Application Of Artificial Intelligent In Safety Analysis Of Embankment

This paper uses artificial neural network, fuzzy neural network and genetic algorithm in artificial intelligent to analysis three problems affecting the safety of embankment, which includes prediction and judgement of piping occurring in embankment, seismic reliability of concrete dams, slope reliability evaluation of levee projects, and gets good results.The first problem analyzed is seepage piping in embankment, the artificial intelligent model of prediction and judgment for piping is proposed. A mechanics model has been presented to predict and judge the possibility of piping occurring based on the analysis of process of piping occurring and the factors affecting on the occurring of piping. A set of factors which have significant effects on the occurring of piping as well as more easily to be observed and measured have been selected as the input of artificial neural network by the developed mechanism model, and the mapping relation has established between the factors affecting on the occurring of piping and the index of occurring of piping. The developed artificial neural network of piping has been trained and predicted by the collected data and the precision of predicted results is high, which proves that artificial neural network is an effective method to predict and judge the occurring of piping. Based on the uncertainty and fuzzy characteristic of occurring of piping, fuzzy neural network of piping is proposed further. The same collected data is used to train and predict the model and gets good results, which proves fuzzy artificial network is also an effective method to predict and judge the occurring of piping.The second problem analyzed is seismic reliability of concrete dam, the artificial neural network model of seismic reliability of concrete gravity dam and concrete arch dam is proposed. In the analysis of seismic reliability of concrete dam, the most performance functions can't give a visible analytical expression and the most used method is geometric method. Because of hypothesis and approximate in the geometric method, it is not a maturity method and very complex even hard in the tackling the nonlinear problem. The artificial neural network has been used to establish the mapping relation between the seismic reliability of concrete dam and the factors affected on which. The collected data has been used to train and test the artificial neural network model. The output result is good, which proves the map ability of artificial neural network and the prediction of seismic reliability of concrete dam is feasible.The third problem analyzed is safety evaluation of embankment slope stability, the genetic algorithm is used to find the minimum safety factor coupled with the minimum reliability index. A safety reliability evaluation method for embankment slope stability under random storm wave actions and erosion-damage effects is proposed, the genetic algorithm has been used as the modern optimization technique applied to the minimization of the reliability index and the safety factor of embankment slope stability. The results obtained by the genetic algorithm are compared with that obtained by the general optimization method, and the output shows that the results from the genetic algorithm are stable and independent of the initial value, which proves the genetic algorithm method more excellent than the general optimization method.