Your question is really 2 questions? The first being the ability of the WWTP to remove BOD, COD, etc., from the refinery effluent before it enters the Bay. This requires a good appreciation for the design of the WWTP to adjust to presumably an increase in reduced carbon without an increase in N or P, or other trace nutrients. There should be much in the literature on this. The second question is the calculated changes on the water quality in Havana Bay, based on the change in the input load (i.e., will there be an improvement in the Bay?). There are good (or at least reasonable models for Bays (San Francisco Bay, Chesapeake Bay, which are much larger), and you could first look at these models to see if you can develop one for Havana Bay - knowing that reaction rates will be faster due to the warmer temperatures. You could start here:

https://www.chesapeakebay.net/what/programs/modeling

or here: https://www.sfei.org/sites/default/files/biblio_files/SEP_Report_Final.pdf

Have a loot at Envirosim's Petwin program. If you cannot afford it you could look at using the underlying model and coding your own solution - which may cost more than buying the program! - by Julia Baker ( https://macsphere.mcmaster.ca/handle/11375/8752 )

Or try this - this approach will work with WRc's STOAT, which is a free download. (I work on STOAT)

Article Modelling of an Oil Refinery Wastewater Treatment Plant