470383 Design and Implementation of Model Predictive Control Strategies for IGCC Power Plant Cycling with Carbon Capture

Thursday, November 17, 2016: 1:06 PM
Monterey II (Hotel Nikko San Francisco)
Xin He and Fernando V. Lima, Department of Chemical and Biomedical Engineering, West Virginia University, Morgantown, WV

In recent years, as attention for energy and greenhouse gas emissions has raised, technologies for power plants with carbon capture have been studied worldwide. Integrated gasification combined cycle (IGCC) power plants with carbon capture, which have higher efficiencies than traditional pulverized coal-fired power plants, are a promising technology for power generation [1]. Additionally, with the increasing interest and focus on renewable energy sources, fossil-fueled power plants that were designed to operate at base-load conditions might need to cycle their load in the future. In this work, we analyze the implementation of model predictive control (MPC) strategies for base-load and cycling scenarios associated with an IGCC power plant equipped with a water gas shift membrane reactor (WGS-MR) unit for carbon capture. A model for this IGCC power plant represented by a system of differential algebraic equations was previously developed [2, 3]. Using this model, centralized control strategies based on linear and nonlinear MPC algorithms are presented to address different power generation scenarios. Given the computational time challenge for online implementation of nonlinear MPC to address high-dimensional systems, approaches for the MPC computational time reduction are also investigated.

Specifically, the linear MPC control strategy is based on a dynamic matrix control (DMC) method. For this method, a multiple-input multiple-output step response matrix is obtained by performing step tests in the IGCC plant. The proposed nonlinear MPC strategy is an extension of NLMPC [2, 4, 5], which is based on the conversion of the differential algebraic equation system into a large-scale nonlinear programming (NLP) problem. The resulting NLP is solved using IPOPT, an efficient interior point-based large-scale nonlinear optimization algorithm [6]. Also, the ADOL-C [7], an open-source automatic differentiation package, is employed for improved accuracy and speed of the calculations of system derivatives. For additional NLMPC computational improvements, sequential quadratic programming (SQP) algorithms are analyzed for solving the posed MPC NLP problem.

The proposed MPC algorithms are implemented for the IGCC-MR power plant as centralized plantwide control strategies. For such implementations, the selected control structure consists of the following controlled output variables (CVs): carbon capture, power generation, temperatures of cooled syngas and cooled permeate streams, steam to CO ratio at the WGS-MR inlet and hydrogen purity in the permeate stream. Also, the selected manipulated input variables (MVs) are: water flow rates for syngas and permeate cooling, steam injection flow to syngas for the facilitation of the WGS reaction, total coal/water slurry, oxygen enriched air flow for the gasifier, sweep flow for the WGS-MR and air flow for the gas turbine. A number of control scenarios for the IGCC power plant are addressed including: (i) trajectory tracking associated with a power generation demand cycling, in which the IGCC power plant demand follows a certain profile from its original power generation setpoint; (ii) disturbance rejection related to the variability in the quality of the coal feed. In this case, the power generation should still track the cycling profile while the carbon content in the coal slurry is reduced. In this presentation, different NLP solving methods will also be discussed when the nonlinear MPC controllers are implemented for this high-dimensional system. Results on the closed-loop responses for the different scenarios will be analyzed and compared considering the advanced linear and nonlinear MPC control strategies.


  1. F.V. Lima, P. Daoutidis, M. Tsapatsis and J. J. Marano. Modeling and optimization of membrane reactors for carbon capture in Integrated Gasification Combined Cycle units. Ind. Eng. Chem. Res., 51(15):5480–5489, 2012.
  2. F.V. Lima, R. Amrit, M. Tsapatsis and P. Daoutidis. Nonlinear model predictive control of IGCC plants with membrane reactors for carbon capture. In Proceedings of the American Control Conference. Washington, DC, June 2013.
  3. F.V. Lima, P. Daoutidis and M. Tsapatsis. Modeling, optimization, and cost analysis of an IGCC plant with a membrane reactor for carbon capture. AIChE Journal, 62(5):1568-1580, 2016.
  4. F.V. Lima, Xin He, R. Amrit and P. Daoutidis. Advanced control strategies for IGCC plants with membrane reactors for CO2 capture. In Materials and process systems for CO2 capture: modelling, design, control and integration, A.I. Papadopoulos and P. Seferlis (eds.), Wiley, 2016.
  5. R. Amrit, J. B. Rawlings and L. T. Biegler. Optimizing process economics online using model predictive control. Computers and Chemical Engineering, 58:334-343, 2013.
  6. A. Wächter and L.T. Biegler. On the implementation of an interior-point filter line-search algorithm for large-scale nonlinear programming. Mathematical programming, 106(1):25-57, 2006.
  7. A. Walther and A. Griewank. Getting started with ADOL-C. Combinatorial Scientific Computing, 181-202, 2012.

Extended Abstract: File Not Uploaded
See more of this Session: Process Control Applications I
See more of this Group/Topical: Computing and Systems Technology Division