Energy from storage: levelised costs and losses

The cost of energy delivered by a storage system depends on lifetime costs and the amount of energy actually delivered. Levelised cost of storage (LCOS) helps compare these quantities. Meaningful comparison requires consistent calculation boundaries, use conditions and treatment of charging energy.

A battery's rated efficiency does not determine the complete cost per kilowatt-hour at the site output. Conversion, auxiliary consumption, degradation, service and future replacements all matter. This article explains which inputs to prepare, how to read efficiency and where losses can be counted twice.

What levelised cost of storage shows

LCOS expresses costs per unit of energy delivered by storage over the analysis period. It is a lifecycle measure, not simply the battery price divided by rated capacity. An installation can charge and discharge repeatedly, but available capacity and operating conditions change.

First distinguish storage power and usable capacity. Power sets the rate of energy exchange; capacity is the reserve under specified conditions. LCOS also requires actual or reasonably forecast delivery over a period. Nominal capacity alone does not establish that volume.

The general logic is LCOS = discounted costs / discounted energy delivered. Initial investment is accounted for at its date; future costs and delivery volumes are discounted using the selected method. Fix the term, rate, currency and time step for comparison.

LCOS is not a sales tariff or a ready conclusion about profitability. Compare it with operating mode and the source of revenue or savings. Investment recovery is discussed separately in storage payback and cash flows.

Which costs enter the calculation?

Define initial investment, operation, repair, capacity maintenance and other included costs. Check the list against the full storage budget. If a major item is omitted, a lower result may reflect incomplete data rather than an equipment advantage.

Pay particular attention to charging energy. Some methods include its full cost; others assess storage costs using a different boundary with losses accounted for separately. Establish the approach before comparing figures. The label LCOS alone does not define the boundary.

State separately whether terminal work, taxes and financing are included and how. The list depends on the method. Comparable options need the same approach so the difference reflects scenarios rather than accounting rules.

Planned replacement may already be in an agreed annual budget or appear as a separate future payment. Do not include it twice. Replacement cost and the capacity forecast must also agree: a capacity restoration should not appear without justification.

Where should round-trip efficiency be measured?

Round-trip efficiency (RTE) is the ratio of energy delivered to energy received for the relevant cycle. Comparison requires agreed metering boundaries and the same state of charge at the start and end. Otherwise part of the difference may reflect a change in stored energy rather than cycle loss.

The boundary may be at the battery, converter or whole installation. Transformer, cables, cooling, heating and controls affect the site result when included in that boundary. A component rating cannot automatically serve as the guaranteed result at the connection point.

Specify operating modes, power, environmental conditions, included auxiliaries and verification method. Establish whether standby periods are included. Verification is discussed in storage testing and acceptance.

Compare auxiliary consumption with the actual schedule. An installation that rarely delivers energy can have a different ratio of auxiliary consumption to output than one used regularly. Explain standby duration and auxiliary modes rather than applying a short test result to an entire year.

For a Sungrow PowerStack 255CS-2H configuration, request inputs from supply and project documentation. The product page identifies the model but does not establish LCOS or whole-system efficiency for every operating profile.

Energy balance example

For an educational example, assume whole-installation round-trip efficiency of 86%. At the same metering boundary and with the same final state of charge, 100 MWh received yields 86 MWh delivered. The difference is 14 MWh. This is a calculation example, not a measurement of a specific product.

100 × 0.86 = 86 MWh; 100 - 86 = 14 MWh.

To deliver 1 MWh in the same simplified scenario requires approximately 1 / 0.86 = 1.163 MWh of charging energy. If the price per unit received is C, charging energy alone costs C / 0.86 per unit delivered. Investment, service and other costs are not yet included.

A model may show 93% for equipment and 86% for the whole installation. These are different boundaries, not independent sets of losses. If 86% already includes auxiliary consumption and conversion, do not deduct them again. The difference is 7 percentage points.

Likewise, do not add charging cost twice through a separate losses line. Check whether the input volume was already increased to compensate for losses. Define the energy balance before applying tariffs and calculating a financial result.

C / 0.86 assumes a given price and efficiency. If charging occurs at different hourly prices or in different modes, calculate costs using the corresponding volumes and conditions. An unexplained average can hide differences between scenarios.

How utilisation and battery life change LCOS

Total delivery depends on cycles, depth, available capacity and operational readiness. A forecast should not assume continuous operation at full initial parameters without verification. Some time may be lost to service, restrictions or absence of the required operating scenario.

Greater delivery, all else equal, spreads fixed costs over more energy. However, extra cycles can change wear and future costs. Do not artificially reduce LCOS by increasing cycle count if that schedule is not technically and economically supported.

The replacement plan must match the resource forecast. Degradation, warranty and capacity maintenance are discussed in storage performance changes and service. The same analysis period does not imply the same available output for different configurations.

For backup installations, part of the resource may remain available for the required load. This limits other operations. The comparison method must match the business's task so a lower cost per unit delivered does not replace the backup-readiness requirement.

Increasing capacity at the same power changes initial cost and potential duration. It does not prove a proportional increase in useful annual delivery. Check whether suitable charging hours, loads or another agreed use exist for the added resource.

Data needed after commissioning

Keep separate records of energy received and delivered at the agreed boundary, operating and downtime periods, events and required auxiliary measures. Compare consistent intervals. A cycle counter or state-of-charge percentage alone cannot verify the complete balance.

For solar with storage, discuss support through solar plant maintenance. Diagnostics for the battery, auxiliaries and specific storage system are agreed separately. Define the required data, access and response to deviations.

1. Collect the balance. Obtain charging and delivered energy and measurement conditions.
2. Compare conditions. Check mode, boundaries and current configuration.
3. Update the model. Explain differences and adjust confirmed assumptions.

A difference between forecast and actual results needs analysis. A changed mode or metering boundary can distort comparison without equipment deterioration. Preserve data history and explanations rather than merely replacing the old result.

How to compare two calculations

Request a table of inputs and methodology. Align term, rate, energy units and cost scope. Check that the denominator uses the same delivery boundary and that capacity and replacement forecasts match the actual configuration.

• Same methodology and treatment of charging cost.
• Agreed efficiency and auxiliary-consumption boundaries.
• Justified output, availability and operating profile.
• Consistent costs, degradation and replacement schedule.
• Common term, rate, currency and scenario assumptions.

The final result should be reproducible from these inputs. Compare not just one LCOS value but why it changes between scenarios. This shows whether the result depends on verified installation properties or assumptions about future use and costs.