Certificates

We are aware of an issue with certificate availability and are working diligently with the vendor to resolve. The vendor has indicated that, while users are unable to directly access their certificates, results are still being stored. Certificates will be available once the issue is resolved. Thank you for your patience.

(118a) Machine Learning-Based Predictive Control of Nonlinear Processes

Authors: 
Wu, Z., University of California, Los Angeles
Tran, A., University of California, Los Angeles
Christofides, P. D., University of California, Los Angeles
Machine learning has attracted an increased level of attention in model identification in recent years. Among many machine learning techniques, the recurrent neural network has been widely-used for modeling a general class of dynamical systems. In this work, we focus on the design of model predictive control (MPC) for nonlinear systems that utilizes an ensemble of well-fitting recurrent neural network (RNN) models to predict nonlinear dynamics. Specifically, RNN models are initially developed based on the dataset generated from extensive open-loop simulations within a certain operation region to capture process dynamics for a general class of nonlinear systems with a sufficiently small modeling error between the RNN model and the actual nonlinear process model [1,2,3]. Subsequently, the Lyapunov-based MPC (LMPC) that utilizes RNN models as the prediction model is developed in a sample-and-hold fashion to achieve closed-loop stability in the sense that the state of the closed-loop system is bounded in the stability region for all times and ultimately converges to a small neighborhood around the origin [4]. Additionally, ensemble regression modeling tools are employed in the formulation of LMPC to improve prediction accuracy of RNN models and overall closed-loop performance while parallel computing is utilized to reduce computation time. Finally, a chemical reactor example demonstrates the effectiveness of the proposed LMPC design using ensemble regression models and the significant improvement of computational efficiency under parallel operation.

[1] Wu, Z., A. Tran, Y. M. Ren, C. S. Barnes, S. Chen and P. D. Christofides. Model Predictive Control of Phthalic Anhydride Synthesis in a Fixed-Bed Catalytic Reactor via Machine Learning Modeling. Chem. Eng. Res. & Des., 145, 173-183, 2019.
[2] Kosmatopoulos, E. B., Polycarpou, M. M., Christodoulou, M. A., and Ioannou, P. A. High-order neural network structures for identification of dynamical systems. IEEE transactions on Neural Networks, 6, 422-431, 1995
[3] Sontag, E. D. "Neural nets as systems models and controllers." Proc. Seventh Yale Workshop on Adaptive and Learning Systems. 1992.
[4] Alanqar, A., H. Durand and P. D. Christofides. On Identification of Well-Conditioned Nonlinear Systems: Application to Economic Model Predictive Control of Nonlinear Processes. AIChE J., 61, 3353-3373, 2015