LAUSR

An effective hybrid strategy‐based protection scheme to differentiate between inverter faults and symmetric/unsymmetric faults on distribution line, in addition to detect and identify that faulty section in a microgrid. The proposed hybrid strategy is combined execution of principal component analysis (PCA) and nomadic people optimizer (NPO) learning process‐based gradient boosting decision tree (GBDT). Hence it is commonly named as PCA‐NPOGBDT strategy. Here, three stages are occupied using fault analysis on microgrid to differentiate among inverter faults on microgrid and symmetric/unsymmetric faults on distribution line, in addition to detect/classify that faults and recognize that faulty section. They are: data set generation, feature removal and training process. In this paper, the first procedure of the proposed strategy is the microgrid parameters such as voltage and current signals at usual and unusual condition data set preparation with PCA technique. The data set based on PCA preparation process has feature removal of the power flow parameters and describes that characteristics nature of the various signals happened with microgrid system. The removed data set is evaluated with the NPO‐based GBDT system to classify that sort of fault happened at microgrid system. The proposed model is performed at MATLAB/Simulink working stage and compared with various existing techniques. Computation time of fault detection using proposed and existing system is analyzed and likened to other existing systems like RF, DT, SVM and DBN under 100, 150, 200, 250, and 500 trails. The computation time under 100, 150, 200, 250 and 500 trails of proposed technique is 48.1740, 51.2133, 71.0483, 60.00126, and 57.80132.