简介:鉴于目前海上风机的安装主要借助自升式风电安装平台,为保证自升式风电安装平台吊装的安全性,开展平台插桩入泥深度的计算方法研究.考虑到相邻土层的影响,提出海底多层土极限承载能力的计算方法,并将其与实际施工记录及有限元分析结果相对比,验证该方法的准确性,为海上风电装备的施工提供参考.
简介:提出了一个全新的动态公交客流预测模型,使用智能卡数据和网络爬取的常州市兴趣点(POI)数据、天气记录数据,建立一种特征融合的长短期记忆递归神经网络(Feature-fusionLSTMRNN)模型,并应用于常州市地面公交客流集散点的智慧卡(SmartCard)刷卡客流预测.将以上模型用于刷卡客流预测实验,结果表明经过良好数据训练后的特征融合的长短期记忆递归神经网络可以在刷卡客流预测中保持较高的稳定性、准确性和泛化能力.
简介:简要介绍了ESP(专门用途英语)和EGP(一般用途英语)理论的特点和两者之间的联系,并结合高职教育中英语课程教学的现状和面临的问题,探讨高职教育中EGP、ESP英语课程的融合对学生职业能力培养的重要性。
简介:Inthenormaloperationcondition,aconventionalsquare-rootcubatureKalmanfilter(SRCKF)givessufficientlygoodestimationresults.However,ifthemeasurementsarenotreliable,theSRCKFmaygiveinaccurateresultsanddivergesbytime.ThisstudyintroducesanadaptiveSRCKFalgorithmwiththefiltergaincorrectionforthecaseofmeasurementmalfunctions.Byproposingaswitchingcriterion,anoptimalfilterisselectedfromtheadaptiveandconventionalSRCKFaccordingtothemeasurementquality.Asubsystemsoftfaultdetectionalgorithmisbuiltwiththefilterresidual.Utilizingaclearsubsystemfaultcoefficient,thefaultysubsystemisisolatedasaresultofthesystemreconstruction.Inordertoimprovetheperformanceofthemulti-sensorsystem,ahybridfusionalgorithmispresentedbasedontheadaptiveSRCKF.Thestateanderrorcovariancematrixarealsopredictedbythepriorifusionestimates,andareupdatedbythepredictedandestimatedinformationofsubsystems.Theproposedalgorithmswereappliedtothevesseldynamicpositioningsystemsimulation.TheywerecomparedwithnormalSRCKFandlocalestimationweightedfusionalgorithm.ThesimulationresultsshowthatthepresentedadaptiveSRCKFimprovestherobustnessofsubsystemfiltering,andthehybridfusionalgorithmhasthebetterperformance.Thesimulationverifiestheeffectivenessoftheproposedalgorithms.