Rotating and non-rotating sensors for improving condition monitoring of wind turbines

Abstract

Wind turbine blades are susceptible to damage from lightning strikes as well as from the repetitive cyclic fatigue loads and the continuously changing temperature and humidity environmental conditions which further increase the failure probability. This paper presents examples of novel vibration measurements on wind turbines that can be utilised to improve condition monitoring methods, further leading to improvements in system reliability. A test rig has been developed to monitor blade behaviour under different transient loads, including the measurement of the main shaft centre line motion in addition to the tower vibration. The novel data acquisition measurement methods include utilizing a slip ring assembly on the turbine hub providing efficient low noise signals for the measurement of the rotating components. Strain gauges have been attached with rotating components and instrumented in the slip ring. Accelerometers and piezoelectric sensors have accompanied the measurement setup in the rotating and non-rotating wind turbine components to assist in giving an improved understanding of the system dynamic behaviour. A servo-motor provides the input transient torque of the main shaft to get the desired velocity as well as velocity control through an encoder coupled to the motor that provides correct phase reference data of the shaft during the rotation. In addition to accelerating the turbine rotation, it has the ability to provide impulsive torsional loading during the operation further exciting the system resonances. Laser displacement sensors have also been instrumented to measure fluctuations of the drive shaft transverse vibration through shaft orbit plots. This paper presents details of the developed test rig components and the novel sensors illustrating the improvements in knowledge that can be gained through comprehensive vibration condition monitoring. Case studies of implanted faults are also provided, clarifying the potential that can be gained for improved fault detection.