Delayed coker on a Maya vacuum residue (screening)

A delayed coker — the bottom-of-the-barrel unit that turns vacuum residue into gas, naphtha, light and heavy coker gas oil and petroleum coke — run on the heaviest two cuts of a real NOAA ADIOS assay (Maya, EXXON, AD00734: API 21.9, characterized at constant Watson K 11.5, so the residue cuts carry SG 1.03 and 1.06). This is a screening model (public correlation, not a licensor model). Yields come from the Gary & Handwerk Conradson-carbon correlation (coke = 1.6·CCR; gas = 7.8 + 0.144·CCR; naphtha = 11.29 + 0.343·CCR; gas oil the rest), split into LCGO/HCGO and fuel gas/LPG on a published commercial yield pattern. Coker yields are set by licensor correlations regressed on vendor data, and none of that is used here. The CCR of 27.8 wt% is the published value for an Isthmus/Maya vacuum residue (OGJ case study), stated as such. That blend is not this assay, so the feed is an approximation, and it sits near the top of the correlation's 10-29 wt% CCR range. Measured against published yields, the correlation's coke is biased high (mean absolute error ~4.4 wt% over three published points). A bench study of a Maya residue reports ~38 wt% coke where this predicts ~44. Read the slate as an order-of-magnitude estimate. The furnace duty is the enthalpy balance between residue at its feed state and products at the drum temperature (coke priced as graphite), with no heat of cracking unless you give one. Sulfur is distributed 30% to coke, 30% to gas as H2S and 20% to HCGO, which puts coke sulfur below the 1.3-1.6x feed enrichment usually reported. Bounded: no drum cycle, no recycle-ratio or pressure effect unless you state an operating basis, no coke quality (needle vs sponge).