Optimisation of controller parameters for grid-tied photovoltaic system at faulty network using artificial neural network-based cuckoo search algorithm

This study exhibits the optimum design procedure to tune controller parameters for grid-connected distributed generation system based on cuckoo search algorithm (CSA). To investigate the effectiveness of proposed algorithm, a grid-tied photovoltaic (PV) system consisting of two power electronic converters controlled by five proportional integral (PI) controllers is chosen. Setting proper values for all the PI controllers is a complicated task, notably when the system is non-linear. In this study, response surface methodology (RSM) is used to develop the mathematical design of the PV system which is required to apply the optimisation algorithm. To minimise the design efforts of RSM, an alternate approach based on artificial neural network is introduced to develop the mathematical model of the PV system which is another salient feature of this research. Moreover, two modifications in the CSA are proposed to extract optimum parameters for the controllers which are found suitable in power system applications. Both the transient and dynamic performances of the system with the optimum values obtained through CSA are studied for different types of grid fault conditions using PSCAD/EMTDC. The design values are compared with values obtained through genetic algorithm and bacterial foraging optimisation. Experimental validation is also given for the proposed method.