Machine learning for energy conservation of microgrid under fault and varying load conditions

Ensuring the optimal operation of Microgrids becomes more challenging due to various reasons such as dynamic load characteristics, presence of supplementary devices and fault situations. Most importantly, the microgrids structure gets updated during each event of a fault in order to isolate the faulty section from the grid. The resultant configuration at post fault does not need to be optimal which requires the support of some special technique to identify the optimal configuration and optimal capacitor location from the updated configuration. The main objective of the proposed work is to reduce the Service Restoration Cost (SRC) along with the elimination of the out-of-service area during faulted conditions. In this paper, the machine learning technique is adapted to find the optimal configuration and the optimal placement of capacitors instantly based on the fault condition. The different optimization algorithms are used to prepare the performance dataset which contains optimal solutions under different fault situations. The proposed approach is validated with the IEEE 33-bus Radial Distribution System.