Heuristically optimized RBF neural model for the control of section weights in stretch blow moulding

The injection stretch-blow Moulding (ISBM) process is typically used to manufacture PET containers for the beverage and consumer goods industry. The process is somehow complex and users often have to heavily rely on trial and error methods to setup and control it. In this paper, a novel identification method based on a radial basis function (RBF) network model and heuristic optimization methods, such as particle swarm optimization (PSO), deferential evolution (DE), and extreme learning machine (ELM) is proposed for the modelling and control of bottle section weights. The main advantage of the proposed method is that the non-linear parameters are optimized in a continuous space while the hidden nodes are selected one by one in a discrete space using a two-stage selection algorithm. The computational complexity is significantly reduced due to a recursive updating mechanism. Experimental results on simulation data from ABAQUS are presented to confirm the superiority of the proposed method.