Coal Oxycombustion Power Plant Optimization using First Principles and Surrogate Boiler Models

Numeric optimization has successfully been applied in the chemical industry to improve process performance, reduce emissions, create advanced control systems and intelligently explore process design alternatives. Regarding power plants, optimization methods provide a systematic approach to balance trade-offs when designing efficient and minimal cost carbon capture systems. Coal oxycombustion power plants are an ideal candidate for optimization, given the complex trade-offs regarding flue gas recycle strategies, heat integration and design of cryogenic separation systems. Many oxycombustion studies consider sensitivity analysis instead of multi-variable optimization, which neglect consequentially important interactions between subsystems. Furthermore, these studies rely on over- simplified boiler models available in commercial process simulators. In this paper we present a hybrid 1D/3D boiler model that balances accuracy and computational expense, making it well suited for optimization studies of an entire coal power plant. This model is then incorporated into an equation-based optimization framework for power plants. Trust region optimization methods are used to ensure convergence. Finally a proof-of-concept case study is presented, and future work is discussed.