When the grid goes down, can the building keep working?

Journey to Resilience, Part 1: Building-integrated PV

Solstrides Projects · Büşra Yılmaz-Depondt

On 28 April 2025, shortly after half past twelve, Spain and Portugal lost power. It was a bright spring day and solar was supplying more than half of Spain's electricity, yet within seconds almost all of it had gone, together with everything else on the grid.

Hospitals switched to diesel generators and postponed whatever could wait, while across both countries rooftop solar systems sat in full sun producing nothing, because that is exactly what they are designed to do when the grid disappears. A few buildings kept going, though. The University of Almería kept its lights and air conditioning running on solar panels and batteries, and three months later Portugal announced €25 million to help hospitals and other critical infrastructure install their own solar and storage.

I keep coming back to that afternoon because it showed two things at once: solar on buildings can keep essential services running, but solar on buildings, as most of it is designed today, does not.

That gap is what this series is about. Journey to Resilience is a set of five articles on how integrated, decentralised solar can help clinics, schools and homes keep working when the grid fails, and on what that really takes. I am starting with buildings because that is where the services are, and because I have spent most of my career on the kind of solar that becomes part of the building itself.

A module that does a wall's job

Building-integrated photovoltaics, or BIPV, are solar modules that form part of a building rather than sitting on top of it. The standards (EN 50583 in Europe and IEC 63092 internationally) use a test I like for its simplicity: if you removed the module, would you have to put something else in its place to keep the building working as intended? If the answer is yes, it is BIPV; if the roof underneath would still be a perfectly good roof, you are looking at building-applied PV, the familiar panels mounted on a rack.

That might sound like a technicality, but it shapes everything about these products. Because a BIPV module is also a roof tile, a pane of glass or a piece of cladding, it has to meet two sets of requirements at the same time, one for electrical safety and performance and another for everything a building product must do, from fire safety and wind loads to impact resistance and keeping out the weather. Anyone who has taken a product through both processes knows how different those two worlds are, and how rarely the people in them talk to each other.

The figure below shows how far the idea goes, from roof tiles and metal roofs to curtain walls, skylights, greenhouse windows, canopies, balconies and balustrades.

Figure: Colouring technologies for BIPV and the building elements a PV module can replace, with the functions it takes over.

Each of these elements keeps the job it always had, whether that is keeping out rain and noise, shading a room or stopping someone from falling off a balcony, and generating electricity is simply added to the list.

Then there is appearance, which is where I spent years of my working life. A façade on a historic street has to look as if it belongs there, and architects will not specify a material they find ugly, however much electricity it produces. The industry has found plenty of ways around this, from coloured cells, encapsulants and foils to glass with interference coatings, printed glass and even photographic images stretched across an entire façade.

Most of these approaches cost some efficiency, and how much depends on the technology, the size of the module, the opacity of the colour layer and the shade itself, so a pale terracotta will generally produce less than a dark grey. I have stopped seeing that as a weakness, because a façade that gets built and produces something is worth more than a more efficient one that never gets past the planning committee.

Is it worth the money?

Measured purely in cost per kilowatt-hour, BIPV usually loses to a standard rooftop array, since façades are vertical and often shaded and integrated products are made in smaller production runs. If cheap electricity were the only goal, I would tell most people to put standard panels on the roof and leave it there.

But a building owner rarely chooses between a BIPV façade and nothing; the real choice is between a BIPV façade and a conventional one made of stone, glass or metal. Researchers at SUPSI and the Becquerel Institute assess BIPV on exactly that basis, looking at the additional cost compared with the material it replaces rather than the full price of the solar product. Seen that way, the numbers can look very different. An IEA PVPS review of BIPV projects found two large solar façades where the additional cost came to just 1–2% of the cost of the whole building, and although those are individual projects rather than a rule, they show how misleading it can be to price the solar part on its own.

Timing matters too. Measurements at ZSW in Stuttgart showed that a solar façade produces more of its electricity in winter and in the mornings and evenings, when the sun is low, while a roof array peaks around midday in summer, so the two complement each other well. Using data from its own building, ZSW modelled a standard office building and found that roof and façade together could cover nearly 40% of its electricity demand over a year, without any battery storage.

So BIPV makes sense when a building is being built or renovated anyway, when the roof is small or already crowded with equipment, when an old housing block needs a new outer skin, or when heritage rules make a conventional array hard to install. It makes much less sense when there is a large, empty, sunny roof and no renovation planned, or when the façade is in shade for most of the day.

For communities working with tight budgets, one more question matters even more: would the same money buy more continuity if it went into a battery and a backup inverter instead? A beautiful solar façade on a clinic that still goes dark in a blackout has not made anyone more resilient.

Europe is about to make many of these decisions at once. Under the recast Energy Performance of Buildings Directive, new public and non-residential buildings over 250 m² will need suitable solar installations from the end of 2026, where this is technically suitable and economically and functionally feasible. Existing public buildings follow in stages up to 2030, depending on their size, and new homes are included from the end of 2029. Thousands of buildings will consider solar over the next few years, and whether those systems are designed around the building, and around the day the grid fails, is still an open question.

Why the panels went quiet in Spain

Back to that afternoon in April. Ordinary grid-connected solar systems switch off when the grid fails, even in bright sunshine, because of a safety function called anti-islanding, which stops an inverter from energising a part of the network that should be dead and so protects the people working to restore it. It is an important rule, but it also means that rooftop solar on its own gives a building no protection in a blackout.

What made the difference at Almería was everything around the panels. To keep running, a building has to disconnect safely from the network and operate as its own small island, which takes an inverter that can form and hold a local grid, a switch to separate the building from the public network, and batteries to carry the load once the sun goes down. It also takes a step people often skip, which is deciding which circuits actually matter: the vaccine fridge, the emergency lighting and the water pump, rather than the whole building at once.

This is where BIPV meets decentralised energy. A solar façade adds generation on surfaces that would otherwise go unused, right where the essential loads are, and with storage, the right controls and a clear answer to “what must stay on?” it becomes part of a building that can keep serving people when the network cannot. Because BIPV is designed together with the building, often years before it is installed, that is also the cheapest moment to prepare for backup operation; doing it later can mean opening walls, rewiring and replacing equipment. I will come back to batteries and microgrids in Part 4.

What the models don't see

There is another reason these options get overlooked, and it sits far from any building site.

The energy models behind national strategies and long-term scenarios have to simplify a very complicated system, and solar is often one of the first things to be simplified, with large solar farms and PV built into buildings treated as a single technology.

When that happens, a solar façade may be charged the full price of a construction product and credited nothing for the cladding it replaces, and its winter and evening output disappears into an average solar curve. Keeping a clinic open during a blackout earns it nothing either, because many models give no explicit value to avoiding an outage in the first place.

Models have to simplify, and that is fine. The problem is what falls out when they do, because options for buildings, and for the people who rely on them, can disappear before anyone has properly compared them. That is where this series will end, in Part 5.

One question worth asking

If you own, design or renovate public buildings, the new solar requirements will reach your desk soon, if they have not already. Before anyone specifies the panels, ask one question: which loads in this building must keep running when the grid goes down? Then choose the inverter, the wiring and the space for a battery with that answer in mind, even if the battery itself comes later, because designing a building to be ready for a blackout costs far less than rebuilding its electrical installation after the first long one.

Next time I will leave the city for the countryside. Part 2 looks at agrivoltaics, where solar and farming share the same land, and at what that could mean for rural communities that are often last in line for reliable power.

The views expressed are those of Solstrides Projects and do not represent the positions of ETIP PV, IEA PVPS or any of its Tasks.

Sources

1. Wikipedia, 2025 Iberian Peninsula blackout. Timeline, solar share and hospital generators.

2. SolarPower Europe, Statement: The Iberian power outage, 30 April 2025. University of Almería example.

3. Government of Portugal, Portugal is now better prepared for a blackout, 28 July 2025. The €25 million support measure.

4. IEA PVPS Task 15, International definitions of BIPV. Definitions and standards.

5. IEC 63092-1:2020, Photovoltaics in buildings – Part 1: Requirements for building-integrated photovoltaic modules.

6. SUPSI and Becquerel Institute, Building Integrated Photovoltaics: A practical handbook for solar buildings' stakeholders, BIPV Status Report 2020.

7. IEA PVPS Task 15, Inventory on existing business models, opportunities and issues for BIPV, 2018.

8. ZSW, Greener office buildings with power from façade and rooftop photovoltaic systems, 24 February 2021.

9. Directive (EU) 2024/1275, Energy Performance of Buildings Directive, Article 10.

10. IEC 62116:2014, Utility-interconnected photovoltaic inverters – Test procedure of islanding prevention measures.