Elman Neural Network Soft-Sensor Model of PVC Polymerization Process Optimized by Chaos Beetle Antennae Search Algorithm

Abstract

The conversion rate of vinyl chloride monomer (VCM) is an important product quality indicator in the process of Polyvinyl chloride (PVC) polymerization. Due to the complexity of the PVC polymerization process and the limitation of site conditions, it is difficult to obtain the VCM conversion rate online in real time.Therefore, this article puts forward a soft-sensor model based on Beetle Antennae Search Algorithm (BAS) to optimize Elman neural network(Elman). Firstly, Multi-Cluster Feature Selection (MCFS) is used to reduce the dimensionality of the high-dimensional input variables, so that we get auxiliary variables of the soft-sensor model. Then, using Elman neural network as a soft-sensor model, and it is trained by the proposed optimization algorithm, which combines the chaotic map and the Beetle Antennae Search Algorithm (CBAS). The simulation results show that the model can significantly improve the prediction accuracy of the VCM conversion rate while realizing the real-time control of the PVC polymerization production process.