Microgrid Control Latency Prediction with PMU Data from Hardware-in-the-Loop Experiments

The commission of a real microgrid may encounter unexpected time latency if no hardware experiment is conducted. Meanwhile, it is difficult to estimate the control signal latency of a microgrid's energy management system (EMS) without an efficient prediction algorithm. Thus, this paper proposes a time-latency prediction method utilizing the grid edge technology of PMU and hardware-in-the-Ioop (HIL) experiments. The latency problem has the following challenges based on unique properties of data: (1) Spatially, PMU and SCADA sensors have different locations and latency. Thus, it's required to approximate the data correlations to estimate the missing data accurately. (2) Temporally, the dynamic transitions in SCADA samples are scarce, urging efficient utilization of the SCADA data to approximate the dynamics. For challenge (1), we employ Deep Neural Networks (DNNs) with high capacities to capture spatial-temporal information to predict dynamic states. For challenge (2), we develop a new mechanism to utilize SCADA data efficiently and eliminate the latency by estimating the latency distribution. The results show that our EM (Expectation-Maximization) algorithm in time synchronization offers superior performance over other methods in detecting and fixing time latency. This advanced algorithm effectively identifies discrepancies in time data by iteratively estimating a statistical model's parameters to ensure synchronization's precision and reliability. The proposed method shows its potential in tackling time latency and data scarcity under 2 different systems in 4 different scenarios in our HIL real-world setup.