Modeling and Control of Renewable Energy in Grids and Microgrids

To save the non-renewable resources and reduce pollution emissions, the integration of renewable energy sources (RES) into the main grid is increasing very fast. However, power systems experience more stability issues that are caused by the higher penetration of RES. The main objectives of this dissertation are to 1-investigate these stability issues using linear analysis, EMT simulation results, and experiment results; 2-provide the suitable solutions with the evaluations. The research results will be presented into two parts based on two modeling approaches. In the first part, RES units are normally modeled as an aggregated power source for the investigation because they are installed in the same farm and have the same structure and parameters. Hence, the investigations in this part are based on the single-inverter aggregated modeling of RES. In the second part, RES units and/or battery energy storage system (BESS) units are integrated into the same microgrid but they can be different types and have different parameters. Hence, each unit needs to be modeled individually. In other words, multi-inverter based modeling is used to investigate the stability issues in microgrids. In each part, both analytical models and testbeds are built using the same corresponding modeling approach. Analytical models are built in the dq frame to produce the linear analysis while the testbeds are built using EMT software or hardware to provide more practical results for validations. The first part of this dissertation investigates stability issues that happened in wind farms connected to the conventional grid using the single-inverter aggregating modeling. Due to the different reasons, the stability issues in this part are also divided into two categories, weak grid, and series compensated networks. The investigations on stability issues in the weak grid are started by two real-world events. In 2016, a subsynchronous oscillation (SSO) was observed in Type-4 wind farms connected to the weak grid in Northwestern China. This subsynchronous oscillation (SSO) also caused the torsional interaction between wind farms and remote thermal power plants. However, the previous instability events in the Type-4 wind farms were normally reported with the low-frequency oscillations such as the event in Texas in 2012. To explore the critical factor behind this difference, a single-inverter aggregated analytical model is built based on Type-4 wind connected to the weak grid. According to the linear analysis, this difference in the oscillation frequency is caused by the dynamics of phase-locked-loop (PLL). This finding is validated using an EMT testbed which is built in MATLAB/SimPowerSystems. Meanwhile, the torsional interaction with the remote synchronous generator is demonstrated using this testbed. To solve the stability issues caused by the weak grid, stability control with two strategies is implemented in two aggregated analytical models. The eigenvalue analysis is used to evaluate its performance. Then, a Type-3 EMT testbed and a Type-4 EMT testbed are built to validate the analytical results. Furthermore, the stability control is also implemented in an FPGA-based hardware testbed. The experiment results show the excellent enhancement of stability. In the series compensated networks, stability issues are investigated by replicating three real-world subsynchronous resonance (SSR) events. From August to October 2017, the Electric Reliability Council of Texas (ERCOT) reported three SSR events in the same transmission system which consists of six Type-3 wind plants and a series compensated line. These events were excited by the line outage but they had different consequences. To provide a reasonable explanation, this dissertation replicates these three SSR events using the EMT testbed built based on the real system. The challenge of replication is the limited information of system parameters. The fragmented information is collected and combined from different public project reports and websites. To estimate some parameters which are not found such as the number of online wind turbines and wind speed, the sensitivity analysis is conducted using both simulation results and analytical results. Note that these results are generated based on the single-inverter aggregated modeling of Type-3 wind. Finally, improved SSR control was implemented in the EMT testbed to mitigate SSR. The second part of this dissertation discusses the stability issues in microgrids. Different than the conventional grid, the investigations on microgrid need to consider not only the characteristics of each distributed generation resource (DER) but also the coordinate control among DERs. In this part, two coordinate controls, voltage-current (V-I) droop and consensus control, will be investigated under both grid-connected mode and autonomous mode. V-I droop control was proposed in 2015 to enhance the accuracy of reactive power sharing in the microgrid. However, our simulation results showed that the oscillation issues happened in a microgrid with V-I droop when small values were selected for droop coefficients. To investigate the effects of droop coefficients, a two-inverter based analytical model is derived with the multi-input multi-output (MIMO) matrix of transmission networks. The linear analysis is carried out to identify the root causes of oscillations under both grid-connected mode and autonomous mode. Analytical results are validated using the detailed testbed which is built in MATLAB/SimPowerSystems. In the testbed, two DERs are modeled separately as well. As the second coordinate control investigated in this dissertation, the consensus control can maximize the efficiency of a microgrid by synchronizing performances of all parallel BESS. Meanwhile, it only requires each BESS to share the limited information with its neighboring BESS. To evaluate the performance of the consensus control, a multi-inverter based analytical model is derived based on the grid-connected microgrid with three BESS. Furthermore, a controller-hardware-in-the-loop (CHIL) testbed is built to emulate the microgrid which has three BESS and is integrated into the IEEE 9-bus system. The entire circuit is simulated by a real-time simulator while three BESS are controlled by three external FPGA-based controllers. This CHIL testbed provides a more realistic environment to evaluate the consensus control because of the full dynamics of microgrid and the large-scale power system. This dissertation has led to four published journal papers and one working paper