Off-grid sizing fails in a predictable way: someone starts from the roof ("I have space for six panels") instead of the loads ("what actually needs to run, for how long?"). Work the other direction and the numbers fall into place. Here is the five-step method our engineers use.
Step 1 — Audit every load
List each appliance with its power rating and hours of use per day. Multiply to get watt-hours, then sum. Be honest about the things that run invisibly: routers, standby loads, well pumps, chest freezers in summer. A typical energy-conscious household lands between 3 and 8 kWh per day; a cabin with lighting, water pump, fridge and electronics can be under 3 kWh.
Step 2 — Size the battery from daily kWh
Battery capacity (kWh) = daily kWh × autonomy days ÷ (DOD × temperature factor). Use 0.8 DOD for LFP and derate 10–20% for cold climates. For one day of autonomy on 4 kWh/day in a mild climate: 4 × 1 ÷ (0.8 × 0.9) ≈ 5.6 kWh. Two days of autonomy doubles it — decide based on how often overcast stretches happen locally.
Step 3 — Size the array from peak sun hours
PV power (kW) = daily kWh ÷ (peak sun hours × system factor). Use a system factor of 0.65–0.8 to account for temperature, soiling, cable and conversion losses. If your worst-month peak sun hours are 3.5 and the load is 4 kWh: 4 ÷ (3.5 × 0.75) ≈ 1.5 kW of modules. Design for the worst month, not the average, or you will own a generator-shaped problem in December.
Step 4 — Size the inverter from loads, not energy
Add up the power of everything that can run simultaneously, then check surge. Motor loads — pumps, fridges, compressors — can draw 2–3× their rated power for a second at startup. An inverter that handles 3 kW running may need 6 kW surge capacity. Undersizing here produces the classic complaint: "everything works individually, nothing works together."
Step 5 — Sanity-check the pairing
Worked example: small off-grid home
- Loads: fridge 1.2 kWh + pump 0.6 kWh + lighting 0.4 kWh + electronics 0.5 kWh + washing 0.6 kWh = 3.3 kWh/day
- Battery: 3.3 × 1.5 days ÷ (0.8 × 0.9) ≈ 7 kWh LFP
- PV (worst month 4.0 PSH): 3.3 ÷ (4.0 × 0.75) ≈ 1.1 kW → install 1.2–1.5 kW
- Inverter: 2.5 kW continuous, ≥5 kW surge; PV input must accept the array's Voc at record low temperature
Details that separate good systems from returns
- Voc at low temperature: panel open-circuit voltage rises as temperature falls. Check the inverter's max PV voltage against the coldest recorded site temperature, not STC.
- Cable losses: keep DC runs short and thick; a 3% loss on 30 metres is real money over ten years.
- Generator input: a hybrid inverter with AC input turns a bad weather week from an outage into an inconvenience.
- Monitoring: remote telemetry catches a failing string weeks before the customer notices the battery draining early.
Frequently asked questions
How many days of battery autonomy do I need?
One day suits sunny climates with generator backup; two to three days suits cloudy maritime climates. More autonomy is cheap insurance for essentials, but expensive if it forces a much larger PV array.
Can I mix panel orientations on an off-grid system?
Yes — east/west splits actually smooth the daily generation curve, which helps charging before and after noon. Use an inverter with multiple MPPTs or separate strings per orientation.
Why does my system underperform in winter even with sun?
Cold panels produce more voltage but short winter days cut total energy; snow, low sun angle and higher loads compound it. Size the array and battery for the worst month and consider a generator input.
Is a bigger inverter always better?
No. Oversized inverters run inefficiently at light loads and waste standby power. Match continuous rating to realistic simultaneous loads with about 25% headroom.
Sources & further reading
- IEA PVPS Task performance data reports — iea-pvps.org
- NREL solar resource and PVWatts methodology — nrel.gov
中文摘要
离网系统按"负载→电池→组件→逆变器"五步走：先做负载审计得日耗电量；电池容量=日耗电×自持天数÷(DOD×温度系数)；组件功率按最差月份峰值日照时数与 0.65–0.8 系统效率折算；逆变器按"同时运行负载"选型并核对电机类负载 2–3 倍启动浪涌；最后核对低温 Voc、线损与发电机输入。