domeet webmaster

In order to predict a complex and uncertain world, you cannot rely solely on past data and data science, you must combine it with the expertise of society.

This is the CEO of Cosmo Tech Hugues de Bantel said. The French company hopes to use this technology to change the power production and distribution through AI in the future.

This seems to be similar to the idea of ​​Tesla’s founder, Musk – and AI team, together to answer future questions.

Cosmo Tech The company chose to create software to help decision makers make decisions in the most complex environments. Software is used to predict the impact of decisions on their business, which is in the area of ​​enhanced intelligence. The idea of ​​enhancing intelligence is to combine human intelligence with artificial intelligence. Once this is done, we have the ability to make predictions, make hypothetical scenarios, and make decisions.

And these technologies can even be applied to power industry.

We all know that solar and wind energy are renewable clean energy and a type of future energy, but they are not guaranteed to provide stable and sustainable predictable energy. Their presence in urban power grids complicates energy balance and optimization. This has led to the loss of renewable energy and the failure of the power supply.

In 2003, a massive blackout due to clean energy in the northeastern United States caused 50 million people to lose power for several days. Today, private residential apartments in Europe and the United States are also installing solar panels to power themselves, but private residential users cannot consume the energy generated by solar panels. Power companies also need to purchase (receive) these excess energy to make these surplus The power needs to be returned to the grid.

The oversupply of personal solar panels is a situation, and there is also a situation where electricity is in short supply. When demand exceeds supply, the power company will start a fossil fuel power plant, which will enable the “peak power plant” to avoid power outages. This process is expensive, and we should try our best to avoid this in terms of consumer cost and environmental impact.

This is a microcosm of smarter energy applications in the grid, which means that the “smart grid” that provides energy on demand will become increasingly important in the coming years.

To this end, the National Grid Corporation of the United Kingdom is in preliminary negotiations with Google’s DeepMind to develop a system to accurately predict demand and more effectively balance the national energy system with the supply of wind and solar.

The US Department of Energy is also working with the National Accelerator Laboratory at Stanford University to develop artificial intelligence and machine learning algorithms to build an autonomous grid that processes data from satellite imagery, utility operations, and other sources..

This is still a case of feeding data to AI and then feeding back humans by AI. The AI ​​initially receives a large number of cases of grid system outages, and then they can gradually learn to distinguish (and accurately classify) normal operational data from defined system faults.

Training, researchers can apply AI to In the current data, instead of the manual processing method. This is the first real-time application of AI to the grid, where it can detect where anomalies or faults occur and define the type and location of the disturbance.

If a power plant fails, the load on other power plants may suddenly peak. The increased load slows the generator and reduces the frequency. This requires a quick decision, quickly referring to less than 500 milliseconds.

And AI is the “person” who can make decisions within 500 milliseconds. It can make decisions in 20-50 milliseconds and have enough time to make adjustments to ensure the normal operation of the grid.