292334 Use Dynamic Simulation to Reduce Startup Flare Emissions for Ethylene Plants

Tuesday, April 30, 2013: 2:30 PM
Presidio C (Grand Hyatt San Antonio)
Jian Zhang, Meiqian Wang, Jie Fu, Chuanyu Zhao, Ha Dinh, Shujing Zhang, Qiang Xu and Kuyen Li, Dan F. Smith Department of Chemical Engineering, Lamar University, Beaumont, TX

Ethylene plants play an important role in petrochemical industry.  The products of  an ethylene plant, such as ethylene, propylene, butadiene, benzene, toluene and other co-products,  are all basic feedstocks in petrochemical industry.  Ethylene production process is one of the biggest and most complicated processes in petrochemical industry.  Startup of an ethylene plant is also very critical.  A typical startup of an ethylene plant could flare 400,000 - 1,000,000 lbs of materials, and generate 200,000 lbs of CO and VOCs (volatile organic compounds).  Reduction flare emissions will help reducing environmental impact and saving materials.

In this paper, rigorous steady-state and dynamic models of an ethylene plant was built.  The models include CGC (charge gas compressor), chilling train, demethanizer, deethanizer, depropanizer, C2 splitter and C3 splitter sections.  Both steady-state and dynamic models have many details so that they are more close to startup procedure.  For example, performance curves were used in CGC section, so that the section can simulate scenarios on different feed and bypass flowrate.  In chilling train section, information of equipment size and heat capacity is considered for simulating precooling process.  For distillation columns, rigorous distillation model and tray sizing calculation are employed.  Both steady-state and dynamic models are tuned and validated to make them close to real plant operations.

The validated dynamic model was used to simulate startup scenarios.  A startup procedure has dozen of steps, and some steps have several flare reduction options.  Through dynamic simulation, all major steps and options can be simulated and evaluated, and the flare emissions can be calculated.  For example, CGC startup is the most critical step and has largest amount of flare vent, because charge gas has to be continuously produced through furnaces and sent to flare to keep inlet pressure of compressors.  We simulated several scenario cases with charge gas of different flowrate, methane and nitrogen, and found the best scenario, which could largely reduce flare emissions of CGC section.  Another case study is regarding the precooling of chilling train.  A chilling train should not be chilled very fast, otherwise it could be damaged by thermal expansion.  On the other hand, chilling train cannot work properly before chilled to certain temperature.  Through dynamic simulation, several chilling scenarios were simulated.  For each scenario, control strategies are employed to control the chilling rate within a certain range.  Simulation results showed some scenarios could even cool down chilling train without any flare emissions.

Flare calculation showed that the total flare emissions could be greatly reduced through dynamic simulation.  The final scenarios have been acknowledged by ethylene plants.

Extended Abstract: File Not Uploaded