Heliene–Caelux (November 2026): tandem PV procurement sensitivity#

Business decision: On an area-constrained PV project, how much additional upfront installed cost per module square metre could a higher-efficiency tandem module justify? Could uncertain long-term degradation reverse an initial efficiency advantage?

This is a technology-procurement sensitivity screen, not investment-grade validation or a warranty.

Primary evidence, provenance and data status#

Claim / input

Value

Evidence class

Source / qualification

Press-release date

10 November 2026

Dated announcement

Heliene/Caelux GlobeNewswire announcement; PV Tech corroboration

Tandem efficiency

Up to 27%+

Manufacturer-reported module STC efficiency

This analysis uses 27.0% as a scenario, not an independently certified floor, observed power yield or warranted efficiency.

Energy gain versus N-type silicon

Up to 20%

Manufacturer claim

Not a measured annual yield, a guaranteed uplift or a verified scenario input.

Delivery schedule

2027

Manufacturer projection

Production announced in 2026; operating fleet evidence not supplied.

Silicon module efficiency

22.5%

Illustrative assumption

Selected for a controlled STC power-density contrast (27/22.5 = 1.2); not a measured competitor or market mean.

Installed module area

10,000 m²

Illustrative assumption

Equal surface for both modules; module area, not roof parcel area.

Effective POA irradiation

1,600 kWh/m²/y

Illustrative assumption

Not a sampled location, live API, TMY, P50, or ground measurement.

AC delivery multiplier

0.80

Illustrative assumption

Shared constant annual losses, not a full PVWatts run or a measured PR.

Silicon annual decline

0.5%

Literature-inspired assumption

Historical mixed-PV median from Jordan and Kurtz (2013), not a specific supplier warranty.

Tandem decline

0.5%, 1.5%, 3%, 5%/y

Hypothetical stress scenarios

No known probability distribution or validated Heliene/Caelux durability curve.

Energy value / discount

$0.04–0.12 per kWh / 8% real

Illustrative business inputs

Not a signed PPA, live price, 2026 quoted procurement cost or project WACC.

Operating horizon / incremental annual O&M

25 years / $0 per m²

Model assumptions

Differential maintenance, insurance and replacements excluded.

No generation, degradation, price, output, or failure dataset is claimed to have been measured or downloaded. The single manufacturer-sourced numerical modeling anchor is the announced up-to 27%+ STC module efficiency.

Reusable method and physical/economic baseline#

Source: the first-order area × effective-irradiation × STC-efficiency energy balance is inspired by, not equivalent to, Dobos (2014), PVWatts Version 5 Manual, NREL/TP-6A20-62641. PVWatts itself models time-resolved effective irradiance, module temperature and inverter conversion that this annual balance omits.

For module area A, annual effective plane-of-array irradiation H, STC module efficiency η, uniform assumed delivered-energy multiplier f, degradation fraction d and project year t (one-indexed):

\[ E_1 = A H \eta f, \qquad E_t = E_1 (1-d)^{t-1}. \]

For real power price p, annual operating cost O and discount rate r, the discounted operating value is

\[ PV = \sum_{t=1}^{T} \frac{pE_t - O_t}{(1+r)^t}. \]

The break-even incremental upfront installed-cost premium per module m², relative to the silicon reference, is

\[ C_{\mathrm{extra,max}}/A = (PV_{\mathrm{tandem}} - PV_{\mathrm{silicon}})/A. \]

For any specified extra initial cost C, the incremental investment NPV is PV(tandem) − PV(silicon) − C. A negative break-even premium means the tandem scenario is worth less even if its upfront cost were identical. This framework assumes the same area, effective irradiation, energy sale price, financial horizon and discount convention. It does not estimate a tandem module’s future observed power or guarantee a 20% annual advantage.

The implementation lives in energy_methods.tandem_pv_screen. The new gallery example generates deterministic annual-yield curves and an energy-value-versus-degradation break-even chart; these are sensitivity curves, not statistical confidence intervals.

Reproduction#

uv sync --locked --all-packages --python 3.13
uv run --no-sync pytest methods/tests/test_tandem_pv_screen.py
MPLBACKEND=Agg uv run --no-sync python examples/analysis/plot_heliene_caelux_tandem_2026.py
uv run --no-sync ruff format --check .
uv run --no-sync ruff check .
uv run --no-sync ty check
uv run --no-sync sphinx-build -W --keep-going -b html docs docs/_build/html

No API credentials or live provider calls are required.

Validation boundaries / next measurements#

Module efficiencies measured at STC cannot establish a product’s annual field performance. Particularly for perovskite/silicon tandems, spectral and thermal response, cell mismatch, inverter clipping, illumination stability, early degradation and failure modes matter; project-specific price, warranty, manufacturing yield and O&M also matter. Without certification, independent outdoor data and product quotation, this case is a business screen, not an investment recommendation or independent validation of the manufacturer’s press-release claims.

Next evidence: obtain datasheets and independent STC certification, actual tandem spectra/temperature coefficients, outdoor monitoring by site, metered P50/P90 AC energy, procurement and warranty quotes, and risk-aware financing. Only then calibrate a project-specific life-cycle economic comparison.