CAJ | 학술논문

基桩缺陷辨识是一个高度的非线性问题，现有分析方法人为干预比较多，难以得到准确的桩身完整性结果。利用MATLAB工具编制了基于BP神经网络模型的基桩缺陷识别程序，该方法具有较好的判别精度。应用神经网络模型结合实测资料对南京地区的一些人工挖孔桩进行完整性分析，取得了较好的应用效果。
기장결함변식시일개고도적비선성문제，현유분석방법인위간예비교다，난이득도준학적장신완정성결과。이용MATLAB공구편제료기우BP신경망락모형적기장결함식별정서，해방법구유교호적판별정도。응용신경망락모형결합실측자료대남경지구적일사인공알공장진행완정성분석，취득료교호적응용효과。
Idendificating the defect of foundation piles is a complicated non -linear problem, there are a lot of artificial interference in the methods that we use , so we are difficult to get the accurate result about integrity of foundation piles. The thesis discusses the BP neural network model used in idendificating the integrity of foundation piles. Compiling programme to idendificate defect of foundation piles base on the model, the result of training and examination show that this model is exact on idendification of integrity. Applying the neural network model in dozens of concrete piles in Nanjing region, the method gets a good application effect.