Research On Parallel Detection Method Of Steel Bar Diameter And Position Based On BP Neural Network

The diameter and location of reinforcement in reinforced concrete（thickness of protective layer and horizontal distance）play a very important role in the safety of built buildings.For the detection of steel bars in reinforced concrete structures,the existing detection methods will have large errors and complex operations.In order to further improve the convenience and accuracy of steel bar diameter and protective layer thickness detection.In this paper,the commonly used electromagnetic induction steel scanner is used to detect the waveform of multiple groups of single and multiple reinforcements,and the BP neural network model is established by analyzing the obtained characteristic parameters.It can be convenient and accurate to obtain the diameter and position of reinforcement in reinforced concrete.The main research contents and conclusions are as follows:1.According to the detection waveform of single steel bar,the influence of steel bar diameter and protective layer thickness on wave peak and waveform width is analyzed.The results show that the wave peak and wave width increase with the increase of steel bar diameter.But with the increase of the thickness of the protective layer,the wave peak and wave width decrease.2.The function relationship between the wave peak and wave width of multiple groups of single reinforcement and the diameter of reinforcement and the thickness of protective layer is established,and the accuracy of this function is verified.The verification results show that the error between the diameter of steel bar and the thickness of protective layer calculated by this function and the actual value is relatively small.3.According to the actual situation of the project,the diameter of the steel bar and the thickness of the protective layer under various working conditions are simulated.Through the analysis of the waveform data,it is found that each steel bar has its own waveform curve when the single row of multiple adjacent steel bars in a certain range.When the diameter of one steel bar and the thickness of the protective layer are fixed,the change of the adjacent steel bar diameter,spacing and the thickness of the protective layer will not affect the wave peak of the steel bar and the curve change trend on both sides of the peak.4.In order to more conveniently predict the diameter of multiple reinforcements and the thickness of protective layer.The four optimal characteristic parameters of BP neural network model are obtained by analyzing the waveform data of single steel bar : wave peak;Wave width corresponding to 10 % decrease of wave peak;The data in the range of 80 mm width curves on both sides of the wave peak are fitted to obtain the fourth-order coefficient A and the second-order coefficient B.The BP neural network model is used to detect the diameter of single and multiple reinforcements and the thickness of protective layer in parallel.When the reinforcement spacing is greater than 80 mm,the error between the test results and the actual value is basically within the range of ±2 mm.