Research On Intelligent Viscosity Model Of Thrombus Based On Rheological Experiment

Today,cardiovascular and vascular diseases are becoming more common.As a common cause of cardiovascular disease,the rhetorical properties of the thrombus have increasingly attracted the attention of experts and scholars.To accurately describe the rhetorical properties of the thrombus,with the advent of computer science,the application of intelligent algorithm for model definition provides a new idea for model definition.In this paper,based on the measured data,the rhetorical properties of the thrombus were studied from two aspects:the classical model consisting of viscosity and the intelligent component viscosity model based on the neural network.The specific research content is as follows:（1）Experimental data on thrombus viscosity and shear rate were obtained through the design of rhetorical experiments.Thrombin has been found to have both shear thinning and thixotropy characteristics.Experimental data were introduced into different viscosity synthesis models,and the advantages and disadvantages of the model were evaluated from two perspectives of adjustment accuracy and parameterization.The results show that the Herschel-Bulkley model has the highest adjustment accuracy,reaching 0.9927.At the same time,the configuration is more in line with the viscosity characteristics of the thrombus,the parameter of initial resistance to yieldτ₀in the Herschel-Bulkley model was discussed,and the thrombus types were pre-divided byτ₀=500 Pa.（2）Taking into account the new modeling method,and in order to avoid defects,some parameters must be determined by adjusting the curve in the process of determining the classical viscosity model component.In this paper,referring to the adjustment of certain parameters in the classical viscosity model we use the measured data as input,introduce a genetic algorithm to optimize the initial network parameters based on the BP neural network,establish the GA-BP neural network model,and realize the regression of the thrombus viscosity.The average error rates of four groups of test samples are 18.25%,12.14,11.42%and 6.91%.The results show that the use of a genetic algorithm can effectively improve the prediction accuracy of the model,and this model method also provides a reference for solving similar complex nonlinear problems.