ANN-Based Adaptive NMPC for Uranium Extraction-Scrubbing Operation in Spent Nuclear Fuel Treatment Process*

This paper addresses the particularities in optimal control of the uraniumextraction-scrubbing operation in the PUREX process. The control problemrequires optimally stabilizing the system at a desired solvent saturationlevel, guaranteeing constraints, disturbance rejection, and adapting to setpoint variations. A qualified simulator named PAREX was developed by the FrenchAlternative Energies and Atomic Energy Commission (CEA) to simulateliquid-liquid extraction operations in the PUREX process. However, since themathematical model is complex and is described by a system of nonlinear, stiff,high-dimensional differential-algebraic equations (DAE), applying optimalcontrol methods will lead to a large-scale nonlinear programming problem with ahuge computational burden. The solution we propose in this work is to train aneural network to predict the process outputs using the measurement history.This neural network architecture, which employs the long short-term memory(LSTM), linear regression and logistic regression networks, allows reducing thenumber of state variables, thus reducing the complexity of the optimizationproblems in the control scheme. Furthermore, nonlinear model predictive control(NMPC) and moving horizon estimation (MHE) problems are developed and solvedusing the PSO (Particle Swarm Optimization) algorithm. Simulation results showthat the proposed adaptive optimal control scheme satisfies the requirements ofthe control problem and provides promise for experimental testing.