Study Of Steel Support Axial Force Based On Artificial Neural Network For Deep Excavation

With the development and utilization of underground space, a variety of deepfoundation pits are constantly emerging, and the steel support technology because ofits subject to the limit of small and easy construction has been widely used in deepfoundation design. People on the steel support system have been the use of thetraditional theory and numerical simulation technology, but these methods are thebasic parameters of the model, under normal circumstances is difficult to obtain theseparameters. The artificial neural network has a strong learning, Lenovo andanti-jamming capability in the prediction analysis has shown great advantage.Therefore, the use of artificial neural network study of steel support system isnecessary.Based on the engineering background of Qingdao subway train station of deepfoundation pit, we analyze steel support axial forces data and get the axial forcevariation regulation with working conditions. Simultaneously we establish theprediction model of steel support axial force based on artificial neural network.Conclusion:(1) Deep foundation construction monitoring program, the axial force on thesteel support on-site monitoring, the focus of the steel support of the variation of theaxial force. The results showed that: steel support after installation is complete, thesupport shaft force showing decreases and then increases until a stable law, but thevalue of the axial force of the phenomenon of repeated changes by leaps and boundsforward; throughout the construction process, four steel support maximum axial forcevalues are far less than the design value of the support system is safe, but support thedesign of the axial force is conservative, you should take full account of allinfluencing factors, to optimize the design.(2) Analysis of influencing factors, a deep foundation to support changes in theaxial force to support the axial force is mainly the result of the interaction of external natural conditions and internal engineering conditions. Various influencing factorsaccording to certain rules, broken steel strut axial forces influencing factors evaluationindex system, and as a steel brace axial force prediction model discriminationindicators. The results showed that: the evaluation on the steel support the predictionof the axial force to good effect.(3) Using artificial neural networks to study the deep foundation steel to supportthe prediction of the axial force. The results showed that: this method of the predictedand measured data agree well, you can more accurately reflect the post-steel supportaxial force trends, therefore, support the use of artificial neural network forecastingsteel axial force is an accurate and effective provide scientific guidance for the actualapplication.