Research On Optimization Of Shield Safety Construction Parameters Based On Intelligent Algorithm

Due to the high degree of automaticity and quick molding,The subway shield tunnel method has become an important method in subway construction.Because the subway is built in the center of the city,and the safety problems caused by the construction mainly are ground settlement,so the safety problems caused by the subway shield construction can not be ignored.But shield construction safety is far from enough.In practical engineering,it is equally important to have a faster construction speed under the premise of ensuring safety.Therefore,This paper takes No.1 Line of Urumqi Rail Transit in Xinjiang as the engineering background,and based on on-site monitoring and construction data,Taking the maximum surface subsidence value and the single ring tunneling velocity as the output target to build BP neural network,and through multi-objective optimization,Searching for decision-making parameters of shield tunneling in accordance with actual engineering requirements,the main tasks are as follows:(1)Firstly,this paper introduces the engineering situation of Xinjiang Urumqi Rail Transit Line 1,the selection of shield machine and the project monitoring scheme,and then analyzes the five subsidence stages of ground subsidence.According to this change law,the influence factors of shield construction settlement and tunneling speed are analyzed.(2)According to the sensitivity analysis,the shield construction is used as a systematic project to establish a shield construction database.The data generated are mainly divided into three categories: environmental parameters,construction process parameters and safety monitoring.The performance of dynamic data such as post-grouting and muck improvement was optimized first,and then the data generated during construction was statistically analyzed and screened.Due to the lack of sample data,we proposed three-spline interpolation of surface subsidence and formation parameter samples.The results show that interpolation results are better.After comparison,it shows that the interpolation of settlement monitoring points is reasonable.(3)Through the detailed design process of the neural network process,and based on the actual construction conditions,select sensitive factors for surface settlement and single-ring tunneling speed as input variables for decision making,and use surface settlement and singlering tunneling speed as output target variables,then BP neural network was selected as a modeling tool.(4)Because surface settlement and excavation speeds affect each other,this paper uses surface settlements within the scope of early warning.Through physical planning,settlement values are planned to fall within the settlement safety range,the target settlement fitness function is based on the safety of settlement and the faster the tunneling speed,the multiobjective genetic algorithm NSGA-II is used to optimize the multi-objective optimization,and the Pareto optimal solution set with decision-maker preference is obtained.That is,under the premise of safety,there are faster decision parameters for the shield tunneling construction.