Modeling of Organic Rankine Cycle for waste heat recovery using RBF neural networks

The Organic Rankine Cycle (ORC) process is promised to significantly recycle the waste heat from medium and low temperature heat sources, achieve better performance to recover low grade waste heat than traditional waste heat recovery processes used in the industrial applications. An accurate ORC model is indispensable for the optimization and control of ORC systems. A new Radial Basis Function (RBF) modelling approach, which combines the node selection based on Fast Recursive algorithm (FRA) and non-linear parameters optimization using the PSO algorithm is proposed to model the ORC system. The experimental results verify that the resultant models can achieve high training accuracy and desirable generalization performance.