Hemodynamic Analysis Of In-Vitro Circulatory System Based On Photoelectric Detection And Neural Network Model

At present,cardiovascular disease becomes a major cause of death in the world.It accounts for 56% of the total deaths worldwide.Common cardiovascular diseases include hypertension,coronary heart disease,atherosclerosis,and heart failure,etc.The occurrence of cardiovascular disease is often accompanied by the changes in some important hemodynamic parameters.Therefore,the hemodynamic analysis runs through various aspects including the occurrence and development of cardiovascular diseases,early prevention,diagnosis,and treatment.Compared with animal and human clinical experiments,the hemodynamic analysis using in-vitro mock loop circulatory system can not only easily adjust various parameters,but also conveniently use photoelectric detection technology to facilitate the real-time measurement.calculation,and analysis of key hemodynamic signals in the circulatory system.In addition,the existing hemodynamic analysis is usually implemented based on the principle of hemodynamics to establish physical models.These hemodynamic physical models often transform the circulatory system to linear system.Although the hemodynamic mechanism of the circulatory system can be effectively described,the simulation results are not accurate and the calculation process is complicated.In this thesis,the machine learning method based on neural network model is used to simplify the calculation of hemodynamic model and realize the fast prediction of hemodynamic parameters.In this thesis,firstly,the in-vivo blood pressure and blood flow waveforms of the subjects at the common carotid artery are obtained,and the input impedance curve describing afterload in the common carotid artery is obtained by Fourier decomposition of the pressure and flow waveforms;Secondly,a five-element lumped parameter model and a neural network model based on the principle of hemodynamics are established based on the pressure and flow waveforms and obtained input impedance curve;Finally,an in-vitro mock loop circulatory system was constructed to simulate the hemodynamic characteristics of common carotid artery.The in-vitro mock loop circulatory system is mainly composed of photoelectric devices and consumables including centrifugal pump,elastic cavity,resistance valve,elastic silicon tube,pressure sensor,laser rangefinder,etc.The centrifugal pump simulates the function of heart pump,elastic cavity simulates the compliance of blood vessels,resistance valve simulates blood flow resistance,elastic silicon tube simulates human artery,pressure sensor is used to measure blood pressure,and laser rangefinder is used to measure the diameter of blood vessels.The control of the working frequency and power class of centrifugal pump is realized using a STM32 controller,which adopts pulse width modulation.The blood pressure and blood flow data measured with the above system are used to train the neural network model,and the prediction results of the neural network model are compared with those of the five-element lumped parameter model.The numerical simulation and experimental results show that the in-vitro mock loop circulatory system designed in this thesis can simulate the blood pressure and blood flow information under different physiological conditions;The results of hemodynamics predicted by neural network model are obviously better than those obtained by five-element lumped parameter model.In conclusion,the hemodynamic analysis using photoelectric detection and neural network model in the in-vitro mock loop circulatory system proposed in this thesis provides new idea and method for further study of the hemodynamic mechanism of the human circulatory system.