Under The Combined Action Of Injury Inducements Inversion Analysis Of Overhead Vertical Wharf

Overhead vertical high-pile pier has the advantages of adapting to large water level amplitude change,good berthing conditions and strong durability,and is widely used in large-scale terminal structures in Chongqing.During the operation of the terminal,it is vulnerable to multiple adverse factors such as instability of the bank slope,uneven settlement of the foundation,out-of-control ship collision,and non-standard mooring,resulting in complex damage mechanisms and difficult identification of the terminal structure.The existing research is limited to the research of wharf structures under a single damage inducement,which is not sufficient to meet the requirements for tracing and early warning of damage inducements of wharves under complex stress conditions in practical engineering.Therefore,this thesis conducts inverse research on the combined effects of uncontrolled ship collision and bank slope instability on wharf structures,making up the gap in inverse analysis of overhead vertical wharves under the combined effects of damage inducements.With the support of the Chongqing major engineering research project "Research and Development and Application of Key Technologies for Long-term Intelligent Monitoring of Structural Safety of Inland River Frame Vertical Wharf",the following research work has been carried out for the inversion of damage inducements of overhead vertical high-piled wharfs under combined actions:(1)According to characteristics of overhead vertical wharves,the indicators such as the action position and action strength of the combination of damage inducing factors are studied.Using the structural finite element numerical simulation method,a mathematical model with the same scale as the physical model of the wharf has been established.The pile foundation of the wharf has been uniformly numbered,the strain data collection area of the wharf structure has been determined,and the strain analysis of the overhead vertical wharf under the combined action of damage incentives has been carried out.The axial stress characteristics of the wharf pile foundation under the combined action of two damage incentives,namely,runaway ship collision and bank slope instability,have been revealed.The sensitivity analysis method has been adopted,Selecting a reasonable pile foundation as the axial strain data collection pile foundation for physical model testing provides strong support for the construction of the inverse model training set.(2)In the face of the need to build a damage inducement inversion test set for the overhead vertical wharf in the reservoir area,the basic idea of inversion was clarified,a physical model test system was established,appropriate physical model geometric parameters and strain data collection system were determined,accurate and effective load loading plans were developed,and appropriate strain sensing methods were selected.By analyzing and comparing the bank slope test and the combined test,it is found that the combined effect of damage inducements is more obvious for the bank slope effect on the rear row pile foundation than for the ship impact effect;The accuracy verification of the physical model and the mathematical model data shows that the overall average relative error is only 15.8%.The feasibility of using the mathematical model data as a training set and the physical model data as a test set is guaranteed.(3)In order to highlight the characteristics of strain data and accelerate the training speed of the inversion model,principal component analysis is used to reduce the sample data set to three dimensions,and network structure applicability analysis is conducted based on existing deep learning methods.The BP neural network model is selected as the algorithm core to build the inversion model under the combined action of damage inducements,clarifying the data sources of the inversion sample training set and the inversion sample test set,By optimizing model parameters through reasonable parameter adjustment methods,a feedforward neural network model with different hidden layers is constructed to meet the different requirements of position inversion and strength inversion,and an algorithm framework for position inversion and strength inversion of an overhead vertical wharf under the combined effects of damage incentives is constructed.(4)The BP neural network inversion model is used to inverse the combined effects of damage inducements on overhead vertical wharves.By adjusting the number of layers of the parameter full connection layer and the size of the parameter Dropout zero setting rate,the maximum accuracy rate of damage inducement location inversion based on the sample data set can reach 0.94.The Hamming loss index distribution is the same as the accuracy index,and the maximum R2 value of strength inversion is about 0.92,The inverse model can well identify the location of combined damage incentives,the strength of ship impact force under combined damage incentives,and the strength of bank slope instability.Finally,based on the damage inducement inversion visualization window,the impact strength and impact location of the ship’s impact force in the combination of damage inducements are accurately displayed,providing a preliminary idea for the development of subsequent visualization software in practical engineering.