Once you have a robust shafting/bearing system design, you must still align it properly. The level of accuracy required is +/- 0.1 mm outside the stern tube bearing and +/- 0.05 mm within the stern tube bearing. This is not being achieved.

In spite of the fact that laser alignment and optical procedures such as Taylor-Hobson have been around for more than a decade, the yards, with Class approval, still depend on piano wire for alignment of the stern tube. The shaft alignment proper is done by hydraulic jack-up tests which involve measuring the load on a jack's dial near the intermediate bearing and engine coupling flange. This procedure has at least four glaring limitations:
  1. Jack up tests can only measure what's happening at the intermediate bearing and the engine coupling. But what we are really interested in is what's happening in the stern tube. In particular, jack up tests tell us nearly nothing about the all important pressure distribution within the aft bearing.
  2. Jack up tests are completely static. No jack measurement can tell us what happens as the shaft revolves. But all the interesting shaft phenomena including the bearing oil film pressure only come into play with the shaft rotating.
  3. Jack up tests are limited to a couple of conditions. Currently, the Rules require only three conditions: in dock, cold along side at extremely low draft, and one condition with the engine hot. The first two conditions are totally artificial and the last is the RPM = 0 "operating" condition. We have absolutely no information on how the bearing reaction forces change with draft or major local load changes such as aft peak tank full, aft peak tank empty.
  4. Jack up tests are notoriously inaccurate and unrepeatable. Due to measurement errors and hysteresis effects, the allowed error range is +/- 20%. In many cases, the error range is larger than the design margins. The yard could use electronic load cells and reduce the hysteresis effect markedly, but the other limitations would still remain.
These limitations are totally unnecessary. For at least the last twenty years, naval vessels have had their shafts aligned by dynamic strain gauging. The results are far more accurate, can cover almost all the important system variables, and can be taken with the shaft operating for a large range of conditions including turning. The down ratchet doesn't always operated by weakening the Rules. Sometimes it works by preventing an obvious improvement to the Rules from being implemented. Any Class that attempts to impose a new procedure which might possibly force an improvement in standards -- no matter how obvious -- will run into exactly the same factors that make the down ratchet function in the first place. (Ever wonder why the Finite Element models didn't get better as the cost of computation went down by a factor of 100,000?) It just doesn't happen.
Given the mini-epidemic of shaft bearing failures and the availability of dynamic strain gauging, we recommend that the Rules require that shaft alignment be checked by the following instrumentation: With such a system,
  1. The propeller forces and bending moments are measured directly. There is no need make any assumptions about the propeller forces.
  2. The displacement, shear force and bending moment of the shaft at both ends of the aft bearing are known. This should be enough to make a simple but reasonable model of the pressure distribution within the bearing.
  3. One can directly measure the shear force and bending moment at the main engine coupling through its complete firing order. This is critical to main engine journal bearing failures inter alia.
  4. These measurements can be taken continuously even while maneuvering and over a variety of drafts, loading conditions, and RPM.
  5. Finally one can leave the system in place and monitor the shaft in actual operation. (The cost of the gauges, the wireless transmitters, the signal conditioners is less than $20,000.) If the situation starts to deteriorate, there a good chance this system will pick it up before the actual failure.
Once this system become standard and routine, the additional cost relative to the current system will be negligible. Shaft alignment will go smoother and quicker. There will be fewer redos. The quality of the alignment will be drastically improved. And the data that will be generated will materially improve our understanding of this critical issue.