Linear in power and voltage

Method

On each measured curve the AC output is read by linear interpolation in the DC input between the measured points, starting from zero output at the power threshold. Between the curves the output is interpolated linearly in the DC voltage. Below the lowest measured voltage the lowest curve applies, and above the highest the highest curve. With one curve only, that curve is used at every voltage; with no curve, the maximum efficiency applies above the threshold. With exactly two measured curves and a part-load curve, the part-load curve applies at every voltage. An OND file with two test voltages gives its main curve as the part-load curve.

The math

$$P_{out}(P_{dc}, U) = (1 - x_U)\,P_{out,k}(P_{dc}) + x_U\,P_{out,k+1}(P_{dc}),\qquad x_U = \frac{U - U_k}{U_{k+1} - U_k} \tag{1}$$

where

Symbol	Meaning	Unit
$P_{out}$	the AC output	W
$P_{dc}$	the DC input	W
$U$	the array voltage at the inverter terminals	V
$P_{out,k}(P)$	the AC output read on the curve measured at voltage $U_k$	W
$U_k, U_{k+1}$	the two measured voltages that bracket $U$	V
$x_U$	the position of $U$ between them	—

References

1. Driesse, Jain & Harrison 2008, Beyond the curves: modeling the electrical efficiency of photovoltaic inverters, IEEE 33rd Photovoltaic Specialists Conference
2. PVsyst 8 help, Inverter model and operating limits

API

inverter.efficiency_interpolation = bilinear

Symbols follow the symbol registry.