One building, one meter, one pergola: efficiency and emergency

In articles about collective self-consumption we always end up saying the same thing: the internal grid wins. It's easy to say and hard to prove with numbers. So we took a specific building, modelled it hour by hour for a full year, and worked out what happens to the bill.

The result holds a surprise. The piece that saves the most money isn't the panels. And the battery, in this specific case, contributes much less than we expected — for a reason that's worth understanding before signing any quote.

The article comes in two parts. The first is about money: what the building costs today, what it will cost, and when the investment pays back. The second is about something else: what happens if the power goes out for three days and someone who can't walk down the stairs lives on the third floor. The two arrive at batteries of different sizes, and explaining why is probably the most useful part of all.

---

Part one: the money

The case

A building of 10 homes on the central Catalan coast. Each home consumes around 3,000 kWh a year and has 4 kW of contracted power. The car park has a 150 m² pergola that's already built —it's been shading the spaces for years— half facing east and half west, with a 15° tilt. It's the typical geometry of a car park roof: it isn't oriented south, but follows the parking layout. The fact that the structure already exists changes the numbers a great deal, and we'll come back to it.

The meter will be registered to a company owned jointly by all the residents, which owns the building. At the end of the article we explain why this piece isn't an administrative detail but the one that underpins the biggest saving of them all.

The building moves from ten individual contracts to a single 10 kW meter.

The pergola can fit around 32.2 kWp of modules: 16.1 on each side. We modelled the sun's position hour by hour for latitude 41.4° N, with a split between direct and diffuse radiation and an 82% efficiency.

First: the term nobody looks at

Let's start at the end of the bill, not the beginning.

Ten homes with 4 kW contracted each add up to 40 kW. Under the 2.0TD tariff in 2026, between tolls and charges, each kW contracted costs around 29.6 € a year. That's 1,183 € a year just for having the power available, before consuming a single kilowatt-hour.

A single 10 kW meter costs 296 €.

All the figures in this article are broken down the same way: the lines above are the taxable base, and the total includes the electricity tax and VAT. Whenever we later refer to «annual cost», it always means the total with taxes, which is what actually gets paid.

1,128 € a year, 17% of the bill, without installing anything. Just by no longer paying ten times over for power capacity that's never used simultaneously.

This works because ten households don't hit their peak at the same time. Our model gives a real aggregate peak of 7.9 kW for these ten homes: contracting 10 kW leaves plenty of margin. This is what's known as the simultaneity factor, and it's the number that gets looked at least and matters most.

One important caveat, though: this saving isn't free in the legal sense. We cover that at the end.

What an east-west pergola delivers

A car park roof can rarely be oriented south. The good news is that at a 15° tilt, the penalty is smaller than it looks:

That's 1,276 kWh per kWp installed. An equivalent installation facing south at the optimum tilt would produce around 1,450: the pergola loses 12%.

In exchange, it gains something that doesn't show up in the annual total: a wider, flatter curve. The east side produces in the morning, the west side in the afternoon, and the midday peak is lower. For a residential building —which consumes little at midday and a lot in the evening— this shape fits better than a peak concentrated at noon.

Notice also the ratio: 41,147 kWh generated against 29,850 consumed. The pergola produces 38% more than the building uses. This isn't a design error: it's what happens when the surface area is a given. And it's the origin of everything that follows.

Second: the panels

With the installation running and no battery at all, the building directly self-consumes 13,709 kWh: 33% of generation and 46% of consumption.

It's worth pausing on one line: variable tolls and charges drop from 1,182 to 632 €. It's not just that you buy less energy — you also stop paying the regulated part of those kilowatt-hours. On a Spanish electricity bill, more than a third of the variable term isn't energy but tolls and charges, and self-consumption saves you both.

The bill goes from 6,472 € to 1,927 €. A saving of 4,545 € a year, 70%.

On surplus compensation. We calculated at 0.04 €/kWh, which is what's actually paid in practice. You'll see offers of 0.06 and even more, but the real figure that ends up on the bill usually moves between 0.03 and 0.05 depending on the retailer and the moment. This is one of the few figures worth checking against the specific offer before deciding anything: between 3 and 5 cents there's 150 € a year in this building.

The investment, with the pergola already built —which is the case we're dealing with here: the structure is already there and has been shading the spaces for years—:

Simple payback: 5.0 years. Over a 25-year service life, with a structure that's already paid for, that's twenty long years of practically free generation.

If the pergola had to be built from scratch, you'd need to add around 26,000 € and the payback would stretch to 10-11 years — still reasonable, but a different conversation. Making use of an existing structure is, after centralisation, the second factor that moves these numbers the most.

Third: the battery, and the wall it hits

This is where we expected the big leap. It's there at the start, and then it dies suddenly.

Before anything else, a clarification that changes how every table in this section should be read: we're talking about a single battery for the whole building, connected behind the single meter. When we say 30 kWh, we mean 30 for the ten homes — three kilowatt-hours per home. It sounds like little. At the end of this section we explain why it's enough.

A battery in a building like this does two jobs, and not at the same time of year. In summer it stores sunshine for the evening. In winter, when there isn't any, it charges from the grid overnight —off-peak hours, at 0.106 €/kWh— and discharges at the evening peak, where the same kilowatt-hour is worth 0.251: 13.4 cents net for every one it displaces.

We modelled both jobs together:

And the marginal value of each block:

The sweet spot is 30 kWh, with the best absolute payback (6.0 years) and the best marginal one (5 years). From here it falls off a cliff.

It's worth noting something counterintuitive: the lower the surplus compensation, the better the battery performs. If they pay you 3 cents for the surplus instead of 6, storing it to use yourself is worth twice as much. Among other things, the battery is a hedge against whatever the retailer decides to pay.

We're battery manufacturers and we're writing this anyway: in this building, going beyond 45 kWh is throwing money away.

The cycles column confirms it. At 15 kWh the battery does 346 full cycles a year —almost one a day, perfect use—. At 60 kWh it does 215. You're buying capacity that sits idle for half the year.

Why it stalls anyway: two things at once

The first is physical, and it's the one that rules. By around 30 kWh, the battery already covers all of the evening consumption, and beyond that point there's nothing left for it to do: in June, a 60 kWh battery behaves exactly like a 30 kWh one, because both reach midnight full, with the building asleep. You can't save more than 100% of a bill, and as you approach that limit, each additional kilowatt-hour is worth less than the one before.

The second is regulatory, and it's the one that surprises. Look at this column:

The more battery, the more money is lost. This isn't a model error: the surplus falls —from 27,438 to 21,147 kWh— but what you're paid falls much faster, from 864 to 448 €. The next section explains why.

The compensation cap, explained

The regulation states that surplus compensation can never exceed the value of the energy you've bought within the same billing period. If in one month you buy 100 € of energy and feed in 150 € of surplus, you're compensated 100 and the other 50 is lost.

Look at June for this building with three different batteries:

The columns, left to right: how much energy the building buys that month, how much it feeds in, what that surplus would be worth at four cents, what the cap is —which is the value of the energy bought, the column everything else depends on—, what it's actually paid, and how much is lost along the way.

The battery does its job: June's purchases fall from 928 to 472 kWh. But the cap is calculated precisely on what you buy, and by reducing it from 84 € to 38 € the battery has cut its own entitlement to be paid for the surplus. It's paid less because it bought less.

Put another way: every kilowatt-hour you store is one you stop buying, and every one you stop buying lowers the ceiling on what you can be paid for the surplus. Part of the battery investment works against the rest.

What does justify a large battery

The numbers here only value self-consumption arbitrage. There are three reasons to install more than these tables suggest, and none of them show up on the bill:

• Backup. Keeping the building running during a blackout isn't measured in euros saved until it happens. April 2025 reminded everyone of that.
• Reducing contracted power. A battery that shaves the peak allows you to drop below 10 kW. We haven't counted this here because the peak is already 7.9 kW and there's margin, but in a tighter building it's worth money.
• Future consumption. If tomorrow there's communal air-source heating or electric cars charging every night, the curve changes and the battery has work to do again. We'll analyse that separately: it deserves its own article.

What doesn't justify a large battery is this table. With this consumption and this generation, the edge is at 30 kWh.

What we would do

In order, and this order matters:

1. Centralise. 1,128 € a year, zero investment. Before anything else.
2. Fill the pergola with panels. An extra 4,719 € a year, payback in under five years.
3. A 30 kWh battery, which is the sweet spot from a financial point of view: it pays back in 6.0 years, and the jump from 15 to 30 is the best of all (5 years). It does both jobs —storing sunshine in summer, arbitraging the tariff in winter— and has capacity left over for both. Above 45 the numbers die: going from 45 to 60 kWh saves eight euros a year. The second part of this article shifts this figure, and explains why.
4. Plan for growth: space in the electrical panel and a modular installation, because the building's consumption won't be the same in ten years' time.

---

Part two: what if the power goes out

So far we've talked about money. Now let's talk about something else.

In April 2025, half the Iberian Peninsula lost power for hours. It came back the same day. The question left hanging in the air is what would have happened if it had been three days, and by 2026 that question is no longer rhetorical for anyone who has to decide on a twenty-year investment.

For a building, the answer isn't technical. It's who lives there.

What exactly we're talking about

A lift that's out of service is an inconvenience for most people. For a resident with reduced mobility on the third floor, it means being shut in at home. If the blackout lasts three days and that person needs to get to hospital, or simply needs to go down to buy food, the lift stops being a service and becomes their only way out.

Every lift is legally required to carry a rescue device that, if the power fails, brings the car to the nearest floor and opens the doors. This stops anyone from being trapped inside. It doesn't stop someone being trapped upstairs, which is a different and much longer-lasting problem.

And there's a second thing, less visible and probably more serious.

In a three-storey building, water doesn't rise on its own. It reaches the homes because a booster pump pushes it up, and that pump runs on electricity. Without power there's no pump, and without a pump no water comes out of any tap except on the ground floor.

The consequences arrive fast and they're no mere inconvenience: you can't wash dishes or hands, you can't flush toilets, you can't cook or clean anything. Within hours, the sanitary conditions of a ten-home block fall apart; within three days, food preparation and basic hygiene become a genuine health problem. And the people who can't walk down the stairs to fetch water are, once again, the same ones.

With that in mind, «autonomy» stops being a word and becomes four very different things:

(with no generation at all; the battery column is nameplate capacity and already accounts for everything lost along the way, which we explain below)

Level D isn't an option. That's nearly half a megawatt-hour of batteries for ten homes: half a tonne of equipment and six-figure numbers. Keeping ten households running normally during a winter blackout isn't a storage problem, it's a generation problem, and the available roof space can't reach it.

The other three are achievable. And they're the ones that matter.

The three services that keep the building going

Lighting doesn't count: 370 Wh a day is less than a fridge.

And energy-wise, neither does the pump: 38 minutes of motor running a day to raise water three floors with sufficient pressure at the most unfavourable tap, 690 Wh in total. That's 15% of the communal services' consumption and the cheapest of the three to guarantee.

It's worth pausing here, because cost and criticality don't line up at all. The pump is the cheapest service to keep alive and the one with the most immediate consequence if it fails: the lift leaves people stuck upstairs, but the pump leaves the whole building without water. With less than a fridge consumes, you keep running water flowing to ten homes during a blackout.

The lift takes up 77% of it. And with two quirks worth knowing before sizing anything.

It consumes more standing still than moving. Of the 3.63 kWh a day, 2.88 is standby —control panel, car lighting, electronics— and only 0.75 is movement. Forty trips a day amount to twelve minutes of motor running. The rest of the time the lift does nothing and uses the same amount of energy.

The problem isn't the energy, it's the start-up. A 7.5 kW hydraulic lift with direct-on-line starting draws 15 kW for one or two seconds. If the pump is running too, 16.1. The grid doesn't notice this, but an inverter working in island mode does.

Note: not all lifts are the same

This study uses a hydraulic lift, which is the worst case. If the building has a traction lift with a frequency drive, the numbers change a lot:

	Energy	Start-up	Battery for 3 days
Hydraulic	3.63 kWh/day	15 kW	47 kWh
Traction with drive	1.93 kWh/day	5.2 kW	38 kWh

The hydraulic lift consumes 88% more and starts up at almost triple the power. For sizing an emergency installation, this difference amounts to thousands of euros, both in battery and in inverters.

Before budgeting anything, you need to know which type of lift is installed. It's the cheapest check in the whole project and one of the ones that moves the result the most.

What makes it possible: the pergola doesn't stop

Here's the piece that changes the scale of the problem. If the inverter can work in island mode —disconnecting from the dead grid and continuing to power the building— the pergola keeps producing for as long as the blackout lasts.

And in December it produces far more than the essential services need:

On a normal December day, the pergola generates almost three times what the whole of level C consumes. The battery doesn't need to store three days of energy: it needs to bridge one afternoon to the next morning, and cover the bad days.

This is what makes this approach viable, and a battery-alone approach not. Without the pergola, three days of level C need 81 kWh of battery. With the pergola, and allowing for a catastrophic December week, it's 47.

Sizing: 3 days, level C

For this study we set the target at three days of level C —lift operational, water at the taps, stairwell lit, fridges running, and light, phone and radio in every home— with the hydraulic lift and counting on the pergola in island mode.

The generation assumption isn't December's average but the worst imaginable week: overcast sky at 15% of clear-sky, 7.1 kWh a day. Designing around the average would mean designing for the day you don't need to worry about.

From nameplate to socket: what's actually left

A 45 kWh battery doesn't deliver 45 kWh, and even less after ten years. Between the nameplate figure and what reaches the socket there are four deductions, and none of them is negligible:

You need to install 1.62 kWh for every kWh that must be available ten years from now. Sizing with the nameplate, first-day capacity is the most common mistake in emergency installations, and it has the particular trait of going unnoticed until the day you actually need it.

The result

In practice, 45 kWh, which is a catalogue size and gives 2.9 days. The next step up, 60 kWh, reaches 3.8.

The inverter: here power rules, not energy

Everything discussed so far is about energy. Sizing the inverter is a completely separate problem, and the lift decides it.

A Victron MultiPlus-II 48/5000 delivers 5,000 VA apparent power, 4,000 W continuous rated at 25°C —3,700 at 40°C— and 9,000 W peak for a few seconds. The figure that counts for sizing is the 4 kW one, not the nameplate figure.

The interlock is a piece of basic logic: don't let the pump start while the lift is moving. The pump has 38 minutes of work a day and can wait twenty seconds. It saves an entire inverter.

Limiting consumption during a blackout

There's a decision that matters as much as the battery size, and costs nothing: allocating what's available.

If during a blackout each home can consume whatever it wants, a single resident switching on a heater can drink up the whole building's autonomy in a few hours. Assigning a maximum to each home radically changes the sizing:

Between giving each home 1.2 kWh a day and giving them 0.6, there's 4,800 € of battery. It's a decision for the residents' meeting, not for engineering, and it's worth making it knowing the price.

More important than limiting energy is limiting power. Ten homes at 500 W each make 5 kW; if the lift starts at the same time, the peak jumps to 12.6. With an interlock that gives priority to the lift, it stays at 7.6 and everything fits with two inverters.

This isn't done with a timer: it's done with continuous metering of every circuit and logic that decides who's entitled to what. It's the same layer that optimises the bill the rest of the year.

The two answers don't match, and that's fine

This is where the two halves of this article collide, and it's worth stating clearly instead of hiding it:

The 15 kWh difference doesn't pay for itself. We saw this in the first part: the jump from 30 to 45 has a marginal payback of 26 years. As an energy investment, that's bad.

But it isn't an energy investment. It's an insurance policy, and it should be valued the way insurance is valued: not by what it yields, but by what it covers and what it costs.

About 2,200 € of difference, spread across ten homes and twenty years of service life, comes to eleven euros per home per year so that, if what happened in April 2025 happens again and that time it lasts three days, the lift works, there's water at the taps, and the food doesn't spoil.

Put that way, the conversation at the residents' meeting is a different one.

The practical conclusion isn't a figure but a way of buying: the 30 kWh that pay for themselves, in a modular installation with space planned in the panel to grow. This building's consumption won't be the same in ten years' time, and making room for it now is far cheaper than doing so later.

Who should hold the meter contract

That covers the numbers. Now the piece that underpins them, and which needs to be settled before looking at any of them.

A single meter for ten homes isn't a technical decision, it's a corporate one. Royal Decree 88/2026, in force since February, requires in Article 5 that the holder of a supply point be «a single natural or legal person».

The usual arrangement is a company jointly owned by all the residents —a cooperative, a limited company, whatever is chosen— which owns the building and holds the supply contract. Facing the utility there's a single customer; behind closed doors, ten homes.

With two conditions that can't be skipped.

The building must be a single unit. If each home has its own land registry reference and its own registered owner, this route doesn't work.

The internal allocation must be an allocation, not a sale. A company that re-invoices its members per kilowatt-hour, at a price comes close to acting as a retailer without being one. What's clean is allocating the cost as just another shared expense, using the criteria in the bylaws and without a margin. The line is thin, and it's the one that separates a shared fee from a bill.

If it's an ordinary homeowners' association, this isn't the route — instead it's collective self-consumption over an internal grid: each home keeps its own meter and contract, and generation is shared out using allocation coefficients. You lose the power saving —the biggest of them all— but the whole photovoltaic part still applies equally.

That's why projects with a corporate structure come out so well: they can do something an ordinary association can't. The 1,128 € from the first section aren't bought by the technology, they're bought by the legal form.

This is as far as we can go, being battery manufacturers and not lawyers: the specific structure must be validated with someone who understands it, and before contracting anything.

The assumptions, and what would shift them

None of the numbers here are a promise. These are the weak points, in order of impact:

The simultaneity factor. We modelled an aggregate peak of 7.9 kW for ten homes. If the real building reaches 14, contracting 10 kW isn't possible and the saving from the first section shrinks. It's the only variable that can't be estimated: it has to be measured, and it's the first thing we'd do before sizing anything.

The hourly profile. We used a standard Mediterranean residential curve, with an evening peak. A building with a lot of people working from home would consume more during the day, and all the battery numbers improve.

The price of energy and of compensation. 0.10 €/kWh for energy and 0.04 €/kWh for compensation. These are retailer assumptions, not regulated values. Compensation is the most sensitive one: between 3 and 5 cents there's a 150 € a year difference, and it's one of the few figures that can be pinned down with a phone call.

The 215 Wp per m². This is a real, framed module. Depending on how the row layout on the pergola works out, the power that fits can vary 10% up or down.

From the second part

The type of lift. All the emergency sizing was done with a hydraulic lift, which is the worst case. With a traction lift the battery drops from 47 to 38 kWh. It's the check that moves the result the most, and the cheapest one to do.

The lift's standby consumption. We assumed a permanent 120 W, which is a typical value but varies a lot between models and generations. Since it represents 61% of all level A consumption, an error here propagates everywhere.

Worst-week generation. 15% of clear sky is an estimate for a catastrophic week, not a measurement. We don't have a historical radiation record for this location.

Fridge consumption. 0.85 kWh a day per home is a modern, efficient fridge. A fifteen-year-old fridge can double that figure, and ten homes with old fridges shift levels B and C noticeably.

The capacity of the existing pergola. We assumed the structure can take the 32.2 kWp and the weight of the modules without reinforcement. On a pergola that wasn't designed to carry panels, this must be checked with a structural calculation before anything else: it's a potential cost that doesn't appear in any table in this article.

---

If you have a building and want to know whether these numbers resemble yours, what we need to start isn't a site visit: it's one bill from each home and the quarter-hourly load curve that the distribution company has on file. With that, the model stops being an example and becomes your building. Ask us for the free analysis and we'll do it for you.