Wind Turbine Sensor Fault Detection via Dual-Reference Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting fault modes in sensors, such as those used in wind turbine generators, are limited in identifying and distinguishing various fault modes beyond optical fiber breakage, including partial or complete loss of power, calibration issues, and signal processing errors, which can lead to increased risk of failure and maintenance costs due to the remote operation of wind turbines.
Innovation Solution
A method that compares data output from a first sensor with reference data and data from a second sensor, which can measure different parameters like bending moment, rotation speed, or environmental characteristics, to determine if the first sensor has entered a fault mode, using low-level and high-level fault detection circuitry to identify and compensate for faults.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a single sensor is used to measure bending moment in wind turbine blades, then the device complexity is reduced, but the reliability decreases due to undetected fault modes
Solution Approach 1:
The fault detection system is segmented into two distinct levels: low-level detection comparing sensor output with reference data, and high-level detection comparing first sensor data with second sensor data. This segmentation allows comprehensive fault detection while maintaining modular system architecture that doesn't significantly increase overall complexity.
Solution Approach 2:
Reference data is pre-established for the first sensor during normal operation, creating a baseline for future fault detection. This preliminary action enables the system to detect faults by comparing against pre-stored expected values, reducing the complexity of real-time analysis.
2Reliability
If multiple sensors and comparison methods are implemented, then the reliability improves through comprehensive fault detection, but the device complexity increases
Solution Approach 1:
The detection system divides fault detection into two hierarchical levels with distinct functions. Low-level detection handles simple reference data comparisons, while high-level detection manages cross-sensor validation. This segmentation prevents the system from becoming overly complex by organizing detection tasks into manageable, specialized modules.
Solution Approach 2:
The second sensor serves multiple purposes: it provides alternative bending moment measurements for high-level comparison, and can also detect faults in the first sensor through data inconsistency analysis. This multi-functionality reduces the need for dedicated components, thereby limiting complexity increase while improving reliability.
3Reliability
If fault detection compares first sensor data with second sensor data measuring different parameters, then the ability to detect various fault modes improves, but the difficulty of detecting and measuring increases
Solution Approach 1:
The system accepts that different sensor parameters (bending moment vs. rotation speed vs. environmental characteristics) will be compared, and uses predetermined relationships to bridge these different parameter types. This approach handles the parameter diversity systematically rather than attempting to standardize all measurements, reducing the apparent difficulty of cross-parameter comparison.
4Loss of time
If comprehensive fault detection is implemented, then the loss of time for technician servicing is reduced through timely fault identification, but the device complexity increases
Solution Approach 1:
The segmented two-level detection architecture enables timely fault identification by assigning specific detection tasks to appropriate levels. Simple faults are caught at the low-level reference comparison stage, while more subtle faults requiring cross-sensor analysis are handled at the high-level stage. This segmentation ensures comprehensive monitoring without requiring all detection mechanisms to operate simultaneously at full complexity.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method (500) of detecting a fault mode of a sensor is provided. The sensor may be, for example, a bending moment sensor (142) and may sense a bending moment of a blade (140) on a wind turbine generator (WTG) (100). The method (500) includes comparing (520, 530) data output by a first sensor (142) with reference data indicating what is expected to be output by the first sensor (142) to produce a first comparison result and comparing (550) data output by the first sensor (142) with data output by a second sensor to produce a second comparison result. A determination of whether the first sensor (142) has entered a fault mode is made (560) based at least in part on the first and second comparison results.