Abstract
Research into fault diagnosis in machines with a wide range of variable loads and speeds, such as wind turbines, is of great industrial interest. Analysis of the power signals emitted by wind turbines for the diagnosis of mechanical faults in their mechanical transmission chain is insufficient. A successful diagnosis requires the inclusion of accelerometers to evaluate vibrations. This work presents a multi-sensory system for fault diagnosis in wind turbines, combined with a data-mining solution for the classification of the operational state of the turbine. The selected sensors are accelerometers, in which vibration signals are processed using angular resampling techniques and electrical, torque and speed measurements. Support vector machines (SVMs) are selected for the classification task, including two traditional and two promising new kernels. This multi-sensory system has been validated on a test-bed that simulates the real conditions of wind turbines with two fault typologies: misalignment and imbalance. Comparison of SVM performance with the results of artificial neural networks (ANNs) shows that linear kernel SVM outperforms other kernels and ANNs in terms of accuracy, training and tuning times. The suitability and superior performance of linear SVM is also experimentally analyzed, to conclude that this data acquisition technique generates linearly separable datasets.
Highlights
Over the last decade, the exponential growth of wind farms around the world has involved significant challenges, in order to improve their efficiency [1] and to reduce their operational costs [2]
A promising architecture to detect bearing failures in wind-turbine gearboxes has been presented that comprises a combination of angular resampling for vibration analysis, monitoring of the wind-turbine power output and data-mining techniques for the classification task
To overcome the lack of datasets with a wide range of conditions of loads and speeds, different experiments have been performed on a test-bed for two common failure mechanisms: misalignment and imbalance, in addition to the non-fault case
Summary
The exponential growth of wind farms around the world has involved significant challenges, in order to improve their efficiency [1] and to reduce their operational costs [2]. Once the sub-assembly is chosen, some steps should be followed: first, the most suitable sensors and their position in the sub-assembly should be chosen; second, the most suitable data-processing techniques should be found for each kind of signal collected from the sensors; third the best data-mining technique to extract as much information as possible from the processed signals should be identified. This Introduction addresses all these questions for a specific wind-turbine maintenance problem: the mechanical chain
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
