Wind Turbine Condition Monitoring using Multi-Sensor Data system

Abstract

Wind turbines are being complex and critical systems. They encompasses interacting mechanical systems (e.g. Gear, Bearings, Blades, Brakes, etc.) that are subjected to variable aerodynamics and environmental effects as well as electrical components (i.e. electrical generator) that are subjected to variable operational conditions. Therefore, the use of single sensor to monitor and evaluate the performance and working condition of the system does not generate reliable results. To tackle this issue and enhance the reliability of such system a multi-sensor fusion system is reported and proposed in the modern condition monitoring systems. This research paper is an attempt to explore the potential of using the wind turbine online and real-time parameters acquisition to monitor the health condition of the system.  These parameters include vibrational, generated torque, and electrical signals (i.e. generated current and voltage). The data are processed in both time and the frequency domains for evaluating the health condition of the wind turbine system and diagnose the possible failure in its major components (i.e. blade, gearbox, and bearings). In addition, for profound understanding of wind turbine dynamics and structural characteristics. Five vibrational signals at different measuring locations are selected to represent the wind turbine vibrational behavior. The effects of the working condition in terms of rotational speed and pitch angle on the measured parameter have been investigated. The results show the good correlation between the wind turbine working conditions and both vibrational and electrical signals. In time domain, both wind turbine vibration and the generated current are increased by increasing the blade rotational speed. However, the overall vibration level is reduced by increasing the blade pitch angle. This could be related to the influence of increasing the air resistance, as more area of the blade will be in contact with the air. Thus, the larger area plays the role of a damper to attenuate the wind turbine vibrational magnitude. The developed air resistance reduce the rotational speed of the blade and as a consequence the generated current and voltage are dropped by increasing the blade pitch angle. In frequency domain, analyzing the peaks of FFT power spectrum for both vibrational and electrical signals (i.e. current and voltage) show a great clearly shows the influence of the wind turbine rotational speed on the resulting peaks. However, the peaks are very clear and with less noise in the electrical signal compared with the vibrational signals. This gives a motivation to examine the application of the both vibrational and electrical signals for wind turbine condition monitoring and fault diagnostics.