Applying Cluster Computing to Enable a Large-scale Smart Grid Stability Monitoring Application

The real-time execution of grid stability monitoring algorithms are critical to enabling a truly smart grid. However, the combination of a high sampling rate for grid monitoring devices, combined with a large number of devices scattered across a grid, result in very high throughput requirements for the execution of these algorithms. Here we define a centralized hardware and software infrastructure to enable the real-time execution of a small signal oscillation detection algorithm using a cluster of commodity nodes. Our research has demonstrated that readings from up to 500 phasor measurement units (PMUs) sampling at 60Hz can be analyzed in real-time by a single 8-core, 2.53GHz machine with 8GB of RAM, and that a cluster of four of these machines can be used to monitor up to 2,000 PMUs in parallel.