Wind Turbine Anomaly Detection via Peer Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Wind turbines experience downtime due to sensor malfunctions or operating conditions outside predefined limits, leading to energy production cessation, necessitating a method for predicting potential fault conditions to prevent unnecessary shutdowns.
Innovation Solution
A system comprising sensors, controllers, and computing devices that monitor and analyze performance data from wind turbines to detect anomalies by comparing individual turbine data with aggregated data from similar turbines, identifying deviations and filtering out unreasonable values to notify potential sensor errors, allowing for proactive maintenance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If sensor monitoring with predefined limits is used to detect fault conditions, then reliability is improved through fault detection, but productivity deteriorates due to unnecessary shutdowns
Solution Approach 1:
The system continuously monitors sensor data and compares it against dynamically updated expected values from peer turbines, creating a feedback loop that adapts to changing operating conditions. This allows the system to distinguish between normal variations and actual faults, reducing false positives that lead to unnecessary shutdowns while maintaining reliable fault detection
Solution Approach 2:
Instead of using fixed predefined limits for fault detection, the system dynamically adjusts detection thresholds based on real-time operating conditions and peer turbine performance data. By changing the reference parameters from static limits to dynamic expected values, the system maintains sensitivity to actual faults while tolerating normal operational variations
2Measurement precision
If individual turbine performance monitoring is performed, then measurement precision is improved for fault detection, but device complexity increases due to data aggregation and comparison systems
Solution Approach 1:
The system combines data from multiple peer turbines to create aggregated expected value profiles, leveraging collective performance data to improve individual turbine anomaly detection. By merging information from multiple sources, the system achieves higher measurement precision without requiring complex individual turbine analysis systems
Solution Approach 2:
Instead of analyzing complex individual turbine data patterns, the system creates simplified copies or representations of expected performance based on peer turbine aggregates. These copied expected value profiles serve as reference models for anomaly detection, reducing the complexity of the monitoring system while maintaining detection accuracy
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A device (135) for use in detecting anomalies in wind turbines is provided. The device includes a sensor interface (220) configured to receive an operating condition at a plurality of times from one or more sensors (120) associated with a first wind turbine (100), a communications interface (215) configured to receive an operating condition at a plurality of times from one or more sensors associated with a grouping of wind turbines similar to the first wind turbine, a memory device coupled in communication with the sensor interface and the communications interface and configured to store a series of performance data samples that include an operating condition. A processor (205) is coupled in communication with the memory device and is programmed to compare a performance data sample of the first wind turbine with a performance data sample of the grouping of wind turbines, produce a sensor error if the performance data sample of the first wind turbine deviates more than a predetermined amount from the performance data sample of the grouping of wind turbines, and display the results of the comparison of the performance data sample.