Limb Blood Flow Velocity Simulation And Flow Velocity Distribution Reconstruction Of A Multi-Electrode Electromagnetic Blood Flowmeter

A multi-electrode electromagnetic flowmeter(MEF)based on non-invasive measurement principle is proposed in this paper,which is applied to reconstruct the blood velocity profile of human limbs by measuring the induced electromotive force generated at the edge of the blood fluid.Cardiovascular diseases such as coronary artery stenosis,coronary heart disease,and atherosclerosis have become more common and severe that endanger human health seriously,meanwhile the morbidity of these involved diseases is increasing year by year.It has been testified that monitoring blood flow rate changes could prevent and control such diseases in advance.Therefore,it is significant to study the multi-electrode electromagnetic limb blood velocity simulation and velocity distribution reconstruction.The finite element analysis method has been used in this paper,and the three-dimensional models of different scale of multi-electrode limb blood measurement system have been established by using the COMSOL Multiphysics5.3a simulation software.The C-type coil excitation system has been simulated and optimized in order to improve the magnetic induction intensity of the measurement cross section.The velocity value of each flow area has been preset to study the induced electromotive force generated by the blood flow.The regional weight function value has been obtained based on the research mentioned above.Considering blood as a non-Newtonian fluid,the Euler-Euler model has also been used to numerically simulate and analyze the solid-liquid two-phase flow of blood laminar flow and RANS turbulence,which provides simulation basis and support for the further study of limb blood flow velocity measurement.The reconstruction of arterial and venous images at the cross-section of human limbs is crucial to calculate the regional weight function accurately.The deep learning method based on convolutional neural network(CNN)has been adapted to the reconstruction of arterial and venous images of MEF.Ten-fold cross-validation has been used to divide a large amount of simulation data into a training set and a validation set.The network structure based on LeNet includes two convolutional layers,two pooling layers,a dropout layer,and fully connected layers.The result of image reconstruction could determine the position,size and shape of arteries and veins.According to the simulation data,this paper develops a combination of truncatedsingular value decomposition and Tikhonov algorithm(TSVD_TIK)and fractional order Tikhonov algorithm(F_TIK)to reconstruct the ill-conditioned matrix of dimensions.Comparing with the velocity reconstruction results of the truncated singular value decomposition algorithm and the Tikhonov algorithm,the algorithm proposed in this paper has small standard deviations which is suitable to velocity reconstruction in future.By analyzing the standard deviation of different fractional order parameters of F_TIK,the best fractional order parameters have been obtained in this paper.