(486m) Dynamic Simulation and Optimization for Startup Operation of Ethylene Oxide Production

Ethylene oxide (EO) is a flammable, colorless, and toxic gas. It is also an important commodity chemical that can be found in the production of solvents, antifreeze, textiles, detergents, adhesives, polyurethane foam, pharmaceuticals, and hospital sterilization of surgical equipment, and etc. The manufacturing of EO involves critical exothermic reactions at high temperature and pressure, whose failure may cause safety problems like runaway reactions, which in turn result in severe air pollutions and tremendous economic loss. Thus, EO production must be strictly controlled under various situations, especially during EO plant startups.

Chemical plant startups are highly nonlinear, complex operations that usually employ multiple auxiliary streams, pipelines and equipments, involve discontinuous, parallel, recycling operating procedures, as well as irregular change of many controller set points. It has been reported 40% of plant accidents occur during plant startups. For EO plant startup, the safety requirement is very serious because it involves critical exothermic reactions at high temperature and pressure. Meanwhile, the reactants of ethylene and oxygen should be well maintained within the safety zone (i.e., Ethylene 25 %, Oxygen 8 %); otherwise, the mixture would explode directly. Note that the explosion not only causes equipment damage and personnel injury/death, but also emits large amounts of toxic EO and highly reactive VOCs, which will cause severe environmental pollution and raw material loss. Therefore, the EO plant startup has significant economic and environmental impacts that need in-depth study. By far, there is still lack of related studies.

In this paper, a general methodology for obtaining optimal startup strategy of EO production has been developed. It employs plant-wide dynamic simulation to virtually check the feasibility and test the safety by systematically examining plant dynamic behaviors and transient response of critical parameters. Based on the validated dynamic simulation model, control strategy has been modified to improve the system startup feasibility, and original plant startup strategy has been further optimized so as to shorten the startup time and smoothen the startup operations. A case study has demonstrated the significant operational and economic benefits of the proposed methodology.