基于两级级联神经网络估计血泵转速及压力变化对血流量的影响

背景：对人工心脏输出流量进行及时精确的检测直接关系到人工心脏在动物实验及临床应用中的效果,但实现起来却比较困难。目的：探究辅助循环过程中,心血管血流动力学参数如何反映血泵的工作状态？通过体外辅助循环实验,结合理论分析法,以得到掌握血泵流量特性的神经网络结果。方法：为了正确地评估和检测心室辅助装置中血泵的工作状态,建立两种不同类型的神经网络级联模型,评估血泵转速、压力的连续变化对流量的影响。在第一级中,运用BP神经网络来评估在血泵不同转速下连续变化的压力对流量的影响;在第二级中,运用径向基网络估计血泵转速连续变化对流量的影响。结果与结论：将经训练的级联网络应用于评估转速、压力对流量造成的影响,其结果与以往的方法相比显示了很好的评估能力。

Effects of rotary speed and pressure changes of blood pump on flow rate base on two cascaded neural networks

BACKGROUND：In clinical applications,the timely and accurate detection of the output flow of artificial heart is very necessary,as it is directly related to the effects of zoopery and clinical application.However,it is difficult to achieve.OBJECTIVE：To explore how cardiovascular hemodynamic parameters reflect the blood pump＇s working status during the process of assist circulation and obtain the neural network results which grasp the characteristics of the output flow of blood pump by experiments and theoretical analysis.METHODS：In order to estimate and test the working condition of the blood pump in ventricular assist device correctly,two types of neural networks were established,and the effects of changing rotary speed and pressure on flow rate of the pump were estimated.In the first order,the flow rate affected by continuously changing blood pressure at different rotational speed was estimated by the BP neural network.In the second order,radial basis function neural network was applied to estimate the flow rate of the pump at different rotational speed of the blood pump.RESULTS AND CONCLUSION：This method showed a better estimation ability to estimate the flow rate according to different rotational speed and blood pressure compared with previous methods.