Allocation of Water in waste heat steam generators 1

Are you superheating also in this same generator ? If so, with BFW on shellside, cross flow baffles will be required to allow for superheating the saturated steam.
Though what is say makes sense , there are a few considerations that may work counter to what gains you expect with this alternate configuration :
a) It is typical to allocate the poorer heat transfer mode to the shellside due to its higher flow turbulence - in this case, the waste heat stream in sensible heat transfer mode will be the poorer one. So, at the outset, you would need less total exchange surface with BFW on the tubeside, given that most of the heating duty is for boiling and much less for superheating.
b) The fouling stream should preferably be on tube side to enable cleaning. Cleaning the shellside of scale, corrosion debris is next to impossible. In this case, the cleaner fluid is most likely the waste heat stream ?

If you do a detailed thermal sizing routine, it is possible to verify (a).

There is considerable feedback from industry on mechanical failures with waste heat steam generators and the like on this website; mostly related to differential expansion stresses, and/ or due to poor welding of tube to tubesheet joints -see if these threads are of interest.

Thanks George for your great inputs, considering only LP of High pressure saturated steam (without superheating) then allocating the waste heat stream on tube side makes more sense to me to optimize the design thickness of the shell side in view of the high design temperature requirement. However, at the same time good engineering practice calls to maintain both shell and tube side design temperature the same, apart from the recommended 10/13 ASME SEC VIII rule to avoid tube rupture overpressure protection.

Your point on the leakages from the tube sheets also makes sense as it is less keen to allocate the waste heat stream on the shell side and we would tend to use floating head heat exchangers which can accommodate differential thermal expansion. It is also now clear to me that in case saturated steam is produced, with the high film heat transfer coefficient allocating on tube side would not be an issue so water-steam generated can be on tube side, yet as you correctly explained, when superheating is desired then we would need to put the steam generated/BFW fed on shell side to accommodate the lower heat transfer coefficient. I hope my understanding is aligned with yours

for the corrosive/slurry/fouling service, there is no doubt we have to allocate it on tube side except if it is highly viscous with laminar flow conditions.

"when superheating is desired then we would need to put the steam generated/BFW fed on shell side to accommodate the lower heat transfer coefficient."

Even when superheating is required, it is preferable to have the water steam stream on tubeside. The combination of lower heating duty contribution for superheating plus the fouling risks would indicate water steam to be on tubeside.