engineer independent of the EPC, or at minimum one whose report carries professional liability and is addressed to you. An EPC's in-house assessment can be perfectly competent, but the incentive alignment is imperfect – they would like the answer to be yes.

The cost is small relative to the project and trivial relative to discovering the problem after the modules are on the roof.