The load audit is the foundation of every off-grid design
A load audit converts a list of appliances into a single daily energy figure in watt-hours. It is the first calculation in any stand-alone power system because every other component is sized from it: the battery bank holds a multiple of it, the array replaces it each day, the charge controller passes the current that replaces it, and the cable carries that current. An audit that is 30% low produces a system that is 30% short in every dimension at once, and the failure appears in the third week of December when nothing can be fixed cheaply.
Two different quantities come out of the audit, and confusing them causes most sizing errors. Energy, in watt-hours per day, is what the battery and array must supply, and it is the product of power and time. Power, in watts, is what the inverter must deliver at a given instant, and time has nothing to do with it. A 1,500 W microwave used for six minutes contributes 150 Wh — trivial energy — but demands 1,500 W of inverter while it runs. A 5 W router contributes 120 Wh a day, which is more energy than the microwave, while demanding almost no power at all.
The third quantity is surge. Induction motors draw several times their running current for the first second or two while the rotor accelerates, and an inverter that cannot supply it will simply fault rather than start the pump. Surge is a power figure that lasts long enough to matter and briefly enough that it never appears in the energy total.
Off-grid design is unforgiving about all three because there is no grid to absorb an error. On a grid-tied house an underestimate just means a slightly higher bill; off-grid it means the lights go out. Feed the result of this audit into the off-grid battery bank amp-hour calculator to size storage.
How the arithmetic works, and where the traps are
For each load, multiply running watts by the hours it actually runs. Add the AC loads together, divide by the inverter's efficiency to get what leaves the battery, then add the DC loads, which bypass the inverter entirely and cost exactly what they consume.
Hours means equivalent full-power hours, not hours switched on. This is the single largest source of error in DIY audits. A refrigerator is plugged in 24 hours a day, but the compressor runs perhaps a third of the time, so its equivalent full-power run time is around 8 hours. Entering 24 overstates its energy threefold. The same applies to any thermostatically controlled load: freezers, water heaters, block heaters, engine-driven pumps. If you can, measure with a plug-in energy meter over a full day and enter the measured watt-hours divided by the running wattage.
Inverter efficiency divides, it does not multiply. To deliver 2,220 Wh to AC loads at 90% efficiency, the battery must give up 2,220 ÷ 0.9 = 2,467 Wh. Multiplying by 0.9 gives 1,998, which understates the draw by 19% and quietly undersizes both the battery and the array. Use an average efficiency for your load profile rather than the datasheet peak: inverters are least efficient at very light loads, and an off-grid house spends most of the night there.
Standby draws add up faster than anything else because they run continuously. A router, a satellite receiver, an inverter's own no-load draw and a couple of always-on chargers can total 40 W. Over 24 hours that is 960 Wh a day, which for a small cabin can be a third of the total energy budget and can cost more in battery and array than the appliances themselves.
Peak power is a worst-case sum of everything that could run at once. Being conservative here is cheap in a design and expensive in a failure: an inverter that trips when the kettle and the pump coincide is not a fixable software problem.
The growth margin is not padding. Off-grid systems reliably acquire loads after commissioning — a second freezer, a starlink terminal, a workshop tool. Fifteen to twenty-five percent is normal practice, and it is much cheaper to allow for it in the initial design than to add a parallel string later.
Worked example: a small off-grid cabin
A cabin has a 150 W AC refrigerator running the equivalent of 8 hours a day, 60 W of LED lighting for 5 hours, 120 W of electronics for 6 hours, a 700 W water pump for 45 minutes, 300 W of miscellaneous AC loads for 2 hours, and a 30 W DC load running continuously. The inverter averages 90% efficiency and the owner allows a 20% growth margin. The pump surges to three times its running current.
- Refrigeration. 150 × 8 = 1,200 Wh.
- Lighting. 60 × 5 = 300 Wh.
- Electronics. 120 × 6 = 720 Wh.
- Pump. 700 × 0.75 = 525 Wh.
- Other AC. 300 × 2 = 600 Wh.
- Total AC at the outlets. 1,200 + 300 + 720 + 525 + 600 = 3,345 Wh.
- AC drawn from the battery. 3,345 ÷ 0.90 = 3,716.7 Wh.
- DC loads. 30 × 24 = 720 Wh, taken straight from the battery.
- Total daily battery draw. 3,716.7 + 720 = 4,436.7 Wh.
- Design load with 20% margin. 4,436.7 × 1.20 = 5,324 Wh a day.
- Monthly consumption. 4,436.7 × 365 ÷ 12 ÷ 1,000 = 134.9 kWh.
- Peak continuous power. 150 + 60 + 120 + 700 + 300 + 30 = 1,360 W.
- Peak surge. 1,360 + 700 × (3 − 1) = 2,760 W.
Two things stand out. First, the 30 W DC load contributes 720 Wh — more than the lighting and the pump combined — purely because it never switches off. Second, the pump contributes only 525 Wh of energy but 700 W of the peak and 1,400 W of the surge, so it drives the inverter selection almost single-handedly while barely touching the battery size. Those two loads pull the design in opposite directions, and neither would be visible if you looked only at the daily total.
Carrying the design figure forward: 5,324 Wh a day at 48 V with two days of autonomy, 50% depth of discharge and a 90% inverter needs 5,324 × 2 ÷ (48 × 0.5 × 0.9) = 493 Ah of lead-acid. That is the direct consequence of this audit, and it is why an error here is expensive.
How to read the numbers
Use the design load, not the audited load, for sizing. The audited figure is your best estimate of today; the design figure carries the margin that makes the system survive tomorrow. Size the battery and array from the design load.
Look at the share table before you look at the total. The category taking the largest share is where a design change pays. On most small off-grid systems that is refrigeration or a continuous DC parasitic, not the appliances people worry about. Replacing an AC fridge with an efficient DC compressor model can remove a quarter of the daily energy and lets you shrink both the bank and the array.
Check the peak and the surge against a specific inverter, not a category. Inverter datasheets quote continuous power, a 30-minute rating, and a surge rating with a duration. Your peak continuous figure must sit inside the continuous rating with margin, and your surge must sit inside the surge rating for the time the motor takes to start.
Run a worst-month case as well as an average one. Lighting hours roughly double between June and December, heating loads appear, and solar production falls at the same time. An audit built on annual averages consistently understates the December problem, which is the one that sizes the system.
Compare your total with your grid bill if you have one. An average US household uses on the order of 30 kWh a day, which is far beyond what a practical off-grid system carries. Off-grid living generally means a load in the 2-10 kWh a day range, achieved by choosing efficient appliances and by heating, cooking and heating water with fuel rather than electricity.
Typical running power and daily energy for common off-grid loads
| Load | Running power (W) | Equivalent hours/day | Energy (Wh/day) | Notes |
|---|---|---|---|---|
| AC refrigerator, 18-20 cu ft | 150 | 8 | 1,200 | Compressor cycles; hours are equivalent full-power |
| DC compressor fridge, 12 V, 3 cu ft | 45 | 8 | 360 | Bypasses the inverter entirely |
| Chest freezer | 120 | 7 | 840 | Better insulated than an upright |
| LED lighting, whole cabin | 60 | 5 | 300 | Doubles in midwinter |
| Laptop | 50 | 6 | 300 | Charger losses included |
| Router and modem | 12 | 24 | 288 | Continuous draw |
| Satellite internet terminal | 50 | 24 | 1,200 | Often the largest single load in a cabin |
| Television, 50 in LED | 90 | 4 | 360 | |
| Water pressure pump, 1/2 hp | 500 | 0.5 | 250 | Surges to 1,500-3,000 W on start |
| Well pump, 1 hp submersible | 1,000 | 1 | 1,000 | Surges to 3,000-6,000 W |
| Microwave, 1,000 W output | 1,500 | 0.2 | 300 | High power, low energy |
| Washing machine, cold fill | 500 | 0.7 | 350 | Hot fill adds enormously |
| Electric kettle | 1,500 | 0.2 | 300 | Consider a stovetop kettle instead |
| Inverter no-load draw | 20 | 24 | 480 | Use the standby mode if fitted |
Running powers are representative nameplate values for common equipment, not measurements of any particular model. Electric space heating, electric water heating, electric ranges and EV charging are omitted deliberately: their daily energy is normally beyond what an off-grid system can carry, and they are better served by propane, wood or a generator.
Audit mistakes and what they cost
- Entering hours plugged in instead of equivalent run hours. Overstates cycling loads like fridges and freezers by a factor of two or three, and inflates the whole system to match.
- Multiplying by inverter efficiency instead of dividing. Understates the battery draw by about 19% at 90% efficiency, in the direction that leaves you short.
- Ignoring standby and parasitic loads. Forty watts of always-on equipment is 960 Wh a day, which on a small system can be a third of the budget.
- Forgetting the inverter's own no-load consumption. Many inverters draw 10-30 W simply being switched on. Over a year that is a real amount of battery and array.
- Sizing the inverter to the daily energy. Energy and power are different quantities. A modest daily total can still contain a 3,000 W surge that decides the inverter.
- Averaging summer and winter. Lighting, heating and pumping all peak when the sun is weakest. Run a December case as well.
- Leaving out the loads you intend to add. The growth margin exists for this. Off-grid systems that are easy to expand are rare and expensive; margin at the design stage is cheap.
What to do with the number once you have it
The design load feeds three downstream calculations in order. Battery capacity comes first: multiply the daily draw by your days of autonomy and divide by depth of discharge, inverter efficiency and the temperature factor. That is exactly what the battery bank amp-hour calculator does, and it needs the figure from this page as its input.
Array size comes second, and off-grid sizing is not the same as grid-tied sizing. A grid-tied array is sized to annual energy because the grid absorbs the seasonal mismatch. An off-grid array must replace the daily load in the worst month, with enough surplus to recover the autonomy it just used and to keep lead-acid fully charged. Use the worst-month peak sun hours, not the annual average, in the solar output calculator, and expect the answer to be considerably larger than a grid-tied design for the same consumption.
Inverter and charge controller come third and are power calculations. The inverter must carry the peak continuous figure with margin and the surge for its duration. The charge controller must handle the array's short-circuit current with the multipliers the National Electrical Code requires for photovoltaic circuits.
Finally, revisit the audit before you buy anything. Reducing the load is nearly always cheaper than generating and storing it: every watt-hour you remove takes with it a share of battery, array, controller and cable. On small systems the payback on an efficient refrigerator or on eliminating a 50 W continuous draw beats the payback on any component upgrade you can name. If you are on the grid and only need outage coverage, the home battery backup sizing calculator is the right tool instead.