Research On Construction Simulation And Optimization Of Complex Long Distance Diversion Tunnel Considering The Impacts Of Risks

Complex long-distance diversion tunnel has the characteristics of hidden construction environment,complex construction conditions and technology,complex inter-process interference and long construction period,which bring great challenges to the construction simulation and optimization of complex long-distance water diversion tunnel.The simulation technology provides an effective way for the analysis of construction progress and the optimization of construction scheme.In the process of tunnel construction,the construction parameters show the characteristics of random variation,while the construction process is faced with many potential risk factors such as construction machinery failure,collapse,water gushing,rock burst and so on,and the construction parameters have the characteristics of dynamic change.However,the comprehensive impact of the above risks on the construction process has not been fully considered in the current tunnel construction simulation studies.Meanwhile,the existing Bayesian updating methods and time series forecasting methods for updating simulation parameters,or assuming the distribution type in advance,and the distribution type remains unchanged in the updating process,or the predicted results of the parameters are fixed values,without considering the randomness of the construction parameters under the influence of risk factors,which lead the results can not fully reflect the dynamic change law of the construction parameters.In addition,the existing research on optimization of construction schemes lacks comprehensive consideration of the impact of risk and dynamics of construction system,and the effectiveness of the optimization method needs to be further improved.Therefore,how to consider the impact of risk factors on the construction simulation and optimization of complex longdistance diversion tunnel comprehensively and scientifically,explore the dynamic variation law of construction parameters,and put forward a scientific and effective optimization method of simulation schemes,are the key scientific problems to be solved in the simulation and optimization of construction progress of complex long-distance diversion tunnel.In-depth scientific research has been carried out around the above issues,and the following research results have been achieved:1.Existing simulation studies on the construction of complex long-distance diversion tunnel fail to consider the impact of risk factors comprehensively and scientifically,and simulation parameters lack of dynamic updating.Therefore,a mathematical model of complex long distance diversion tunnel construction simulation considering the impact of risk factors is proposed.In the construction of complex long-distance water diversion tunnel,there are random variations in activity time,construction machinery failure,and risk factors such as collapse,water inrush and rock burst,which will lead to changes in activity time parameters,construction machinery parameters and construction logic.Current tunnel construction simulation researches focuses on the analysis of the impact of single risk on the simulation process,and most studies only consider the impact of risk factors on simulation parameters,which lack comprehensive and scientific consideration of the comprehensive impact of various risks on simulation parameters and simulation logic.In addition,the existing researches on dynamic updating of simulation parameters,whether using Bayesian updating method or time series prediction method,both lack effective consideration of the impact of construction risk factors.In view of the above problems,a mathematical model of complex long-distance diversion tunnel construction simulation considering the comprehensive influence of risk factors is proposed.Based on the layered simulation model of tunnel construction coupled with CPM(Critical Path Method)and CYCLONE(Cycle Operation Network),the effects of various risk factors on simulation parameters such as activity time parameters,construction machinery parameters and simulation logic are comprehensively analyzed.At the same time,the dynamic updating of simulation parameters under the influence of construction risk factors is explored.2.Current simulation modeling of complex long distance diversion tunnel construction lack of comprehensive consideration of the influence of risk factors on activity time parameters,construction machinery parameters and construction logic.In light of the problem,a simulation modeling method of complex long distance diversion tunnel construction considering the comprehensive influence of risk factors is proposed.Firstly,the existing simulation modeling methods of complex long distance diversion tunnel construction mostly focus on the impact of single risk,lack of comprehensive consideration of the impact of risk factors on activity time parameters,construction machinery parameters and construction logic,and the current simulation modeling methods can not fully reflect the risk characteristics.Therefore,a simulation modeling method for complex long distance diversion tunnel construction considering the comprehensive influence of risk factors is proposed.Secondly,in view of the random characteristics of activity time parameters under the influence of risk factors are reflected by single distribution such as normal distribution and beta distribution,which fails to fully reflect the random variation of time parameters,a method for determining the probability distribution of activity time parameters based on Gauss mixture model is proposed.Without assuming the distribution type in advance,the method of K-means-EM-BIC coupling is used to solve the Gauss mixture model.The advantage of the solution of the Gauss mixture distribution function is also reflected in the ability of incremental learning.When new data is added,only new knowledge is needed to modify the original distribution,and no need to use all data to re-fit the distribution.Thirdly,in view of the fact that the simulation modeling considering the faults of construction machinery mostly assumes that the whole faulty activity will affect the construction progress,ignoring the problem that the faulty machinery can not be repaired before the end of the original unit activity,and that the residual fault treatment activity has no effect on the construction progress,a fine analysis method for the faults of construction machinery of complex long distance diversion tunnel is proposed.Under the condition of tracing the operation trajectory of construction machinery,the time and type of mechanical failure are determined,and the calculation formula of unit activity delay time under different types of mechanical failure is deduced to analyze the influence of construction machinery failure.Lastly,in view of that the impact of risk events such as landslide,water gushing and rock burst on simulation logic is neglected in current studies,and a few studies considering the impact of risk events use expert experience or statistical analysis to obtain the probability of occurrence of risk events,failing to quantitatively calculate risk on the basis of comprehensive consideration of geological conditions,design level,construction level,management level and their interaction,and considering the advantages of Bayesian network in calculating risk probability quantitatively under the condition of considering the dependence among variables,a risk event analysis method for tunnel construction based on Bayesian network is proposed.Considering the influence of geological conditions,design level,construction level,management level and correlation on risk events,the method establishes Bayesian networks for analysis of landslides,water gushing events and rock bursts,respectively,in order to quantitatively calculate the probability of risk events in each excavation cycle,aiming at the characteristics of landslides,water gushing events and rock bursts.The Bayesian network is embedded in the CYCLONE simulation model,and the influence of potential risk events on the logic of tunnel construction is considered in each excavation cycle.3.Current research lack of considering the random variation of parameters under the influence of construction risk factors of simulation parameters updating.In light of this,a dynamic updating method of simulation parameters for complex diversion tunnel construction considering the influence of risk factors is proposed.At present,the commonly used Bayesian updating method for simulation parameter updating needs to assume the distribution type of parameters in advance,and the distribution type remains unchanged during the updating process;the prediction results of rare time series forecasting methods are deterministic values,which can not reflect the random fluctuation of construction parameters under the influence of risk factors.To solve the above problems,a dynamic updating method of simulation parameters considering the influence of risk factors is proposed.Firstly,the original Gauss mixture distribution model of simulation parameters is obtained based on historical data fitting to fully reflect the random characteristics of simulation parameters under the influence of risk factors.Secondly,the incremental learning ability of the Gauss mixture distribution model is given,and the updated formulas of the mixture weight,mean and covariance of each Gauss component are deduced,so that when adding new samples,the original model is adjusted directly by using the knowledge obtained from the new samples,so that the dynamic updating of simulation parameters can be realized.This method does not need to assume the distribution type in advance,and allows the distribution type to change during the updating process,and each updating result is a new Gauss mixture distribution,so it can reflect the stochastic characteristics of simulation parameters under the influence of risk factors.4.Current research on construction scheme optimization lack of considering the comprehensive influence of risk factors and face the problem that the model solving method is easy to fall into local optimum and the decision-making process can not reflect the hesitation of experts.In light of this,a dynamic optimization method for complex long distance diversion tunnel construction scheme considering the influence of risk factors is proposed.The existing optimization research of construction schemes lacks the comprehensive influence of risk factors such as random activity time parameters,construction machinery failure,collapse,water gushing,rock burst and the dynamic characteristics of construction system.Meanwhile,in the process of solving the optimization model,the commonly used particle swarm optimization algorithm still needs to improve the search accuracy and the ability to jump out of the local optimum.In addition,the existing research fails to reflect the hesitation of experts in the process of program evaluation,and lacks in-depth research on different expert opinions fusion methods.In view of the above problems,a method for optimizing construction schemes of complex long distance diversion tunnel considering the influence of risk factors is proposed.Firstly,an optimization model of construction scheme is established,which comprehensively considers the random activity time parameters,construction machinery failure,collapse,water gushing,rock burst and other risk factors,so as to realize the optimization of construction scheme considering the dynamic changes of construction parameters.Secondly,a Levy Flight Adaptive Chaotic Particle Swarm Optimization(LFACPSO)algorithm is proposed for multi-objective optimization of construction scheme,and then the Pareto scheme set is obtained.This algorithm imports Levy flight to improve the particle updating method in particle swarm optimization,so as to overcome the shortcomings of particle easily falling into local optimum.Meanwhile,chaotic algorithm is used to initialize particles,and the inertia weight coefficient of particles is adjusted adaptively to improve the global and local search ability of particles.Finally,an Intuitionistic Fuzzy Entropy Weighted Power Average(IFEWPA)method is used to select the final scheme from the scheme set.This method not only reflects the hesitation of experts in decision-making process through the intuitionistic fuzzy set,but also considers the mutual support of data in the process of expert opinion fusion,which makes the fusion process more objectivity.5.The feasibility of proposed theory and method are verified by conducting construction simulation and optimization of a real complex long distance diversion tunnel project considering the comprehensive influence of risk factors.Taking a complex long distance diversion tunnel as example,the proposed theory and method are applied to real project:Firstly,considering the random variation of activity time,construction machinery failure,as well as the comprehensive influence of risk factors such as collapse,water gushing and rock burst,a layered simulation model of tunnel construction coupled with CPM and CYCLONE is established,and simulation parameters such as Gaussian mixture distribution of activity time and construction machinery failure parameters are determined.At the same time,Bayesian network is used to calculate the probability of risk events such as collapse,water inflow and rock burst to reflect the impact of risk factors on the logic of tunnel construction.Secondly,considering the influence of risk factors on simulation parameters and simulation logic,the optimization of tunnel construction simulation scheme is carried out.First,the Pareto scheme set is obtained by multi-objective optimization method of simulation scheme based on LFACPSO algorithm;then,the multi-attribute group decision-making method of simulation scheme based on IFEWPA is used to select the final construction scheme from the selected scheme set,considering the hesitation of expert evaluation and the objective and effective integration of expert opinions.Furthermore,based on the optimized construction simulation scheme,the tunnel construction process simulation considering the comprehensive influence of risk factors is carried out.Several simulation calculations were carried out by means of parallel computing,and key routes,construction period,construction intensity and completion probability were obtained.Finally,the dynamic updating method of simulation parameters based on the incremental Gauss mixture model is used to update the time distribution of activities such as survey setting-out,charge blasting,ventilation and smoke,safety inspection,shotcrete and anchor support,and the dynamic optimization of tunnel construction simulation scheme considering the influence of risk factors is carried out based on the updating results and current construction features.