Alphabet's latest lunar project is to solve the dilemma of renewable energy grids

Alphabet-owned X Lab is working on grid-related tools to help utilities monitor and predict activity on the grid. Officials mentioned that the research project is part of the Moonshot project, which aims to develop and test the prototype of the electrical network monitoring system.

The power grid supplies power for many important infrastructure equipment, but this is an engineering design for more than a century, the purpose is to let electricity flow from fossil fuel power plants to towns, with the development of science and technology, countries around the world are pushing renewable energy, so the power supply of renewable energy has also begun to flow into the grid, and consumers are gradually using electric vehicles, or installing solar panels, these changes in power generation and electricity consumption, so that the original grid balance is impacted, X Lab mentioned that the challenge of grid balance will only become bigger and bigger, Because to combat global climate change, it is necessary to achieve 24/7 complete carbon-free electricity.

As the project involved a complex grid project, the research team enlisted Audrey Zibelman, the chief executive of former Australian energy operator AEMO, who weathered the Australian grid safely through wildfires in 2019 and 2020 and was involved in the restructuring of new York's power system after Hurricane Sandy hit hard.

The fundamental obstacle to monitoring and predicting renewable energy power is the inability to determine the amount of renewable energy flowing into the grid, the researchers mentioned that grid operators need to ensure that the power supplied and used in the grid can maintain a balance every day, but there are many uncertainties in management and decision-making, the biggest of which is the inability to grasp the actual situation of intermittent electricity flowing into the grid such as wind power and solar power.

It is also impossible to get an immediate picture of how energy flows, and there is a lack of a consistent global view map and an end-to-end summary view map for all suppliers, so it is impossible to see the actual situation of power plants, even rooftop solar panels in the average home, and these electricity enter the grid. Moreover, the needs of different units are different, and the methods of mastering the power grid are also different, from utilities to system operators, different tools and system models are used at each stage of management, construction and regulation.

For these reasons operators face enormous obstacles, especially as millions of devices in the grid begin to contribute and intercept power to the grid. The researchers mentioned that one of the entry points to solve this complex situation is to build a single view map for the fragmented power system, and the researchers believe that machine learning techniques can help with this task, predicting the power of the overall grid by predicting various situations.

The research team discussed the key points of several grid projects, including the possibility of creating detailed and real-time grid view maps, even to accurately predict the weather, know the time of sunny days and winds, to further predict power generation, and also hope to predict and simulate future events in the grid.

At a time when countries are desperately trying to decarbonize, the research team mentioned that operators need to redesign the way they update and manage the grid, and this matter cannot be solved by a single organization, so the cross-Alphabet team implements the lunar project and integrates internal and external resources to work together to solve this problem.