Rock Bolt Finite Element Model Based On ANSYS And Intelligent Prediction Method

With the increasingly wide application of bolting technology, the problem at quality detection of anchoring system appears. But due to the highly concealed nature of its construction technology, it is difficult to the quality problems and dealing with these incidents will be more difficult. However, the quality problems will bring enormous damage to people and property. As scholars continue study on anchoring system quality testing methods, the use of non-destructive testing techniques in bolt anchoring system quality testing has become a major trend. Today as more and more widespread application of information technology, realize intelligent bolt anchoring system state prediction matters. This paper has done the following researches:(1) Elaborate the finite element analysis software ANSYS/LS-DYNA and establish three-dimensional models in different anchor states of bolt anchoring system,analyze propagation and dynamic response the stress wave in different anchor state systems and calculate the length of anchor, blot and the defect location of different bolt anchoring systems, verified the accuracy of the finite element model.(2) Present experiments and laboratory equipment for non-destructive quality testing with bolt anchoring system. In this article, experiments with anchor bolt system engineering model is done by using quality nondestructive testing instrument.Introduce the experimental principle, experimental instruments and experimental procedures in detail, and analyze the experimental results.(3) It describes the probabilistic neural network, selecting value of smoothing parameter method through experience and trial and error method, using wavelet packet decomposition and reconstruction and wavelet packet energy spectrum extracted signal characteristics as input of the probabilistic neural network. Predict the state of anchor system with analog model and experimental model using the probabilistic neural network and analysis the forecast accuracy.(4) Describes an improved probabilistic neural network, using differential evolution algorithm for optimizing smoothing parameter in order to obtain a betterprediction quickly and accurately. Predict the anchor state of experimental models and simulation models using the improved probabilistic neural network and comparing the predict results with the predict results obtained by basic probabilistic neural network, verified the better prediction results caught by improved probabilistic neural network. The results show that the prediction of the anchor states and defect location of the defected models meets the original design of anchor states and defect location in a higher degree, proved the reliability of the intelligent forecasting methods.