Sustainable Energy Solutions Task 2.0: Wind Turbine Reliability and Maintainability Enhancement through System-wide Structure Health Monitoring and Modifications to Rotating Components

Janet M Twomey, PhD

关键词: ACCURACY;    ACOUSTICS;    BEARINGS;    COMPOSITE MATERIALS;    DISTRIBUTION FUNCTIONS;    EVALUATION;    FIBERS;    MODIFICATIONS;    MONITORING;    NEURAL NETWORKS;    PROBABILITY;    RELIABILITY;    ROLLER BEARINGS;    SERVICE LIFE;    SIMULATION;    TESTING;   

DOI  :  10.2172/991640
RP-ID  :  DOEGO88149 3
PID  :  OSTI ID: 991640
Others  :  TRN: US201110%%442

美国|英语

来源: SciTech Connect

【 摘 要 】

EXECUTIVE SUMARRY An evaluation of nondestructive structural health monitoring methods was completed with over 132 documents, 37 specifically about wind turbines, summarized into a technology matrix. This matrix lists the technology, what can be monitored with this technology, and gives a short summary of the key aspects of the technology and its application. Passive and active acoustic emission equipment from Physical Acoustics Corp. and Acellent Technologies have been evaluated and selected for use in experimental state loading and fatigue tests of composite wind turbine blade materials. Acoustic Emission (AE) and Active Ultrasonic Testing (AUT), were applied to composite coupons with both simulated and actual damage. The results found that, while composites are more complicated in nature, compared to metallic structures, an artificial neural network analysis could still be used to determine damage. For the AE system, the failure mode could be determined (i.e. fiber breakage, delamination, etc.). The Acellent system has been evaluated to work well with composite materials. A test-rig for reliability testing of the rotating components was constructed. The research on the types of bearings used in the wind turbines indicated that in most of the designs, the main bearings utilized to support the shaft are cylindrical roller bearings. The accelerated degradation testing of a population of bearings was performed. Vibration and acoustic emission data was collected and analyzed in order to identify a representative degradation signal for each bearing to identify the initiation of the degradation process in the bearings. Afterwards, the RMS of the vibration signal from degradation initiation up to the end of the useful life of the bearing was selected to predict the remaining useful life of the bearing. This step included fitting Autoregressive Moving Average (ARMA) models to the degradation signals and approximating the probability distribution function (PDF) of remaining useful life based on the results of Monte-Carlo simulation of the ARMA models. This step was performed for different percentages of the degradation signal of each bearing. The accuracy of the proposed approach then was assessed by comparing the actual life of the bearing and the estimated life of the bearing from the developed models. The results were impressive and indicated that the accuracy of the models improved as more data was utilized in developing the ARMA models (we get closer to the end of the life of the bearing).