(39f) Use of Process Simulation In Ethylene Plant APC and Dynamic Optimization

Showa Denko owns and
operates a large Ethylene plant in
Oita, Japan.  Depending on the economic environment, plant
objective is a combination of Energy Efficiency and Production Capacity.  In order to achieve these objectives, Showa
Denko had contracted Yamatake and Honeywell to design, configure and commission
APC and Dynamic Optimization on the Ethylene Plant.  Such a project was completed in 2002 and the APC and
Optimization applications have been continuously operational for the last
decade.

The solution at Showa Denko
utilized Multivariable Predictive Control applications that have linear models.  The Dynamic Optimization application inherited
the APC models which maintained complete consistency between APC and
Optimization.  This approach however
required a strategy for handling non-linearities that
are inherent in the operation of an Ethylene plant.   A gain updating strategy that utilized
custom calculations of yield prediction and energy consumption were commissioned. 

Since the initial
commissioning, Showa Denko has continued to enhance the applications.  Additional variables that accounted for all
major energy usage and associated gain scheduling have been added.  The operator workflow has been made effective
by a hierarchy of displays in Japanese, by a continuous display of active
bottlenecks and by daily optimization reports.         

An upgrade
completed in 2010 by Yamatake and Honeywell has converted the custom
calculations for gain updating by standard software that uses a commercially
available Process Simulator.  A benefit
of this approach is that the applications can now be more easily maintained and
enhanced by Showa Denko.  This paper
summarizes the existing APC and Optimization solution and discusses the details
of the gain updating strategy that was recently implemented.