Panic Mode asked the right question (almost)...
- What, exactly, are you trying to measure?
- How do you plan to make the measurment? (This is not about what sensor.)
- What kind of precision are you looking for?
If you are simply trying to measure the width of the pipe, then, depending on the required precision, you should be able to get away with using a simple analog ranging laser. However, one side of the pipe has to be up against a hard-stop when it is measured. The hard-stop would be a known distance from the laser.
As in...
1. A pipe moves in for measurement.
2. The pipe is clamped against a hard-stop.
3. A measurement is taken.
- The distance between the laser and the hard-stop would be the base-distance.
- The distance between the laser and the clamped pipe would be the over-all offset.
- The difference between the two would be the width of the pipe.
Depending on the nature of the pipes (curves, bends, and what-not) and the precision needed, it might not be necessary to stop the pipe for the measurement. If both the clamp and the hard-stop consist of a set of rollers, you could, possibly, measure the width of the pipe "on-the-fly".
The hard-stop rollers would have to be hard, as in, very hard rubber, or hard plastic, or perhaps steel. The clamp rollers could be softer but they need to provide the necessary friction to keep the pipe moving.
The sensor could be self-adjusting (self-zeroing, self-taring) to accomodate for wear on the hard-stop rollers. Between pipes, the sensor reads the distance to the hard-stop roller - that is the base distance.
It would really help if you gave a better description of what you are really trying to measure.
EXCUUUUSE MEEE! (I'm off on a Tangent now...)
I always find it very disappointing when a poorly described problem is posed, and then a wide range of suggestions are made ("shot-gun" style), and then, after a re-post or two, the original poster disappears without answering questions and without describing the "solution" he has decided to use, and why.
THE single biggest reason for poorly designed process-control is poorly defined problems. EVERY SINGLE STEP of developing a process is a PROBLEM! And if a problem is poorly defined, the resulting "solution" will be equally poor. Unless, of course, one makes a very general query and someone just happens to "make a guess" and come up with a workable solution - not necessarily THE appropriate solution, but alas, a solution that might sorta-kinda work, now and then. However, for whatever reason, it appears to satisfy the original poster.
Those kind of people are "Takers"... not "Givers".
Prince... in all the years you have been coming to this site, I don't recall you being much of a "Giver" (contributor)... you are primarily a "Taker". And you keep coming back, over and over again.
"Coming back" is fine! Not a problem! But... golly, gee, darn-it... at the very least, can't you PLEASE provide some feedback?
Yes, I understand it has only been one day. But... your record is on record!