基于BP神经网络的场道脱空检测方法及实验

发布时间：2019-02-08 17:08

【摘要】：为研究场道脱空检测方法,进行室内模型试验,获得冲击荷载作用下道面板加速度响应时程曲线,利用Matlab小波变换工具箱提取加速度曲线特征值,分析脱空对振动信号的影响规律.通过优化荷载级数、筛选输入向量,建立了场道脱空的BP(back propagation)神经网络预测方法.为检验理论研究结果的正确性,利用重锤式弯沉仪在机场进行跑道脱空测试,通过场道取芯脱空观察评价BP神经网络预测结论的可靠性.结果表明,荷载级数、输入向量、训练次数、训练强度和算法对BP神经网络预测准确性影响较大;脱空影响下场道加速度信号可作为BP神经网络脱空预测的输入向量,取芯后场道脱空状况同BP神经网络预测结果一致.
[Abstract]:In order to study the detection method of field channel void, the time-history curve of acceleration response of track panel under impact load was obtained by indoor model test. The eigenvalue of acceleration curve was extracted by Matlab wavelet transform toolbox. The influence of void on vibration signal is analyzed. By optimizing load series and selecting input vectors, a prediction method of field channel void based on BP (back propagation) neural network is established. In order to verify the correctness of the theoretical research results, the reliability of the prediction results of BP neural network was evaluated by using the heavy-weight deflectometer to test the runway clearance at the airport. The results show that load series, input vector, training times, training intensity and algorithm have great influence on the prediction accuracy of BP neural network. The acceleration signal of field track can be used as input vector of BP neural network to predict void under the influence of void. The condition of field track void in core is consistent with the prediction result of BP neural network.
【作者单位】：冻土工程国家重点实验室;中国民航大学机场学院;
【基金】：国家自然基金资助项目(51178456) 冻土工程国家重点实验室开放基金资助项目(SKLFSE201409) 中央高校基本业务费资助项目(3122016D019)~~
【分类号】：V351;TP183

【参考文献】