Research On Safety State Evaluation System For Deep Foundation Pit Construction Of Metro Station

With the rapid development of urban rail transit,the scale of metro station deep foundation pit is getting to enlarge,which is placing growing risks.During the construction process,monitored data can directly reflect the situation of engineering process.By analyzing these data,it is possible to timely discover risks and develop corresponding countermeasures to ensure the safety of foundation pit construction.Due to massive uncertain factors involved in construction process,these risks are correlated,ambiguous and dynamic.When evaluating the safety state of foundation pit construction,multiple monitoring indicators must be considered comprehensively.In this paper,multiple deep foundation pits in compound soil area of a coastal city are taken as the background material,safety state comprehensive evaluation of foundation pit construction is taken as the research objective.The safety state evaluation system for foundation pit construction based on entropy fuzzy comprehensive evaluation theory is constructed.The main research contents are as follows:(1)According to monitored data of deep foundation pit,the internal law and sensitivity of monitoring indicators during the construction process are analyzed.The monitored data of four metro station foundation pits in compound soil area are summarized and analyzed.Distribution of the ratio of deep horizontal displacement of wall and maximum settlement of surrounding surface,settlement of surrounding buildings is calculated.The correlation between deep horizontal displacement of wall and settlement and horizontal displacement of the top of wall is analyzed,and the relationship is fitted.Statistical analysis is carried out on the actual application situation of monitoring indicators and the usable rate of monitoring points in engineering.The main evaluating indicators were finally selected.(2)The entropy method is adapted for weight analysis of eight indicators and it is concluded that the weights of six main evaluating indicators in foundation pit safety evaluation are significantly greater than the other two indicators.We briefly describe the basic process of using entropy method to determine the weights of main evaluating indicators,determining the hierarchical control value of indicators and rate of change,and applying fuzzy comprehensive evaluation method to quantitatively evaluate the safety state of foundation pit construction by scoring.As a result,the safety level of foundation pit is classified into four levels:"safety",“attention","warning" and"danger",which makes the evaluation result more practical.(3)The MATLAB based Neural Network prediction model is constructed and its training parameters are set.Taking monitored data of a deep foundation pit as training samples,the main evaluating indicators are dynamically predicted.The comparison shows that the predicted value has good agreement with measured value,which fulfills the actual engineering requirement.(4)The evaluation system is applied to actual engineering practice to quantitatively evaluate the safety state of foundation pit and judge the safety level of foundation pit.The evaluation result has successfully reflected the actual situation of the construction process.The safety state of foundation pit construction in next stage is pre-judged.According to the membership of each indicator under the "danger" level,combined with the internal law and sensitivity analysis of monitoring indicators.The safety risks existing in engineering are pre-judged,helping to control risks during the next stage of construction,improve construction safety and reduce unnecessary waste of resources.