Output divided by input is meaningful only for the documented matched cycle and boundary. The calculator cannot verify those conditions from two numbers.
Define the energy boundary first
Use existing manufacturer monitoring or documented metering records; do not open batteries, panels or live conductors, add instruments or change dispatch settings. Identify timestamps, system, register meanings and whether measurements are at AC system ports or DC battery terminals.
Both totals must cross the same defined boundary. Do not mix AC charging with DC discharging, whole-home import/export or solar energy passing directly to a load. Record whether inverter and auxiliary consumption are inside the boundary; do not add a loss again if it is already included.
Check that the record closes the storage cycle
The NREL research procedure requires matched initial/final SOC for its round-trip calculation. This page is record interpretation, not an instruction to reproduce that test.
Use only a documented comparable initial/final stored state. An identical rounded BMS percentage is not proof of identical stored energy. If the state differs, Ein − Eout includes that change: 0.7 kWh difference could be 0.4 kWh left in storage plus 0.3 kWh elsewhere, not 0.7 kWh of loss. Do not invent a correction from SOC × nameplate capacity or force a deep cycle to obtain a number.
NREL 2017 research preprint: matched-state requirement (English PDF; not a home test instruction or compliance claim) ↗Calculate energy ratios, not capacity
EIA defines round-trip efficiency as the share of supplied storage electricity later retrieved. Its historical utility-scale averages are not a target or acceptance limit for a household system.
Fictional Ein = 5 kWh and Eout = 4.3 kWh give return = 4.3 / 5 × 100 = 86%, difference = 5 − 4.3 = 0.7 kWh and unreturned share = 14%. This is not a measured usable capacity, state of health or runtime estimate. If output exceeds input, check states, sources, meters and interval instead of clipping the result to 100%.
EIA: round-trip efficiency definition (2021 US utility-scale context; not a household benchmark) ↗Keep operating conditions and uncertainty visible
Retain cycle dates/duration, temperatures, load/rate conditions, initial/final state evidence, boundary and meter resolution. Standby time and auxiliary demand belong in the record. A difference near the measurement uncertainty cannot be interpreted as a precise loss or a fault diagnosis.
For several comparable closed cycles sum energies first. A second 10 kWh in / 8 kWh out cycle gives 80%; pooled return is (4.3 + 8) / (5 + 10) × 100 = 82%, not (86 + 80) / 2 = 83%. Keep different boundaries or conditions separate. Do not alter BMS limits, discharge reserves, protections or manufacturer test procedures; this arithmetic does not approve an installation or warranty claim.
A traceable cycle record
| Record | Keep together | Do not substitute |
|---|---|---|
| Boundary | System, AC/DC ports, gross counters, auxiliaries | Net home import/export or mixed ports |
| Cycle | Exact interval and initial/final state evidence | Unmatched days or rounded SOC alone |
| Conditions | Duration, load/rate, temperature, meter limits | A capacity or health certificate |
Worked example
Fictional matched record, not a performance target
A documented comparable-state cycle has 5 kWh entering and 4.3 kWh returning at the same boundary. Values are illustrative, not expected performance of any product.
R = 4.3 / 5 × 100 = 86% U = 5 − 4.3 = 0.7 kWh u = 0.7 / 5 × 100 = 14%
The recorded return is 86%; 0.7 kWh or 14% is unreturned on that boundary. Without the matched-state and boundary evidence, do not call the difference a storage loss.
What to have ready
- Same storage-system boundary and gross registers
- Matching cycle interval and comparable stored states
- AC/DC basis and auxiliary loads documented
- Conditions and measurement limits retained
- No capacity, health, warranty or safety conclusion inferred
Common mistakes
- Mixing AC input and DC output.
- Treating net home meters as storage charge/discharge.
- Calling a change in stored energy a loss.
- Averaging percentages without energy weighting.
- Forcing a battery cycle or changing protections for a reading.
Use and limitations. Sources include US public agencies. Principles are general; regulations and design values must be checked where you live.Worked examples use stated hypothetical inputs.