Numerical Simulation Of Compressed Column Of Wet Outside Coated Steel And Artificial Neural Network Prediction Of Ultimate Bearing Capacity

Comparing with newly-build construction,the repairing and strengthening of the structure have the advantage of short-term,less cost and high benefit etc. So it can be predicted that the emphases of architecture in our country will be turned to repairing and strengthening of the exited construction in future. There are many methods of repairing and strengthening of reinforced concrete (RC) structure, of which the methods of wet outside-coated steel is wide spreading used for its convenient construction,cutler workload in spot,less changes of section dimension of the member,less increment of weight,remarkable enhancement of bearing capacity,prominent improvement of rigidity and ductility of the member section. The theoretical and experimental study of wet outside-coated steel method is a scientific problem with practical use and academic value.Recently,many achievements of the study on wet outside-coated steel method have been obtained,most of them fail to consider the effect of stress history. Although the bearing capacity calculating formula of strengthening RC members deducted by some scholars have considered it,the coefficient in the formula is not easy decide,and have great impact on the result,so the formula is not widely accepted.Using non-elastic finite element method,an appropriate numerical simulating model is established. In the model,Willam-Warnke five-parameter model simulates concrete,linear strengthened model simulating the materials of bar and angel bar. The model is fully used to carry out the numerical simulation test of strengthened members,to reconstruct the bearing process and properties,cracking occurrence and propagation of cracks and the final destructive morphology in different stress history,space between slabs and diameter of bar. It is found in contrast with the testing results,that the testing results are in good agreement with those calculated by the formula. Accordingly,it is one of the effective methods to recognize the strengthening mechanism and effectiveness of strengthened members.As described above,non-elastic finite element method is one of an effective approach to ascertain the bearing capacity of strengthened members. However,its calculation is complicated and can't be easily mastered by the designer,aiming at these disadvantage,the article establish artificial neural network BP model,with the calculating results of bearing capacity as training sample and goal sample to train network,with trained network to predict the capacity of strengthened members. The results shows that the established network model is reliable,can exactly predict the bearing capacity of wet outside-coated steel column while the parameter (such as stress history,space between slabs,bar diameter,etc) changes,and will offer scientific basis and technique support for engineering design.