A Novel Fault Detection and Classification Strategy for Photovoltaic Distribution Network Using Improved Hilbert–Huang Transform and Ensemble Learning Technique

Due to the increased integration of distributed generations in networks, their development and operation are facing protection challenges that traditional systems incapable addressing. These problems include variations fault current during various modes, diverse network topology, high impedance faults. Therefore, appropriate reasonable detection is highly encouraged improve dependability network. This paper proposes a novel technique employs an improved Hilbert–Huang Transform (HHT) ensemble learning techniques resolve these photovoltaic First, HHT utilized extract energy features from signal. Second, variational mode decomposition (VMD) applied intrinsic function zero component Then, acquired feature input for classification. The proposed implemented using MATLAB software environment, including classification learner app SIMULINK. An evaluation results conducted under normal connected (NCM) island (ISM) radial mesh-soft normally open point (SNOP) configurations. accuracy bagged trees higher when compared narrow-neural network, fine tree, quadratic SVM, fine-gaussian wide-neural presented depends only on local variables has no requirements connection latency. Consequently, faults technology reasonable. simulation demonstrate superior neural support vector machine, achieving 100%, 99.2% 99.7% accurate symmetrical unsymmetrical throughout NCM, ISM, dynamic mode, respectively. Moreover, developed protects DN effectively mesh-SNOP topologies. suggested strategy can detect classify accurately with/without DGs. Additionally, this precisely low within 4.8 ms.