一、引言 在工程实践中,我们经常碰到某些问题,这些问题所需分析的区域与周围介质相比很小,而周围介质对这些区域的影响变化缓慢,把这些问题的边界看成是无限边界,是更为合理而且处理也很方便。 用有限元分析无限边界常需要划分单元到较远的区域,变无限域为一有限域,而划出多大的计算边界范围才能获得较好精度,一直是个不明确的问题。此外,这种方法要求单元、节点数较多,从而计算量较大,费用也高。为克服这一不足,P.Bettess提出 I. INTRODUCTION In engineering practice, we often encounter certain problems. The areas needed to analyze these problems are small compared with the surrounding medium. The influence of surrounding media on these areas changes slowly, and the boundary of these problems is considered to be Unbounded borders are more reasonable and easy to handle. The finite element analysis of infinite boundary often requires dividing the element into a more distant area, and changing the infinite field into a finite field. It is an unclear problem to determine how large the computational boundary is. In addition, this method requires more units and more nodes, resulting in larger calculations and higher costs. In order to overcome this deficiency, P. Bettess proposed