A method for evaluating grid stability with high penetrations of renewable energy and energy storage




Johnson, Samuel Caleb

Journal Title

Journal ISSN

Volume Title



The rapid growth of electricity generation from variable renewable resources like wind and solar has greatly impacted wholesale energy markets and raised questions about future grid stability. With this paradigm shift, some existing coal, natural gas, and nuclear generators have encountered financial struggles, which has led to widespread retirements and tight capacity margins in some regions. Although this change could lead to reduced carbon emissions, synchronous generators provide some important reliability benefits to the grid that other technologies cannot easily replace. To assess the impact of an energy transition away from synchronous generation (e.g. fossil fuel fired power plants) and towards non-synchronous generation (e.g. wind and solar), future grid stability was investigated in the following three studies: (1) evaluating rotational inertia as a component of grid reliability with high penetrations of variable renewable energy, (2) determining the impact of non-synchronous generation on grid stability and identifying mitigation pathways, and (3) quantifying the regional economic and stability impacts of grid-scale energy storage.

First, a method was developed to assess grid stability with increasing penetrations of non-synchronous renewable energy generation to determine when an electric grid might be more vulnerable to frequency contingencies, such as a generator outage. Unit commitment and dispatch modeling was used to quantify system inertia, an established proxy for grid stability. A case study of the Electric Reliability Council of Texas grid was used to illustrate the method. Results from the modeled scenarios showed that the Texas grid is resilient to major grid changes, even with relatively high penetrations (~30% of annual energy generation compared to 19% in 2018) of renewable energy. However, retiring nuclear power plants and private-use networks in the model led to unstable inertia levels in our results. When the system inertia was constrained to meet a minimum threshold in our model, multiple coal and natural gas combined-cycle plants were dispatched at part-load or at their minimum operating level to maintain stable system inertia levels. This behavior is expected to expand with higher renewable energy penetrations and could occur on other electric grids that are reliant on synchronous generators for inertia support.

A method was also developed for assessing the impacts of stability support from inverter-connected resources. In this analysis, a fully disaggregated, inertia-constrained unit commitment and dispatch model was used to study the stability of future grid scenarios with high penetrations of non-synchronous renewable energy generation. As before, the Texas grid (the Electric Reliability Council of Texas – ERCOT) was used as a test case and instances when the system inertia fell below 100 GW·s (the grid's current minimum level) were found, starting at an annual renewable energy penetration (including both synchronous and non-synchronous renewable resources) of ~30% in our model. At an ~88% renewable energy penetration, the average system inertia level also fell to 100 GW·s. When the modeled critical inertia limit was reduced to 80 GW·s, no critical inertia hours occurred for renewable energy penetrations up to 93% of annual energy. The critical inertia limit could drop to 60 GW·s if the largest generators in ERCOT (two co-located nuclear plants) were retired, but this had the same effect as reducing the limit to 80 GW·s and keeping these generators online, since the nuclear plants contribute a large portion of the grid's system inertia. Emissions also increased by ~25% in the modeled scenarios where these nuclear plants were retired. If the critical inertia limit was kept the same (100 GW·s), adding 525 MW of fast frequency response from wind, solar, and energy storage could reduce the number of critical inertia hours by 86% with a response time of 15 cycles. Therefore, while the transition to a grid with mostly non-synchronous energy generation should be handled with care, many feasible pathways for integrating inverter-connected technologies and maintaining a stable grid exist.

Building on the prior two methods, a third method was developed to evaluate the impact of energy storage systems on grid stability and system cost. While many grid-scale energy storage projects have been built and several have been announced, energy storage is costly and could negatively impact grid stability if systems are connected non-synchronously. Three different energy storage technologies with varying durations, ramp rates, and costs were modeled using a linearized dispatch model with discrete transmission zones and sub-hourly intervals (i.e. 15 minutes). Small penetrations of these technologies were modeled in a grid dominated by non-synchronous generation (51% wind and solar) to identify optimal storage zones. Transmission zones in the North, Northwest, West, Far West, and Panhandle regions were found to be the most favorable for building grid-scale storage from an economic standpoint. Next, higher energy storage penetrations were modeled to analyze the impact of storage on system inertia and the system cost. These high penetration scenarios focused primarily on storage divided across the optimal storage zones in proportion to their system cost impact. The modeling results showed that flywheels were able to maintain higher system inertia levels. Even so, the system cost was much lower when compressed air energy storage systems were modeled, demonstrating that high-duration energy storage technologies provided the most value to the grid. Energy storage was also more effective at maintaining grid stability and reducing costs than peaking plants. As a result, our model showed that new peakers might not be revenue sufficient in zones with high penetrations of renewable energy and energy storage.

Many options exist for reliably integrating high penetrations of variable renewable energy generation, including an inertia market, synthetic and virtual inertia, and grid-scale storage, but few of these solutions are available today. Together, each of the analyses presented in this dissertation communicate when grid stability issues might occur and how low system inertia levels could be avoided.


LCSH Subject Headings