LIFETIME MAXIMIZATION OF WIRELESS SENSOR NETWORKS IN A WIND TURBINE

Abstract

Wind turbines are installed in remote places and operated continuously which demands continuous condition monitoring systems for effective utilization. The condition monitoring of wind turbine is using machine learning approach is an established area. While the initial cost of the wind turbines are higher, the additional cost introduced by the condition monitoring system makes the situation worse. The fault diagnosis technique is employed using wireless sensor networks (WSN) wherein one sensor node is attached to every wind turbine in which the fault diagnosis is done. A sink node receives the signals from all the wind turbines and all the operations are carried out in the sink node. A major issue in WSN is that the battery of the sensor nodes will reduce gradually with time, as the location of the sensor nodes are very far from the sink node in many wind turbine plants. The effective operation of the WSN based fault diagnosis system is greatly dependent on number of wind turbines per base station of the wireless sensor network. In the present work, the machine learning technique is used for the extracted vibration signals from the wind turbine bearing. The wind turbine plant is equipped with the wireless sensor network and found that when the base station is in the middle of the wind turbine plant, three wind turbines has to be kept for one base station. The simulation is done using Matlab©.