Communication anomaly detection method and system for wind turbine variable pitch control system
By performing time-sequential processing and adaptive decomposition of the CAN message stream of the wind turbine pitch control system, and combining it with the isolated forest model for anomaly scoring, the problem of insufficient accuracy and robustness of communication anomaly detection in the existing technology is solved, and early fault identification and predictive maintenance of the communication system are realized.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- BEIJING HUANENG XINRUI CONTROL TECH
- Filing Date
- 2026-03-24
- Publication Date
- 2026-07-10
AI Technical Summary
Existing methods for detecting communication anomalies in wind turbine pitch control systems cannot effectively predict communication degradation trends, resulting in insufficient accuracy and robustness in communication anomaly detection. This makes it impossible to identify potential faults in a timely manner, increasing system complexity and maintenance costs.
By performing time-series processing on the original CAN message stream, extracting the inter-frame arrival time sequence, and combining it with SCADA data to determine the operating conditions in real time, the system performs time-series decomposition that is adaptive to the operating conditions. It then uses a pre-trained isolated forest model to perform anomaly scoring and cumulative evaluation, and outputs the health status of the communication system.
It significantly improves the accuracy and robustness of communication anomaly detection in wind turbine pitch control systems, enabling early identification of potential faults, reducing false alarms, and enhancing the system's predictive maintenance capabilities.
Smart Images

Figure CN122372465A_ABST