Implementation of ANN on CCHP system to predict trigeneration performance with consideration of various operative factors

[Display omitted] •ANN modeling tool was implemented on the CCHP system.•The best ANN topology was detected 10–8–9 with Levenberg–Marquadt algorithm.•The system is more sensitive of CC outlet temperature and turbine isentropic efficiency.•The lowest RMSE=3.13e−5 and the best R2=0.999 is related to lambda and second law efficiency terms, respectively. A detailed investigation was aimed based on numerical thermodynamic survey and artificial neural network (ANN) modeling of the trigeneration system. The results are presented in two pivotal frameworks namely the sensitivity analysis and ANN prediction capability of proposed modeling. The underlying operative parameters were chosen as input parameters from different cycles and components, while the exergy efficiency, exergy loss, coefficient of performance, heating load exergy, lambda, gas turbine power, exergy destruction, actual outlet air compressor temperature, and heat recovery gas steam generator (HRSG) outlet temperature were taken as objective output parameters for the modeling purpose. Up to now, no significant step was taken to investigate the compound power plant with thermodynamic analyses and network predictability hybrid in such a detailed oriented approach. It follows that multilayer perceptron neural network with back propagation algorithm deployed with 10–8–9 configuration results in the modeling reliability ranged within R2=0.995–0.999. When dataset treated with trainlm learning algorithm and diversified neurons, the mean square error (MSE) is obtained equal to 0.2175. This denotes a powerful modeling achievement in both scientific and industrial scale to save considerable computational cost on combined cooling, heating, and power system in accomplishment of boosting the energy efficiency and system maintenance.