现有机械时变可靠性模型主要以概率论为基础，依赖于基本随机变量的大样本空间，当没有足够的数据来验证这些随机变量的概率密度正确性时，分析结果的准确性和置信度值得怀疑，概率可靠性方法存在局限性。在航天机构中，由于成本高，在复杂服役环境中可观测性差、检测难，积累信息难、统计类信息往往难以获得，主要存在区间类信息，为保证分析结果更加精确，考虑时变区间变量，研究并提出非概率时变可靠性分析方法：将时变区间转化为服从截断正态分布的随机过程，求出在时变区间内的上穿率，针对正态截断分布变量，利用改进的非概率PHI方法，求得累积失效概率。解决区间变量随时间变化下的非概率时变可靠性问题。利用算例和工程实例验证本方法的准确性和可行性。

Several approaches have been proposed object to solve time variant reliability problems in recent decades based on classical probability theory dealing with problems in practice. Indeed the statistical information of complex mechanical systems,such as aircraft or spacecraft,is always hard to be collected. Interval data are encountered frequently in practical engineering problems. An improved non-probabilistic PHI method is proposed in order to guarantee the accuracy of results with time-variant interval-defined. A time variant interval is converted to a stochastic process distributed the truncated normal as well as the up crossing rate is calculated. Considering the variables with the truncated normal distribution,the cumulative failure probability can be estimated using the improved non-probabilistic PHI method. For demonstration and validation purpose,the method is applied in a case representing a mechanical system (a beam) placed in an environment that can have exceptional configuration.