Large-Scale Generation and Validation of Synthetic PMU Data

In spite of the challenges associated with obtaining actual PMU measurements for research purposes and analytic methods testing, it remains crucial that experimental input data exhibits similar quality features of real measurements for proper grid assessment and planning. The objective of this paper is to generate and validate large sets of synthetic, but realistic, PMU datasets obtained from complex grid models. A study of different variability components in PMU measurements is first presented followed by the proposed steps in generating synthetic datasets. Random variations of resource inputs are used in a simulation platform to generate prior voltage data from a synthetic 2,000-bus system, followed by a data modification process to infuse further realism into the dataset. The validation process used to assess the accuracy of the generated voltage dataset utilizes a variability metric to determine the level of inherent variations in individual measurements, and further applies a dimension reduction technique to identify the extent of electrical dynamics retained in the overall synthetic dataset.