Artificial neural network simulator for supercapacitor performance prediction

Artificial neural network was used to calculate the performance of a model supercapacitor as signified by the power density, energy density and utilization to the synthetic, intrinsic and operating characteristics. A four-layer neural net having two hidden layers having 6 and 15 nodes was found to be well capable of simulating the capacitor performance with the convergence achieved often a relatively small number of epochs. As for the input parameters, crystal size, surface lattice length, exchange current density of the capacitor active material and the cell current employed while utilization, energy density and power density were the outputs.