六米複合材料風力葉片之可靠度評估

Abstract

目前大部分研究著重於風機整體的完整性分析，考慮材料變異對於葉片可靠度影響的探討較為稀少。本研究目的為同時考慮材料常數與風速為隨機變數下的可靠度，針對型號NACA4418的六米風力葉片進行評估。 以伯努力定律的簡化公式將風速轉換為等效風壓，平均分佈於葉片蒙皮表面，以有限元素法計算葉片蒙皮應力。分析得知隨著材料常數的值增加，葉片產生挫屈的臨界風速會跟著提升。可靠度分析中，將上蒙皮恰發生挫屈視為極限狀態。考慮材料常數與風速為隨機變數，找出上蒙皮在不同材料常數下發生挫屈時的風速(挫屈風速)，每個材料常數值都對應一個挫屈風速值，將所有點線性連接，以此建立極限狀態方程式。為得到材料常數之機率分佈，將製作葉片的方法同理應用於製作試片，進行標準拉伸試驗得到材料常數資料；風速的機率分佈則由中央氣象局取得蘭嶼地區近二十年的風速資料。兩種資料統計整理後，得到平均值與標準差，轉換為韋氏分佈中的純量與形狀參數，以此繪製機率分佈曲線，將曲線與統計的結果比較，兩者結果接近。根據不同的極限狀態結合韋氏機率密度函數，經直接積分法分段計算每個材料常數區間的可靠度值，將所有結果相加後得到葉片可靠度値。 最後將葉片改良，考慮蒙皮層數與腹板數量兩設計參數，建立四種模型並計算其重量。對四種不同的葉片進行可靠度分析，經由考慮葉片重量與可靠度的關係，比較不同設計參數對於可靠度提升的程度。經過比較可發現增加蒙皮確實可以增加葉片可靠度，而雙腹板葉片相較於單腹板葉片對於可靠度增量有更好的效果。

Recently many research works have focused on the reliability analysis of entire wind turbine. The research about the effects of material property variation on wind blade reliability is relatively rare. The purpose of this thesis is to treat both material property and wind speed as random variables and study how these random variables affect the reliability of a 6-M wind blade with NACA4418 airfoil cross section. The Bernoulli’s principle is used to determine the uniformly distributed wind load on the blade surface. The nonlinear finite element method is used to predict the mechanical behavior of the blade. The buckling failure mode is treated as the limit state of the blade. In the reliability analysis, both material property (Young’s modulus) and wind speed are treat as random variables. In attaining the probability distribution of the Young’s modulus, the curing parameters used in blade manufacturing are also adopted to make the specimens for material testing. The test results are used to find the probability distribution of Young’s modulus. The maximum instantaneous wind speeds in the past 20 years provided by the Central Weather Bureau are used to determine the probability distribution of wind speed. The results of a series of buckling analyses together with a piecewise-linear approximation technique are used to build the limit state curve of the blade. A numerical integration technique is then adopted to calculate the blade reliability. The improvement of the blade is studied by means of analyzing four distinctive blade designs. The effects of blade design parameters such as skin thickness and number of spar web on the blade reliability are investigated and discussed. The comparison results have shown that the increase in skin thickness can increase blade reliability and two-spar blade is more reliable than one-spar blade.