SVM And Information Fusion Applied In Deep Foundation Construction Safety Real-time Monitoring

In recent years, with the vigorous development of China’s infrastructure andconstruction in terms of expanding the scale of investment, development andutilization of underground space has attached increasingly importance, and thustechnology for the construction of deep excavation and safety management put forwardhigher requirements. As a consequence, in order to take timely and effective earlywarning management measures to ensure the smooth progress of construction,real-time monitoring of their safety in deep foundation construction process andpre-assessment in security state have become one of the key issues which need to besolved.This paper summarizes the research results of predecessors and destruction in theform of deep foundation to identify three deep foundation construction’s real-timemonitoring indicators. Besides, during the procedure of monitoring, in the sphere ofthe organizational structure, monitoring instruments and methods, monitoring steps andfrequency and monitoring data are all articulated. Because of long duration, enormousmonitoring data, various random factors may impact the accuracy of data. Therefore, inthis paper, singular value test and inspection systems are used to eliminate invalid dataerrors; and on this basis, by the least squares vector machines (LS-SVM) regressionalgorithm with timing scrolled trend monitoring data to predict the future. In this paper,each monitoring point regarding as a sensor, using DS evidence theory fusionalgorithm to predict its outcome, it then can calculate supporting structure tilted earthpressure, security status, and various levels of pore water pressure indicators. Theneach of its three indicators can be as a sensor fusion, thus supporting structure, thesecurity status of the two secondary indicators of the surroundings can be figured out. Through the various monitoring indicators multi-level fusion, the state of the globaldeep foundation construction’s safety can be ultimately assessed in this model duringthe construction process. In addition, the model also can determine the weak link in theprocess of construction. According to different security conditions, five distinguishingwarning levels develop in this essay. After a brief introduction of how warningmanagement works, then technically support in terms of the instability, bottom uplift,surface subsidence and other aspects are stated and measures are taken to effectivelytackle them, such as establishing security control points and educating workers to beaware of strengthening safety. Finally, real-time monitoring of deep foundationsystems used in construction of Xi’an International Port Area project, and the finalassessment of the security status of the project is secondary, which means safe and ingood condition. Practice shows that this system has a good safety control in real-timeand accurate performance, and it played an important role in ensuring the smoothconduct of deep foundation construction.With real-time monitoring of the deep foundation construction security, deepfoundation construction process safety state can be informed timely, and the instabilityfactors can be found, so effective measures can be taken to ensure the safety of thatdeep foundation, its surrounding buildings and structures and safety various municipalfacilities,. Also it can provide a reference for the design and construction of deepfoundation. It has a very important theoretical and practical significance.