Corrosion in the upper portion of the Coker fractionator is normally due to ammonium chloride salts depositing in the tower. If that is the case for your unit, raising the top temperature of the fractionator above the salt point will prevent corrosion in the tower. One of the products from the thermal decomposition reactions is ammonia. Salts in the feed to the unit will be hydrolyzed and generate HCl that will form ammonium chloride salts at a low enough temperature. At typical conditions for the fractionator overhead system, moderate concentrations of chlorides result in relatively high salt points. My normal recommendation is to keep the fractionator top temperature above 250 F (120 C). This assumes that the chloride level in the overhead water is less than 5 ppm. With good desalting in the upstream crude unit, the chlorides in the overhead water of the Coker fractionator will be well below 5 ppm. If the chloride concentrations are higher, then salt point is higher and those ammonium chloride salts are extremely corrosive.
Tramp amines can form amine chloride salts as well. One of the decomposition products of some of the H2S scavengers is MEA (mono ethanol amine) that forms a very corrosive amine chloride salt at lower temperatures than ammonium chloride salts.
If you have ammonium chloride or amine chloride salt corrosion occurring in the Coker fractionator, raise the fractionator top temperature or lower the chlorides in the overhead system by improving desalting in the upstream units. Some crude towers inject an imidazolene based filming amine into the top of the tower to move the salts out of the tower and into the pump around systems. The chemical vendors also sell an ammonium chloride salt dispersant chemical that has had good performance in FCC and other main fractionators. You could consider using one of those, if it is confirmed you have ammonium chloride salt deposition.
For the polymerization, the most common catalyst is oxygen. If the coke drums are incompletely purged of oxygen before being put back on line, that oxygen can polymerize dienes. I like to use steam injected from the bottom of the coke drum over a period of 20 minutes to purge the oxygen out of the drum. The steam purge rate should be high enough to change over the volume of the coke drum six times in that time. With six volume changes, the theoretical oxygen content of the drum will be 0.05% by vol. Typically, only a few ppm oxygen is enough to polymerize dienes. A thorough purge of the coke drums down to concentrations of about 0.1% is normally low enough to prevent the oxygen from the coke drums from being a significant contributor to fouling.
The corrosion products from ammonium chloride salts can polymerize dienes as well, so mitigating the corrosion in the tower is part of resolving the tower fouling problem.