Investigation of Renewable Energy Integration Challenges and Condition Monitoring Using Optimized Tree in Three Phase Grid System

Abstract

The stochastic nature of distributed renewable energy resources (DREs) poses many operational and technical challenges in integrating renewable energy resources. Voltage instability is one of the challenges due to integration of renewable energy resources in which the supply is less than the demand of energy at the distribution side. The major challenge associated with integration of renewable energy resources is the islanding scenario which needs to be addressed. In this paper, an efficient and accurate machine-learning algorithm is developed for the islanding detection and classification process. For this purpose, simulation and compilation perform on a 12kW 3Ø grid-connected photovoltaic (PV) model in Typhoon Hardware-in-Loop (HIL) simulator in which data is extracted for low voltage ride through (LVRT) grid fault. Furthermore, this data is trained with the proposed machine-learning algorithm for the islanding classification mechanism. The proposed Optimizable Decision Tree (O-DT) technique has a training accuracy of 99.7% and can efficiently predict islanding classification with a speed of about 140000bs/sec.