Modernizing Hydropower with Digital Twins

In April 2019, the 850-year-old Notre Dame cathedral was destroyed in a fire. 

A mammoth effort is now underway to rebuild the 12^th century monument in time for the 2024 Olympics in Paris. To restore the cathedral to its original glory, scientists and designers from around the world are creating a “digital twin” of the building.  

The digital twin integrates tens of billions of 3D coordinates, archived photograph collections, and construction sketches to create a database of scientific and technical data about the cathedral to aid in its restoration. 

The digital twin concept is also being used to solve more complex science and engineering problems. One of the most significant applications in the works is using digitalization to modernize the nation’s aging hydropower plant fleet. 

Pacific Northwest National Laboratory (PNNL) and Oak Ridge National Laboratory (ORNL) are collaborating on a project, sponsored by the U.S. Department of Energy’s Water Power Technologies Office, to create the Digital Twins for Hydropower framework. Over the next five years, PNNL and ORNL are expected to develop a virtual, open-access platform that creates a digital representation of hydroelectric plants using virtual and real-time data and feedback to optimize operations. 

The platform will serve as a place for the hydropower industry to evaluate and replace mechanical components, accelerate technology development, and improve hydropower operations and performance. These improvements reduce service downtime and shutdowns, which interfere with delivering electricity to the grid. 

“The goal here is for hydropower to remain an efficient, affordable resource to own and operate, and to address the ever increasing demand of providing grid resiliency,” said Osman Ahmed, a PNNL mechanical engineer and advisor leading the Digital Twins project. 

Aging hydropower fleet  

There are 50 hydropower plants in the United States that have been in service since 1908, according to the U.S. Energy Information Administration. These plants represent the oldest electric generating facilities in the country. 

As the hydropower fleet ages, the industry faces the challenges of operational stress to older mechanical components and associated costs of degrading functions.

With an average life expectancy of 40 to 60 years, hydropower plants need preventative maintenance much like an aging car. By predicting and planning for components that need to be replaced or updated, the hydropower industry can maintain flexible and reliable energy services. 

Modernizing hydropower 

Digital Twins uses next-generation technologies, including artificial intelligence, machine learning, and virtual reality to simulate hydropower generation, transmission, and distribution systems. The simulations can predict plant performance under various types of market demand and complexities.

“If you have the digital twin, it can help you to decide what kind of investment is needed before you spend money,” Ahmed said. “Maybe you kick the virtual tire and take it for a virtual test drive. It’s a big area of opportunity.” 

PNNL talks with hydropower industry 

In this first year of the project, PNNL is designing an open platform framework of a Digital Twin needed to modernize the hydropower fleet. The PNNL team is engaging with owners, operators, utility companies, and technology providers in the hydropower industry to get their feedback and incorporate the elements and components that should be included in the platform. For example, industry stakeholders are weighing in on what elements in the platform would help a hydroelectric dam perform more efficiently, flexibly, and affordably.

PNNL has already discussed with five different hydroelectric plant owners and they all have communicated that they want tools to help improve the facility performance. 

“What we’re hearing is that the industry needs a multifaceted design that accounts for operations, predictions of failure, how to avoid downtime, and scenario planning with the different complexities,” Ahmed said. “If you create a good Digital Twin, they can all learn from each other.” 

ORNL will use their capabilities in modeling and simulation to develop the prototype of the platform to demonstrate to the hydropower industry how it works. The platform is expected to be built next year and should be customizable for users within the next three to four years. Users will be able to pick and choose à la carte mechanical components to add to or change their plant. 

In the long-term, the platform is expected to create a library of hydropower plants across the U.S. fleet so that users can adapt their plant based on others they’ve seen work effectively.