中文摘要：

为改善液压滑阀开启过程中阀芯的工作性能,基于流体-固体耦合理论和动网格技术,建立滑阀开启过程三维流体-固体耦合数值分析模型,对阀芯在驱动力、弹簧力及液动力作用下的动作过程进行了模拟分析。针对阀芯受力分析结果,改进滑阀内部流道结构;以改进后阀芯槽口和凸台的4个结构参数为优化对象,利用最小二乘拟合和反向传播神经网络构建最大冯米塞斯应力与液动力峰值的目标函数,借助遗传算法确定了槽口和凸台的4个结构参数最优值。研究结果表明：优化后阀芯所受液动力和最大冯米塞斯应力的峰值分别减小了16.3%和22.0%;优化设计阀芯的结构参数可明显提高滑阀开启性能。

英文摘要：

To improve the working capability of the spool during the opening process of hydraulic spool valve, a three-dimensional numerical model is established based on fluid-structure interaction theory and dynamic mesh technology, and the movement of spool under the action of driving force, spring force and flow force is simulated. The internal flow channel structure is improved in view of stress on spool. The objective functions of peak value of the maximum von Mises stress and flow force are set up through least squares fit and BP neural network, with four structural parameters of improved spool notches and convex being taken as the optimization object. The optimal values of four structure parameters of notches and convex platform are determined by using genetic algorithm. After optimization, the maximum von Mises stress is decreased by 22.0% , and the maximum flow force is decreased by 16.3%. Results show that the spool valve opening performance is obviously improved through the study of the optimization design of structural parameters of spool.