The wind energy industry has been growing rapidly during the past decades.Along with this growth,engineering problems have gradually emerged in the wind power industry,including those related to the structural reliabi...The wind energy industry has been growing rapidly during the past decades.Along with this growth,engineering problems have gradually emerged in the wind power industry,including those related to the structural reliability of turbine towers.This study proposes a rapid seismic analysis methodology for existing wind turbine tower structures.The method is demonstrated and validated using a case study on a 1.5 MW tubular steel wind turbine tower.Three finite element(FE)models are developed first.Field tests are conducted to obtain the turbine tower’s vibrational characteristics.The tests include(1) remotely measuring the tower vibration frequencies using a long range laser Doppler Vibrometer and(2) monitoring the tower structural vibration by mounting accelerometers along the height of the tubular structure.In-situ measurements are used to validate and update the FE models of the wind turbine tower.With the updated FE model that represents the practical structural conditions,seismic analyses are performed to study the structural failure,which is defined by the steel yielding of the tubular tower.This research is anticipated to benefit the management of the increasing number of wind energy converters by providing an understanding of the seismic assessment of existing tubular steel wind turbine towers.展开更多
针对滑坡位移具有高度非线性和复杂性,难以利用传统优化算法结合人工智能方法进行更合理、准确的预测建模的问题,本文提出一种Lévy飞行策略的混沌麻雀优化算法(CLSSA)-变分模态分解(VMD)-支持向量回归(SVR)的滑坡位移预测模型。首...针对滑坡位移具有高度非线性和复杂性,难以利用传统优化算法结合人工智能方法进行更合理、准确的预测建模的问题,本文提出一种Lévy飞行策略的混沌麻雀优化算法(CLSSA)-变分模态分解(VMD)-支持向量回归(SVR)的滑坡位移预测模型。首先利用CLSSA优化VMD分解参数对滑坡位移时间序列进行分解,其次采用CLSSA-SVR模型对VMD分解子序列进行预测,最后通过叠加子序列预测数据求出累计位移预测。以白水河滑坡为例,对该模型进行验证,实验结果表明,所提方法在最终累计位移预测结果中MAE为2.24 mm, RMSE为3.37 mm, R^(2)为0.995,相对于麻雀优化算法-变分模态分解-支持向量回归(SSA-VMD-SVR),所改进的优化算法增加了VMD的自适应能力,提高滑坡位移各分量预测效率。展开更多
基金National Natural Science Foundation of China under Grant No.51208382Shanghai Science Foundation under Grant No.12ZR1433500+4 种基金Shanghai Pujiang Scholar Program under Grant No.13PJ1407900Specialized Research Fund for the Doctoral Program of Higher Education under Grant No.20120072120001State Key Laboratory of Power Transmission Equipment&System Security and New Technology under Grant No.2007DA10512711414State Key Laboratory of Disaster Reduction in Civil Engineering under Grant No.SLDRCE14-B-02Tongji University Testing Facility Funding under Grant No.2012096
文摘The wind energy industry has been growing rapidly during the past decades.Along with this growth,engineering problems have gradually emerged in the wind power industry,including those related to the structural reliability of turbine towers.This study proposes a rapid seismic analysis methodology for existing wind turbine tower structures.The method is demonstrated and validated using a case study on a 1.5 MW tubular steel wind turbine tower.Three finite element(FE)models are developed first.Field tests are conducted to obtain the turbine tower’s vibrational characteristics.The tests include(1) remotely measuring the tower vibration frequencies using a long range laser Doppler Vibrometer and(2) monitoring the tower structural vibration by mounting accelerometers along the height of the tubular structure.In-situ measurements are used to validate and update the FE models of the wind turbine tower.With the updated FE model that represents the practical structural conditions,seismic analyses are performed to study the structural failure,which is defined by the steel yielding of the tubular tower.This research is anticipated to benefit the management of the increasing number of wind energy converters by providing an understanding of the seismic assessment of existing tubular steel wind turbine towers.
文摘针对滑坡位移具有高度非线性和复杂性,难以利用传统优化算法结合人工智能方法进行更合理、准确的预测建模的问题,本文提出一种Lévy飞行策略的混沌麻雀优化算法(CLSSA)-变分模态分解(VMD)-支持向量回归(SVR)的滑坡位移预测模型。首先利用CLSSA优化VMD分解参数对滑坡位移时间序列进行分解,其次采用CLSSA-SVR模型对VMD分解子序列进行预测,最后通过叠加子序列预测数据求出累计位移预测。以白水河滑坡为例,对该模型进行验证,实验结果表明,所提方法在最终累计位移预测结果中MAE为2.24 mm, RMSE为3.37 mm, R^(2)为0.995,相对于麻雀优化算法-变分模态分解-支持向量回归(SSA-VMD-SVR),所改进的优化算法增加了VMD的自适应能力,提高滑坡位移各分量预测效率。
