Abstract:
Power system operators meet constantly fluctuating electricity demand through coordinated operations of power plants, transmission lines, and other critical infrastructure. Even with physical redundancy built-in and emergency protocols in place, extreme weather events regularly overwhelm these measures and disrupt the tenuous balance between electricity supply and demand, resulting in outages and dramatic increases in prices in wholesale electricity markets (the economic institutions that oversee the production and sale of electricity in most of the U.S.). At the same time, it is increasingly accepted that the electric power sector (responsible for 27% of greenhouse gas emissions in the U.S.) must expand and decarbonize by 2050. System operators and utilities are tasked with managing the effects of growing renewable energy penetration on wholesale electricity market dynamics and physical reliability, while contending with growing exposure to drought (e.g. reduced hydropower production); extreme temperatures (e.g. spikes in heating/cooling demand); wildfire (e.g. transmission line impacts and power shutoffs) and flooding (i.e. prolonged outages and damaged equipment). Considering these short and long term sources of uncertainty and modeling system dynamics at sufficient resolution to capture real world behavior is increasingly important for characterizing physical and financial risk for grid participants. This talk will focus on open source modeling approaches for simulating electric power system operations under uncertainty, especially hydroclimatic extremes, and exploring system vulnerabilities from an environmental, engineering reliability, and financial/economic perspective.
Bio:
Jordan Kern is an assistant professor at North Carolina State University in the Department of Forestry and Environmental Resources. Jordan is a three-time graduate (BS in Environmental Science, MS and PhD in Environmental Science and Engineering) from the University of North Carolina at Chapel Hill, where he was a research faculty from 2016-2018 before joining NC State. His research has been supported by multiple National Science Foundation funding programs (Innovations at the Nexus of Food, Energy, and Water Systems; Coupled Natural Human Systems) and Department of Energy funding programs (ARPA-E, Bioenergy Technologies Office, Office of Science). He is an institutional lead on the DOE Office of Science funded Integrated Multi-sector Multiscale Modeling (IM3) project lead by Pacific Northwest National Laboratory (PNNL). As part of IM3, Jordan’s team at NC State is collaborating with PNNL researchers on the development of open source grid operations models for each of the three major electric power interconnections in the U.S. (Western, Eastern, and the Electric Reliability Council of Texas). These models will be used to support large, integrated modeling experiments that evaluate the current and future vulnerability of the nation’s power grids to population change, technology adoption, climate change and extreme weather.
Summary:
Focus:
Uncertainty about the impact of climate change on society
Emphasis on the overall power grid, local details out of scope
Climate change impacts
Wildfires (can be caused by environment or caused by power systems)
Increased demand due to temperature extremes
Wind/flooding damage due to hurricanes
Droughts affect usable capabilities of different types of power plants (e.g. hydro)
These increase societal costs of climate change
More damage
More stuff to damage
Losses due to disasters are already rising for these two reasons
Decarbonize the grid mostly via wind and solar
Challenge: these depend on meteorological conditions
People are far more likely to support decarbonization if it doesn’t affect their cost of energy
However, damage due to climate change puts more stress on power grids, which can increase the cost of supplying power and put stress on their finances
US electricity grid is organized into
Physical grids: West, East and Texas
Markets:
7 regions where there is a competitive wholesale market for power
Rest of country uses regulated pricing
Power sources can be compared via:
Size (MW)
Cumulative capacity (MW)
Marginal cost (e.g. fuel cost) ($)
Lifetime cost ($)
Can plot a supply curve & demand curve
Demand: largely insensitive to price
Marginal cost of power: the most expensive power source needed to satisfy the last (most expensive) bit of power demand
Rises with increasing power demand
Infinite if power demand exceeds supply
Open source simulation: https://github.com/romulus97
Commercial alternative: https://www.energyexemplar.com/plexos
Approach:
Unit commitment economic dispatch
Python: PYOMO modeling language
Inputs: time series of demand, availability of renewable energy
Output: optimal supply schedule, costs, emissions
Power grid is a graph of assets (transmission, generation, load centers)
Network reduction algorithm to merge network into a simplified topology (needed since solvers are expensive)
Solver: mixed integer + linear constraints
Model must be run for 1200 simulated years to understand system behavior under alternative scenarios
Model of renewable energy depends on datasets of weather, streamflow, etc.
Validated by comparing model’s predictions against historical data
Future: synthetic weather generation
Resampling past weather data
Use samples to
Create long time series
Create different distributions with different properties (e.g. more wind, higher temperature)
Simulation can be used to understand
How different climatic conditions affect power demand and supply
The power grid conditions that would lead to different distribution of damages (e.g. pollution)
Observation: in CA availability of hydro power (i.e. water) strongly impacts pollution, power availability, etc.
Under climate change: smaller snowpack, so less water flows into reservoirs
Impact: more West Coast-wide blackouts
If we triple renewable energy, energy prices will drop
Different distributions of power sources will result in different power prices under the same weather distributions (e.g. in 2020 a low wind year is fine but in 2050 it causes prices to rise)
Design financial products to quantify cost impact due to extreme weather events so they can be hedged and used as insurance.
Parametric/index-based insurance:
Correlate an entity’s financial risk to environment
Payouts:
If environmental conditions exceed threshold, there’s a payout
Challenge: including market conditions into payout rule