Commercial Solar Installation

How a commercial rooftop project actually runs

If you manage a building — as an owner, FD, operations lead or facilities manager — and you are trying to understand what putting solar on it actually involves, this page walks the process end to end: the stages, the sequence, who does what, and which item quietly sets the whole timetable. A commercial rooftop project of 30–250 kWp is a construction project, not a large domestic install: it carries CDM 2015 duties, method statements, controlled deliveries and a programme measured in weeks — typically £30,000–£250,000 of capital paying back in around six years on a daytime-heavy load.

One scope note before the detail. This page explains the process. For sector-by-sector depth — factories, warehouses, schools, farms, and the procurement questions specific to each — our specialist commercial solar installation partners cover the commercial side in full, and that is the right resource if you are comparing installers rather than understanding the work. What follows is the programme itself, stage by stage.

Stage by stage: feasibility to energisation

A well-run project moves through seven stages. The order is not decorative — several stages exist precisely to de-risk the ones after them.

Stage 1: desk feasibility from your meter data

Serious design starts with twelve months of half-hourly meter data, not a satellite photo of your roof. Half-hourly data shows when your building actually uses electricity — the daytime base load, the seasonal shape, the weekend troughs — and the economic system is sized to that consumption pattern, exporting little. Roof area is a constraint, not a target: filling the roof beyond what the building can absorb produces exported units at a fraction of the value of self-consumed ones and stretches the payback. A desk feasibility from real data, before anyone visits, establishes whether the project stacks up and roughly at what size.

Stage 2: the roof survey — structure, covering, asbestos

The site survey is dominated by the roof, and it asks three questions. Structure: can the purlins and frame carry the array — panels, mounting and (on flat membranes) ballast — with a structural engineer’s confirmation, not an assumption? Covering: what condition is the roof covering in, and what is its remaining life? Installing 25-year panels on a 10-year-old failing roof is the classic false economy — the honest recommendation is sometimes “recover the roof first”, and a process that surfaces that now saves a six-figure mistake. And on any building constructed before 2000: the asbestos register is checked before survey work proceeds, as the Control of Asbestos Regulations 2012 requires — asbestos-cement sheet roofs change the method entirely or rule the roof out.

The survey also covers the electrical infrastructure — the supply capacity, switchgear, and where the array will connect — and access: where scaffold, edge protection and deliveries will go without stopping your operation.

Stage 3: the G99 application — submit it early

Any system at this scale needs a G99 connection application to the District Network Operator, approved before energisation, and this stage belongs immediately after feasibility because it sets the timetable more than the panels do — the next section explains why. A competent process files the G99 with the design still at scheme stage, precisely so the DNO’s response arrives in parallel with detailed design rather than after it.

Stage 4: detailed design and the CDM framework

Detailed design fixes the layout (respecting roof zones, walkways, rooflights and plant), the string design, inverter positions, cable routes and protection — and, under CDM 2015, the safety architecture of the build: who acts as principal designer and principal contractor, the construction phase plan, and the risk assessments and method statements (RAMS) the crews will work to. On a commercial project these are legal duties with named holders, and the client holds duties too — a professional installer will tell you yours rather than letting you discover them.

Stage 5: procurement and programme

Panels, mounting, inverters and switchgear are ordered against the approved design, and the programme is fixed: scaffold and access dates, delivery slots, crane or telehandler bookings, and the sequence of roof zones so your operation keeps running underneath. Lead times at commercial scale run weeks, which is another reason the G99 filed at stage 3 rarely delays a well-ordered project — the approval and the kit tend to arrive in the same window.

Stage 6: the installation weeks

On site, the pattern of a controlled build: inductions and RAMS briefings, roof access established, banksman-controlled deliveries, materials distributed to the roof in planned lifts. Mounting first — fixed to purlins through the covering with sealed, engineered fixings on metal roofs, or ballasted frames on membranes — then panels zone by zone, DC strings made off and tested as they complete. Electricians run the DC to inverter positions, build the AC side through new switchgear to the connection point, and test progressively rather than saving everything for the end. The roof work on a 30–250 kWp project typically runs two to six weeks, inside the longer programme the paperwork set.

Stage 7: commissioning, witness and handover

Commissioning at this scale is procedural: full test documentation to BS 7671, inverter and protection-settings configuration to the G99 agreement — the DNO may witness-test protection settings before granting final permission to energise — then metering, monitoring and export registration. Handover should include the as-built drawings, test certificates, structural and CDM documentation, the O&M (operation and maintenance) manual, warranty registrations and the monitoring platform, with your team trained on what normal generation looks like. A project without a proper handover pack is not finished, whatever the roof looks like.

Why the G99 application sets the timetable

The single most useful thing to understand about the process: the DNO, not the installer, usually controls the critical path. A G99 application at commercial scale triggers a network study — can the local network absorb your export, does protection need specific settings, is reinforcement needed — and the study plus connection agreement can run eight weeks to many months depending on the network’s headroom in your area. The panels can be on the roof in a fortnight; the right to switch them on is what takes the time.

This is why “when do you file the G99?” is the sharpest question to ask any commercial quoter. Filed at feasibility, the DNO clock runs concurrently with design and procurement and often costs the programme nothing. Filed after installation, the same clock runs with a completed system sitting dead on the roof — capital spent, generating nothing, possibly facing an export limit the design never anticipated. The DNO can also attach conditions: export limitation, protection requirements, occasionally reinforcement costs. Discovering those at feasibility changes the design; discovering them after installation changes the business case.

Sizing from half-hourly data, not roof area

Worth expanding, because it is the most common commercial design error. The value of a solar unit depends on who consumes it: used on site, it displaces your import price; exported, it earns a fraction of that. So the economic optimum sizes the array to the building’s daytime base load — the consumption floor the building never drops below during generating hours — rather than to the roof’s capacity. A distribution unit drawing 60 kW through every working day supports a very different system from a two-shift factory or a weekend-quiet office with the same roof.

Half-hourly data answers this precisely, which is why a process that starts with your meter data is credible and one that starts with your roof dimensions is selling panels. Typical outcomes at this scale: 30–250 kWp generating 27,000–230,000 kWh a year, displacing 6–53 tonnes of CO2 annually, with paybacks around six years where self-consumption is high. East-west layouts frequently beat south-facing on flat commercial roofs — flatter daily output matches operating hours better and packs more capacity within ballast limits.

On the financial side, note one correction to a claim that circulates widely: solar is special-rate expenditure for capital-allowance purposes, so the correct 100% year-one route is the Annual Investment Allowance (up to £1m), not “full expensing” — a distinction your accountant should confirm for your structure. The grants and funding routes page covers the detail.

Planning, CDM and the compliance stack

Non-domestic rooftop solar in England is Class J permitted development — panels up to 1 m above the roof plane, with conditions — so most commercial rooftop projects proceed without a planning application; listed buildings and some designated settings are the exceptions, and the England-only caveat matters: Scotland and Wales run their own regimes, so multi-site operators should check per nation.

The rest of the stack, in one list: CDM 2015 (named duty-holders, construction phase plan, RAMS); BS 7671 with full test certification; the G99 connection agreement; structural sign-off on the loaded roof; the Work at Height Regulations 2005 governing access and edge protection; and the Control of Asbestos Regulations 2012 wherever the building predates 2000. None of this is optional dressing — it is the difference between a construction project and an oversized domestic job done in the wrong setting, and the documentation should all reach your handover pack.

A programme scenario: 45 kWp on a distribution unit

A modelled scenario, not a client project. A distribution business with a single-ply membrane roof and a daytime-heavy load commissions a 45 kWp array — 102 panels in a ballasted east-west layout, chosen because the wind-uplift calculation per roof zone would not accommodate south-facing rows within the ballast the deck could carry, while the east-west profile put 96% of a theoretical south array’s yield on the same roof. The structural engineer confirms the dead load; the installer acts as principal contractor under CDM 2015.

The programme runs fourteen weeks end to end. The G99 application, filed in week one alongside detailed design, returns a connection agreement in week three. Procurement lands the kit by week eight; the roof installation itself takes eight days, with banksman-controlled deliveries and the depot operating throughout; commissioning, witness testing and handover close the programme. Modelled generation is 40,500 kWh a year at roughly 85% self-consumption on the daytime load, worth £9,500–£11,500 annually at current prices, for a payback around 5.5 years. The design decision — east-west within the ballast budget — was the project; the panels were the least of it.

Process questions

How long does the whole process take?

Typically three to six months from instruction to energisation for a 30–250 kWp rooftop project: feasibility and survey inside a fortnight, the G99 study eight weeks to several months depending on your network area, procurement running in parallel, then two to six weeks on the roof and a commissioning close-out. The DNO study is usually the critical path — which is why it is filed first, not last.

Will the installation stop my operation?

A properly-planned project should not. Deliveries are banksman-controlled and scheduled around your operation, roof zones are sequenced so work stays clear of critical areas, and the electrical connection — the one genuine interruption — is a planned, agreed shutdown, typically brief and often out of hours. The construction phase plan should state all of this in writing before anyone mobilises.

What does a commercial rooftop system cost?

£30,000–£250,000 covers the typical 30–250 kWp range, with per-kW pricing falling as scale rises. The programme items domestic buyers never see — structural engineering, CDM documentation, access and edge protection, switchgear — are real lines, and their absence from a cheap quote is a warning, not a saving. See the cost guide for the breakdown, and note the Annual Investment Allowance position above.

Who handles the G99 application?

The installer should — preparing, submitting and managing it with the District Network Operator, and building the DNO’s response into the design. Your sharpest procurement question is when they file it: the credible answer is at feasibility or scheme-design stage, concurrent with everything else. “After installation” is the wrong answer, and it costs months with a dead system on the roof.

Can panels go on our asbestos-cement roof?

No — drilling asbestos cement releases fibres and the sheets cannot safely carry point loads, so fixing through them is not acceptable practice. The honest options are replacing or overcladding the roof first (often the right moment anyway on a covering that old), or looking at other mounting locations. Any pre-2000 building’s asbestos register is checked before survey — a quoter who has not asked for it is skipping steps.

Related pages

Much of the engineering on commercial membrane roofs is shared with the flat roof solar installation page — ballast, wind zoning and dead load in more depth. Commercial storage increasingly enters the same programme, and the coupling and grid-paperwork logic on the solar battery installation page scales up directly. For numbers see typical costs and payback, for the capital-allowances position see grants and funding routes, or request a quote — and ask every quoter when they file the G99.