On October 11, at the scientific and technological achievements appraisal meeting organized and held by the China Electricity Council, the "research and engineering application of key technologies for wind turbine diagnosis and early warning" completed by Longyuan (Beijing) Wind Power Engineering Technology Co., Ltd was appraised to be on par with the international advanced level.
Under the background of Longyuan Power's integration of IT application with industrialization and regional maintenance, the company's research on fault diagnosis and early warning technology aims to timely and accurately find out the abnormal conditions of the equipment and give early warning to the operators before the equipment failure according to the change trend of the equipment operation status or parameters, so as to reduce the possibility of further expansion of wind turbine failure, reasonably arrange the operation inspection plan to improve the generator operation reliability and benefits. In view of the many difficulties of the company, such as the large number of machine types, the difficulty of deep integration of wind power mechanism and artificial intelligence, and the difficulty of closed-loop landing of early warning results, the company has developed a big data diagnosis and early warning platform to realize the fault diagnosis and early warning of wind turbine generator, and carried out engineering application. Based on the wind turbine design and operation mechanism, machine learning, statistical analysis and other analysis technologies, the platform has constructed four early warning models, including unit control, large unit components, unit capacity reduction and unit sensors, so as to realize unit fault early warning. At the same time, by using the design operation parameters, operation data and fault cases of wind turbines, the knowledge base of operation parameters of different models and the real-time detection module of model accuracy are constructed, which laid the foundation for the generalization and promotion of diagnosis and early warning model, so as to realize the large-scale engineering application of diagnosis and early warning and the whole life cycle management of diagnosis and early warning model.
At present, the achievements of the project have been successfully applied to 13,100 units of the company, with an early warning accuracy of 75.3%, and the cumulative economic benefit has increased by 93.335 million yuan since the platform was launched and tested in 2018.
