Neural Network Based Explicit MPC for Chemical Reactor Control
This work addresses process control challenges in chemical engineering by providing a neural network-based method to simplify MPC implementation, though it appears incremental as it applies existing neural network techniques to a known control problem.
The paper tackles the problem of implementing explicit model predictive control (MPC) for chemical reactor control by using a deep neural network to approximate MPC behavior, maintaining constraints on states and control inputs regardless of weighting matrices, and demonstrates this in a simulation of a continuous stirred tank reactor.
In this paper, we show the implementation of deep neural networks applied in process control. In our approach, we based the training of the neural network on model predictive control. Model predictive control is popular for its ability to be tuned by the weighting matrices and by the fact that it respects the constraints. We present the neural network that can approximate the behavior of the MPC in the way of mimicking the control input trajectory while the constraints on states and control input remain unimpaired of the value of the weighting matrices. This approach is demonstrated in a simulation case study involving a continuous stirred tank reactor, where multi-component chemical reaction takes place.