结构工程师
結構工程師
결구공정사
STRUCTURAL ENGINEERS
2014年
4期
105-111
,共7页
高速公路%路面不平度%识别%RBF 神经网络
高速公路%路麵不平度%識彆%RBF 神經網絡
고속공로%로면불평도%식별%RBF 신경망락
freeway%road surface roughness%identification%RBF neural network
为了发现高速公路上路面不平度产生的汽车动载对路面破坏影响,建立了7个自由度汽车振动模型,以得到车身质心垂直加速度和俯仰角加速度作为神经网络理想输入样本,路面不平度作为网络理想输出样本,识别了 B 级和 C 级路面不平度。识别的 B 级路面和 C 级路面相对误差的最大值分别是0.24%和0.42%。结果表明,该方法具有较理想的识别精度和较强的抗噪声能力,C 级路面不平度的相对误差比 B 级路面不平度的相对误差大。识别出来的路面不平度可为分析高速公路路面动力响应研究提供理论基础。
為瞭髮現高速公路上路麵不平度產生的汽車動載對路麵破壞影響,建立瞭7箇自由度汽車振動模型,以得到車身質心垂直加速度和俯仰角加速度作為神經網絡理想輸入樣本,路麵不平度作為網絡理想輸齣樣本,識彆瞭 B 級和 C 級路麵不平度。識彆的 B 級路麵和 C 級路麵相對誤差的最大值分彆是0.24%和0.42%。結果錶明,該方法具有較理想的識彆精度和較彊的抗譟聲能力,C 級路麵不平度的相對誤差比 B 級路麵不平度的相對誤差大。識彆齣來的路麵不平度可為分析高速公路路麵動力響應研究提供理論基礎。
위료발현고속공로상로면불평도산생적기차동재대로면파배영향,건립료7개자유도기차진동모형,이득도차신질심수직가속도화부앙각가속도작위신경망락이상수입양본,로면불평도작위망락이상수출양본,식별료 B 급화 C 급로면불평도。식별적 B 급로면화 C 급로면상대오차적최대치분별시0.24%화0.42%。결과표명,해방법구유교이상적식별정도화교강적항조성능력,C 급로면불평도적상대오차비 B 급로면불평도적상대오차대。식별출래적로면불평도가위분석고속공로로면동력향응연구제공이론기출。
In order to find the effect of vehicle dynamic load caused by road surface roughness on the pave-ment destruction on the highway,a 7 DOF vehicle vibration model was established,the vertical acceleration and pitching angular acceleration of vehicle body centroid were got,which were regarded as neural networks i-deal input sample,the corresponding road surface roughness was regarded as neural networks ideal output sam-ple.The level B and level C road surface roughness were identified.The maximum of relative error of level B and level C road surface was respectively 0.24% and 0.42%.The results show that the method has ideal iden-tification accuracy and better ability of anti-noise,the relative error on the level C road surface was greater than the level B road surface.The road surface roughness of the identification can provide a theoretical basis for analyzing the dynamic response of the freeway road surface.