Modified Robust Optimization Over Time for Process Parameter Optimization in Pre-Oxidation Process of Carbon Fiber Production

The pre-oxidation process of PAN fiber is an important step in the practical production of carbon fiber, which has a significant impact on the performance and production efficiency of carbon fiber. In this article, in order to solve the problem that it is difficult to obtain high-quality carbon fiber while reducing energy consumption in the practical production of the pre-oxidation process, we build a dynamic multi-objective model about energy consumption and PAN modulus. We then propose the modified robust optimization over time (mROOT) to solve it and optimize the process parameters, which aims to maximize PAN modulus and minimize energy consumption while decreasing the average switching cost of solutions by reducing the switching times of solutions, hoping to provide some theoretical guidance for the pre-oxidation process of carbon fiber in practical production. The experimental results show that compared with TMO and RPOOT, the solutions obtained by mROOT have longer average survival time and lower average switching cost, and compared with traditional time series prediction models, the combination of online predictor proposed in this article and mROOT is a more practical and effective method for solving the model.