Palladium alloy thin films measure hydrogen elegantly, with excellent parameters. But the whole industry faces one honest problem: short-cycle lab data is abundant, while real 10-year oil-immersed field data is scarce.
Before any gas can be measured, it has to be driven out of the oil. Vacuum, headspace or membrane directly sets the error, the speed and the long-term reliability of the reading — and the difference is often overlooked.
A directly oil-immersed hydrogen sensor removes the degassing step — but not the conversion step. Turning activity into a ppm reading still depends on a coefficient tied to oil type and temperature.
Photoacoustic, NDIR, FTIR and TDLAS all fail physically on H2 — hydrogen has no infrared absorption. Yet almost every 9-gas online DGA claims to measure it. Here is where that hydrogen number actually comes from.
A 2026 study finds acetylene does not mix evenly inside a transformer, so the oil sampling point decides whether an online DGA monitor ever sees a developing fault.
Three words appear on almost every DGA sensor datasheet — cross-sensitivity, selectivity and specificity — and only one of them is a number you can check.
Both sensing elements read the same palladium-hydrogen reaction, but through different quantities. What a resistive thin film and a MOS junction are each good at, and how to buy Pd alloy online monitoring hardware that fits the duty.
Hydrogen and oxygen are both invisible to optical DGA. Measuring them properly means not extracting them at all — here is the physics and the engineering.
In 2026, online DGA converges with AI and digital twins: fault diagnosis, deep-learning reviews, vendor platforms. What is real, what is research, how to pilot.
A transformer dissolved hydrogen sensor is now a compliance tool: 2026 regulations set a 450 µL/L hydrogen limit for UHV transformers. Why online H2 sensing is the earliest tripwire.
