Physics-informed recurrent neural networks and hyper-parameter optimization for dynamic process systems

Many of the processes in chemical engineering applications are of dynamic nature. Mechanistic modeling of these processes is challenging due to the complexity and uncertainty. On the other hand, recurrent neural networks are useful to be utilized to model dynamic processes by using the available data. Although these networks can capture the complexities, they might contribute to overfitting and require high-quality and adequate data. In this study, two different physics-informed training approaches are investigated. The first approach is using a multiobjective loss function in the training including the discretized form of the differential equation. The second approach is using a hybrid recurrent neural network cell with embedded physics-informed and data-driven nodes performing Euler discretization. Physics-informed neural networks can improve test performance even though decrease in training performance might be observed. Finally, smaller and more robust architecture are obtained using hyper-parameter optimization when physics-informed training is performed.