The loss nobody puts in the model

Financial models for solar assets in West Africa are careful about irradiance, degradation, availability and tariff. Soiling usually appears as a single conservative percentage, set once at financial close and never revisited. In practice it is the most variable line in the whole yield stack, and the only one that can be reversed in a single working day.

Dust does not damage a module. It sits on the glass and takes light out of the system before the cell ever sees it. Because the loss is gradual and uniform across an array, nothing alarms. Nothing fails. The plant simply produces less, month after month, and the shortfall is quietly absorbed into performance ratio.

Why the dry season changes the maths

Between December and February, the Harmattan carries fine mineral dust south across West Africa. Particle size matters more than volume here: fine dust adheres to glass, resists being blown off, and forms an even layer that scatters light rather than blocking it in patches. It is the worst possible optical outcome for a photovoltaic surface.

It is also the period when there is no rain to help. Operators who rely on wet-season rainfall as a de facto cleaning cycle find that the strategy simply stops working for a quarter of the year — exactly when clear-sky irradiance is at its highest and every lost percentage point is worth the most.

The question is not whether an array is soiled. It is whether anyone has measured how much it is costing this month.

Turning a percentage into a number

Soiling loss becomes a management conversation the moment it is expressed in energy and currency rather than percent. Take installed capacity, a realistic specific yield for the site, the loss you are carrying, and your effective tariff or displaced-diesel cost. The result is a monthly figure that can be set directly against the cost of a cleaning cycle.

For most commercial and utility assets, the arithmetic settles the argument on its own. A cleaning cycle that recovers even a portion of a mid-single-digit loss typically pays for itself well inside a month of restored production. The uncomfortable part is not the cost of cleaning — it is realising how many months went past before anyone ran the numbers.

Setting a cycle you can defend

A fixed calendar cycle is better than nothing, but it is rarely optimal. Sites soil at different rates depending on proximity to unpaved roads, agriculture, construction, coastline and industry. Two plants forty kilometres apart can carry very different losses on the same day.

The defensible approach is to establish your own soiling rate: clean a reference section, compare its output against the rest of the array over several weeks, and derive the daily loss rate for your site. From there the cleaning interval is an optimisation problem with a clean answer, and one you can put in front of a board.

How the cleaning is done matters

Two things damage modules over a plant's life: abrasion of the anti-reflective coating, and load applied to frames and rails by people and equipment moving across the array. Both are introduced by the cleaning method rather than by the dust.

Non-contact washing with deionised water avoids both. Nothing abrades the coating, nothing bears on the structure, and because deionised water leaves no dissolved solids behind, the surface dries without the mineral film that gives the next dust layer something to bond to. On rooftop and carport arrays there is an additional benefit that has nothing to do with yield: nobody has to be up there.

What to do next

If you operate a solar asset in Ghana and cannot state your current soiling loss as a number, that is the first thing to fix. Measure it, price it, and only then decide how often to clean. The cost of the decision is small. The cost of not making it accrues every clear-sky day.

Separating the two activities

High-access cleaning is habitually priced and discussed as one activity. It is two. There is the cleaning — water, chemistry, technique, dwell time. And there is the access — getting a person and their equipment to a surface thirty metres above a public pavement, keeping them there safely, and getting them down.

Once you separate them, the picture changes. Cleaning is the cheap, low-risk part. Access carries the scaffold design, the erection and dismantle programme, the inspection regime, the anchor testing, the rescue plan, the standby cover, the permits, and the insurance loading. It also carries essentially all of the fatality risk in the category.

What the access layer really costs

On a tall or complex elevation, access frequently accounts for the majority of the total price and almost all of the programme duration. A scaffold that takes weeks to erect and dismantle for a wash that takes days is not unusual. During that period the building is wrapped, the ground floor is obstructed, and in many cases a section of pavement is closed under permit.

Rope access reduces the erection burden but replaces it with a different set of obligations: anchor testing and certification, technician certification and currency, supervision ratios, and a rescue capability that must be present for the duration of every working shift.

Removing the access layer does not reduce the risk of working at height. It removes the exposure entirely, because nobody is at height.

Residual risk when the work is flown

Flying the work is not risk-free, and it would be dishonest to present it that way. It introduces a different and much smaller set of exposures: an aircraft above a ground area, water run-off, third-party property beneath the flight path, and airspace obligations.

Each of these is controlled by measures that are well understood and cheap to apply. A physical exclusion zone under the flight path. A spotter alongside the pilot. Weather thresholds with hard abort criteria. Run-off planning agreed before the flight. Regulatory permission obtained before mobilisation. None of these require a person to be suspended above a street.

How to compare the two properly

When you next tender high-access cleaning, ask for access and cleaning to be priced separately. It is a reasonable request and it changes the conversation immediately, because it exposes how much of the quotation is scaffold, supervision and rescue rather than the work you actually want done.

Then ask the second question: what is the consequence, for your organisation, of an incident involving a person working above a public area on your site? For most boards, once that question is asked out loud, the comparison stops being purely commercial.

Where rope access still wins

There are jobs a drone should not take. Work requiring physical contact — sealant replacement, mechanical repair, panel refixing, anything needing hand pressure or torque — needs a person there. Very confined light wells, deep recesses and heavily obstructed elevations can defeat a flight path. Sites where airspace permission genuinely cannot be obtained are simply not available to fly.

The honest position is that drone cleaning removes the largest and most repetitive category of at-height exposure — routine cleaning — and leaves rope access for the work that genuinely needs hands on the building. That is a better allocation of risk than sending someone over the parapet twice a year to wash glass.

Who regulates this

Unmanned aircraft operations in Ghana fall under the Ghana Civil Aviation Authority. Commercial operations — anything flown for hire, reward or business purpose, which includes cleaning and inspection work on your building — are treated differently from recreational flying and carry their own approval requirements.

For the site owner, the practical consequence is straightforward: the obligation to be properly approved sits with the operator, but the consequence of an unapproved flight on your property lands on you as well. That makes verification part of procurement rather than an afterthought.

Location matters as much as the operator

Approval is not only about the company. It is about where and when they intend to fly. Proximity to aerodromes, controlled airspace, restricted or sensitive installations, and operations over people or public areas all change what permission is required for a specific site on a specific day.

In practice this means that a contractor who is properly approved in general may still require site-specific clearance for your particular location. A contractor who says approval is not necessary for your site should be asked to put that in writing.

Five questions to ask before you sign

• Can you show me your current operator approval and its scope?
• Who is the pilot in command for this job, and what is their licence status?
• What site-specific permission does this location require, and has it been obtained?
• What is your liability insurance limit and does it cover UAV operations specifically?
• What are your weather abort criteria and who makes the call to stop?

A competent operator will answer all five without hesitation and will offer the documents before you finish asking. Hesitation on any of them is the finding, not the answer.

What your own team should prepare

Site owners carry a share of the work. Before a flight day, useful preparation includes: confirming who controls the ground area beneath the intended flight path, identifying any external events or deliveries scheduled that day, notifying tenants or occupants, and flagging air intakes, opening windows and any sensitive equipment near the elevation.

Where the site sits inside a larger estate, a port, an airport perimeter or a secured industrial zone, the estate's own permit process usually applies on top of the aviation permission. Starting that conversation early is the single most common way to avoid a wasted mobilisation.

Keeping the record

Ask for the flight log and the permissions to be issued as part of the job pack alongside the before and after imagery. It costs the operator nothing and it means that if anyone asks — an insurer, a landlord, a regulator, a board — the answer exists in your asset file rather than in someone's memory.

Why exterior cleaning stays reactive

In most buildings, exterior cleaning is triggered by an event rather than a schedule: a complaint, an inspection, a VIP visit, or a letting agent's photograph. The result is a lumpy, unplanned cost that is hard to defend at budget time and impossible to compare year on year.

The reason is usually not negligence. It is that nobody owns a number. Interior cleaning has a contract, a frequency and a monthly cost. The exterior has none of those, so it competes for attention against items that do.

Start with exposure, not the calendar

Buildings soil at very different rates depending on where they stand. In Accra, proximity to the coast introduces salt loading. Roadside and junction-adjacent buildings pick up exhaust film and brake dust. Sites near construction or unpaved routes carry a heavier particulate load. Sheltered inland stock in a low-traffic setting can look acceptable for far longer on the same specification of glass.

The practical starting point is a condition survey that records the current state of every elevation and identifies which faces are driving the perception of the building. Very often one or two elevations carry the problem, and a targeted cycle on those faces outperforms an infrequent full wrap of the whole building.

Build the line from three inputs

A defensible annual figure needs three things: the area to be cleaned by elevation, the cycle frequency for each elevation, and the unit approach for the access method. Once those are fixed, the number is arithmetic and can be presented as such.

The frequency conversation is where most of the value sits. Two cycles a year on the two exposed elevations plus one full pass may cost less than a single reactive full clean commissioned in a hurry, and it keeps the building continuously presentable rather than alternating between neglected and freshly done.

A building that is always acceptable costs less than a building that is periodically rescued.

What to include beyond the wash

A cleaning cycle is also the cheapest condition survey you will ever run, because the equipment is already covering every square metre of the elevation. Specifying that condition imagery and findings are returned with each cycle turns a cleaning line into an asset-management input at no additional mobilisation cost.

Over three or four cycles this accumulates into something genuinely valuable: a dated visual record of how the building is ageing, which supports capital planning, dilapidations discussions, insurance conversations and warranty claims.

Presenting it internally

Finance approves recurring lines more readily than surprises. Present the cycle as a fixed annual figure with a defined scope, a condition report attached to each cycle, and a year-on-year comparison. Attach the alternative — the reactive cost, plus the access risk carried each time a crew goes over the parapet — and the decision usually makes itself.

A skyline built in one generation

Walk from Airport City through Ridge and down to the Cantonments edge and you pass more structural glazing than existed in the whole of Ghana twenty-five years ago. Bank headquarters, insurance towers, serviced offices, mixed-use blocks with retail podiums. Almost all of it curtain wall or unitised glazing, almost all of it above the height where a ladder stops being an answer.

The buildings arrived quickly, and the specification arrived with them: imported systems, sealed units, aluminium capping, spandrel panels. What did not arrive at the same pace was a maintenance market sized to look after them. Facility managers inherited elevations designed for a service regime that the local supply chain could not price competitively.

Why the corridor soils the way it does

Accra's glass belt sits in an unusually hostile position for a curtain wall. It is close enough to the coast to take salt-laden air on the sea breeze. It runs alongside some of the busiest arterial roads in the country, which means exhaust film and brake dust. It is surrounded by active construction, which means cement dust and fines. And it takes a full Harmattan season of mineral dust every year.

Those four loads behave differently. Salt is hygroscopic, so it holds moisture against the glass and the gasket line. Exhaust film is oily and binds the rest to the surface. Construction fines are abrasive if wiped dry. Harmattan dust is fine enough to hold as an even veil rather than settle in patches. Together they produce the specific look that a lot of Accra glass has by the end of the dry season: not dirty in spots, but uniformly flat and grey.

The building does not look neglected because of one bad month. It looks neglected because nothing has been budgeted for the elevation since handover.

What owners are actually paying

In practice, three patterns dominate the corridor. Some towers were built with a building maintenance unit on the roof and now cannot use it, because certification, spares or a trained operator are not available. Some rely on rope access teams brought in when a tenant complains or a landlord's inspection is due. Many do nothing above the podium at all, and clean only what a person can reach from the ground.

All three are expensive in different ways. An unused BMU is capital sitting idle plus a recurring inspection liability. Reactive rope work is the most expensive way to buy cleaning, because mobilisation, supervision and rescue provision are charged whether the job is two days or ten. And doing nothing has the slowest, largest cost: staining that sets, gasket degradation that accelerates, and a letting proposition that quietly weakens against newer stock.

The letting argument nobody makes out loud

Grade A tenants in Accra have more choice than they did five years ago. When a prospective tenant is comparing two towers at similar rents, the exterior is the first and cheapest signal they read. It is read before the lift lobby, before the floor plate, before the service charge schedule.

This matters more for older stock than new. A tower that opened in 2012 competes directly with one that opened last year, and the newer building has an unearned advantage for a few years simply because nothing has had time to weather. Cleaning is the only lever that closes that gap quickly, and it is far cheaper than any other intervention on the list.

Building the line item

The corridor's exposure profile is consistent enough that frequency can be reasoned about rather than guessed. Elevations facing the sea breeze and the main roads carry the heaviest load. Sheltered elevations behind other structures carry much less. In most buildings we assess, two elevations drive the entire perception of the tower.

That points to an obvious efficiency: cycle the exposed faces more often than the sheltered ones. A programme of two cycles a year on the driving elevations plus one full pass typically costs less than a single reactive full clean commissioned under pressure, and the building never enters the visibly-neglected phase at all.

What to do this quarter

If you manage an asset in the corridor, three things are worth doing before the next dry season. Establish which elevations actually drive the building's appearance. Get a dated condition record of all of them, so that next year's argument is evidenced rather than remembered. And convert whatever you currently spend reactively into a fixed annual figure with a defined scope.

None of that requires committing to a method. It requires knowing what you own and what it currently costs you, which is a surprisingly rare position for a Ghanaian commercial asset to be in.

Look down on the free zone

From the air, the Tema industrial area and the free zone enclave read as a field of very large rectangles. Warehousing, light manufacturing, cold chain, packaging, logistics. Roof plates of five, ten, twenty thousand square metres each, mostly profiled metal sheet, mostly with translucent skylight banks running in strips, mostly untouched since practical completion.

Ask a plant manager what the roof cost to build and there is an answer. Ask what it costs to maintain and there usually is not one, because it has not been maintained. Not through negligence, but because access to it is a permit exercise that nobody wants to open.

Why nobody goes up there

Walking a large industrial roof in Ghana means confronting several things at once. Profiled sheet and older composite panel are fragile in the spans between purlins. Skylight panels degrade with UV and will not reliably take a person's weight after a few years, which is the single most common cause of serious falls on roofs anywhere in the world. Edge protection is often absent. And on a hot afternoon, the surface temperature makes work unpleasant and judgement worse.

So the honest reason most Tema roofs are dirty is not cost. It is that the access conversation is difficult enough that it never gets started, and the roof falls off the maintenance schedule entirely.

What that actually costs

Three costs accumulate quietly. The first is corrosion. Dust and biological growth hold moisture against coatings, fixings and laps. In a coastal industrial setting with salt in the air, that is exactly the condition under which coated steel starts to go. The damage begins in the places you cannot see from the ground.

The second is drainage. Valleys, gutters and outlets fill with a mixture of dust, vegetation and windblown material. Water then sits where the structure was never designed to hold it, or finds a path into the building. Most roof leaks reported inside a Tema warehouse start as a blocked valley that nobody inspected.

The third is lighting load. Skylight banks are specified to reduce daytime lighting demand across a large floor plate. Once they are heavily soiled, that benefit disappears and artificial lighting runs through the day on a plate of several thousand square metres. On a large site with tariffs where they are, that is not a rounding error.

A roof that has never been cleaned has also never been inspected, because nobody has been close enough to see it.

Cleaning as the cheapest survey you will run

The argument that lands with plant managers is rarely the appearance one. It is that cleaning the roof is the only affordable way to find out what condition it is in. Once the surface is clear, coating breakdown, corrosion at fixings, damaged laps, cracked skylight panels and standing water are all visible and photographable.

That turns an unscheduled expense into a planning input. You come away knowing which roof planes need attention this year, which can wait, and whether the drainage regime is coping. That is a defensible basis for a capital request, which the previous position — nobody has been up there — is not.

Scheduling around production

The other objection is downtime, and here the free zone's operating pattern helps. Flying the roof does not require a shutdown, edge protection, scaffold, or a permit to work at height for a cleaning crew. It requires an exclusion zone on the ground beneath the flight path, coordination with vehicle movements in the yard, and agreement on which shifts to work around.

For most sites that means the line keeps running. For sites with genuinely continuous processes, it means working the roof plane by plane around the yard's rhythm rather than stopping anything.

A reasonable first step

If you run a site in the Tema industrial area or the free zone, the useful first move is not a quotation. It is a survey pass over the roof to establish what is actually up there, followed by a decision about frequency based on what the imagery shows. Most sites we assess need less intervention than they fear and more than they have been doing, which is nothing.

Not one problem, four

Exterior soiling gets discussed as a single thing — the building is dirty. On the Ghanaian coast it is more useful to separate it into four loads, because they arrive on different schedules, attack different components, and respond to different cleaning intervals.

Salt

The sea breeze carries salt aerosol inland across Greater Accra every day. How far it reaches depends on wind, topography and what is in the way, but buildings within a few kilometres of the shoreline take a measurable load, and beachfront hospitality takes a heavy one.

Salt matters because it is hygroscopic: it pulls moisture out of humid air and holds it against whatever it is sitting on. On glass that means a persistent film. On aluminium capping, fixings and gasket lines it means a wet, chloride-rich micro-environment, which is the classic condition for pitting and for accelerated failure of anodised and powder-coated finishes.

What it argues for

Frequency, and rinsing rather than aggressive treatment. Salt is highly soluble, so it comes off readily with a clean water wash. The damage comes from leaving it in place through a long dry period, not from the difficulty of removing it.

Harmattan dust

From roughly December to February the Harmattan brings fine mineral dust south. The particle size is the problem: it is fine enough to form an even veil across a surface rather than settling in visible patches, and fine enough to work into gasket lines, drainage channels and the head of window openings.

On glass it is an optical loss and a cosmetic one. On solar it is a direct yield loss. In drainage channels it combines with the first rains into a paste that blocks weep holes, which is a much more expensive problem than a dirty pane.

Traffic film

Buildings on the arterial roads and around the major interchanges take exhaust particulates and brake dust. This load is oily, and that changes the chemistry: it binds the dust and salt already present into a film that resists a plain rinse and needs the mechanical action of water at the right pressure to lift.

It is also directional. The elevation facing the road takes almost all of it, which is why one face of a building often looks years older than the others.

Biological growth

In the wet season, humidity and shade let algae and mould establish on render, stone, ACM joints and the north-facing elevations that never dry properly. Growth holds water against the substrate, stains porous surfaces, and on some finishes will etch permanently if left across multiple seasons.

Each load has a season. A cleaning cycle set against the calendar rather than the exposure profile will always be in the wrong place at the wrong time.

Reading your own building

You can usually diagnose the dominant load by looking at which elevation is worst. If the sea-facing face is worst, salt dominates and you need frequency. If the road-facing face is worst, traffic film dominates and you need pressure and technique. If the shaded face is worst, growth dominates and you may need targeted chemistry. If all faces are equally flat and grey, dust dominates and you are looking at a Harmattan cycle.

Most Accra buildings show two of these at once, and the mix tells you where the money should go first.

What a sensible year looks like

For a typical coastal commercial building, a defensible cycle is a pass after Harmattan clears — which recovers the appearance and clears the dust out of drainage before the rains — and a second pass on the exposed elevations later in the year. Beachfront hospitality and anything within sight of the water usually needs more.

The specific answer belongs to the specific building. But the framing is transferable: stop asking how often the building should be cleaned, and start asking which load is doing the damage and when it arrives.

The exterior is the first review

Hospitality is the sector where building appearance converts most directly into revenue. A guest forms an opinion in the approach from the road, photographs the building, and — increasingly — that photograph ends up attached to a review. The elevation is doing marketing work whether or not anyone has budgeted for it.

This is sharper on the Ghanaian coast than in most markets, because beachfront and near-beach properties take the heaviest salt load in the country. A hotel two hundred metres from the water ages visibly faster than the same building would inland, and it does so on the face that everyone photographs.

Why it does not get cleaned

General managers know all of this. The obstacle has never been the argument, it has been the intervention. Conventional high-access cleaning at a hotel means a scaffold going up across the frontage, or a cradle running past guest room windows, or rope technicians working above the pool terrace and the arrival court.

Each of those has a real occupancy cost. Rooms on the affected elevation become hard to sell. Terrace and pool areas lose their setting or close. The arrival experience runs past a hoarding. Events cannot be sold into the affected weeks. So the work slides to the low season, and then to the following low season, and the building keeps ageing on the face that matters most.

The cleaning was never the disruption. The access was.

What changes when the work is flown

Removing scaffold and cradle from the equation changes the scheduling problem rather than solving the cleaning problem, and for a hotel that is exactly the right change. There is nothing standing on the frontage for weeks. There is no cradle passing guest windows. The arrival court and the terrace are affected only for the period the exclusion zone is actually live beneath the flight path.

In practice that means an elevation can often be worked in early morning sessions before the terrace fills, elevation by elevation, with the exclusion zone moving as the work moves. Rooms are not taken out of inventory. The pool does not close for the season.

What the property needs to prepare

Three things make the difference between a smooth mobilisation and a difficult one. First, a decision about window openings and balcony doors on the working elevation for the session — guests need notice, and housekeeping needs a checklist. Second, air intakes and any external kitchen extract identified in advance. Third, agreement on the ground area: which part of the terrace, car park or arrival court sits under the flight path, and who manages it.

None of that is unusual for a hotel operations team. It is a smaller ask than the pre-mobilisation meeting for a scaffold.

Cycles, not projects

The properties that stay ahead of the salt do not commission a big clean every few years. They run a cycle, usually weighted toward the sea-facing and road-facing elevations, timed around their own occupancy pattern rather than a contractor's calendar.

That converts a lumpy capital-feeling expense into a predictable operating line, and it means the building is never in the state where a full rescue job is required. For a property whose exterior is part of the product, that is the whole argument.