Calculate system losses as a chain
Shows losses as a cascaded chain from STC to the load. The factors combine multiplicatively, not additively — PVGIS prescribes this explicitly. The incorrect additive figure is reported alongside for comparison.
The formulas behind the calculator
Every number above can be recomputed: the full calculation path, all assumptions and the data source with retrieval date — plus cross-validation against independent references. Disclosed, not claimed.
An estimate based on the stated assumptions. The final design must be checked by a qualified professional against the rules that apply where you are.
| Step | Formula | Value | Provenance |
|---|---|---|---|
| System efficiency | Π eta_i | 0.82935 | exact |
| Power after all losses | P_STC * eta_total | 829.35 W | exact |
| Added up instead | 1 - Σ(1 - eta_i) | 0.82 | exact |
- Formula
η_total = Π η_i- Valid for
- Stationary view with annually averaged factors, multiplicatively linked (PVGIS convention). Reference cases: three stages 0.97·0.95·0.90 = 82.94 % (additively wrong: 82.00 %); full chain with defaults 72.7 % (additive: 69.0 %); routed through the battery 65.4 %. Contradictions between controller type and efficiency are named instead of silently computed.
- Not covered
- Load dependence of inverter efficiency across the day, partial shading and the temporal correlation of generation and consumption. Battery round trip is only applied when energy actually passes through the battery — the share-based split by daily profile is computed hourly by the self-consumption calculator. The temperature factor remains a visible assumption; calculator 2 computes it site-specifically for the design month.
Frequently asked questions
How do I correctly calculate the total losses of a PV system?
The stage efficiencies — reflection, soiling, temperature, mismatch, cabling, controller, inverter, degradation — are multiplied, not added: η_total = Π η_i. Instead of "5 % cable + 3 % mismatch + 10 % inverter = 18 %", the correct result is 0.95 · 0.97 · 0.90 = 0.8294, i.e. a 17.06 % loss. PVGIS explicitly prescribes the multiplicative method; the incorrect additive figure is shown alongside for comparison.
Should battery round-trip efficiency always be part of the loss chain?
No — it is only applied when the energy actually passes through the battery. Energy consumed directly during the day does not, and applying the factor across the board is a double-counting error. For the same reason the inverter stage is dropped in DC-only systems. A round-trip value of exactly 1 is physically impossible; typical values are 0.85 to 0.95.
How accurate is the loss chain across a day or a year?
It is a steady-state analysis using annually averaged factors. The load dependence of inverter efficiency over the day, partial shading, and the timing correlation between generation and consumption are not covered. Degradation enters the chain as (1 − annual degradation) raised to the number of years.
Why is the battery not always in the chain?
Because its round-trip loss only occurs when energy actually passes through the battery. Direct daytime consumption (fridge running while the sun shines) bypasses the battery entirely — the “routed through battery” switch adds or removes the stage accordingly. Real systems sit in between; the self-consumption calculator provides the share.
Which stages are typically the biggest levers?
With PWM controllers, the charge controller (up to 20 percentage points); otherwise usually battery round trip (6% for LiFePO₄, 20% for lead) and temperature in summer. Reflection, spectrum and mismatch are individually small (1–3%) but add up. The waterfall marks the two largest losses of your specific system in red — addressing them yields the most, and it is rarely the stage you would guess.
Why is the additive comparison wrong — and why show it anyway?
Losses act multiplicatively on the respective remainder: 0.95 × 0.97 × 0.94 … Adding the percentages systematically overestimates the total loss, because later stages only act on the smaller remainder. The calculator shows both numbers side by side because the additive calculation is the most common mistake in forums and rules of thumb — visibly refuted beats asserted.
How large is the additive error for a full chain really?
With this page's defaults (nine stages): multiplicative 72.7 % system efficiency, additively computed 69.0 % — almost four percentage points apart. With only three stages it is 82.94 % versus 82.00 %. The error grows with the number of stages, which is exactly why the calculator outputs the wrong value as a comparison instead of merely avoiding it.
Why does the calculator warn for PWM with a high efficiency?
Because controller type and efficiency describe the same physics and can contradict each other: a PWM controller sits typically at 70 to 85 %, the 97 % default belongs to MPPT. Switching only the type while leaving the efficiency models a controller that does not exist. Since 5 August 2026 the calculator names this contradiction instead of silently computing with it.
What does the “routed through battery” switch do?
It inserts the battery's round-trip loss into the chain — but only for the energy share that actually passes through the battery. At 90 % round trip the system efficiency drops from 72.7 to 65.4 % in the default case. The associated efficiency is adjustable since 5 August 2026; before that it sat unreachably at 1.0, making the switch do nothing.
Why is the round-trip loss not always applied?
Because it only occurs when energy really passes through the battery. Power consumed directly during the day bypasses it; applying round trip to everything double-counts — the same error as in the runtime calculator, just in the other direction. The honest version is a share-based calculation from the daily profile, which the self-consumption calculator does hourly.
How does degradation enter the chain?
As an additional factor (1 − rate)^years, 0.5 % per year by default. At year 0 the chain shows the new state; set to 20 years the factor drops to 0.905 and the chain shows what the system delivers when aged. Both fields are adjustable — the degradation rate via the form since 5 August 2026 as well.