Which Hundred Megawatt-Hours Did You Buy?

Nameplate, Usable, Guaranteed and Delivered Energy in a Battery Storage System — Why the Measurement Point and the Definition Matter More Than the Percentages, and How to Compare Bids That Are Not Comparable

1. Executive Summary

A battery energy storage system marked one hundred megawatt-hours does not deliver one hundred megawatt-hours to the grid. That much is well understood, and the loss chain — state-of-charge limits, degradation, temperature, conversion losses, auxiliary consumption — appears in every explainer on the subject.

The explainers are directionally right and operationally useless, for three reasons. They present a single stack of percentages as though every project shares it. They apply deductions that, depending on how the vendor defines its nameplate, may already be inside the number being deducted from. And they treat the answer as an engineering estimate when it is a contractual definition: the only figure that matters commercially is the energy the supplier guarantees, at a named measurement point, at named conditions, under a named duty cycle, in a named contract year.

That distinction is not pedantry. Two suppliers can both offer one hundred megawatt-hours and mean numbers that differ by twenty percent, because one is quoting installed direct-current energy at beginning of life and the other is quoting guaranteed alternating-current energy at the point of interconnection in year ten, net of auxiliary load. Comparing the headline figures compares nothing. Worse, an owner who normalises the bids incorrectly can select the more expensive system while believing the opposite.

This paper works through what each capacity definition actually means, how the measurement point changes every number downstream of it, where the common loss stack double counts, why degradation is a duty-dependent curve rather than a percentage, how augmentation and oversizing trade against each other, what temperature genuinely affects, the difference between a one-way conversion loss and round-trip efficiency, the auxiliary load metering boundary that decides who pays for the parasitic energy, and the throughput limits buried in every warranty. It closes with a bid normalisation checklist, a testing and verification section, and a twenty-five question FAQ.

The sentence to carry into every procurement

Do not compare nameplate. Compare guaranteed energy, at the contract measurement point, at reference temperature, under the specified duty cycle, in a specified contract year, net of auxiliary load, with the availability guarantee stated separately.

If a bid cannot be expressed in those terms, that is itself the finding.

2. The Question Is Which Hundred Megawatt-Hours

The industry uses the same unit for at least six different quantities, and vendors are not being deceptive when they use different ones — they are quoting the number their product is naturally described by. The problem is that the buyer receives them all as though they were the same measurement.



An integrator quoting installed direct-current energy is describing the sum of the cell capacities as manufactured. A supplier quoting usable energy is describing what the battery management system will actually permit to be withdrawn, at beginning of life, under stated conditions. A supplier quoting guaranteed energy is making a commercial commitment about what will be available after a specified period of operation under a specified duty. Each is a legitimate description. None of them is the others.

This is the same class of problem as comparing a solar plant on module nameplate rather than on modelled annual production, except that storage has more definitional layers and less standardised practice. It is why the first question in any storage technical review should be what basis the number is quoted on — and why the answer is often "we would need to ask the vendor."

3. Six Different Capacity Numbers

The single most useful discipline an owner can adopt is to require every capacity figure in every document — proposals, datasheets, models, the interconnection application, the offtake agreement — to carry its basis in the same sentence. A number without a basis is not a specification.

4. Where You Measure Changes the Answer

Every deduction in the loss chain sits between two physical points, and naming those points is what makes the deduction meaningful. A guarantee expressed without a measurement point is unenforceable in any useful sense, because the supplier and the owner will measure in different places and both will be right.

The provision that resolves most disputes

State the guarantee at the revenue metering point, and state explicitly whether auxiliary consumption is inside or outside that boundary.

Everything else — conversion efficiency, transformer losses, cable losses — then falls inside the supplier’s scope to manage, which is where the incentive belongs.

5. The Stack, Corrected

The familiar waterfall runs nameplate, then state-of-charge limits, then degradation, then temperature, then conversion, then auxiliary. Reordered and defined, it is a useful frame; taken literally, it produces wrong answers in two ways.

The first error is double counting. If the quoted nameplate is already usable energy — which it frequently is for modern lithium iron phosphate products — then applying a further state-of-charge deduction subtracts the same reservation twice. The second error is treating separable effects as sequential multipliers. Temperature does not reduce usable energy by a fixed percentage in a thermally managed system; it increases auxiliary consumption, and at extremes it limits available power and accelerates degradation. Deducting it as a flat energy haircut and then separately deducting auxiliary load counts part of the same physics in two places.

A defensible framing separates the chain into three questions rather than one stack.

• How much energy can be withdrawn from the battery, in this contract year, under this duty cycle and at this temperature? That is the battery question, and it combines the operating window, the degradation curve, and the conditions.
• How much of that reaches the metering point? That is the electrical question — conversion, transformers, cables — and it is a straightforward efficiency calculation at the actual operating power.
• How much is consumed by the plant itself, and on which side of the meter? That is the auxiliary question, and it is as much a commercial boundary as an engineering one.



Answering those three separately is more work than multiplying six percentages, and it produces a number that survives contact with a vendor.

6. State-of-Charge Limits and Double Counting

Cells are not operated across their full theoretical range. Charging to the absolute upper limit and discharging to the absolute lower limit accelerates degradation sharply and, at the top of the range, raises safety risk. Every system therefore operates within a window managed by the battery management system.

What varies between products is where that window is applied and whether the buyer can see it. Some suppliers quote installed energy and separately state the usable window. Others quote usable energy directly, with the reservation already taken. Both are honest; they are simply different bases. An owner who applies a standard percentage deduction to a quoted number without establishing which basis it is on will be wrong in one direction or the other, and the error is of the same order as the differences between competing bids.



Two further reservations are worth asking about explicitly, because they sit inside the usable window and reduce what is actually dispatchable. Systems providing frequency response or reserve services need state-of-charge headroom in both directions to respond, which is an operational reservation the energy model must reflect. And some control strategies reserve capacity for state-of-charge management, so that the system can return to a target state between events. Neither appears on a datasheet.

7. Degradation Is a Curve, Not a Percentage

Presenting degradation as a single percentage is the most consequential simplification in the popular treatment, because degradation is the variable that determines whether the asset performs over its life and it is the one most sensitive to how the asset is used.

Capacity fade has two components. Calendar fade proceeds with time and is driven principally by temperature and by the state of charge at which the cells are held. Cycle fade proceeds with energy throughput and is driven by depth of discharge, rate, and again temperature. The two are not additive in a simple way, and the resulting curve is characteristically steep in the first year and shallower thereafter.

The consequence for procurement is that a degradation figure is meaningless without the duty cycle it assumes. A guarantee premised on one full cycle per day at moderate temperature describes a different asset from the same equipment operated at two cycles per day in a hot climate, and the second case will not be covered by the first warranty. Every lithium warranty contains throughput or cycle limits, and exceeding them is the most common way an owner discovers the guarantee no longer applies.

Questions that separate a real guarantee from a marketing figure

What duty cycle does the guaranteed capacity assume — cycles per year, depth of discharge, average and maximum rate, temperature profile?

What are the annual and cumulative throughput limits, and what happens commercially if operation exceeds them?

Is the guarantee a capacity level in each year, or a single end-of-term level? Is it measured, or calculated from a model?

What is the remedy if a measured test falls short — augmentation, liquidated damages, or replacement — and on whose schedule?

8. Augmentation Versus Oversizing

Because capacity falls with time, holding a contracted capacity for a twenty-year term requires either starting with surplus or adding capacity later. Both strategies are legitimate and they have different engineering and commercial profiles.

Oversizing installs more capacity at the outset than the contract requires, so that the degraded capacity still meets the obligation in the final year. It is simple, it avoids future construction, and it means paying at the start for capacity that sits unused for years — and, because the cells are held at a state of charge and temperature the whole time, calendar fade proceeds on equipment that is not earning.

Augmentation adds capacity at planned intervals. It defers capital, it captures falling battery prices, and it introduces obligations the design must anticipate: physical space and foundations for future enclosures, electrical capacity in the collector system and auxiliary supply, thermal and fire separation distances that satisfy the applicable requirements, control system provision for mixed vintages, and — the one most often missed — the technical consequences of connecting new cells alongside aged ones with different impedance and capacity characteristics.

Whichever strategy is chosen, it should be visible in the electrical design from the first drawing set. Retrofitting augmentation capacity into a plant laid out without it is expensive, and discovering at year seven that the collector system has no spare positions is a familiar outcome.

9. Temperature: What It Actually Affects

Temperature appears in the popular loss stack as a flat energy deduction, which is the least accurate way to represent it. In a thermally managed system its effects are real but they act through different mechanisms.

• Auxiliary consumption. The thermal management system is the dominant auxiliary load, and its energy use is a strong function of ambient temperature and of the heat the cells generate while cycling. A hot site with heavy cycling can consume several times the auxiliary energy of a mild site with light cycling.
• Available power. At temperature extremes the management system derates charge and discharge power to protect the cells. That affects the ability to deliver rated power, which for a duration-based product is equivalent to affecting the deliverable energy in the window it matters.
• Degradation rate. Sustained high cell temperature accelerates both calendar and cycle fade, so temperature enters the capacity question through the degradation curve rather than as an independent deduction.
• Capacity measurement. Measured capacity varies with temperature, which is why a capacity test procedure must state the reference temperature and the correction method. A test performed on a cool morning and a test performed on a hot afternoon are not comparable results.

The design consequences follow: thermal management capacity sized for the actual site extremes rather than a standard rating, auxiliary supply sized for the resulting load, and the energy model reflecting the site’s temperature profile rather than a reference condition.

10. Conversion Losses and Round-Trip Efficiency

The popular stack shows a conversion loss on discharge. That is correct as far as it goes and it is not the number that matters commercially, because energy has to be put in before it can be taken out.

Round-trip efficiency is the ratio of energy delivered on discharge to energy taken on charge, and where it is measured determines what it includes. A direct-current round trip through the cells is high. Adding two passes through the power conversion system, two passes through the inverter step-up transformer and the collector system, and the auxiliary energy consumed during charge, discharge and idle produces a considerably lower figure at the revenue meter. For contemporary lithium iron phosphate systems with liquid thermal management, alternating-current round-trip efficiency measured at the point of interconnection typically lands well below the direct-current figure, and the gap is where the arbitrage margin goes.

Three details make published efficiency figures difficult to compare.

• The measurement point, as above.
• The operating power. Converter efficiency varies with loading and is usually quoted near its optimum. A system that spends its life at partial load does not achieve the quoted figure.
• Whether auxiliary energy is included, and whether idle auxiliary consumption between cycles is counted. For a system that cycles rarely, idle auxiliary energy can be a large fraction of throughput even though it is small in absolute terms.

For an owner, the practical requirement is to specify round-trip efficiency at the revenue meter, at a stated power level and duty cycle, including auxiliary consumption — and to require it to be demonstrated by test rather than warranted on a datasheet.

11. Auxiliary Load and the Metering Boundary

Auxiliary load is thermal management, the battery management system, controls, communications, fire detection and suppression, lighting, and any pumps or fans in the plant. Its magnitude depends on climate and duty, and its treatment depends on a decision that is commercial as much as electrical.

Where auxiliary load is supplied from the plant side of the revenue meter, it reduces the energy that reaches the meter and it is the supplier’s and owner’s problem in the way the loss stack implies. Where it is supplied from a separate station service arrangement on the utility side, it appears as a load rather than as a reduction in delivered energy, and it is billed at retail rather than netted against wholesale sales. The two arrangements produce materially different economics and materially different guarantee language, and the choice is frequently made by default during design.



Two further practical points deserve attention. Auxiliary consumption continues when the plant is not cycling, so a system dispatched rarely still consumes; the energy model must include idle consumption over the full year rather than as a percentage of throughput. And auxiliary supply reliability is a plant availability issue — loss of thermal management does not merely reduce efficiency, it takes the plant offline, which makes the auxiliary supply arrangement and its backup a design decision rather than an afterthought.

12. Duty Cycle, Rate, and Throughput Limits

The duty cycle is the specification that ties everything above together, and it is frequently the least defined element of a storage procurement.

Discharge rate affects deliverable energy. A system discharged at a high rate delivers somewhat less than the same system discharged slowly, through internal resistance losses and voltage effects; the magnitude is modest for lithium iron phosphate but it is not zero, and a duration guarantee is only meaningful at a stated power. Rate also affects heat generation and therefore auxiliary consumption and degradation.

Cycle count and depth of discharge drive cycle fade. Two full cycles a day is not twice the degradation of one, and shallow cycling is disproportionately gentler than the throughput alone suggests. That is why warranties are expressed with both an annual cycle or throughput limit and a depth-of-discharge basis.

The failure mode is a mismatch between the duty the asset is contracted to perform and the duty the warranty assumes. A system procured for daily energy shifting and then dispatched heavily for ancillary services may accumulate throughput far faster than the warranty contemplated. The remedy is to define the intended duty in the technical specification, to require the guarantee to be premised on it, and to monitor actual throughput against the warranted limits from day one rather than discovering the divergence at a claim.

13. Availability and Energy Over a Year

Everything to this point concerns what the system can deliver when it is working. Delivered energy over a period also depends on whether it is working, and availability is a separate guarantee with its own definitions and its own room for ambiguity.

The questions that matter are what counts as available — full capacity, or any capacity at all — how partial availability is treated when some enclosures are out of service, whether scheduled maintenance is excluded, how force majeure and grid-caused outages are handled, and what the remedy is. A system with a high energy guarantee and a weak availability guarantee can deliver less than one with the reverse.



For an owner’s model, the useful discipline is to build delivered energy from the bottom up — capacity in the relevant year, at the metering point, net of auxiliary, multiplied by the dispatch profile, multiplied by availability — rather than applying a single loss factor to nameplate. The bottom-up build makes each assumption visible and each one negotiable.

14. What to Compare Between Bids

Bids arrive on different bases. Normalising them is a defined exercise, and the following are the parameters that must be fixed before any comparison means anything.

1. Capacity basis: installed, usable at beginning of life, or guaranteed — stated explicitly for every number in every document.
2. Measurement point: battery terminals, converter terminals, or revenue meter, with auxiliary treatment stated.
3. Contract year: capacity in year one, in the middle year, and in the final year, not a single figure.
4. Duty cycle: cycles per year, depth of discharge, charge and discharge rates, and the temperature profile assumed.
5. Throughput limits: annual and cumulative, with the consequence of exceeding them.
6. Round-trip efficiency: at the revenue meter, at a stated power, including or excluding auxiliary and stated which.
7. Auxiliary load: expected annual consumption at the site’s temperature profile and the intended duty, and which side of the meter supplies it.
8. Capacity maintenance strategy: oversizing, augmentation, or a combination, with augmentation scope, timing, and cost stated and with the design provisions it requires identified.
9. Availability guarantee: definition, exclusions, measurement, and remedy.
10. Test regime: acceptance test at commercial operation and periodic tests thereafter, with procedure, reference conditions, correction method, and remedy for a shortfall.



Expressed on a common basis, the comparison becomes cost per guaranteed megawatt-hour delivered at the metering point in the relevant years — and the ranking frequently changes from the ranking on headline price.

15. Testing and Verification

A guarantee that is never tested is a marketing statement with a signature on it. The verification regime should be specified before contract, not negotiated afterwards.

• Capacity test. A discharge from the defined full state to the defined empty state at a stated power, measured at the contract point, with ambient and cell temperature recorded and a defined correction to reference conditions. The procedure, including how full and empty are determined, must be agreed in advance because it materially affects the result.
• Round-trip efficiency test. A complete charge and discharge measured at the contract point, with auxiliary energy measured over the whole test period including any idle time, at a stated power.
• Auxiliary consumption measurement. Separately metered and recorded across a range of ambient conditions and duty states, because a single measurement on a mild day tells you nothing about the annual figure.
• Response and power tests. Verification that the system can deliver rated power across the state-of-charge range and at the temperature extremes, since a capacity test alone does not demonstrate it.
• Periodic re-test. An annual or biennial capacity test on the same procedure, which is what converts a degradation guarantee from an assertion into an obligation with a measurement behind it.
• Data retention. The test records are the evidence for every subsequent claim, and they are also the baseline against which a later dispute about duty cycle or throughput is resolved.



Recognised measurement and test methods exist for storage system performance parameters and should be referenced in the specification rather than each party proposing its own. Where a project-specific procedure is necessary, it should be written and agreed during procurement, when both parties still have leverage.

16. Reading the Graphic Correctly

The lesson at the bottom is the right one

Never comparing on nameplate alone, and comparing instead on energy guaranteed at the point of connection, is exactly correct and is the most useful sentence in any of these summaries. The rest of this paper is an expansion of it.

The percentages are not a template

The specific figures in any illustrative stack are one plausible set for one configuration. Applying them to a different product, climate, or duty produces a number with false precision. The value of the stack is the list of things to ask about, not the arithmetic.

Some of the deductions may already be inside the number

If the quoted nameplate is usable energy — common for current lithium iron phosphate products — then applying a further state-of-charge deduction subtracts the same reservation twice. Establish the basis before deducting anything.

Temperature is not a flat energy haircut

In a thermally managed system, temperature acts mainly through auxiliary consumption, through power derating at extremes, and through the degradation rate. Deducting it as an independent energy percentage and then separately deducting auxiliary load counts part of the same effect twice.

Conversion loss is not round-trip efficiency

The stack shows a one-way discharge loss. The commercially meaningful figure is round-trip efficiency at the revenue meter including auxiliary energy, which is a different and lower number, and it is where arbitrage margin is actually won or lost.

Degradation is not a number

It is a curve, its shape depends on the duty cycle, and the contractual question is what capacity is guaranteed in which year under which duty, with what throughput limits and what remedy. A single percentage cannot express that.

Two things are missing entirely

The measurement point, which changes every figure downstream of it and is the most common source of bid comparison error; and availability, which determines delivered energy over a period regardless of how much the system can deliver when running.

17. Keentel Electrical Power Engineering Services

Keentel Engineering supports storage projects from technical specification through interconnection and commissioning, on the engineering questions that decide whether the asset performs as procured.

17.1 Owner’s Engineer and Technical Due Diligence

• Bid normalisation: restating competing proposals on a common capacity basis, measurement point, duty cycle and contract year so that they can actually be compared.
• Technical specification development covering capacity basis, guarantee structure, measurement point, duty cycle, throughput limits, auxiliary boundary, augmentation strategy, and test regime.
• Performance guarantee and warranty technical review, including the assumptions the guarantee is premised on and the design provisions the capacity maintenance strategy requires.
• Design review of EPC and vendor submittals and QA/QC of third-party study and model packages.

17.2 Electrical Design

• Collector system, medium-voltage distribution, and point-of-interconnection design with electrical provision for the augmentation strategy built in from the first drawing set.
• Auxiliary power system design and load assessment across the site temperature profile and intended duty, including supply arrangement, metering boundary, and backup.
• Transformer and conversion equipment specification including converter-duty considerations and loss evaluation.
• Grounding, bonding, and lightning protection design, and substation design.

17.3 Power System Studies and Interconnection

• Interconnection application support, study-phase technical packages, and coordination with the utility, transmission provider, and system operator.
• Short-circuit, protective coordination and selectivity, and arc-flash studies with inverter-based resources represented as current-limited sources.
• Load flow, reactive capability, harmonic and power quality studies, grid strength assessment, and transient stability analysis.
• Electromagnetic transient modelling for control interaction, weak-grid stability, ride-through verification, and switching and overvoltage transients.
• Effective grounding and ground fault overvoltage assessment at the point of interconnection.

17.4 Commissioning, Compliance, and Performance

• Commissioning specification and test procedure development including capacity, round-trip efficiency, auxiliary consumption, and power capability tests, with reference conditions and correction methods defined.
• Witness testing, performance test evaluation, and independent assessment where measured results fall short of guarantee.
• NERC compliance support including model verification and validation, ride-through and disturbance monitoring obligations, and protection and control coordination.

Keentel Engineering holds a Florida Certificate of Authorization and maintains offices in Tampa, Austin, Sacramento, and Baltimore, supporting projects across the interconnections.

18. Frequently Asked Questions

References and Further Reading

The following are referenced by subject in the body of this document. The current published edition of each standard governs its own requirements, and the specific manufacturer data and contract terms govern any project decision.

Performance, Parameters, and Testing

• IEC 62933-2-1, Electrical energy storage systems — Unit parameters and testing methods, and the related parts of the IEC 62933 series covering terminology, planning, safety and environmental requirements  —  International Electrotechnical Commission
  https://webstore.iec.ch/
• IEEE Std 1679, Recommended Practice for the Characterization and Evaluation of Energy Storage Technologies in Stationary Applications, and IEEE Std 1679.1 for lithium-based energy storage  —  IEEE Standards Association
  https://standards.ieee.org/
• IEEE Std 2030.2.1, Guide for Design, Operation, and Maintenance of Battery Energy Storage Systems, Both Stationary and Mobile, and Applications Integrated with Electric Power Systems  —  IEEE Standards Association
  https://standards.ieee.org/
• US Department of Energy protocol for uniformly measuring and expressing the performance of energy storage systems, developed through the national laboratories — a widely referenced basis for capacity, efficiency and response testing  —  US Department of Energy / Sandia National Laboratories
  https://www.sandia.gov/ess/

Safety and Installation

• UL 9540, Standard for Energy Storage Systems and Equipment, and UL 9540A, Test Method for Evaluating Thermal Runaway Fire Propagation in Battery Energy Storage Systems  —  UL Standards & Engagement
  https://www.shopulstandards.com/
• NFPA 855, Standard for the Installation of Stationary Energy Storage Systems — including separation and spacing requirements relevant to augmentation planning  —  National Fire Protection Association
  https://www.nfpa.org/
• NFPA 70, National Electrical Code — including the articles addressing energy storage systems and their disconnecting means, overcurrent protection and working space  —  National Fire Protection Association
  https://www.nfpa.org/

Interconnection and Grid Performance

• IEEE Std 2800, Standard for Interconnection and Interoperability of Inverter-Based Resources Interconnecting with Associated Transmission Electric Power Systems  —  IEEE Standards Association
  https://standards.ieee.org/ieee/2800/10453/
• IEEE Std 1547, Standard for Interconnection and Interoperability of Distributed Energy Resources with Associated Electric Power Systems Interfaces, where the installation is distribution-connected  —  IEEE Standards Association
  https://standards.ieee.org/
• IEEE Std 519, Standard for Harmonic Control in Electric Power Systems, and IEEE Std 3002.8 for harmonic studies  —  IEEE Standards Association
  https://standards.ieee.org/
• NERC Reliability Standards — including model verification and validation obligations and the ride-through and disturbance monitoring requirements applicable to inverter-based resources  —  North American Electric Reliability Corporation
  https://www.nerc.com/pa/Stand/Pages/ReliabilityStandards.aspx

Notice and Disclaimer

This document is original technical content prepared by Keentel Engineering LLC for general professional information. It is not project-specific engineering advice, commercial advice, or legal advice, and it does not constitute a technical specification, a performance assessment, or a contract review for any project. Capacity, efficiency, degradation and availability commitments must be established from the specific manufacturer data and the specific contract terms applicable to a project.

Capacity definitions, loss mechanisms, and comparative relationships described here are general engineering discussion. Any numerical relationships used are illustrative and demonstrate the method rather than describing any product. Actual performance varies with cell chemistry, product design, thermal management, site conditions, duty cycle, and system configuration, and must be taken from the manufacturer data and test results for the equipment under consideration.



Keentel Engineering LLC is an independent engineering consultancy. Reference to any standard, code, industry organisation, regulator, or equipment category in this document does not imply affiliation with, endorsement by, or sponsorship from any such organisation or manufacturer.