Power grid fault detection using an AMR network

Smart grids have been introduced in many developed countries and facilitate more efficient and economical utilization of generated power. The components of such a grid are usually remotely monitored and selected components are controlled either manually or through advanced software programs. This is called an Advanced Metering Infrastructure (AMI) and the objective is to globally optimize available resources. In this paper we assume that an Automated Meter Reading (AMR) network is available. AMR is the precursor to an AMI system. AMR meters periodically report electrical energy consumption information to a centralized location and can also (since they contain batteries) transmit indications of power outage as well as power restoration. This information may also detect failed network components. The trade-offs considered are, the time taken to detect the failure, the accuracy of the fault decision and the amount of data required to perform these tasks.