Chris, thanks for the above. You always have some amazingly good information. I appreciate that.

Bickford suggest that the measurment method has an accuracy of +/-3%. No mention if the errors are attributed to reading the instrument, measurement error, the calibration of the instrument, the parallelism of the faces being measured, a combination or something else. If you practice good method and you have a good feel, I believe one should be able to achieve +/- 3%. Just my 2 cents.

The problem with the angle method, in my opinion, is getting an accurate indication for the angle. If you can get past that, then you should be OK.

The only experience I have in ultrasonic measurement is that once, I was looking at different methods for improving pre-load accuracy for a reactor flange, and I considered a company that offered that service. They mentioned that they needed to calibrate the machine for pre-load. They could apparently get close with just the diameter, length and material type, but for best accuracy, they wanted us to measure the stretch so that they could zero in on that corresponding preload. I didn't get to actually work with them and shortly I got laid off so, incomplete project.

The plant I currently work for has a Siemens steam turbine. The turbine casing uses that system you mention, Walt, a heater rod is used to get the bolt to X degrees. The service engineer uses a 1/8" rod that is inserted into a hole on the side of the nut. Using only a couple of fingers, if the nut turns, then it's ready and is removed by hand. If it doesn't, they heat it a little more until it does.

Alfonso,

I would generally agrre that with practice you can improve typical figures. I was only trying to suggest typical variations.

I am sure yiou woauld agree that when you carry out an Analysis of Variance for any skill based technique the most significant variable will always be the operator.

Many years ago I did some work on looking at the scatter of Yield Point determination on manually operated tensile testing machines and comparing them with a new generation of computer controlled machines and the differences were staggering.

Help to sell a bagful of new test machines.

I think a modern torque wrench with an angle plate is not too bad and should give reasonable results with -as you say- practice.

I must say that if you could get at both ends of the bolts I just use stretch and not ultrasound but on large plant maybe they want a traceable record and some form of printout that is easier with this equipment.

I used to be involved with the design and manufacture of on line Torque Measuring systems fitted to Steam and Gas Turbine driven process compressors (for around 9 years). We installed pahse shift torque measuring systems on the machine couplings and designed all the guarding and coupling spacers. The largest ones werfitted to Frame 9 Gas Turbines.

We also used to make high speed rotating machinery test rigs from 60MW to 250 000rpm - Clearly not at the same time!!

Yes, practice makes perfect, so they say


Yes, physical measurement is my preference, but there were other things I needed to consider such as there were 200 - 1" bolts, 25 ft diameter vessel. Other factors too, but not worth going into on this forum. Even though ultrasonic may not be the most accurate, for a large job my perception of its advantages are significant accuracy improvement over the torque method and significant time efficiency over physical measurement. Perhaps a happy median?

I have not worked on a Frame 9, that's on beautiful monster I have not gotten to see in person, but have worked on a frame 7, did the coupling from compressor rotor to the 3 stage turbine wheel. We used stretch.