Dynamic fluid pressure on a tug boat rudder

[RBPrice]

Could someone point me toward a source of info regarding the water pressure on a rudder used to steer a commercial tug boat.

The tug suffered a failure of the welded joint between the upper shaft and the pintle shaft which is the tubular pipe that runs down the axis of the rudder. I need some idea of how much pressure was on the rudder to calculate the torsional load. The link shows a photo of the rudder and the failure was at the joint above the plates at the top of the rudder. The rudder dimensions are approx. 65 in. wide and 115 in. tall and you are looking up at the stern of the tug.

Thanks a bunch

Bob Price
A*G*M[/b]

 

[MikeHalloran]

Well, it broke, so you can back-calculate the strength of a joint that wasn't quite up to the task, and work from there.

A Google search on "hydrodynamic force on a rudder", without the quotes, produced a lot of potential reading material.

I'm not sure that just designing for hydrodynamic forces alone would be sufficiently conservative for a _TUG_boat. The condition of the paint on the rudder, especially near the waterline, suggests that the rudder regularly impacts trash, or ice, or both, probably while going forward and while backing and while sitting still at full throttle in a whirlpool of its own making while attempting to move something large and massive.

Also design for fatigue. I have been aboard yachts where the rudder bearings were loose, and/or the rudder servo didn't have enough deadband, so the rudder was wobbling back and forth and whacking the bearings hard enough to feel the shock throughout the after third of the boat. No way is a tugboat going to be taken out of service for a little extra vibration.

Has that tug's propeller been polished manually, or has it been clawing through shallow bottom muck? That's another design load case to consider.



Mike Halloran
Pembroke Pines, FL, USA

 

[RBPrice]

Thanks Mike - the existing welds failed because the design was grossly undersized and poorly executed. In order to design what is needed, knowing the potential torque requires knowing the water pressure on the rudder. Once one knows that, one needs to allow for the fact that the rudder is "balanced" i.e., there is some of the rudder forward of the C.L. of the pintle tube. And, even though the rudder is a simple plate, when it is turned so it is not aligned with the prop wash and the boat, there is a low pressure area on the leeward side of the blade that would tend to increase the unit force. That is what makes airplanes fly. I have already calculated a design that is much better than the one that failed, but it is based on an assumption of the unit pressure on the blade acting on the centroid of the net area times the radius to the C.L. of the pintle tube.

And yes, it failed while the tug was moving through ice so a SF of at least 5 is needed since a failure of the rudder could create a very serious condition for not only the tug but the vessel it was working with or some shore facility nearby.

Without some believable source, I cannot proceed. I was hoping for some sort of reference book but I did not use "hydrodynamic" in my search so will go do that now.

Thanks again.

Bob Price
A*G*M

Bob Price
A*G*M

 

[dhengr]

RBPrice:
I see several different pressure loads on that rudder. One is pulsating and a serious fatigue loading, from the prop wash flow. A fairly high frequency impulsive loading. The other is a velocity pressure from water moving past the rudder as the tug moves through the water. A fairly constant uniform pressure. These are both dependant upon the angle of the rudder as it is not aligned with the beam of the tug, to turn the tug. These load will vary from side to side as the angle of the rudder changes. I don’t have these loads or formulations off the top of my head, but I would start by looking at my Fluid Mechanics text books. Then there are axial loads, shearing loads and moments on the rudder support points. Show us some photos of the failure and the failure area.

 

[LittleInch]

Is that vertical bar at the end of the rudder triangular and presents a flat face to the water flow? I can only imagine the force and disruption to flow given the size of that propeller. I agree that there are a whole lot of forces both static and pulsing going on here. The central tube looks a bit weedy to me...

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[desertfox]

Hi RBPrice

Try this link and look at page 54 on the document and carry on down it might help:-

http://www.shipmotions.nl/DUT/LectureNotes/ShipHydromechanics_Intro.pdf

 

[RBPrice]

I'll take this in pieces - dhenger > I'll process some photos later this morning and you will see how "weedy" things were.

LittleInch - the tube [pintle shaft] is a piece of 8 in. Sch. 80 pipe to which, both blades are welded plus of course, the ribs. The extra plates on the aft end of the rudder, one on each side, are just to reinforce the edge from collision with docks etc.

DesertFox - I will go look at that link.

Things are a bit discombobulated here this morning as we had 14 inches of snow yesterday and since my office is in my home and we have 1,000 ft. of driveway I spent some time clearing it out. But the wind is blowing hard so that will have to be done several more times today. Done with a 1983 Intl. 484 with 50 HP and a 75 inch wide blower. Lots of fun actually.

Thanks for all the help guys

Bob Price
A*G*M

Bob Price
A*G*M

 

[LittleInch]

Thanks - will be interesting. what I meant was that if that vertical stiffening rib at the end is actually a flat edge, then when you have lots of water flowing past it on an angle when the rudder is turned away from straight ahead, it will exert a substantial force trying to pull the rudder away from the mounts and could have a substantial shear force at the top and bottom of the shaft.

8" sch 80 doesn't sound like anywhere near enough to handle the combination of forces to me.

What sort of bollard pull / shaft power are we talking about here?

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[rb1957]

i'd've thought the force on the rudder was the force required to balance the applied moment and the motion of the tug ... sort of like what we do for elevator force to balance the pitch manoeuvre of the plane.

Quando Omni Flunkus Moritati

 

[RBPrice]

Here are some JPG pictures of the failed parts. The square plates are the upper palm plate to which was welded the stub shaft. You can clearly see the very small fillet weld that was between the OD of the stub shaft and the plate and the view of the end of the shaft which was welded to the ID of the hole. Note that there is no evidence of any beveling of the entrance edge of the hole or the end of the stub shaft to add to the weld area attaching the shaft to the palm plate.

The history of the tug Margot is typical in that it has had at least three owners over the last 20 years, much of its service was in salt water and like all tugs there was no real preventive maintenance: it was fixed when something broke. And this is a non-trivial issue. When the rudder failed, the boat was pushing a barge through Jamaica Bay immediately south of JFK airport on the south shore of Long Island east of Manhattan. The boat was towed to Staten Island and placed in a drydock for repairs. The bill came to $84,000 for about four days work. The insurance company is denying the claim on the basis that the bushings which guided the stub shaft were preferentially worn so that the stub shaft could cock along its vertical axis and that eccentric load caused the weld failure and the policy does not cover "normal wear and tear" from normal use.

My view is that the welds failed on the stub shaft palm plate because they were grossly inadequate to the loads and cracks propagate through the welds over the course of at least five years leaving the joint ready to fail which it did when the rudder hit some ice. How do I upload more than one file??

Bob Price
A*G*M

 

[rb1957]

probably can't help sounding like a snit but, i guess now you're learning the value of maintenance costs.



Quando Omni Flunkus Moritati

 

[MintJulep]

Ah so, it's about the insurance.

You need to find some textbook evidence of fatigue failure, and a clear initiation point in the weld - an undercut, lack of fusion, whatever.

Then your argument becomes "This is a clear latent manufacturing quality defect that YOUR INSPECTOR should have found before you issued the policy. Pay up!"

 

[rb1957]

you should've maintained it better ... there's clear evidence of wear, contributing to the problem; nay, causing the problem ... eff off.

look out! incoming bread roll ...

Quando Omni Flunkus Moritati

 

[desertfox]

Hi RBPrice

I would like to see more photographs. If can scan them into your computer and save the scans as a multi paged document, you should then be able to load them on here as a single file.

I have some reservations about your view regarding the failure, firstly you say the welds are not big enough, well how big should they be? Also from your post you say the welds have lasted for at least 5 years, how long were they supposed to last? It doesn't sound like grossly inadequate welds to me, particularly when you see the rust and general state of the components on the photograph as stated by others.
Looking on the photograph it appears to me, that the weld is generally still in place and the area where the crack is, if that's what it is, starts at the edge of the hole where there is no weld and then passes through the weld at right angles to the weld axis. My understanding of fillet weld failure is that no matter how you load them, they generally fail in shear across the throat of the weld, I cannot see any evidence that the weld as failed in shear. See pages 3 and 4 of the link below.

http://www.egr.msu.edu/~harichan/classes/ce405/chap6.pdf

To do an analysis on the strength of the weld you would need the size of fillet weld and what electrode was used, as electrodes come in different strength grades, alternatively the parent metal yield strength might also be a good indicator for a rough analysis. If you have the above information we could estimate what maximum torque loading the weld can withstand as a starter however this would be a static loading and not anything to do with fatigue which is a completely different ball game.
For a fatigue analysis we would need to know the cycling of the rudder and the stresses involved at different attack angles to establish the stress range.

I doubt unfortunately you will convince the insurance company that they should really pay up but I wish you luck.

 

[RBPrice]

Hello desertfox and thanks for your comments -

My conclusion about the welds is based on my assumption of a value for the water pressure on the rudder which worked out to be approx. 200,000 lb-in of torque on the welded joint. I then designed a joint that had a SF of 5 which was drastically more robust than the failed joint. The old joint did not have a chamfer on the entrance edge of the hole in the palm plate, there was no chamfer on the leading edge of the stub shaft and from the photo the fillet weld on the OD of the stub shaft was half the size of mine. I suspect that one aspect of a correct joint would have been the need to mill the mating face of he stub shaft palm plate after welding to remove any distortion.

I am acutely aware that there are a zillion strength and grades of welding wires and electrodes, acutely aware of the difference between hot rolled steel plate and A36 steel plate, know enuf about the problems of stress cracking to be cautious, and used my favorite welding book, The Design of Weldments by Omar Blodgett, the founder of Lincoln Electric as a guide.

The insurance inspector pointed out some nasty cracks in both the palm plate and the stub shaft, found lots of cracks in the stub shaft and in one of his photos, pointed out old cracks (very rusty) and the new crack which was the one that undoubtedly failed when the ice hit the rudder.
I found out today that the on the old joint, the perimeter joint around the outside of the two palm plates was welded together in a misguided effort to strengthen the joint. That was not done on the repaired assembly - see the attached photo.

Bob Price
A*G*M

 

[3DDave]

I am always suspicious of torsion joints carrying the torque via friction. It may be they were welded together because the bolts kept coming loose.

In case of mechanical overload, can the rudder be back-driven until it is against a hard stop? If so, the joint only needs to be that strong. If not, make sure the load carried by the new joint isn't enough to break something more costly.

 

[desertfox]

Hi RBPrice

Thanks for your post, I apologize if I've covered ground that you already know but and its an interesting topic this failure.

Can you load of anymore pictures of the failed parts? I ask because you mention that their are cracks in the stub shaft, would these cracks be at an angle of roughly 45 degrees to the stub shaft axis? If the cracks are angled that might well be fatigue cracks growing under tensile stresses due to torsional loading, see this site for failure modes of a shaft:-

http://www.colorado.edu/engineering/CAS/courses.d/Structures.d/IAST.Lect07.d/IAST.Lect07.pdf

page 7 in the above link.

From the bending moment on the rudder, can, or have you worked backwards to see what the effect this would have had on the actual shaft, welds etc. that have failed and see what margin of safety was used in the original design?

The insurance company seem to have their case wrapped up and I'm not sure what it is you need to show them to convince them to change their minds, other than proving that the weld broke first, which caused the stub shaft to float off vertical but reading the last paragraph of your latest post it seems to me there was numerous cracks around the shaft and palm plates already, prior to the final failure. Have the guide bushings for the stub shaft been measured and inspected for wear by yourself or by the people that repaired it?

regards

desertfox

 

[RBPrice]

Hello 3DDave - the rudder is hydraulically controlled with a single cylinder and there are double limit switches at the stops.

Theoretically, the torque was shared by a half round key that mated with a slot in the lower palm plate but that was really useless. So yes, the torque was taken by the six 1.250 Dia. Gr.8 bolts. Notice that there is no sign of the plates slipping since none of the holes are elongated. Does hat say that the torque was taken by the friction between the two palm plates or by the perimeter weld? The bolts are a close fit in the holes meaning the hole are about 0.002 smaller than the nominal size of the bolt threads. And I have no knowledge of them ever coming loose. But who knows what has transpired since the boat was built in 1958.

But it has just occurred to my addled brain that if I knew the hydraulic pressure and bore of the cylinder and the length of the torque arm, I would know the torque used to move the rudder. Got some emails to send out.



Bob Price
A*G*M

 

[zeusfaber]

Couple of thoughts on that last approach.

Your hydraulic pressure for that calculation is usually defined by the setting of your cross-line reliefs (even on systems where the HPU output is limited to a lower pressure, wave loads may still lift the reliefs)

Transient loads will be higher - there can be a lot of inertia in the tiller.

It looks like you've got external hard stops fitted (which is a bit of a novelty to me). You might want to check whether there are internal hard stops as well - once the sea/ice pushes the tiller onto a hard stop, the torque in the stock isn't going to be limited by the hydraulics any more.

A.

 

[desertfox]

I think its difficult to know whether its the friction between the palm plates or the six bolts that transmitted the torque, for example imagine each bolt central in its bolt hole and tightened up, theoretically then the friction would transmit the torque, of course getting six bolts central would be impossible in assembly terms, so in practice I guess its a combination of both friction and bolt shank assuming of course the key is useless has you stated.

It would be interesting to see what the maximum operating torque was available to operate the rudder as it would also give credence to the 250000 lb.-in torque you calculated for the rudder.

regards

desertfox