How do fiber optic pressure sensors work?

It depends on the type of sensor. The most popular is probably the diaphragm design. They are basically just a regular fiber with a diaphragm on the end that is normal to the fiber. Light is sent to the end of a fiber via a fiber coupler and received by a photodiode/phototransistor. As the pressure increases, the diaphragm deforms and causes more light from the fiber to deflect thereby lowering the return intensity/optical path.

Here's a white paper about some commercial ones pressure sensors:
http://www.fiso.com/modules/AxialRealisation/img_repository/files/documents/2007/Miniature%20Fiber%20Optic%20Pressure%20Sensor%20for%20Medical%20Applications%20IABP%20therapy_2005%20OFS-17.pdf [Broken]

 

Old post, but updating in case anyone encounters this.

The previous post seems to describe a sensor based on physical dimming due to a diaphragm reducing light reflected back into the fiber. There may be a sensor like that, but the more common types are based on Fabry-Perot etalon or Mach-Zehnder interferometers, and interference patterns. There is an element in these that is sometimes called a "diaphragm", but it is related to changing the distance light travels on a leg of the interferometer. Pressure changes vary the path length, which cause the light/dark bands of the interference pattern to shift. A photo sensor detects the number of light/dark changes moving past it, and the count is then translated into a pressure.

I remember an optics lab experiment where I used a two leg fiber based Mach-Zehnder interferometer. One fiber leg of the device was exposed to the pressure to be measured, and the other shielded. The pressure changes on the exposed fiber cause internal reflection differences resulting in path differences. I think the F-P devices are more common in industry than the M-Z based, especially where small size is important.