Application of key technologies for equipment health status early warning and diagnosis under new energy multi-scenario smart operation

 Driven by my country's carbon peak, carbon neutrality strategic goals and "new infrastructure", new energy on the power generation side has become an important part of the new power system and will become the main force on the energy supply side in the future; on the energy consumption side, new energy electric The rapid development of automobiles and charging facilities has made carbon emission reduction the norm. On the one hand, after large-scale new energy is connected to the new power system, due to the complex operating conditions of the equipment, the coupling and correlation between fault characteristic parameters and the fault evolution process, and the uncertain demand for source-load coordination in multiple scenarios, it faces low operating efficiency, Problems such as low utilization of data assets, difficulty in early warning of equipment health status, and difficulty in diagnosis; on the other hand, there are more and more fire accidents in electric vehicle charging. As the service life of the vehicle piles declines, it is difficult to accurately predict the safety warning boundary, facing charging challenges. There are problems such as the difficulty of safe real-time and accurate detection, vehicle-pile coordinated early warning, diagnosis, and active safety protection.

 Under the joint research of the National Natural Science Foundation of China, State Grid Tianjin Electric Power Company, Guoneng Star Technology Co., Ltd. and other industry-university-research science and technology projects, this project has carried out new energy network control consisting of wind power, photovoltaic, energy storage, vehicle piles, etc. Systematic research, through theoretical analysis of the operational stability of complex network systems, carries out joint industry-university-research-application research from the three dimensions of "system theory-key technologies-demonstration applications". It took 8 years to develop and improve the new energy network switching control system, propose a modeling method that combines equipment mechanism models and multi-source data drivers in new energy multi-scenarios, and develop digital twins of systems and equipment suitable for multi-scenarios. Model, a cloud platform-based new energy operation optimization and fault warning diagnosis system, charging safety intelligent monitoring and diagnosis equipment and a big data security protection platform were developed to solve the problems of difficult real-time monitoring and intelligent early warning and diagnosis of charging faults, forming a A complete set of digital solutions improves operation and maintenance efficiency and has been promoted and applied. The main innovation points are as follows:

 (1) Proposed a new energy network switching control system including wind power, photovoltaic, energy storage, vehicle piles and other nodes, explained the event triggering control mechanism of the network switching control system, designed a zero-obstacle function based on security protection requirements, and analyzed the network Stability of switching systems under the framework of multi-Lyapunov theory.

 (2) A modeling method that combines equipment mechanism models and multi-source data drivers in multiple scenarios is proposed, and a digital twin system model of equipment in multiple scenarios is established, which enhances the model's situational awareness and improves the model accuracy of new energy equipment. The full-cycle operation rehearsal of the equipment is realized through different speed accelerations.

 (3) A full life cycle health assessment early warning and diagnosis system for equipment oriented to wind power, photovoltaic, vehicle piles and other scenarios was constructed, and a new energy equipment full cycle health assessment and diagnosis method was proposed that improves the particle swarm algorithm to optimize the deep confidence network, and is accurate. It depicts the intrinsic information and fault evolution mechanism of the equipment, improving the practicality of the equipment health assessment, warning and diagnosis system.

 (4) A charging pile fault monitoring and diagnosis technology that improves RNN neural network is proposed, forming an intelligent early warning library with self-updating capabilities; a vehicle-pile integrated safety status evaluation model that improves gray correlation fuzzy evaluation is constructed, forming a cloud Edge collaborative active security protection system.

 The project has been authorized to have 18 invention patents, 15 utility model patents, and 23 software copyrights; it has published 88 papers, including 43 SCI indexed papers and 35 EI/Chinese core articles; it has trained 10 doctors and 45 masters.

 The project generates direct economic benefits of approximately 3 billion yuan and new profits of approximately 500 million yuan. On the new energy power generation side, the project results have been promoted and applied in more than 120 wind farms and more than 20 photovoltaic power stations in 9 power generation groups in 21 provinces and regions across the country; on the power distribution side, the project results have been demonstrated and applied in Tianjin and promoted to Shandong, There are more than 150 large-scale charging stations in 21 provincial-level administrative regions across the country, including Hebei. The project results in the past three years have increased power generation by 290 million kilowatt hours, reduced carbon emissions by approximately 250,200 tons, and saved approximately 92,700 tons of standard coal. The project promotes the sustainable development of the new energy industry and "new infrastructure" and provides support for the national strategic goal of "carbon peaking and carbon neutrality".