(96g) Flare Minimization Via Dynamic Simulation

Flaring in chemical process industry (CPI) is an important method to protect equipment and personnel safety during the process upset and the chemical plant turnaround.  However, excessive flaring emits lots of carbon dioxide as well as air pollutants, which cause negative environmental and social impacts. Besides, excessive flaring reduces company profits due to the tremendous material and energy losses which should be minimized.

CPI has set up flare minimization as a goal in their daily operations. Root-cause-analysis (RCA) has been used in chemical plants to identify causes of major flaring; but RCA does not offer a method for flare minimization. Up to now, flare minimization in chemical plant depends almost exclusively on experienced operators, engineers, administrators, and a well planning, scheduling, and training in the plant. But this is not enough for CPI today because of increasingly strict environmental regulations and economic competitions.  

This presentation describes a general methodology we have developed for flare minimization through plant-wide dynamic simulation. The methodology starts with establishment and fine-tuning of plant-wide steady-state and dynamic simulation models. The fine-tuned dynamic simulation model is then used to virtually run and examine the startup operating procedures for the plant. If an unstable or a dangerous situation is identified, the plant operational procedures will be modified and then checked again with the dynamic simulation model. This modification-and-check can be repeated until optimal operating procedures with realistic constrains are concluded. The dynamic simulation provided an insight into the process dynamic behaviors, which is crucial for a plant to minimize the flaring; meanwhile, maintain the operating safety.

The effectiveness of the developed methodology has been demonstrated by field tests in ethylene plants. Two types of startup procedures, with and without total recycles, were tested. Results of unstable conditions were identified and dynamic responses of new startup procedures were examined. Three field test cases will be presented.