(488b) State Estimation of Energy Integrated Systems with Time-Scale Multiplicity

Energy integration requires dynamic interactions between different process systems units that give rise to challenges associated with the process operation. In particular, process integration results in models with time-scale multiplicity that are characterized by stiff systems of differential equations [1, 2]. For example, networks with large energy recycle, when compared to input/output energy flows, result in a system in which the individual units operate in a fast time scale as opposed to the whole network that operates in a slow one [3]. In the past, control strategies to address time-scale multiplicity have been studied and applied to several process systems [1-3]. However, state estimation approaches for systems with time-scale multiplicity are scarce [4]. In this presentation, we introduce a state estimation framework for energy integrated systems with multiple time scales. The developed framework relies on the time-scale separation of the system, in which state estimators are implemented to match the system dynamics associated with each of the time scales. For example, recursive and optimization-based estimators are investigated for the fast and slow time scales, respectively.

To demonstrate the effectiveness of the proposed framework, two case studies with time-scale multiplicity are addressed: (i) a chemical reactor system [5] as a motivating example. In particular, this example illustrates the challenges on the implementation of previously proposed state estimation techniques for low-dimensional systems with time-scale multiplicity; and (ii) an energy-integrated system that corresponds to a reactor-feed effluent heat exchanger (FEHE) network [2]. In addition to the presence of multiple time scales, the FEHE system is also high dimensional with more than 2,000 state variables, which introduces additional challenges to the proposed approach. To address these challenges, a framework with multiple extended Kalman filters (EKFs) for different time scales will be initially presented. Then, an implementation strategy with extended EKFs for the fast time scale and moving horizon estimators (MHEs) [6,7] for the slow time scale will be discussed.

References

[1] S. S. Jogwar and P. Daoutidis. Dynamics and control of vapor recompression distillation. J. Proc. Cont., 19(10):1737â??1750, 2009.

[2] S. S. Jogwar, M. Baldea, and P. Daoutidis. Dynamics and control of process networks with large energy recycle. Ind. Eng. Chem. Res., 48(13):6087â??6097, 2009.

[3] S. S. Jogwar, M. Baldea, and P. Daoutidis. Tight energy integration: Dynamic impact and control advantages. Comput. Chem. Eng., 34(9):1457â??1466, 2010.

[4] S. C. Patwardhan, S. Narasimhan, P. Jagadeesan, B. Gopaluni, and S. L. Shah. Nonlinear Bayesian state estimation: A review of recent developments. Control Eng. Prac., 20(10): 933â??953, 2012.

[5] C. Pantea, A. Gupta, J. B. Rawlings, and G. Craciun. The QSSA in chemical kinetics: As taught and as practiced. In N. Jonoska and M. Saito, editors, Discrete and Topological Models in Molecular Biology. Springer, 2013.

[6] F. V. Lima and J. B. Rawlings. Nonlinear stochastic modeling to improve state estimation in process monitoring and control. AIChE J., 57(4):996â??1007, 2011.

[7] F. D. Rincón, G.C. Le Roux, F.V. Lima. The autocovariance least-squares method for batch processes: Application to experimental chemical systems. Ind. Eng. Chem. Res., 53(46):18005-15, 2014.