Research On The Methods For Predicting Foundation Settlement In The Base Pit Vicinity

With the development of the policy of opening up and reform, the economic development of coastal regions becomes very critical. Infrastructure has become an important index for measuring a region's economic development level. Unfortunately, the soil of coastal regions is generally muddy soft clay, which is the main factor hindering infrastructure. Under the condition that there is no serious enough problem leading to buildings'settlement, how to design a reliable foundation treatment program and have a reasonable construction schedule in defined period becomes a topic of great importance which is in urgent need of being solved in the field of soft ground engineering. Carrying out foundation settlement prediction is an inevitable approach to work out this topic.In recent decades of years, scholars from various countries have conducted many studies and already made a great achievement. Some studies have been applied to real projects. However, there are still many studies which failed to be applied in reality, because of the complexity of soft soil project itself and the influence of upper load and upper structure.Available foundation settlement prediction method is based on some certain hypothetic condition which is inconsistent with reality to a great extent, which often leads to a big difference between the obtained settlement prediction result and actually measured settlement value. Therefore, studies on foundation settlement prediction method are needed to be further developed. Based on a review of studies on the theory of foundation settlement prediction at home and abroad, this thesis provides a detailed theoretical analysis on grey prediction model, grey Verhulst prediction model and poisson convolution model life in terms of respective characteristics, application mechanism and modeling methods and steps. In order to correctly evaluate these three prediction methods, this thesis especially quotes some settlement observation data from a typical real project. Through selecting data, we got 15 groups of typical observation data. Three methods are employed to do modeling calculation of this project. Through comparison between obtained calculation value and actually measured value, we could get prediction effects of each prediction model which will be reflected by fitting curve, variance ratio and small error probability, so each method would be offered a justice evaluation including their own merits and demerits and their own application prospects and corresponding prediction and indication could be supplied to the next step of research.