The prospects of "artificial intelligence + energy Internet" are promising

China Energy Storage Network: Artificial intelligence will play a strong supporting role in promoting the intelligentization of energy production and consumption. In turn, "Internet +" smart energy will also support the development of artificial intelligence.

Building a safe and efficient energy internet

The key tasks of the "New Generation Artificial Intelligence Development Plan" are to build a ubiquitous, safe and efficient intelligent infrastructure system. Vigorously promote the construction of intelligent information infrastructure, improve the intelligence level of traditional infrastructure, and form an infrastructure system that meets the needs of intelligent economy, intelligent society and national defense construction. Accelerate the promotion of digital and networked information infrastructure with information transmission as the core, and transform it into an intelligent information infrastructure that integrates integrated perception, transmission, storage, computing and processing. Coordinate the use of big data infrastructure, strengthen data security and privacy protection, and provide massive data support for artificial intelligence research and development and widespread application. Build high-performance computing infrastructure and enhance the service support capabilities of supercomputing centers for artificial intelligence applications. Build a distributed and efficient energy Internet, form a new energy network that supports the coordinated complementarity of multiple energy sources and timely and effective access, promote intelligent energy storage facilities and intelligent power facilities, and realize real-time matching and intelligent response of energy supply and demand information.

At the infrastructure level, the "Plan" clearly proposes to build a distributed and efficient energy Internet. Gao Feng, executive vice president of the Energy Internet Innovation Institute of Tsinghua University, pointed out that the development of energy is a process of long-term separation and integration, and long-term integration and separation. In the early days of energy, it was a micro-energy network, where energy was generated and used locally. As the load became larger and larger, the energy network became more and more complex and large, forming a situation of centralized energy supply, and some problems emerged. The energy Internet is the energy system from centralized to distributed, and then to decentralized, to achieve a ubiquitous state. Distributed energy also requires energy stations. When it reaches a ubiquitous state, energy stations are no longer needed. Ordinary people have changed from a pure energy consumer to a producer and seller, consuming energy, producing energy, and possibly storing energy.

"Artificial intelligence will play a big role in building a distributed and efficient energy internet and forming a new energy network that supports coordinated complementarity of multiple energy sources and timely and effective access." Gao Feng said, "The abbreviation of 'Internet +' smart energy is energy internet. The two are equivalent concepts. Efficiency is a very important feature of energy internet. Efficiency refers to maximizing the energy efficiency of the entire distributed system, maximizing equipment utilization, maximizing investment efficiency, maximizing social benefits and maximizing user utility."

Gao Feng analyzed that "multi-energy coordination and complementation, timely and effective access" is the state of the energy system with Internet form. At this stage, the energy Internet is divided into two categories. The first is the energy system with Internet form. Just like the Internet has local area network, wide area network, World Wide Web, etc., the energy Internet has micro energy network, local energy network, global energy network, energy is optimized in different ranges, can be connected to the network, or disconnected and run independently when the external energy system fluctuates. The second category is energy services with Internet form, such as Didi taxi and shared bicycles, which are Internet sharing economies.

"The traditional energy system is to build power grids for electricity supply and heat networks for heat supply. Power grids are built to meet the peak demand of the power grid, and heat networks are built to meet the peak demand for heat, resulting in a huge waste of resources. In fact, heat and electricity are interchangeable. The coordinated and complementary nature of multiple energy sources is a variety of collaborative concepts in the energy Internet, which realizes the comprehensive optimization and complementarity of different energy sources and achieves the highest overall investment efficiency." Gao Feng said.

Timely and effective access is another goal of the Energy Internet. Just like the Internet system, users can access the Internet at any time by entering a password. In the future, energy will also learn from the Internet concept to achieve plug-and-play and timely and effective access. "Timely and effective access to new energy networks requires many intelligent means to ensure non-differentiated access to the system and good operation, which requires the support of artificial intelligence technology. This is the direction that the Energy Internet is committed to researching and promoting." Gao Feng introduced, "The construction of the Energy Internet requires the promotion of intelligent energy storage facilities and intelligent power facilities to achieve real-time matching and intelligent response of energy supply and demand information."

Artificial intelligence and energy internet support each other

There are more and more distributed energy infrastructures, and the energy curve of production is constantly changing. At the same time, the energy consumption of users is also constantly changing. In the era of centralized energy, dispatchers cannot timely, effectively and accurately judge and control the supply and demand curve of distributed energy. At this time, an effective auxiliary tool is needed—artificial intelligence, which can replace the human brain to optimize, analyze, judge, make decisions and issue instructions for massive data. Therefore, in the energy Internet, artificial intelligence will be well applied to support the development of the energy Internet.

"The distributed coordination of the energy internet requires artificial intelligence to support its distribution, and the distributed optimization coordination does require a lot of artificial intelligence means." Gao Feng pointed out. How to analyze massive amounts of data? How to pre-judge accidents and perform maintenance in advance? How to judge user behavior and provide better services to users? Solving these problems requires the application of artificial intelligence technology. Gao Feng introduced: "Artificial intelligence technologies such as image recognition and voice recognition have been well applied. They are emerging technologies that the energy internet is currently focusing on promoting and have very broad prospects."

Artificial intelligence requires information transmission, massive computing, and energy support. Gao Feng introduced that artificial intelligence and energy Internet support each other, and in turn, distributed energy Internet will also play a strong supporting role in the development of artificial intelligence. "The development of artificial intelligence has a greater demand for energy. For example, data center storage and supercomputing centers perform massive computing, which require a large amount of energy, and the energy needs to be efficient and reliable without interruption. The development of energy Internet and artificial intelligence is a complementary process." Gao Feng said.

"Energy Internet + Artificial Intelligence" can achieve comprehensive optimization and control of energy. For example, a coffee shop has multiple energy sources such as lights, freezers, and coffee machines. In the past, only the total amount of electricity could be collected. Now, through "Energy Internet + Artificial Intelligence", it can analyze where the load of the continuously used lights is, where the load of the intermittent coffee machine is, and where the load of the refrigerator compressor is, and on this basis, load management can be carried out. Activating existing resources is the core concept of the Energy Internet.

Gao Feng pointed out that there are some bottlenecks in the development of artificial intelligence in the field of energy Internet. Now, AlphaGo and some other successful cases of artificial intelligence have clear rules and massive training samples. It is difficult to find a suitable sample with massive data and definite rules. For example, how to judge whether a person is tired? This kind of fuzzy and biological state recognition is difficult. The success of AlphaGo also lies in the innovation of the algorithm, not simply using artificial neural network for training.

"Data barriers are a very important constraint on the development of artificial intelligence. If the government wants to create a good environment for the development of artificial intelligence, it must create an open and shared environment, open up relevant data, and achieve flourishing and mass innovation, so that breakthroughs can be made in artificial intelligence," said Gao Feng.