Solar for Hospitals and Clinics: Designing Power Systems Where Failure Costs Lives

In most buildings, a power cut is an inconvenience. In a health facility it is a clinical event.

An outage during a caesarean section is a surgical team working by phone torch. An outage in a vaccine store is a cold chain excursion that may have quietly destroyed the entire stock, and nobody will know until the doses fail to protect anyone. An outage in a ward with oxygen concentrators is a patient whose supply stops. These are not hypothetical scenarios. They are routine across a large part of the continent.

The scale of the problem is documented and it is worse than most people assume. The World Health Organization reports that roughly 15% of health-care facilities in sub-Saharan Africa have no access to any electricity supply at all, and that only about half of hospitals in the region report reliable access. Globally, close to one billion people are served by health facilities with no electricity or with a supply that cannot be depended upon.

Solar is the obvious answer, and it is being deployed at scale. But solar for hospitals and clinics is a harder engineering problem than solar for an office, and health facilities are littered with installations that worked for a season. This article covers how to design one that does not.

Why Health Facilities Are a Different Design Problem

Three characteristics separate a health facility from every other building type, and each of them breaks a normal solar design assumption.

The load runs around the clock and does not follow the sun. A commercial building consumes most of its energy in daylight, which is why solar suits it so well. A hospital's cold chain runs all night. Maternity and emergency admissions do not schedule themselves for solar noon. Inpatient wards need lighting and fans through the night. The result is a load profile with a substantial overnight component, which means storage is not optional and cannot be trimmed to save budget.

Failure has consequences that are not financial. In a factory, an outage means lost production, which is a number. In a health facility, an outage during a procedure is a different category of event entirely. This changes the design philosophy: the correct question is not "what does this system deliver on an average day" but "what happens on the worst day, and what survives when something fails".

The maintenance environment is unusually hostile. Many facilities are remote, staffed by clinical personnel with no technical training and no budget line for maintenance, and part of a public system where a fault report may take months to become a site visit. A design that assumes competent local maintenance will not survive. A design that assumes none, and is built to be robust and simple, has a chance.

Start With a Critical Load Hierarchy, Not a Building Total

The most consequential decision in health facility solar design is made in the first hour, and it is not about panels. It is about deciding what must never lose power, what should stay on, and what can be shed.

Tier one, the never-fail loads. Vaccine and blood cold chain. Theatre lighting and essential theatre equipment. Oxygen concentrators and suction. Emergency and corridor lighting. Communications and, in facilities that depend on it, the sterilisation equipment schedule. These loads justify the full engineering treatment: storage autonomy, redundancy, and monitoring that alerts somebody when it fails.

Tier two, the important loads. Ward lighting and fans, laboratory equipment, the pharmacy, water pumping, general administration and computing. These should stay on in normal circumstances but can be shed in a genuine emergency without immediate clinical harm.

Tier three, the deferrable loads. Staff accommodation, non-clinical air conditioning, laundry, external lighting and general convenience power. These can be shed automatically when the battery reaches a threshold, and shedding them is what protects tier one.

This hierarchy is not a document. It should be wired. Tier one loads belong on a separate, clearly identified essential distribution board fed from the protected supply, physically distinct from general circuits. Without that separation, the automatic load management has nothing to act on, and the first time someone plugs a kettle into a theatre socket, the hierarchy exists only on paper.

The physical separation matters for another reason. Health facilities grow. Somebody will add a load in three years, and the only thing preventing them from putting it on the essential board is that the essential board is labelled, locked and obviously different.

Planning a solarisation programme across multiple facilities? Our renewable energy engineering team designs to critical load hierarchies rather than to building totals.

The critical load hierarchy that should be physically wired into a health facility, not merely documented. Tier one belongs on its own essential distribution board. Source: MIMAH engineering design practice.

Sizing for Autonomy, Not for Average Days

Commercial solar design optimises for financial return, which means sizing storage to the economic optimum and accepting occasional shortfalls. Health facility design optimises for the bad day.

Autonomy is the number of days the system can carry its critical load with no meaningful solar input. For a commercial site, one day is often sufficient. For a health facility, the design target is typically two to three days of tier-one load, because the situations that knock out solar generation, meaning heavy cloud, dust storms, or damage, tend to be the same situations that make a site hard to reach.

That autonomy figure is calculated on the tier-one load alone, not the whole facility. This is what makes the number affordable. A facility might draw 40 kWh across a full day but need only 8 kWh to hold the cold chain, essential lighting and oxygen through the night. Sizing three days of autonomy on 8 kWh is a manageable battery. Sizing it on 40 kWh is a proposal nobody will fund.

Depth of discharge and temperature then determine what nameplate capacity actually delivers. A battery's usable energy is not its rating, and its life depends heavily on the temperature it lives at. This matters acutely in health facilities because battery rooms are frequently the worst room in the building: a windowless store, often under an uninsulated metal roof, sometimes chosen precisely because it was the only lockable space available. Sustained high temperature is the fastest way to destroy an expensive battery bank, and it voids warranties.

Generator interaction needs deciding explicitly. Most health facilities already have a diesel generator. The right approach is almost never to remove it, and almost never to leave it as it was. It should become the third line of defence behind solar and storage, starting automatically only when the battery reaches a defined reserve, and it should be sized and controlled so that it can also recharge the battery during extended poor weather. Our guide to hybrid solar systems covers the dispatch logic in detail.

The Cold Chain Deserves Its Own Design

Vaccine refrigeration is the load where solar most obviously earns its place, and where the engineering is most often done badly.

A vaccine refrigerator has a narrow temperature band and a limited tolerance for excursions. Its compressor draws a starting current several times its running current, which is exactly the load characteristic that trips undersized inverters. And it does its most demanding work at night and in the hottest part of the afternoon, precisely when a poorly designed solar system is weakest.

Several design decisions follow. The inverter must be sized on starting surge, not running load, and if several refrigeration units share a supply their compressors should not be free to start simultaneously. Cold chain equipment should sit on the tier-one board with autonomy calculated specifically for it. Temperature monitoring with alarming should be independent of the power system, so that a failure is detected by something that does not depend on the thing that failed.

There is also a strong argument for solar direct-drive refrigeration in the smallest and most remote facilities. These units store energy thermally, in ice or phase-change material, rather than electrically in a battery. They remove the battery from the critical path entirely, which removes the component most likely to fail and most expensive to replace in a facility with no maintenance budget. They are not right everywhere, but where they fit they solve the durability problem elegantly.

Water is the load that gets forgotten. A health facility without running water cannot maintain hygiene, and in many rural facilities water comes from a borehole with an electric pump. That pump belongs in the load hierarchy, and the same solar pumping engineering discussed in our solar water pumping guide applies directly, including the point that a storage tank is cheaper and more robust than extra battery capacity.

The scale of the health facility electrification gap. Sources: Electricity in Health-Care Facilities, World Health Organization, 2023; WHO news release, January 2023.

What the World Health Organization's Own Assessment Says

The WHO and World Bank's Energizing Health report is worth reading in full by anyone specifying these systems, because its central finding is not about technology. It is about what happens after commissioning.

The recurring failure in health facility electrification is not that the wrong equipment was installed. It is that nobody was funded to maintain it. Programmes are frequently structured as capital projects: a donor funds installation, a contractor installs, a ribbon is cut, and operational responsibility passes to a facility with no technical staff and no maintenance budget. The system then degrades on a predictable curve. Panels are not cleaned, output falls, staff assume the system is failing, and when a battery reaches end of life there is no mechanism to replace it. Within a few years the facility is back on diesel or back on nothing.

This is an institutional failure, not an engineering one, but engineering choices can mitigate it substantially.

Design for the maintenance that will actually happen, not the maintenance specified. Simpler systems with fewer failure modes outlast sophisticated ones in unsupported environments.

Standardise across the programme. A hundred facilities with the same inverter, the same battery and the same array configuration can be supported by one spares holding and one trained team. A hundred facilities with whatever was cheapest at each tender cannot be supported at all.

Instrument remotely. A monitoring link that reports output and battery state to a central team converts an invisible degradation into a ticket. Without it, nobody knows a site is failing until a clinician reports that the fridge is warm.

Fund operations from the start. The most useful thing a programme can do is contract multi-year maintenance as part of the original scope, so that cleaning, inspection and eventual battery replacement have a funding route that does not require a new approval.

The failure modes here are the same ones we catalogue in why solar systems fail early, amplified by the fact that health facilities are usually the least able to absorb them.

What Delivery at Scale Looks Like

Between 2024 and 2025 MIMAH delivered health facility solarisation across Sudan under a UNDP programme: 110 sites totalling 1.39 MW, commissioned in 2025, with 30 sites in Elgadarif, 22 in Kassala, 6 in the Red Sea state and the balance distributed across other states. The work was executed during active conflict, and every site remains under MIMAH operations and maintenance. A further 13 cold-chain solar sites for NMSF are in progress.

The lessons from that programme are mostly logistical rather than technical. Standardised packages by facility size make spares viable. Local operator training at each site, rather than a central team expected to travel, is what keeps arrays clean. Structured reporting from every site is what makes an O&M obligation across a hundred facilities administrable at all, because manual paperwork simply does not scale past a few dozen sites.

And the single most important design decision, repeatedly, was the critical load hierarchy. Facilities where tier-one loads were properly separated onto their own board rode through problems that took down general power. Facilities where everything shared a board did not.

Want to see what health facility solarisation looks like in practice? Our project portfolio and impact record document the programme in detail.

Why autonomy is sized on the tier-one load alone: the same three-day target becomes an affordable battery instead of an unfundable one. Source: MIMAH worked example.

Frequently Asked Questions

How large a solar system does a rural clinic need? It depends entirely on the load, but the sizing should start from the critical load rather than the building. A small clinic with vaccine refrigeration, lighting, a few small devices and a water pump often needs a system in the low single-digit kilowatts with storage sized for two to three days of tier-one autonomy. A district hospital with theatres, laboratory equipment, imaging and inpatient wards is a different order of problem and needs a full load study.

Should the existing generator be removed? No. It should be repositioned as the last line of defence, starting automatically when the battery reaches its reserve threshold, and configured so it can recharge the battery during extended poor weather. Removing a working generator from a facility where power failure has clinical consequences removes a layer of protection for no engineering gain.

What is the biggest cause of health facility solar systems failing? Not equipment. It is the absence of a funded maintenance route after handover, compounded by battery installations in hot, unventilated rooms. Soiling reduces output invisibly, batteries degrade faster than expected at high temperature, and with no monitoring and no budget the decline is not noticed until the system stops.

Can solar run an operating theatre? Yes, provided the theatre load is properly characterised, including the starting behaviour of every motor-driven device, and provided the theatre sits on a tier-one board with adequate autonomy and a generator behind it. What is not acceptable is treating a theatre as an ordinary load on a general circuit.

How much autonomy should we design for? Two to three days of tier-one load is a reasonable target for facilities that are remote or hard to reach, calculated on critical loads only. Facilities in urban areas with reliable generator fuel supply and quick maintenance access can justify less. The figure should be an explicit design decision with a stated rationale, not a residual of the budget.

Is remote monitoring worth the cost on small sites? On a programme of any size, yes. The alternative is discovering failures through clinical incident reports. A simple monitoring link that reports daily yield and battery state converts silent degradation into a maintenance ticket, and it is what makes a multi-site O&M obligation deliverable.

Reliability Is the Deliverable

The temptation in health facility electrification is to measure success in installed kilowatts, because kilowatts are easy to count and easy to report. They are also close to meaningless. A megawatt of installed capacity across a hundred sites, half of which are underperforming and nobody knows which half, has not solved the problem it was funded to solve.

What matters is whether the fridge was cold every hour of every day for the last five years, whether the theatre lights stayed on during the caesarean at two in the morning, and whether somebody will be there to replace the battery when it reaches end of life.

That is an engineering problem, a logistics problem and an institutional problem at the same time, and it is only solved by treating it as all three: critical loads separated and sized properly, standardised equipment that can be supported with one spares holding, monitoring that makes failure visible, and a funded maintenance route that outlives the capital project.

Planning a health facility solar project or reviewing one that is not performing? Talk to our engineering team. We will start with the critical load hierarchy and the autonomy requirement, and be straightforward about what the site needs to stay reliable in year five, not just year one.