Inmarsat uses geostationary satellites so if the signals were received by them then triangulation based on signal strength would be impossible at that distance
Agreed that the "triangulation" that would be used would likely
not be based on signal strength because the difference in power attenuation would be miniscule between satellites, the atmospheric fade would be different, and because power is measured not to an adequate level of precision. But
time can be and is measured to exquisite levels of precision. Differences in time of receipt of signal from satellites is how GPS works. To the extent that location can be inferred, it would be inferred from the difference in time that each satellite received the signal from the aircraft. But unlikely that more than two satellites received a signal, so you would not be working with triangulation, which can resolve to a point on a plane. You would be working with biangulation (is that a word?), which can only resolve to a line. But that would still be extremely useful.