Study On The Prediction Of Monitoring Information For Deep Foundation Pit Based On Intelligent Algorithm

Deep foundation pit engineering has the characteristics of high cost, difficult construction and instability, so monitoring system is very important for its stability. But, deep foundation pit is the complex space system of multimedia combined with bracing system and adjacent soil and its deformation and security are affected by geological conditions, characters of rock mass and soil body, climate and underground water dynamic factors. Regular statistic method can not make prediction by modeling with monitoring information but due to good non-linear fitting capability, the BP neural networks has been widely used in the analysis and prediction of complex non-linear system. However, BP networks also have disadvantages. Genetic algorithm, simulated annealing algorithm and mutative scale chaos optimization algorithm are used to optimize BP networks, which forms a series of intelligent algorithm. On the basis of monitoring information, nonlinear implicit model should be set up to predict stability development trend of foundation pit. The case is the south anchor deep foundation pit engineering of Jiangsu Runyang Yangtze River Bridge which adopted frozen ground and pile-support technology.To introduce the basic principles of intelligent algorithms such as BP neural networks, genetic algorithm, simulated annealing algorithm and mutative scale chaos optimization algorithm;to design related programs with MATLAB7.0.To research the capability factors of BP networks and "automatic search algorithm" is built against the problem how to decide the number of hidden nodes of BP networks and also design related programs with MATLAB7.0.To research the possible combination between neural networks and genetic algorithm, simulated annealing algorithm and chaos optimization algorithm and forms a series of intelligent algorithm also design programs with MATLAB7.0. This intelligent algorithm prediction model can be used in not only stability prediction of deep foundation pit engineering which is based on the monitoring data but also other prediction.The freezing and row-pole methods adoption in south anchor deep foundation pit engineering is the first example in same engineering type at home so the difficulty can not be imagined. It is necessary for stability to keep the real-time monitoring information feedback. Intelligent prediction is made to south anchor deep foundation pit monitoring information by above mentioned algorithms. The fact shows that this intelligent algorithm has high accuracy to the prediction of foundation pit monitoring information, so it will be a good guidance to the next construction.