Good question….without digging deep into the details, I would provide the following ideas.
1. Small coke sizes are a result of incomplete cracking, especially at the top of the coke bed. Certain types of feeds require more energy to drive the reactions. If the cracking activity decreases, the resulting coke will be less cross linked. We sometimes call this “sandy” coke. Adjusting COT could help.
2. Alternatively, molecules which crack very easily will increase vapor velocity and reduce bonding leading to sandy coke. Has there been a decline in the COT to drum inlet to drum vapor temp ratios with the problem coke?
3. Drastic changes in resid sulfur could indicate a significant change in crude blend. What familys or types of crude do you typically run? Naphthenic, parrafinic, oil sands, synbit, orinoco, arab heavy, etc.
4. Unless you operation the unit in different modes, eg anode vs shot, with different feeds, the decoking and coke handling system should not dramatically effect fines generation when compared to crude slate. That being said, you can have additional fines generation with worn cutting tools, lower cutting water pressure, and more reworking of the coke pile with the handling equipment.