As electricity systems become more dependent on wind and solar power, engineers face a problem that cannot be solved simply by adding more generating capacity. Renewable output varies with the weather, electricity demand changes with temperature, and transmission networks determine whether power can reach the places where it is needed. Decisions about where new infrastructure is built may therefore be just as important as decisions about how much is installed.
Researchers at the Massachusetts Institute of Technology have developed a planning framework intended to help energy developers and grid operators account for those interactions. By combining detailed power-system simulations with fine-scale meteorological modelling, the approach examines how future weather conditions could affect electricity generation, demand, storage and transmission at the same time.
Most energy infrastructure is designed using historical weather data. That has traditionally provided engineers with a reasonable basis for estimating patterns of demand and generation over the lifetime of a project. Wind farms, solar installations and transmission lines being developed today, however, are likely to remain in operation for several decades, during which regional weather patterns may differ from those used in their original design.
This creates a particular challenge for electricity systems with a high proportion of renewable generation. Wind and solar output can fall over several days, while high or low temperatures may simultaneously push demand upwards. Transmission constraints can then prevent available electricity from reaching the places where it is needed.
Advertisement
The MIT team wanted to understand these effects as a connected engineering problem rather than examining each part of the grid in isolation. Previous studies have often focused on individual technologies or relied on global climate models that provide limited regional detail. The new framework instead links local weather conditions with the physical layout and operation of an energy system.
The researchers applied the method to future electricity systems in New England and Texas, two regions with different climates, demand patterns and network structures. They compared infrastructure designed around historical weather with systems planned using projections of future conditions.
Their results suggest that relying on historical climate patterns could leave both regions more exposed to supply shortfalls. Under some scenarios, systems designed without considering future weather experienced up to five times as many periods in which demand could not be met by 2050. These shortfalls were driven less by the performance of individual power plants than by several factors occurring together, including prolonged drops in renewable generation, changing demand and limitations in the transmission network.
The location of new infrastructure made a significant difference. Sites that appeared most suitable for wind and solar generation under historical conditions were not always the locations that best supported the grid under future weather patterns.
In New England, the simulations indicated that future climate-related disruptions could increase the need for solar capacity and transmission infrastructure closer to cities and other centres of demand. In Texas, the main constraint was the ability of the transmission system to move electricity between regions, with the climate-informed designs placing greater emphasis on wind generation in West Texas. The researchers found that these changes could improve resilience with little or no additional cost.
Advertisement
The framework also accounts for compound weather events, in which several parts of the energy system are affected simultaneously. An extended period of unusual weather might reduce renewable generation while increasing heating or cooling demand, placing greater strain on the network than would be expected if each effect were considered independently.
Michael Howland, senior author of the study and Associate Professor of Civil and Environmental Engineering at MIT, said the largest impacts became apparent when generation, demand and transmission were modelled together rather than as separate problems.
Although the high-resolution modelling used in the study remains computationally demanding, the researchers hope to develop faster tools that can be used in routine planning. They argue that treating future climate conditions as another engineering design parameter, alongside cost, network capacity and expected demand, could help reduce the risk of future blackouts without significantly increasing the cost of building a more resilient electricity system.