Artificial intelligence empowers new power systems and fosters the cultivation of new economic drivers

This year, the "Government Work Report" introduced the "AI Plus" initiative for the first time. "AI Plus" leverages technologies such as large models, big data, and high computing power to achieve deep integration of artificial intelligence into various industries, creating a new development ecosystem. Starting today, the Industry and Economics section will launch a series of reports, focusing on how artificial intelligence empowers sectors like electricity, express delivery, and manufacturing, and helps cultivate new economic drivers. Please stay tuned.

Facing the significant challenge of carbon neutrality, the most effective methods are to increase the proportion of clean energy, enhance energy efficiency, and intensify the substitution of electric energy. However, clean energy and electric energy substitution are long-term processes. For most industries and enterprises, enhancing energy production and usage efficiency through digital technology remains an effective path towards green and low-carbon development.

As an advanced form of digital technology development, artificial intelligence is transforming the modes of electricity production, operation, and transmission, supporting the construction of new power systems and new energy systems. With the accelerated penetration of digital and intelligent technologies, how will artificial intelligence empower the development of the electricity industry? How will risks and challenges be addressed?

Increasing Coverage

In May, photovoltaic power generation continued to climb, posing a challenge on how to better consume this clean green electricity. If we can monitor and accurately predict the operational status of photovoltaic power stations in real time, integrating a vast number of photovoltaic power stations will be a crucial step.

In Jiangsu Province, a major province for distributed photovoltaic installations, State Grid Jiangsu Electric Power Co., Ltd. leverages artificial intelligence and utilizes a distributed photovoltaic observable and measurable system to monitor and accurately predict the operational status of over 610,000 low-voltage distributed photovoltaic systems with a capacity of more than 25 million kilowatts in real time. By analyzing the patterns of photovoltaic output curves and combining information such as weather forecasts and cloud images, it achieves minute-level real-time observation and short-term predictions for the next 10 days at the county level. Compared with the actual output of distributed photovoltaics, the real-time estimation accuracy reaches 97%, and the day-ahead prediction accuracy reaches 95%. This provides a strong basis for grid dispatching predictions and power balance, better serving the local consumption of distributed photovoltaics.

Power inspection has also entered the intelligent stage. On May 10th, a strong wind blew through Shizuishan City, Ningxia, and a plastic strip more than one meter long was blown onto an insulator of a substation. The central monitoring room, located more than 30 kilometers away from the substation, immediately received an alarm message and accurately displayed the real-time image and location information. A processing instruction was issued immediately, and the plastic strip was quickly removed, eliminating the safety hazard in a timely manner.

Liu Jiang, the monitor team leader of the Substation Operation and Maintenance Center of State Grid Shizuishan Power Supply Company, told reporters that currently, the 46 substations under the jurisdiction of the company have remote intelligent inspection capabilities, with a 100% coverage of inspection points. This allows for precise identification of defects such as foreign objects, heating, and cracks at all times and from multiple perspectives, significantly improving the efficiency of defect verification. In the past, it took about 4 days for a team of 2 people to complete routine inspections of 46 substations, but the AR "Smart Inspection Eye" can complete them all in 2 hours.

Not only in power dispatch and maintenance, intelligent applications have also expanded to cover the entire energy governance domain. In Xiaoshan District, Hangzhou City, Zhejiang Province, the power department has deeply integrated digital technology with power business, leveraging local resource endowments to create a comprehensive energy management and control platform. This ensures grid safety while forming a broader and more extensive influence, serving economic and social development.

Lv Bin, the deputy director of the Operation and Inspection Department of State Grid Xiaoshan District Power Supply Company in Hangzhou, explained that the comprehensive energy management and control platform integrates multiple resources, bringing together vast resources such as charging piles, enterprise microgrids, and grid-side energy storage to achieve intelligent collaboration. It also gathers terminal sensing devices such as intelligent fusion terminals and user-side terminal acquisition devices to monitor the real-time status of the grid, diagnose anomalies, and evaluate operations.

Overall, AI-related technologies have penetrated into multiple aspects of power generation, transmission, transformation, and distribution, with application directions mainly including fault detection and prevention, equipment maintenance and optimization, grid optimization and management, automated reporting and document generation, virtual assistants, and customer service.

Leveraging the First-Mover Advantage

Under the "dual carbon" goals, the energy system has undergone profound evolution and changes: energy categories have become more diversified, the number of power sources has increased significantly, the grid structure has become more complex, and energy consumption has become flexible and variable. To successfully complete the green and low-carbon transformation of energy while ensuring energy security and low energy costs, it is necessary to intelligently transform the traditional energy system.

According to the International Energy Agency's prediction, the large-scale application of digital technology will reduce the production cost of oil and gas by 10% to 20%, and by 2040, the curtailment rate of solar photovoltaic power generation and wind power generation will decrease from 7% to 1.6%. Research shows that digital technology can help China reduce carbon dioxide emissions by 1.4 billion tons annually. Currently, the energy industry as a whole is in the process of moving from informatization to intelligence. Due to differences in characteristics, various energy systems are at different stages of development. Among them, the power system has already entered the stage of intelligence.

According to a report released by the State Grid Energy Research Institute, the digital transformation of the power industry contributes more than 70% to the energy sector, primarily due to the early advantage of the power industry in having a high proportion of electronic equipment. This advantage should be fully leveraged to promote the innovative development of power big data, digital technology, and digital business models.

On the generation side, whether it's traditional thermal power plants or new energy power stations, digital technology can empower smart power generation, reduce operation and maintenance costs, enhance energy conversion efficiency, help power plants generate more electricity, and support the large-scale and high-proportion grid connection and consumption of new energy. On the grid side, digitalization can enhance the intelligent interconnection capability for efficient resource allocation. On the user side, digital technology enables extensive interaction, full sharing, and value mining of end-user data, improves comprehensive perception and smart interaction capabilities of end-user energy consumption status, supports efficient and convenient access of various energy consumption facilities, and thus meets the personalized, diversified, and interactive energy consumption needs of various users.

"Artificial intelligence technology is of great significance for accelerating the construction of a new power system and a new energy system," said Jiang Chengling, deputy director of the Safety Operation Department of the Digitalization Department of State Grid Jiangsu Electric Power Co., Ltd. She believed that artificial intelligence, as an important support for the digital transformation of the power industry, plays an irreplaceable role in promoting data business integration, mining the value of massive data, and driving business optimization and upgrading.

Jiang Chengling said that in the past, due to factors such as difficulties in data acquisition, insufficient computing and analysis capabilities, and reliance on manual decision-making and prediction, the power industry faced pain points such as difficulties in matching power supply and demand, low efficiency in equipment operation and maintenance, and low efficiency in new energy consumption. Artificial intelligence technologies such as computer vision, deep learning, and neural networks provide new ways to solve these problems with their powerful data processing and learning capabilities.

Reliability needs to be improved

The application of artificial intelligence technology has brought benefits to the power industry, but it has also brought unavoidable problems and challenges.

Currently, many intelligent products of energy and power enterprises have problems such as poor application, poor experience, and low value creation. There are still bottlenecks in the research and development layout of key technologies. Enterprises generally reflect that the research and development of intelligent technologies such as big data analysis, artificial intelligence, and blockchain are not closely integrated with industry application scenarios, and system compatibility issues are widespread. Many important systems still lack mature alternative products or solutions.

The reporter found through interviews that the application of artificial intelligence in the power industry is still in its initial stage of testing. Most of it is only a spark on a point, and it has not yet formed a prairie fire. There is still room for improvement in the level of intelligent application.

Meng Zhenping, Secretary of the Party Committee and Chairman of China Southern Power Grid Co., Ltd., mentioned that AI is currently in a critical period of rapid evolution from perceptual intelligence to cognitive intelligence and decision-making intelligence. It is necessary to deeply cultivate the development of large models in the power industry and accelerate the formation of large-scale application supply capabilities. Leveraging the advantages of the energy and power industry, such as large-scale demand, complete industrial support, and numerous application scenarios, we should deepen the application of AI in production operations, customer service, emerging businesses, and other fields, explore more high-value business scenarios, and drive the accelerated iteration and upgrading of AI technology products.

It is worth noting that as AI is gradually promoted in the energy and power industry, the digital technology architecture of the industry is becoming increasingly complex, and the risks of stable and secure operation of information systems are more concentrated. On the one hand, in the face of complex and changing on-site operational environments and application requirements, whether AI's performance indicators such as recognition rate, false alarm rate, and missed alarm rate can meet application requirements, and whether AI's computing power can meet on-site needs, all require thorough testing. On the other hand, AI technology also provides new attack methods for criminals. Data poisoning, algorithm backdoors, adversarial sample attacks, and other AI-based techniques pose new challenges to grid information security.

Li Yufeng, a professor at the School of Artificial Intelligence, Nanjing University, stated that while AI brings convenience, it also poses certain security risks. A comprehensive and effective governance mechanism should be established, and a robust theoretical method system for machine learning should be developed, to better apply AI technology safely in the power sector.