Design of an Edge Computing System for Wind Turbine Condition Monitoring and Diagnosis of Variable-Speed Operating Conditions

Abstract

As wind turbines continue to grow in size and power density, issues such as drive system misalignment and the reliability of critical components have become increasingly prominent, posing major risks to the safe operation of wind farms. Traditional condition monitoring systems face bottlenecks such as closed data structures, fixed algorithms, poor monitoring timeliness, and difficulties in feature extraction under non-steady-state conditions. This paper designs and implements a browser-server (B/S) architecture-based edge condition monitoring device for wind turbines, aiming to establish an intelligent closed-loop monitoring system characterized by “data interconnection, autonomous algorithms, and proactive early warning.” The research includes the development of a hardware module supporting 24-bit high-precision synchronous acquisition of multi-source signals, and proposes a feature extraction method based on instantaneous angular velocity analysis and angular domain resampling for variable-speed operating conditions. Furthermore, through a collaborative edge-cloud architecture, the system enables dynamic algorithm updates and efficient access to massive amounts of data. Application validation at actual wind farms demonstrates that the system significantly improves fault detection accuracy and early warning response speed under complex operating conditions. This research not only provides a highly adaptable solution for the intelligent operation and maintenance of wind power equipment but also offers theoretical and practical support for the application of edge computing technology in industrial equipment.