Fault detection in the microgrid is a crucial task due to the diversified
fault conditions, and it must be rapidly identified to reduce any serious
hindrance to the system. The fault current behaviour due to change in fault
resistance of the touching point can damage the switches of integrated
converters. In addition to the above, sporadic conditions can also affect
the profile of voltage and current, in the system. Traditional protection
schemes need modification to prevent relay maloperation of the microgrid.
This paper presents a protection scheme based on support vector machines to
detect faults under such tedious conditions. In this protection scheme,
acquired samples of voltage and current from selected bus have been used and
processed through the data processing tool discrete wavelet transform. The
protection scheme is operating in two dissimilar operating modes, where
initially mode was identified then fault detection/classification was done.
Section identification task was performed to identify faulty sections under
varying operating scenarios. The uncertain conditions of the renewable
sources can affect the performance of the system, therefore some random
cases have been considered to validate the protection scheme.