AC or DC? The 18% Gap Hiding in Every BESS Quote You'll Receive

Most BESS tenders in India now specify capacity at the AC terminals. Most BESS datasheets state capacity at the DC nameplate. Those are different quantities, and the difference is not a rounding error — on a four-hour LFP container it is close to 18%.

This matters because the gap is invisible in a side-by-side comparison. Two vendors can quote "5 MWh" against a 200 MWh AC requirement, both be describing real equipment, and both be wrong about what the plant will deliver at the meter. The buyer discovers it at the capacity test, which is the most expensive possible moment.

What follows is the arithmetic, in the order it applies, using our own published figures for the SIE-BESS5000.

Two haircuts, applied in order

Nameplate capacity is the energy the cells hold when full. AC-deliverable capacity is the energy that leaves the plant through the MV terminals. Between those two numbers sit two independent reductions.

Step one — depth of discharge. A lithium iron phosphate system is not cycled between 100% and 0% state of charge. Operating depth is a contracted limit, and ours is stated at 90%:

4,915.2 kWh nominal  ×  0.90  =  4,423.7 kWh usable DC

That figure is on the product page, not derived here. The 90% basis is also the basis on which the cycle life is rated — ≥6,000 cycles at 90% DoD and 25 °C — which is why the two numbers belong in the same sentence.

Step two — conversion loss. Energy leaving the battery passes through the power conversion system before it reaches the AC terminals. The datasheet quotes round-trip efficiency on two bases: ≥86% AC-AC at system level and ≥94% on the DC side. Those two figures together are what let the conversion path be separated from the battery:

conversion round-trip   = 0.86 ÷ 0.94 = 0.914894
one-way conversion      = √0.914894   = 0.95650
AC deliverable          = 4,423.7 × 0.95650 = 4,231.3 kWh AC

So one container delivers 4,231.3 kWh at the AC terminals — 86.1% of the 4,915.2 kWh nameplate.

The assumption that carries the result

The square root in step two is doing real work, and it rests on an assumption that should be declared rather than buried: charge and discharge conversion losses are taken as symmetric, so the one-way factor is the square root of the round-trip factor.

This is the standard convention, and it is the reason a datasheet quotes both AC-AC and DC figures in the first place. But it is an assumption. If a power purchase agreement defines efficiency one-way, or measures at a different point in the circuit, the factor changes and the container count changes with it. A proposal that shows this line is a proposal you can check. One that states an AC figure without showing how it got there is asking to be taken on trust.

Where the 18% comes from

Now apply it to a procurement scenario. A tender asks for 200 MWh AC-deliverable. A vendor quotes containers described as "5 MWh".

200,000 kWh AC ÷ 4,231.3 kWh AC per container = 47.27 containers

Forty-eight containers, not forty. The "5 MWh" description overstates the AC-deliverable content of each unit by:

5,000 ÷ 4,231.3 = 1.18  →  roughly 18%

Eighteen per cent of a 200 MWh contract is 36 MWh of storage that was never going to be there. On a capacity test with liquidated damages attached, that is not a technical disagreement — it is a commercial event.

The error is rarely deliberate. "5 MWh" is a reasonable shorthand for a 4,915.2 kWh container in a conversation. It stops being reasonable the moment it is set against a requirement expressed in AC terms, and nobody restates the basis.

Why tenders say AC, and why they are right to

An AC-deliverable specification is the correct way to write a storage requirement, because it describes what the grid receives. A system operator does not care how much energy is sitting in cells; it cares how much arrives at the point of interconnection. Writing the requirement in AC terms puts the conversion losses, the operating depth and the auxiliary draw where they belong — on the supplier, who chooses the equipment.

It also removes an argument. If the requirement is DC nameplate, every vendor is free to select its own DoD limit and its own efficiency basis, and the quotes become incomparable. An AC figure is a single number measured at a single place.

Two consequences follow for anyone writing or answering one of these documents.

The metering point has to be named. Station auxiliaries — cooling, HVAC, fire systems, controls, lighting — draw from the same side of the meter as the delivered energy, and at plant scale they run at 2–3% of rated output. Whether the AC-deliverable figure is measured gross or net of those auxiliaries moves the container count. It is a one-line clarification before contract and an expensive discovery after.

The capacity test method has to be agreed before signature. The 200 MWh figure is only as firm as the method that measures it: the test point, the state-of-charge window, the ambient correction and the tolerance band each move the result. Agreeing them at contract is what makes a stated margin meaningful. Discovering them at site is how capacity disputes begin.

The check you can run on any quote

You do not need a model to catch this. You need one question and one multiplication.

Ask the vendor: is the capacity you have quoted DC nameplate or AC-deliverable, and at what depth of discharge and what efficiency basis?

Then check the arithmetic yourself:

AC deliverable ≈ nameplate × DoD × √(AC-AC RTE ÷ DC RTE)

If the vendor cannot give you the two efficiency figures, the one-way factor cannot be derived and the AC number they have quoted is not checkable. That in itself is the answer to your question.

A useful sanity bound: for a four-hour LFP system at 90% DoD and the efficiency figures above, AC-deliverable lands near 86% of nameplate. If a quote implies materially better than that, something in the basis is different and it is worth finding out what before the numbers reach a contract.

What this changes about comparing bids

Once every bid is restated in AC terms, two things usually become visible at once.

The first is that bids which looked equivalent are not. A vendor quoting a higher nameplate at a lower DoD limit can land below a vendor quoting less nameplate at a deeper operating window. The headline number was never the comparison.

The second is that the container count moves the balance of the plant with it. Power conversion blocks, transformers, cable trench and switchgear are all sized against the count, so an eight per cent error in capacity is not an eight per cent error in cost. This is also why capacity sizing is better argued per commercial basis than settled with a single number — the same reasoning that applies to levelised cost of storage, where the basis of the calculation decides the answer more than the inputs do.

For buyers assembling a specification, the related questions are covered in our BESS procurement guide, and the supplier-side version — what to ask and what a straight answer sounds like — is in five questions to ask any BESS supplier. If you are costing a project rather than specifying one, the per-MWh and per-kWh price bands are set out on our BESS cost page.

The short version

A 4,915.2 kWh container delivers 4,231.3 kWh AC. That is 86.1% of nameplate, and the two reductions that produce it — 90% depth of discharge, then one-way conversion loss — are both ordinary, both disclosed, and both frequently omitted from the comparison a buyer is actually making.

Ask for the basis. Do the multiplication. It takes a minute and it is the difference between a plant that passes its capacity test and one that does not.

Talk to our team if you would like a configuration sized against an AC-deliverable requirement rather than a nameplate one.