Modelling and Control of Reactive Polymer Composite Moulding Using Bootstrap Aggregated Neural Network Models

Abstract

This paper presents using bootstrap aggregated neural networks for the modelling and optimisation control of reactive polymer composite moulding processes. Neural network models for the degree of cure are developed from process operational data. In order to improve model generalization capability, multiple neural networks are developed from bootstrap re-samples of the original data and are combined. Reliable optimal heating profiles are obtained by solving an optimisation problem using the bootstrap aggregated neural network model and incorporating model prediction confidence intervals in the optimisation objective function. The proposed method is applied to both simulated data and real industrial data.