The Smart distribution system initiative requires an increased usage of the distribution feeder-level communication infrastructure to improve automation. Using a distributed sensor network for monitoring the distribution system is proposed by various researchers. Such distributed sensor communication architecture requires information to be received within an allowable delay and a minimum processing time at the control center. Increasing the number of sensors in the network also increases the data flow in the communication medium. Therefore, to reduce the burden in the communication medium, an event driven communication protocol could be utilized. This communication architecture assumed that the sensors used a local fault detection system to detect the abnormal event before communicating with the control center. This work focusses on local detection of faults in a distributed sensor network using a Hidden Markov Model considering a minimum processing time.