My experience with partial crack repair is at best DISMAL!!!
Even if you open the entire crack and reweld it properly with all the pre and post heat treatments, peening, etc.  the repair may STILL fail again.  Proper repairs of this sort performed by competant personal are the only way to go but they are still NOT 100%---Partial repair is only a stop gap and will result in future failure, possibly greater, than the original.  If , indeed, the existing crack is causing no problems (?) leave it be and keep an eye on it until such time as repair concerns override distortion concerns.(boy, now that's a hillbilly approach if I ever heard one.)

Rod

With a partial crack repair, the heat of welding can cause the crack to enlarge or go deeper.  Extensive preheating may reduce it somewhat, but it is still a gamble.  I'd only try it if you back it up with extensive NDT (angle beam UT or radiograph) to verify what you've got after the repair.

I have carefully read the posts and note that you have stated that "The cracking is associated with original weld repairs of the casting..."  I've got to side with Rod on this one.  I think it will keep cracking, regardless of what you do, full, partial or no repair, because either the casting has manufacturing stresses or defects of which the crack is only a symptom, or the original weld repair introduced them.  If the crack is below critical size, why not leave it alone and monitor it?

Blacksmith

Thanks for the replies and a few comments.  The equipment is operating in the creep range and therefore a full weld repair is itself a temporary repair which should be postponed for as long a possible.  

The cracking has not progressed into the parent material and there is evidence that the crack is dormant at this stage.  

We believe that the crack is below the critical size (currently performing calculations) however we do not know if the crack is below the threshold size at which fatigue crack propagation may occur.  This may become a problem if the operation of the machine changes and it starts to cycle.  There is also a significant cost in monitoring the crack.

The thinking behind the partial weld repair is that the partial repair would reduce the stress intensity at the crack tip and that may bring the overall stress intensity below the fatigue threshold value.  

It was also believed that even if the repair was a failure we would be in no worse position than we currently are and could perform a full repair at a later date.

We can quantify the crack after the repair with NDT for future monitoring.

We have not come to a final decision yet and we may yet leave the crack and monitor it.  However, I am interested in experience which as shown that, for example as betulja mentioned, that the heat of welding has caused the crack to enlarge or that a partial weld repair (properly carried out) has caused a bigger problem than the original.

Well Gordon, you asked for my experience with this sort of repair and I gave it to you based on my limited time in the trade (just a bit under 40 years at retirement).  I will add this---the ultimate cost must be factored into mix here---How much will it cost if it doesn't work and the  part fails completely?

Rod

Why not just grind out the crack to reduce the sress intensity at the crack tip, and leave it at that.  It the material can withstand the imposed stresses with the crack already there, the grinding should not weaken it any, yet could provide extra service life if it keeps the crack from propogating.  It would be easier to monitor this way also, as if the crack is ground out, it will be obvious if it starts to crack again.

The vanadium in the casing enhances the properties of the other alloys. In particular it improves the fatigue resistance of the base metal.Your weld repair may have included the wrong consumables ie welding rods.
There are very specific terms relating to PWHT of CrMoV.
Can your unit be de-rated?

The stress at the root of a flaw or crack is more intense than the nominal applied stress.  This stress intensity is called K, the stress intensity factor.  The basic relationship between stress intensity factor, nominal stress and flaw size proposed by Irwan is:
                   K=ó*(Pi*a)sqrt
                     
    At some value of stress or at some length of crack, the crack will become unstable and propagate quickly until the component fractures or the remaining cross section is reduced to the point that rupture occurs.  That critical stress intensity is called the fracture toughness.  It is of special note that fracture toughness is a material property just like Young’s modulus or density.  The roman numeral subscript ‘I’ designates mode I fracture which is the crack opening mode.  Mode I fracture is the most prevalent mode of fracture and is the only mode of concern in the discussion.
                  Kic= ód*(pi*Ac)sqrt

Where: Kic is the critical stress intensity or fracture toughness,           
          (material property),
       ód is the design stress normal to the direction of crack opening,
       Ac is the critical flaw size

If the flaw size is less than critical and the stresses are below allowable, grind out flaw and leave it alone.
 
see
H.L. Ewalds, R.J.H. Wanhill, Fracture Mechanics, Delft, The Netherlands, 1984, pp.23
Richard W. Hertzberg, Deformation and Fracture Mechanics of Engineering Materials, 2nd Ed, John Whiley & Sons, 1983, pp.304

I would second the motion to grind out the crack to reduce the stress concentration.  So long as the man with the grinder doesn't remove excessive amounts of material, contouring the area to reduce the stress concentration can only help.

The reference I made earlier with regard to partial repair spreading cracks was in reference to 316 SS that had chloride stress cracks.  Your material and cause of cracking is different, so my experience may or may not be relevant.

There is a second risk with a "partial" repair which is associated with a mechanism known as strain age embrittlement. I don't know if your alloy is susceptible to this but in those which are, the strain and temperature cycle imposed on the buried crack during the welding operation reduces the toughness of the material at the crack tip and what is now a safe situation may change dramatically. The Wells Wide Plate Tests which were carried out at the Welding Institute in Cambridge back in the 1960's and 70's were performed to explain a failure in just such a situation on a pressure vessel where a partial repair was done.

Thanks again for all the replies.

Boo1, the idea for the partial weld repair did come from looking at the fundamental equation for K.  

As you have written, in general terms (ignoring geometry factors)
                   K=ó*(Pi*a)sqrt
If you look at the formulae for K for an embedded crack, the value of crack length 'a' used is one half of the length of the embedded crack.  So all things being equal, it is better to have an embedded crack than a surface breaking crack of the same size.

Also, the stress in the formulae above is the net stress normal to the crack face.  As the weld pool cools, the thermal contraction of the weld metal will tend to close the crack, and after welding, there will be a net residual stress.  When system stresses are applied to the body, the net stresses for K calculations at the deepest point of the crack will be the sum of the system stresses (causing crack opening) and residual stresses (causing crack closure).  The sum of the two will be less than the system stresses alone.

Therefore, as I see it, from a fracture mechanics viewpoint, a partial repair reduces the value of K at the crack tip and is preferable.  However, from a materials point of view, this may not be so.  The strain age embrittlement mentioned by shooter45 is associated primarily with carbon steels that have been strained (cold worked) and then subjected to elevated temperatures.  Therefore, for a cold worked carbon steel, a partial weld repair may reduce the fracture toughness of the material and is therefore not desirable.  In this instance however, the material is a low alloy CrMoV casting that has been normalised and is not susceptible to strain age embrittlement.  

Calculations have shown that the crack is tolerable and we will monitor it in future. We have not undertaken a partial weld repair due to the surrounding uncertainty, rather than an understanding of the factors involved.  I am sure this situation will arise again and any additional comments will be greatly appreciated.

Thanks for all the discussion.

Consider shot peening the crack and adjacent material to induce a compressive stress, as the crack tip will only propagate under tension.

Gordon
My experience in this field suggest if you are going to partially reweld your crack, counter sink a hole at both ends this will stop the crack from chasing.
When it comes to the welding side make sure you weld your repair with stringer beads keeping them as small as possable, 2.5mm eletrodes.( Do not weave, as this will have the opposite effect to what you are trying to achieve)
This will keep the crystallisation very small in effect the weld will be stronger,large crystalls create weak stress lines hope my imput is of some help, regards.

Has anyone thought of contacting Metalock for a non welding solution to the problem?
See : Metalstitch repair service
      www.inplace.com