As you can see at the attached image of a specimen we used for forging experiments, the type K thermocouples are very tough. We used three thermos inside the specimen and one on the surface for an in-situ measurement of the temperature field starting from room temperature->heating->forming->cooling. The initial cross-section of the specimen was circular. All the thermos did their job throughout the hole measurement! After finishing the measurement we cut them off and repaired them by laser-welding, thus we could reuse them. The Thermo on the surface was also welded to fix it on the correct position. So don`t worry about the thermocouples, they can take a lot!

It is possible to weld or peen the thermocouples into a surface.  Remember, the voltage is generated by the change in temperature, so a small difference in the shorted end generally is not of concern.  

Try this https://www.youtube.com/watch?v=qEp-6sSCkPE min 1-5 walk you thru the resistance weld process.

https://www.youtube.com/watch?v=qEp-6sSCkPE

You have to welt on the surface for correct temperature value.

Yes and check the spacing of the wires

Make grooves on the bottom welding plate, and insert the thermocouple wires inside the groove. Ensure that the wire is present inside the groove, during the welding process.

Sir,

Can we use small magnet to fix thermocouple end sensor on aluminium plate surface to measure temp during fsw?

If practical hit the surface you intend to secure the thermocouple with a semi fine abrasive (60 grit). if your experiment allows you would like to be on bare parent metal. You can wipe area with acetone and dry before attempting spot weld. Depending on parameters you can use GTAW or resistance welder. Should be video in this thread to demonstrate.

@Nikhil B Ram ?Aluminum magnet???