基于本征正交分解方法的螺旋槽干气密封瞬态响应分析  

Spiral Groove Gas Face Seal Transient Response Analysis Based on Proper Orthogonal Decomposition Method

作　　者：尹源[1] 黄伟峰[1] 刘向锋[1] 胡松涛[1] 刘莹[1] 

机构地区：[1]清华大学摩擦学国家重点实验室,北京100084

出　　处：《机械工程学报》2017年第21期79-85,共7页Journal of Mechanical Engineering

基　　金：国家重点基础研究发展计划(973计划;2012CB026003);国家科技支撑计划课题(2015BAA08B02);国家科技重大专项(ZX06901)资助项目

摘　　要：将本征正交分解方法应用于螺旋槽干气密封瞬态响应计算。通过直接数值模拟求解瞬态雷诺方程和动力学方程,计算出一组指定的压力阶跃变化参数下干气密封的瞬态响应作为采样源,进而由本征正交分解方法得到基函数,建立干气密封动力学降维模型,该模型可用于计算任意压力阶跃变化参数下的瞬态响应。应用此模型求解大量算例,并与直接数值模拟的结果进行比较。结果表明,降维模型求解结果的误差度随基函数的增加总体呈下降趋势,而在被计算算例参数远离采样源算例参数时迅速增大。在采用合适的采样源算例和足够的基函数数目时,本征正交分解方法可以得到精度较高的结果,且其计算速度相对于直接数值模拟法可大幅提高。A proper orthogonal decomposition method is applied to transient response analysis of spiral groove mechanical gas seal. Transient responses of gas face seal under a set of specified pressure step is simulated directly as samples by solving transient Reynolds equation and dynamic equation. The reduced-order model is established by basic functions generated from samples by proper orthogonal decomposition. The model can be used for calculating transient response under widespread pressure step. The comparison of results from reduced-order model and direct simulation comes to following conclusions. Error shows a decreasing trend with increasing number of basic functions, and increases as the example calculated differs greatly with samples. Proper orthogonal decomposition method can get result of high accuracy with appropriate sample and enough basis functions, and the speed is higher than direct simulation significantly.

关 键 词：干气密封 本征正交分解 雷诺方程 瞬态响应 

分 类 号：TH136[机械工程—机械制造及自动化]