A Study Of Brain Injury Evaluation Based On Vehicle Collision Accident Inverse

With the continuous growth of car ownership and the frequent occurance of road traffic accidents,the research of human injury evaluation and protection have aroused wide attention in the field of vehicle safety and injury biomechanics.As an assessment standard,injury criteria is significant for the protection of human body and the improvement of vehicle performance.However,due to the complexity of road traffic accidents and the limitation of vehicle collision data,it is usually impossible to accurately evaluate and predict brain injury in the collision accident,which restricts the improvement of vehicle protection performance.Therefore,according to a large number of crash test data,the establishment of evaluation criteria and prediction models of brain injury has become an important means to evaluate the degree of brain injury in vehicle collision accidents.Taking a case of real vehicle crash accident as the research object,this dissertation seeks to do some research and exploration with practical significance on the reconstruction of vehicle collision accident,the identification of material parameters of brain tissue,the evaluation critiera and prediction of brain injury and the improvement of occupant constraint system.Fully utilizing the advantigates of multifields and multi-disciplinary,this study integrates high fidelity finite element models,the compression test of brain tissue,the computational iniverse technique and intelligent algorithms.According to the above description,on the basis of literature research,the major works and research contents of this dissertation are listed in the subsequent elaboration.(1)The uncertainty inverse of vehicle collison accident considering correlation.Firstly,based on Nataf transformation technique,it can transform the related variables into independent variables.Then,by using the point estimation method,the uncertain inverse problems is transformed into several deterministic inverse problems to calculte the statistical moments of collision parmaters.In order to obtain more collision information,according to statistical moments of the inverse parameters,this dissertation estimates their probability density functions(PDFs)by the application of the maximum entropy principle.Finally,comparing with the calculation results of Monte Carlo simulation(MCS),it demonstrates the presented method is suitable and realible for the vehicle collision accident reconstruction.These inverse parameters are used to the finite element model of vehicle collison.The calculation results show the simulation deformations are in good agreement with the actual deformations of accident vehicle.(2)The identification of visco-hyperelastic material parameters for brain tissue based on the uniaxial compression test.Due to the soft and sticky material characteristics of brain tissue,it is difficult to make the standard specimen of brain tissue.The specimen shape of brain tissue is very irregular,which is not conducive to the reconstruction of brain specimen model.Therefore,by the application of 3D scanning technique,the finite element model of special specimen with complex configuration can be established accurately.To improve the accuracy of solution,the built joint simulation and intelligent optimization algorithm is used to identify the material parameters of brain tissure.In order to validate the accuracy and reliability of the inverse results,the finite element model of other specimen is reconstructed by using the identified material parameters.According to the comparsion of simulation results and experimental results,it indicates two curves match well.In addition,by comparing with the simulation results with the corresponding material paramters from some references,it further demonstrates the identified material parameters are more accurate.It is noted that the identified results can characterize reasonably the relaxtion of brain tissue,but the material parameters is powerless in the references.(3)The evaluation of brain injury based on artificial neural network learning(ANN-L).To obrain the influence factors of brain injury,this dissertation assesses the correlation of kinetic and injury by the application of Spearman's rank correlation coefficient according to large number of vehicle crash test data.The analysis results indicate there is a strong correlation between major kinetic parameters(Maximum Resultant Velocity,Maximum Resultant Acceleration,Maximum Resultant Angle Velocity and Rotational Injury Criterion)and comulacation strain damage measure(CSDM).By the application of K-NearestNeighbor(KNN),some singular data are elimilated to enhance the correlation between kinetic parameters and injury parameters.And it can obtain the optimal weight coefficients of major kinetic parameters by using genetic algorithm.Thus,the new brain injury index(BII)can be formulated.Compared with the tranditional brain injury criteria,it can consider comprehensively the influence of each kinetic parameter and improve the evaluation accuracy of brain injury.In order to quantitatively describe the relationship of kinetic parameters and brain injury,this dissertation carries out maximum-minimum normalization for data preprocessing,and establishes the prediction model of brain injury by the application of ANN-L.The reliability of prediction model is checked by test data.The result shows the accuracy of prediction model of brain injury is high.This approach makes it possible to quantitatively describe the relationship of kinetic parameters and brain injury.(4)Interval multi-objective optimization of occupant restraint system based on brain injury criteria.Firstly,it establishes the incorporate finite element model on vehicle body and occupant restraint system.Then,to assess the influence of important parameters for occupant restraint system,this dissertation proposes a new global sensitivity analysis method based on high sensitivity indices decomposition and derivative-integral.Compared with the tranditional Sobol' sensitivity measure,it is more accurate to evaluate the importance of parameters by using the proposed method.Finally,in order to improve the protection performance of occupant restraint system,this dissertation treats the brain injury criteria HIC and BII as the objectives,regards the other brain injury criteria BrIC and RIC as the constraints to obtain the optimal parameters according to the results of sensitivity analysis.It provides an effective method to improve the protection performance of occupant restraint system under the effects of uncertainty.