Parametric analysis and optimization of regenerative Clausius and organic Rankine cycles with two feedwater heaters using artificial bees colony and artificial neural network

The present work concerns the parametric study and optimization of regenerative Clausius and organic Rankine cycles (ORC) with two feedwater heaters. For the parametric optimization, thermal efficiency, exergy efficiency and specific work are selected as the objective functions, so the mentioned parameters are calculated for different values of the outlet pressures from the second and third pumps by using EES (Engineering Equation Solver) software. Aiming at optimizing these functions, a procedure based on artificial neural network (ANN) and artificial bees colony (ABC) is proposed. The procedure includes two stages. According to the obtained data from the parametric analysis, in the first stage three different multi-layer perceptron neural networks are trained. In the next stage, three distinct artificial neural networks are used to optimize the specific network, the thermal efficiency and the exergy efficiency. Variables and fitness functions in these algorithms are the inputs and the outputs of the corresponding trained neural network, respectively. This optimization process is applied to water for a Clausius Rankine cycle and also to R717 for an ORC. It is shown that some interesting features among optimal objective functions and decision variables involved in this power cycle can be discovered consequently.

► The present work concerns the parametric study and optimization of regenerative Clausius and organic Rankine cycles. ► For the parametric optimization, thermal efficiency, exergy efficiency and specific work are selected as the objective functions. ► A procedure based on artificial neural network and artificial bees colony is proposed. ► This optimization process is applied to water for a Clausius Rankine cycle and also to R717 for an ORC. ► It is shown that some interesting features among optimal objective functions and decision variables involved in this power cycle can be discovered consequently.