you are absolutely right. your reply helps me confirm my apprehensions.
the thing is , I cannot use the paint, so is there any other option? other temp. measurement technology?
Rgds
Which brings us right back the original question.
Reflectivity is roughly the inverse of emissivity. So, with varying reflectivities, you would need a sensor with variable emissivity (and some way to determine either the reflectivity, or the emissivity) to get accurate readings.
A special coating (or just about any uniform coating or surface treatment), could be used to fix the emissivity to a specific value, making measurement more accurate.
Of course, that wouldn't be conducive to welding over, but it depends on what you're measuring. Is it possible to get temperature readings from the back side?
Keep this in mind though:
For high values of emissivity (which can range from 0 to 1, so .9 would be a high value), changes in emissivity would register as small changes in temperature readings.
Unfortunately, for low values of emissivity (like say .2, which is reasonable for shiny metal), even tiny changes in emissivity can result in vast changes in temperature readings.
So, the shinier your surface, the harder it will be to collect consistently accurate readings.
As an example:
I've got a Raytek mini-temp thermometer (its a small handheld IR thermometer, with a fixed emissivity of .95, but a maximum temperature reading way to low to use for welding).
If I put two pots on the stove, filled with boiling water, the anodized aluminum put would measure around 205F, but the shiny stainless pot might only read as 120F.
If you have a variable emissivity sensor, and some way to calibrate it, it can work well. You just need a good sample to measure off of, that reflects like the weld you will be reading later, but at a known temperature (room temp is ok, but the closer to working temps, the better).
So, if you could determine the emissivity off of a sample welded part taken from an oven, you can get pretty close to the mark, so long as the real part "looks" the same, and with the cleanliness of TIG, that should be a reasonable assumption.