From Grid Optics
The Future Power Grid Initiative (FPGI) will deliver next-generation concepts and tools for grid operation and planning and ensure a more secure, efficient and reliable future grid. Building on the Electricity Infrastructure Operations Center (EIOC), the Pacific Northwest National Laboratory’s (PNNL) national electric grid research facility, FPGI will advance the science and develop the technologies necessary for meeting the nation’s expectations for a highly reliable and efficient electric grid, reducing carbon emissions and our dependence on foreign oil.
To tackle these issues, the FPGI will address three of the most pressing areas of research: Large-Scale Real-Time Simulation, Stochastic and Uncertainty Modeling, and Data-Driven Decision Support.
Large-Scale Real-Time Simulation: Today’s grid operation functions include state estimation, contingency analysis, automatic generation control, and economic dispatch along with other market functions. The speed at which the grid functions are running (typically every 3-5 minutes) needs to be dramatically increased so that the analysis of contingencies is both comprehensive and real time. Today’s online grid operation is based on a grid model which can only provide a static snapshot of current system operation status, while dynamic analysis is conducted offline because of low computational efficiency. The offline analysis uses a worst-case scenario to determine transmission limits, resulting in under-utilization of grid assets. There is a great need to bring dynamic grid analysis into online applications to foresee the grid status with more transparency for better reliability and asset utilization. This is especially important given the increasing penetration of variable renewable energy and smart loads into the grid, which significantly increases the complexity of grid reliability management and requires a real-time dynamic view of the grid.
Stochastic and Uncertainty Modeling: Today’s power grid model describes the grid connectivity and associated parameters at the bulk transmission level (typically above 100kV). Distribution network models exist but are not integrated with transmission network models for holistic analysis. The transmission grid model size is typically in the order of 104 components for an interconnection-scale grid. The premise of modeling only the bulk transmission system is that loads residing in the distribution-level grid are passive devices and their behaviors are mostly predictable and can be described as some simple aggregated models without losing accuracy. This is no longer true with the penetration of smart loads and distributed generation (especially small wind turbine and roof-top photovoltaic panels). The load side is actively participating in and responding to grid dynamics, and the electricity flows two ways between the transmission grid and the lower voltage level grid; creating a greater need for modeling the lower voltage levels in the grid. However, the traditional first-principle-based modeling approach will no longer work for the 109 number of load devices, not even mentioning the uncertainties associated with this vast number of devices. In addition, the current separation of operation and planning models creates silos and limits grid management capabilities. The grid of the future needs fundamentally new modeling approaches to integrate transmission and distribution models as well as operation and planning models that will capture emerging behaviors and consider uncertainty quantification and optimization. Advanced algorithms for optimization and under uncertainty will allow the stakeholders to gauge at once the effects of policies over the entire set of outcomes.
Data-Driven Decision Support: Both measurement data and simulation data will grow by many folds in the future grid. The key challenge is how to integrate a variety of data sources and convert the large volume of data to actionable information. The analysis and visualization of multi-scale, multi-data, multi-model systems in a user-centric venue will facilitate credible policy analyses and formulation for the nation’s future power system that are actionable by the owners. This decision support system, integrating both current and historical data, must provide situational awareness and analysis across temporal and spatial scales. Operators and planners must not only be able to assess grid conditions at time horizons ranging from seconds to years, but also geographic areas from the individual consumer to the nationwide grid status. With these tools, operators, planners, and policy makers can then explore potential future states of the grid visually.
We anticipate that our research will allow operators to see grid performance in near real time across a wide service area and under emerging contingency situations. Planners will have the ability to view configurations of the grid and adjust those options against national goals and performance objective. In doing so, our approach is to combine PNNL's distinctive capabilities in power systems, data intensive computing, high performance computing, and visual analytics to address the complex problems from real-time and large-scale challenges in three Focus areas: Networking and Data Management; Modeling, Simulation, and Analysis; and Visualization and Decision Support.
With the anticipated advancements, the Initiative’s core product is GridOPTICS – Grid Operation and Planning Technology Integrated Capabilities Suite, a tool suite that is able to securely collect data in real time, use data to drive modeling and simulation, and convert large volumes of data to actionable information. These concepts and tools will show and analyze grid performance at an unprecedented speed, scale, and resolution and will support operational and policy decision making for the grid of the future.