Indirect Training with Error Backpropagation in Gray-Box Neural Model: Application to a Chemical Process

Gray-box neural models mix differential equations, which act as white boxes, and neural networks, used as black boxes, to complete the phenomenological model. These models have been used in different researches proving their efficacy. The aim of this work is to show the training of the gray-box model through indirect back propagation and Levenberg-Marquardt. The gray-box neural model was tested in the simulation of a chemical process in a continuous stirred tank reactor (CSTR) with 5% noise, responding successfully.