Modeling of an Industrial Hydrotreating Process for Diesel Production under Uncertainty

Hydrotreating is the process by which impurities are removed from gas-oil fractions by reacting those specific impurities with hydrogen over a CoMo or NiMo catalyst under normal operating conditions:  290-370°C, 14-128 bars. The gas-oil fractions contain many impurities (e.g., sulfur, nitrogen, olefins, and aromatics) that produce gases with negative environmental impact when combusted, or decrease diesel engine performance. These input fractions are characterized using detailed fingerprint information from experiments; two of which come from normal and sweet crude oils and the remaining from refinery outputs (i.e., fluidized catalytic cracker, coker, and vis-breaker). A reactor model of the diesel hydrotreatment process can be simulated using the programming capabilities of MATLAB software; the MATLAB code used in our studies is a replicate model of a commercially used diesel hydrotreater catalyst monitoring program.  The goal of this research is to analyze the current feed definition and reaction kinetics by completing uncertainty quantification analysis. This goal will be achieved by implementing a Monte Carlo algorithm for the stochastic simulation of the error in analytical measurements used for the feed definition process (e.g., density, wt% atomic sulfur, ppm atomic nitrogen, bromine number, and the ASTM D86 distillation curve). Key outputs are the overall conversions of species containing sulfur, nitrogen, aromatic carbon, and olefinic carbon atoms. Our analysis will reveal the impact of the instrument errors of the inputs on key output species conversions utilized during hydrotreater performance monitoring.