Urban Energy Planning Gets a Real-World Algorithm
Finding somewhere to put a new power plant or utility-scale battery inside a city is partly an electrical engineering problem. It is also a question of zoning, noise, land availability, historic buildings, tourism, development policy and what the people running the city actually want it to become.
Researchers at Oak Ridge National Laboratory have developed an algorithm intended to bring those competing considerations into the same planning process. Instead of identifying a technically attractive point on a distribution grid and treating the surrounding city as a secondary constraint, the method incorporates external factors while determining the type, location and size of new generation and storage.
The work addresses a practical weakness in conventional generation siting models. Electrical optimisation can identify locations offering suitable connections, voltage performance and capacity, yet the resulting site may be difficult or impossible to develop. In a dense urban environment, apparently available land can sit beside housing, protected architecture, tourist attractions or other uses incompatible with a substantial energy installation.
According to ORNL researcher Rodney Itiki, zoning and noise restrictions alone can remove 20 to 30 percent of potential urban sites in some circumstances. Once those restrictions are recognised at the beginning rather than after an electrical solution has been produced, the optimisation problem looks very different.
Briefing
- ORNL researchers have developed an algorithm for locating and sizing utility-scale generation and energy storage in urban environments.
- The method combines power-system modelling with zoning, policy, social, economic and land-use considerations.
- Zoning and noise restrictions can reduce the number of potentially available urban sites by 20 to 30 percent, according to researcher Rodney Itiki.
- Grid performance is assessed through a 24-hour simulation of changing electricity demand and power flow.
- The researchers believe the approach could eventually be adapted to data centres, critical-minerals projects, disaster response and major international events.
Putting the City Inside the Grid Model
The research, published in the journal Energies, starts from a limitation in existing generation and storage station placement methods. Much of the established optimisation work is suited to rural or relatively homogeneous environments, where candidate locations can be assessed primarily through the electrical network and associated economics. Cities introduce external constraints that are much harder to treat as an afterthought.
A utility-scale generating plant or storage facility occupies physical space and changes the environment around it. Municipal legislation may determine what can be constructed, while architectural, historic and tourism considerations can effectively exclude otherwise attractive locations. Residential development brings further restrictions, particularly where noise, visual impact or intensive industrial activity are concerned.
The ORNL approach combines a power-grid simulator with a weighting system through which policy, regulations and stakeholder priorities can influence the result. Urban planners, energy policymakers and industrial-development authorities can contribute priorities rather than leaving the model to optimise against electrical criteria alone.
โThe tool is revolutionary because previous study approaches were not considering the real world,โ Itiki said. โThey just started with a diagram of the energy system and picked a location based on that infrastructure, without incorporating the needs of the community.โ
A mathematically optimal location has little practical value if planning restrictions make construction unrealistic.
Different Cities, Different Answers
There is deliberately no universal definition of the best energy project within the model. A city trying to attract manufacturing and residents might place greater weight on affordable electricity and employment. An area supporting artificial intelligence data centres and a military installation could give considerably more weight to reliability and energy security.
Those priorities can alter both what should be built and where it should go. The algorithm is designed to evaluate the appropriate generation or storage technology, location and capacity while incorporating the priorities assigned to the project.
The electrical network still has to work. The method models grid capacity over a 24-hour cycle, simulating electricity demand and power flow as loads change throughout the day. Hourly voltage behaviour can therefore be considered alongside the non-electrical constraints governing individual locations. A residential district coming alive at the end of the working day creates a different load profile from an industrial zone or data-centre cluster.
Translating social, economic and policy priorities into mathematical weights introduces its own difficulty. The researchers acknowledge that quantifying such parameters is complex, even though they argue that doing so is necessary to make the optimisation more realistic. The resulting recommendation therefore depends partly on the priorities supplied to the model. Two authorities applying different policy preferences to the same underlying network could reasonably obtain different answers.
That does not make the optimisation less useful. It makes explicit a judgement that conventional electrical models can leave outside the calculation.
From Electrical Optimisation to Infrastructure Planning
New electricity demand does not necessarily appear where spare grid capacity, suitable connections and developable land conveniently coincide. Data centres provide an obvious example: their demand can be substantial and concentrated, while the associated grid reinforcement, substations, generation and storage require their own land and connections. Industrial expansion can present similar difficulties.
Infrastructure planning can consequently become an iterative process as electrical requirements, possible connections, land availability and planning constraints are reconciled. Incorporating external restrictions during optimisation offers a way of eliminating unsuitable locations before engineering effort converges on a site that proves difficult to develop.
The model does not remove the political element from infrastructure planning, nor does it attempt to decide what a community should value. Someone still has to determine the priorities assigned to competing objectives. Engineering establishes feasible options within a wider collection of economic, environmental, planning and social constraints.
A Research Tool Rather Than a Planning Shortcut
The approach should not be confused with an automated planning authority. Its usefulness depends on the quality of the constraints, policies and stakeholder preferences supplied to it, while a model cannot capture every issue that emerges during permitting, land acquisition, detailed engineering or public consultation.
The underlying work nevertheless provides a framework for testing choices before substantial resources are committed to a particular location. Because the weighting system can be adjusted, the researchers say it can respond to changes in government policy and local priorities rather than embedding one fixed definition of an acceptable development.
The peer-reviewed study was produced by Rodney Itiki, Qianxue Xia and Suman Debnath at ORNL’s Energy Science and Technology Directorate. The research was funded by the US Department of Energy’s Integrated Energy Systems Office.
The broader possibilities extend beyond electricity generation. Itiki has suggested that the same type of approach could eventually be applied to the siting of critical-materials mining operations, AI data centres and disaster-response infrastructure, as well as infrastructure associated with large international events such as the FIFA World Cup and Olympic Games.
Those applications would require their own constraints and modelling, but they share the same underlying problem. Infrastructure rarely occupies an empty map. A technically attractive location already has neighbours, roads, buildings, regulations, economic functions and competing claims on land. Treating those conditions as part of the engineering problem rather than inconvenient details encountered afterwards could make optimisation considerably more useful in the places where infrastructure is hardest to build.
Key Industry Questions
- What has Oak Ridge National Laboratory developed?ย ORNL researchers have developed an algorithm for determining the location, size and type of generation or energy-storage installations within urban power distribution systems while accounting for external planning constraints.
- How does the algorithm differ from conventional power-system siting models?ย It incorporates factors outside the electrical network, including zoning, noise restrictions, policy priorities, architectural and historic considerations, tourism and stakeholder preferences.
- Does the algorithm still model electrical grid performance?ย Yes. It includes power-system simulation and assesses changing demand and power flow over a 24-hour period, including hourly voltage variations.
- How much can urban restrictions affect available sites?ย ORNL researcher Rodney Itiki said zoning and noise regulations intended to protect residential areas and tourism can reduce available sites by 20 to 30 percent.
- Can local authorities change the priorities used by the model?ย Yes. The weighting system is designed to incorporate different policy and stakeholder priorities, allowing the model to be adapted as local objectives or government policies change.
- Is the system limited to renewable energy?ย No. The method is intended to evaluate generation and storage options rather than being restricted to a single energy technology.
- Could the approach be used for AI data centres?ย Potentially. The researchers have identified AI data-centre siting as one possible future application of the wider planning approach.
- Does the algorithm replace conventional planning and permitting?ย No. It provides a method for incorporating relevant constraints earlier in site selection and optimisation. Development would still be subject to the applicable planning, permitting, engineering and consultation processes.
- Could similar modelling be used outside cities?ย The research is specifically concerned with the additional external constraints encountered in urban and semi-urban environments, although the underlying approach of combining technical optimisation with external priorities could have wider applications.
Strategic Takeaways
- Urban energy capacity cannot be planned solely from the electrical network when viable development land is constrained by existing uses and regulation.
- Incorporating planning restrictions during optimisation can eliminate unsuitable locations before detailed project development begins.
- A weighting system allows technically feasible infrastructure choices to reflect different municipal and economic priorities without assuming those priorities are universal.
- Concentrated loads such as data centres and industrial developments can make coordination between energy and land-use planning increasingly important.
- The underlying approach could extend beyond electricity infrastructure wherever technical site optimisation collides with competing uses of land.
