Time-of-Use Battery Storage for European C&I Sites: Model the Real Price Spread

Introduction

You have probably seen the calculation in an early BESS meeting: buy electricity at €0.105/kWh, avoid buying it later at €0.245/kWh, and keep the €0.140 difference. It fits neatly on one slide. It is also where many Time-of-Use Battery Storage business cases begin to drift away from the customer’s real bill.

The awkward questions arrive later. Does the €0.245 line include charges the battery cannot avoid? Is €0.105 the price the site actually pays, or a wholesale chart copied from the internet? Who pays for the extra energy lost through the PCS, transformer and cooling system? And if the battery must hold 25% state of charge for backup, why did the savings model spend that same energy every evening?

Those are not small spreadsheet details. They decide whether the owner buys a useful energy asset or an expensive cabinet that spends much of the year waiting for a price window that never appears. European price signals are becoming more granular, but more granularity does not automatically mean more margin. The European Union’s Directive (EU) 2019/944 defines dynamic contracts by reference to spot-market variation at intervals at least as frequent as market settlement. Germany’s Bundesnetzagentur also explains that a dynamic contract’s energy price can change several times per day. Neither source says that a public wholesale price is the same as a factory’s marginal import price.

So let us answer the question a project team actually has to sign off: after the contract, losses, battery wear, operating limits and competing duties are included, how much of the visible spread is still yours? The method below lets an owner challenge a sales model and lets an EPC build a quotation that finance can follow line by line.

When Does Time-of-Use Battery Storage Have a Real Commercial Signal?

Open the electricity contract and ask a blunt question: which number actually changes with time? If nobody in the meeting can point to the clause, formula or tariff calendar, the project does not yet have a TOU signal—it has an assumption. Three very different arrangements are commonly pushed into the same spreadsheet:

There is another trap here. A salesperson sees a demand charge on the bill, converts it into €/kWh and adds it to the energy spread. The result looks better, but the units no longer describe the same service. Demand or capacity charges are tied to a billed power metric in kW or MW. TOU energy value is tied to kWh moved between intervals. One battery can address both, but the model must reserve power and energy for each duty instead of counting the same dispatch twice. If your team is still separating those objectives, use the practical comparison of peak shaving and load shifting.

Build the C&I Price Stack Before Calculating the TOU Spread

This is the part owners rarely receive in a first proposal. The quotation says “electricity price,” while the invoice contains energy procurement, supplier margin, network charges, taxes, capacity charges, fixed fees and sometimes an export settlement. A battery does not make that whole stack disappear. It changes selected metered imports at selected times.

Take one real invoice and one representative settlement day. For every line item, ask: if the site imports one less kWh at 18:00, does this charge actually fall? Then ask the reverse question for charging at 02:00. This simple exercise stops a polished ROI model from claiming savings on money the customer would pay anyway.

Eurostat’s non-household electricity price dataset separates price components at an aggregate level. It is useful context, not a substitute for the site contract. A German industrial customer’s connection level, negotiated supply arrangement and network tariff can produce a very different marginal price from a small commercial site in another EU member state.

Reconcile Germany’s SMARD curve with the actual retail contract

Germany’s SMARD market-data platform is a valuable source for wholesale-market history. It is not the customer’s invoice. The Bundesnetzagentur’s dynamic-tariff guidance explains that the retail energy price may be oriented to the exchange price and can change several times daily. The phrase “oriented to” matters: supplier margins, risk premiums, fees, taxes, network treatment, caps and floors must still be mapped.

Here is the pain buyers feel in practice: someone downloads a day-ahead curve, circles a negative hour and tells management that the battery will be paid to charge. Six months later, finance cannot find that “negative price” anywhere on the bill. The hardware may already have been specified around a price event the customer never receives. Fix this before procurement: build a timestamped marginal import-price series from the supplier’s exact formula, then reconcile several calculated days against an invoice or settlement statement. If the numbers do not tie, do not size the battery from them.

Use a Real-Spread Waterfall, Not the Headline Price Difference

Two suppliers can use the same tariff and still show very different savings. Often the difference is not better engineering; it is a quieter modelling boundary. One quote measures efficiency at the battery’s DC terminals. Another measures energy at the AC bus. The owner, meanwhile, gets paid—or saves money—at the site meter. Until those points are aligned, the comparison is not apples to apples.

Choose the commercial meter first, then follow every kWh from charging to delivery. Include the PCS, transformer, cables and auxiliaries that sit inside that boundary. If a supplier excludes cooling or control power, show it as a separate cost rather than letting it vanish from the model.

Round-trip efficiency (RTE) must match the commercial boundary. The US Department of Energy’s Battery Energy Storage System Evaluation Method treats energy-in, energy-out and the chosen analysis period as defined measurement quantities. For procurement, that principle translates into a simple requirement: every efficiency value should identify its meter points, operating conditions and included auxiliary consumption.

The six steps in the waterfall

1. Value the avoided import. Use the marginal retail price that would have applied to the displaced interval.
2. Value the charging energy. Include the actual purchase price or the opportunity cost of on-site generation.
3. Convert through the metered RTE. A delivered kWh requires more than one purchased kWh.
4. Include marginal degradation. Use the supplier’s warranted-throughput and degradation model, not a universal €/kWh assumption.
5. Include variable operating costs. Add market, aggregation, metering or maintenance costs that rise with dispatch.
6. Price competing uses. Reserving energy for backup, peak control or another contracted service has an opportunity cost.

Worked C&I Example: From a €0.140 Spread to the Net Contribution

Important: the following values are illustrative engineering assumptions, not current European market prices, a supplier quote or a savings guarantee. Replace every input with the customer’s contract, measured load and equipment data.

To deliver 1,000 kWh at 88% RTE, the site purchases approximately 1,136.36 kWh. The calculation is:

• Avoided import value: 1,000 × €0.245 = €245.00
• Charging-energy cost: 1,136.36 × €0.105 = €119.32
• Degradation allowance: 1,000 × €0.035 = €35.00
• Variable O&M and other cost: 1,000 × €0.007 = €7.00
• Illustrative net contribution: €245.00 − €119.32 − €35.00 − €7.00 = €83.68 per eligible cycle

The headline difference was €0.140/kWh, but the modelled contribution is approximately €0.0837 per delivered kWh. The break-even avoided price under these assumptions is about €0.1613/kWh. This is contribution before fixed O&M, financing, taxes, EPC cost and CAPEX recovery; it is not project profit.

Do not multiply €83.68 by 365 without testing eligibility. If only 180, 240 or 300 days pass the price and physical gates, the illustrative annual contribution would be approximately €15,062, €20,084 or €25,105 respectively. The annual model must decide which intervals qualify; the analyst should not decide in advance that every day cycles.

Match the TOU Price Window to the Site Load and Battery Duty

A perfect evening price is useless if the factory is barely importing at that time. This sounds obvious, yet it is a common route to oversizing: the analyst studies the tariff first, chooses a four-hour battery, and only then checks whether the site has four hours of load to offset.

Reverse that order. For every expensive interval, take the lowest of four limits: avoidable site import, permitted discharge power, PCS capability and energy available above the reserved SOC. Keep power and energy separate. kW tells you how much import can be replaced now; kWh tells you how long the replacement can last.

Check the discharge envelope

If a facility imports 420 kW during a two-hour expensive window but must preserve 170 kW of grid import for operational reasons, the theoretical battery dispatch ceiling is 250 kW. The delivered-energy ceiling is then 500 kWh before considering load variation, PCS limits, battery SOC, availability and reserve obligations. A 1 MWh nameplate battery does not automatically double the saving; half its energy may have no load or price window to serve.

Prove the recharge window

Now work backwards. At 88% RTE, delivering 500 kWh requires more than 500 kWh of charging energy. This is where a design that looked comfortable can suddenly fail. The battery has enough nameplate energy, but the factory’s night shift keeps the transformer busy; the EMS either misses tomorrow’s peak window or charges harder and creates a new demand peak.

The practical fix is to overlay charging power on the site’s real overnight load, not its monthly average. Apply the import ceiling interval by interval, include charging losses, and confirm the required SOC is reached before the next expensive period. If it is not, change the duty, extend the charging window, reduce the daily discharge target or address the connection constraint. Buying more kWh does not create more grid headroom.

The solution is not automatically a larger PCS. A higher charging rate can breach an import limit or increase a demand charge. Test the charging schedule against the simultaneous facility load and connection constraint. For the broader power-versus-energy workflow, use the industrial BESS sizing guide; for the operating logic, see the C&I load-shifting guide.

Only then screen a product architecture

Once the dispatch duty is known, equipment can be screened without pretending that one capacity suits every site. A smaller commercial property with a modest shift may investigate a compact 100 kWh cabinet. A space-constrained industrial site may evaluate a modular 418 kWh liquid-cooled block, while a larger repeatable duty may justify assessing a 1 MWh containerized configuration.

These are candidate architectures, not recommendations from energy capacity alone. The final choice still depends on AC power, usable SOC window, end-of-life capacity, temperature, auxiliary consumption, grid connection, fire strategy, acoustic limits, footprint, warranty and local permitting. The broader C&I storage portfolio is useful only after those inputs are defined.

Make the EMS Spend Battery Life Only When the Net Spread Clears the Threshold

Owners do not need an EMS that looks busy. They need one that knows when not to cycle. A controller can correctly identify that 02:00 is cheaper than 18:00 and still make a poor decision if the spread does not cover losses and wear. Every unnecessary cycle uses throughput, occupies thermal capacity and may leave less SOC for a more valuable peak or an outage.

Set a minimum net-margin threshold, not merely a “low price” and “high price.” The EMS should dispatch only when the forecast value clears that threshold and every higher-priority constraint remains satisfied.

Define a project-specific dispatch hierarchy

A C&I hierarchy might be:

1. Safety limits, protection states and grid-connection requirements;
2. Minimum backup reserve for agreed critical loads;
3. Contracted import ceiling or peak-control target;
4. TOU dispatch above the net-margin threshold;
5. PV self-consumption, export optimisation or optional market services.

That order is an engineering example, not a universal standard. A hospital, data centre, cold store and shopping centre will assign different values to backup reserve and load risk. The important point is that reserved SOC and reserved PCS power are explicit. Selling the same kWh simultaneously to backup, peak shaving and TOU arbitrage creates paper revenue stacking, not a deliverable operating plan.

Require forecast logic, reason codes and a fallback schedule

“AI optimisation” is not an acceptance test. If the site manager asks, “Why did the battery discharge at 15:30 and buy the energy back at 17:00?”, the operator needs a reason code, not a black-box answer. Poorly configured control can charge, discharge and recharge inside the same broad window, creating wear while the promised spread slips away.

The acceptance specification should require a historical replay using the same tariff, load and constraints as the feasibility model. Each dispatch should carry a reason code—for example tariff arbitrage, peak limit, reserve recovery or manual override. If a price feed or forecast fails, the local controller should move to a documented safe schedule rather than continue on stale data.

Why European TOU Battery Economics Must Be Modelled Country by Country

“Does TOU storage work in Europe?” is too broad to be useful. A German factory, a Spanish retail portfolio and a British logistics site can all see time-varying prices while facing different contracts, network charges and settlement rules. Copying the spreadsheet from one country into another may preserve every formula and still produce the wrong answer.

EU law provides shared market principles; the money on the invoice is still shaped by national rules, the local network and the customer’s contract. The United Kingdom is a separate jurisdiction and must not be folded into an EU-27 assumption.

This distinction is especially important in Germany. The Bundesnetzagentur’s 2026 AgNes update discusses a future framework and timetable for dynamic network tariffs for storage facilities. It should not be inserted into a current business case as an already-effective final tariff. A credible model separates current contractual rules from proposed or future arrangements and reruns the downside case if policy changes.

Stress-Test the Business Case Before Requesting Hardware Quotes

If the business case needs a perfect Tuesday to work, it does not need prettier charts; it needs a downside test. Production shuts down. Tariff windows move. Summer auxiliaries rise. Backup reserve is called on. Some days simply do not contain a spread worth cycling for.

Use at least twelve consecutive months of interval load and price data when available, aligned to the same timezone and settlement convention. Test both beginning-of-life performance and the agreed end-of-life usable-capacity condition. Management should see how many cycles the model rejects, not only the cycles it accepts.

Minimum scenario set

• Base case: contracted prices, expected availability and the warranted performance model.
• Compressed-spread case: lower high-price periods, higher charging prices or both.
• Low-efficiency/high-auxiliary case: adverse temperature and partial-load operation.
• Reduced-capacity case: end-of-life usable energy and the same operational reserve.
• Restricted-recharge case: less connection headroom or a shorter cheap window.
• Multi-service conflict case: backup or peak-control reservation reduces arbitrage availability.

Do not let bad timestamps create good returns

Interval models fail quietly when the meter and the price file disagree about time. The spreadsheet still produces a number, so nobody notices. Check UTC versus local time, daylight-saving transitions, repeated or missing intervals, holidays, weekend schedules, price caps and floors, contract effective dates and units. Record how gaps are treated. A duplicated autumn hour mapped to the wrong price can make the dispatch look clever when it is simply misaligned.

The final feasibility output should report capturable delivered energy, purchased charging energy, maximum charge and discharge power, eligible and rejected cycles, gross spread, net contribution, break-even avoided price and the reason each rejected interval failed. It should also state what the model does not include—such as financing, taxes, fixed O&M or EPC cost—so contribution margin is not mistaken for project return.

Make Every BESS Supplier Quote Use the Same Economic Boundary

This is a familiar procurement headache: three suppliers receive the same monthly bill and return three battery sizes, three savings forecasts and three payback periods. The buyer is left comparing confidence levels instead of engineering.

Put the same data pack and the same duty on every bidder’s desk. Otherwise, the quotation with the largest claimed saving may simply assume more cycles, a wider SOC window, a different price series or less backup reserve. The cheapest €/kWh can also be the most expensive option if its usable AC boundary is smaller.

For acceptance, replay a defined historical period through the delivered EMS or a validated digital representation. Compare scheduled and measured charge/discharge energy, SOC, import limit compliance and dispatch reasons. Set tolerances in the contract. This turns “smart optimisation” from a sales phrase into a testable control function.

Time-of-Use Battery Storage FAQs

What is time-of-use battery storage?

It is a battery operating strategy that charges during lower-cost intervals and discharges when avoiding grid import is more valuable. For C&I sites, the relevant signal is the customer’s marginal retail price and contract—not merely a public wholesale price.

How much price spread is needed for battery arbitrage?

There is no universal threshold. The avoided price must exceed the charging price divided by the measured RTE, plus degradation, variable O&M, transaction costs and the value of competing battery uses. Financing and fixed project costs must then be assessed separately.

Does a negative wholesale price mean a business is paid to charge?

Not necessarily. The retail contract may include supplier adders, network charges, taxes, caps or a different settlement formula. Confirm the actual marginal retail price and any restrictions before assigning a negative charging cost.

How does round-trip efficiency affect TOU savings?

RTE determines how much energy must be purchased to deliver one usable kWh later. At 88% RTE, delivering 1,000 kWh requires approximately 1,136 kWh of charging energy. The quoted efficiency must identify its meter boundary and included auxiliaries.

Can one battery handle TOU savings, peak shaving and backup?

Potentially, but the services compete for energy, PCS power and operating time. The EMS needs explicit power and SOC reservations, a priority order and a fallback mode. Nameplate capacity alone does not prove simultaneous availability.

What interval data is required for a commercial TOU model?

Use data at the billing or dispatch resolution required by the contract, ideally covering at least twelve consecutive months. Load, prices and operating calendars must use aligned timestamps, units and daylight-saving treatment.

Can the battery recharge completely during the off-peak period?

Only if the cheap window, charging power and spare connection capacity are sufficient after simultaneous site load is included. The model must also account for charging losses and the target SOC reserve.

Is TOU battery storage profitable across Europe?

There is no Europe-wide profitability answer. Results depend on the national and site-specific contract, network charging, taxes, load shape, price volatility, connection limits, equipment performance, degradation, financing and competing uses.

Move From a Visible Spread to an Auditable BESS Decision

A dramatic market-price chart is easy to sell. A model that survives an invoice check, a winter production week, an end-of-life test and an EMS replay is much harder—and far more valuable.

Before asking suppliers for a firm price, collect the electricity contract, twelve months of interval data, single-line diagram, connection limit, operating calendar, PV/export terms and backup requirement. With those files, the first conversation becomes concrete: which charges move, how much load can be served, whether the battery can recharge, and how many days actually clear the net-margin threshold.

If the answer is “the spread is too thin,” that is still a useful result—you have avoided buying hardware on a weak assumption. If the signal survives, the same model becomes a clean starting point for an EPC design basis and a supplier RFQ.

Sources and Evidence

• European Union, Directive (EU) 2019/944, including the definition of a dynamic electricity price contract and Article 11.
• Eurostat, Electricity prices for non-household consumers—price components.
• Bundesnetzagentur, Dynamic electricity tariffs, contract structure, price linkage and smart-meter information.
• Bundesnetzagentur, SMARD electricity-market data.
• Bundesnetzagentur, 2026 AgNes network-tariff reform update, including the stated future timetable for dynamic network tariffs.
• US Department of Energy, Battery Energy Storage System Evaluation Method, measurement and performance-evaluation boundaries.
• International Energy Agency, The Value of Demand Flexibility, system value and the role of flexible demand.