From the way you mention that empty casings have also failed and that the problem appears to be at the welded section between conductor lengths suggests that you have a fatigue failure at the welds. Do you have access to a fatigue analysis that considers both the cyclic stresses which occurred during driving and those generated by wave loads on the conductor. If you dont have a fatigue analysis in place, I would suggest you get one performed to look to quantify the issue, please contact me if you need more information on this.
If you have the driving records for the conductors, they are also worth reviewing, if they conductors have been near refusal towards the end of the driving process then you can generate a high cyclic stress from the last period of hammering, all of which is passed through the marine section which will then see the wave loads.
For fatigue related rather than corrosion issues at the welds, you may need to install some form of clamp which provides local support at the problem location and acts as a load transfer mechanism from one joint to another taking the load away from the weld. These can be either bolted, welded or grouted in place.
If you have any future conductors you plan to install, possibly consider the use of premium threaded connections rather than welding the conductor. The installation will be faster and any conductor to connector weld will have been performed in a factory rather than at site and will have a far better fatigue performance.
Please let me know if you have any questions on the above and I will look to answer.
Regards,
Stewart