Digital grid empowers the construction of a new power system in three steps

China Energy Storage Network: As the world enters the digital age, traditional technology systems can no longer meet the needs of building digital power grids and smart power grids. How to make full use of new-generation information and communication technologies such as big data and artificial intelligence to accurately "describe" the complex laws of power system operation, establish a power system learning model from data to knowledge, from knowledge to decision-making, and ensure the safe, reliable, green, efficient, and intelligent operation of the power system has become the focus of exploration in the power industry.

Traditional "knowledge expression" cannot meet the needs of new power system construction

随着以新能源为主体的新型电力系统加快构建，大规模新能源并网和电力市场开放后，电力系统形态将发生重大变化，电力网络、信息网络和社会网络之间的耦合关联性显著增强，新型电力系统呈现出非线性、强随机、快时变的复杂巨系统特点。在这种情况下，单纯离线建模和仿真技术难以满足复杂电网实时运行分析与精准前瞻调控的要求，同时直接运用传统的调控模型与算法体系也面临海量电力系统中资源分散分离和构成功能耦合及最优快速决策等挑战。

Therefore, the construction of a new power system faces some urgent problems in the source, grid, load and storage links. Among them, on the source side, it is necessary to provide more flexible access technology and interface methods to ensure the large proportion of new energy consumption; on the grid side, it is necessary to build faster computing capabilities and control methods to adapt to the trend of high-proportion power electronics in the power system; on the load side, it is necessary to explore more flexible interactive technologies and communication channels to fully mobilize the enthusiasm of the demand side to participate in system regulation; on the storage side, it is necessary to achieve more efficient dynamic balance and optimized adjustment to improve the stability and control level of the power system.

In the face of the above challenges, the digital power grid that integrates multiple "knowledge expressions" will provide a more core technical approach and make the role of the power grid as a resource allocation platform and electricity-carbon economic service platform more prominent.

Digital grid plays an important role in supporting the construction of new power system

The role of digital grid in supporting the construction of new power systems is mainly reflected in the following three aspects:

First, data and its measurement. In the era of the Internet of Everything, there is no decision-making and no operation without data. Full data collection and processing are the basic conditions for ensuring the large-scale grid connection and consumption of new energy. Among them, data has become the primary factor to ensure that the power system is "observable, measurable, and controllable", and it is also the key foundation of the power grid command system and decision-making center.

Therefore, to achieve comprehensive observability of the new power system, it must be based on sufficient and effective measurements, and the digital grid has extensive data acquisition and processing capabilities. Through the massive sensors deployed in the power system, the physical structure of the power system can be accurately grasped, thereby gaining insight into the performance, operation mode, real-time status, operating efficiency, health status and environmental protection level of each component unit and the whole.

Second, the comprehensive application of intelligent algorithms and computing power. The organic integration of effective intelligent algorithms for specific fields and powerful heterogeneous computing power is an important means to adapt to the new form of the power grid and meet the new requirements of planning, operation and management.

The dynamic behavior of new power systems is more complex, and higher accuracy and speed of calculations are required. Among them, the main body of new energy means that the characteristics of double high (high proportion and high power electronic equipment) are obvious. Due to the short time sequence of state change, wide frequency domain distribution of sequence signals, and mixed variables affecting dynamic processes, it is difficult to describe and solve the system using traditional static models with fixed parameters as the core. It is necessary to establish high-performance simulation computing capabilities that are suitable for large-scale strong random systems.

Third, rapid coordination. The new power system places high demands on rapid coordination capabilities. As the interaction between upstream and downstream entities of the power grid is strengthened, the content and form of power grid management will change frequently. It is necessary to grasp the main line of data and improve the flexibility and openness of the enterprise's digital operation system to achieve full-chain perception and comprehensive connection of planning and construction, material supply, safe production, asset finance, etc., improve business efficiency, and promote management changes.

Based on years of observation, induction and deduction, the power industry has accumulated rich experience, rules and knowledge, which can describe the external structure of power infrastructure, changes in the state of system electrical quantities, topological connection relationships, etc., and integrate this knowledge into artificial intelligence algorithm models to form a new type of intelligent algorithm driven by data, guided by knowledge and physically modeled. It also uses "knowledge expression" to characterize the laws contained in the data, and then form a "human-machine collaboration" model. This depends on building a complete "knowledge system" covering massive multi-source data, algorithms and applications of the power system.

Digital Grid "Knowledge Expression" System

The high-dimensional, dynamic and uncertain nature of the new power system poses huge challenges to the safe and stable operation of the power grid. Traditional methods are difficult to accurately and completely characterize and control the huge power system in real time. In contrast, the multiple knowledge expressions of the digital power grid will promote the new power system to become "observable, measurable and controllable" a reality.

Through the multiple "knowledge expressions" of the digital power grid, the characteristic laws of the physical power grid can be extracted, the morphology of the physical power grid equipment, the trend of system operation, and the correlation between the human-machine-object ternary space can be accurately described, thus achieving optimal decision-making and control of the physical power grid.

Based on the three AI 2.0 knowledge expressions proposed by Pan Yunhe, an academician of the Chinese Academy of Engineering (the figurative expression of knowledge, the linguistic expression of knowledge, and the deep neural network expression of knowledge), the new power system supported by the digital power grid is concretized and enriched. There are four main forms of multiple "knowledge expressions": the figurative expression of digital power grid knowledge is mainly used to describe the form of physical power grid equipment; the functional expression of digital power grid knowledge is mainly used to describe the physical laws of the time series changes of various data of electrical and non-electrical quantities in the power system; the linguistic expression of digital power grid knowledge is mainly used to describe the relationship between man, machine, object and environment in the power system; the deep neural network expression of digital power grid knowledge is used as an effective data-driven tool to supplement and support the above three types of applications, thus forming an artificial intelligence model that unifies "data-driven, knowledge-guided and physical modeling".

(Huang Wenqi is the head of the artificial intelligence and intelligent software team at the Digital Power Grid Research Institute of China Southern Power Grid; Sun Lingyun is the vice dean of the School of Computer Science at Zhejiang University; Wu Fei is a professor at the School of Computer Science at Zhejiang University)