An Autonomous Data Acquisition and System Identification Framework for In-Service Wind Turbine Using p-LSCF Method

Abstract

Abstract Wind energy has increased its importance among renewable energy sources in recent years due to its high energy production capacity and the return of investment cost in a short time and has become a better alternative to conventional energy sources. The share of relatively large wind turbines in the renewable energy market, which can produce more energy in a short time, is increasing rapidly. Being able to accurately and consistently define the dynamic characteristics of turbines is crucial to ensure that these structures remain operational throughout their economic life and even serve more than their planned lifespan. In this study a novel autonomous data acquisition and system identification framework is developed. The subject of the work is an in-service 2.5 MW horizontal axis wind turbine with three rotor blades in Izmir/Turkey. Vibration, temperature, and relative humidity data is recorded through data acquisition system designed distributed along the turbine tower height (80 m.). Data acquisition system contain three types of sensors at different levels: One tri-axial, three uniaxial accelerometers, temperature, and relative humidity sensors at foundation level; two uniaxial accelerometers at 20-meter level; two uniaxial accelerometers and temperature sensor at 40meter level; two uniaxial accelerometers at 60-meter level; two uniaxial accelerometers, temperature, and relative humidity sensors at 80-meter level. Preprocessed vibration data is transmitted from wind farm to the university campus with operational and environmental data (e.g., wind speed, wind direction, temperature, humidity, rotor speed, nacelle direction, and pitch angle) collecting from turbine SCADA system synchronously. Modal parameters are estimated using p-LSCF method providing more clear stabilization chart than other operational modal analysis methods. A clear stabilization chart makes more easier to choose stable system poles that is significant to automate system identification. Whole data acquisition and system identification process is done without any user interference. A preliminary correlation work is done between the dynamic characteristics of the turbine and operational/environmental factors.