(640g) Plantwide Controller Design for IGCC with Co-Production of Hydrogen and External Steam

An IGCC plant is an assimilation of operating units or subsections which
share characteristics such as tight energy integration, similar process
objectives and/or time scales. These plants, in the past, have been perceived as
a typical chemical process plants rather than conventional power plants. To
provide substantial market for these next-generation plants, operational
flexibility for coproduction of hydrogen (production of chemicals, usage in
fuel-cells etc.) and/or IP/LP-steam (industrial/home heating). With multiple raw
materials (coal, water, air) and numerous operational costs involved (including
logistics), the problem to cope up with overall plant optimization/scheduling
problem becomes overwhelming. From a control layer scenario, the plant
robustness is determined by the range of setpoint values which a controller can
effectively handle. In this study, we attempt to understand the inherent
limitations, which may arise with various combination of abovementioned
"higher-level" setpoint values and demands, many of which could be unrealistic.

This paper extends our previous study¹
on decentralized control of Air Separation?Gas Turbine/Combustor?Gasifier power
loop to now include a Heat Recovery Steam Generator (HRSG) and Steam Turbines.
The controller, based on decentralized model predictive control (MPC), is
designed to meet total power demand, extracted from the sum of gas turbine (GT)
and steam turbine (ST) work. It is worth mentioning that during turndown
conditions, the GT's thermoregulation severly limits the fuel-air flowrate ratio
and hence the extracted work from each of these turbines become mutually
interdependent through GT-exhaust temperature. In addition, the overall control
structure is modified (from the previous study) to accept co-produced hydrogen
and external intermediate-pressure (IP) steam demands as setpoints. It is shown
that plant responses relies heavily on conventional plantwide hierarchical
control layer design, and has numerous lower-level PID regulators placed
whereever "fast" and stabilizing responses are required. A rigorous
pressure-driven dynamic model in Aspen Plus/Dynamics is presented and
performance comparison between a decentralized MPC vs. PID design has been shown
in response to various combination of ramp changes in total power, hydrogen and
IP-steam demand.

References

¹P. Mahapatra and B. Wayne
Bequette. Comparison of Different Controller Designs for Plantwide IGCC Control.
Presented at the 2009 AIChE Annual Meeting, Philadelphia, PA, November, 2009.