机电设备可靠性验证试验统计方案设计方法研究  

Study on the Design of Statistical Schemes for Reliability Demonstration Test of Electromechanical Equipment

作　　者：汪亚顺 张书锋 陈循[1,2] WANG Yashun;ZHANG Shufeng;CHEN Xun(College of Intelligent Science and Technology,National University of Defense Technology,Changsha 410073;Laboratory of Science and Technology on Integrated Logistics Support,National University of Defense Technology,Changsha 410073)

机构地区：[1]国防科技大学智能科学学院,长沙410073 [2]国防科技大学装备综合保障技术重点实验室,长沙410073

出　　处：《机械工程学报》2021年第12期247-257,共11页Journal of Mechanical Engineering

基　　金：国家重点研发计划(2017YFC0806302);国家自然科学基金(51875570)资助项目。

摘　　要：提出机电设备可靠性验证试验统计方案设计方法,推导指数分布场合及Weibull分布场合接受概率方程组,并给出求解方法流程。据此分析样本量及Weibull分布形状参数对传统可靠性验证试验统计方案(GJB899A—2009中典型方案)实际抽样风险的影响。研究表明,指数分布场合下样本量较小时,采用允许失效的传统方案时,生产方与使用方风险均与名义风险相差较大,且对使用方非常不利。随着样本量增大,生产方风险不断上升而使用方风险不断降低,逐渐接近于名义风险。Weibull分布场合下,采用传统方案时样本量越大机电设备通过可靠性验证试验被接收的概率越大,且Weibull分布形状参数越大,理论抽样特性越接近于理想的陡峭"台阶"形状。可为工程实际中开展机电设备可靠性验证试验及减少误判风险提供借鉴。A statistical scheme design method for the reliability demonstration test of electromechanical equipment is proposed. A pair of acceptance probability equations under the occasion of the exponential distribution and Weibull distribution are derived, and solution flow is given. Accordingly, the effect of sample size and Weibull distribution shape parameters on the actual sampling risk of the traditional statistical scheme for reliability demonstration tests(typical schemes in GJB899A—2009) is analysed. It has been shown that, with small sample sizes under exponential distribution, both producer and consumer risks differ from the nominal risk and are very unfavourable to the consumer when the traditional scheme with failures is used. As the sample size increases, the producer risk increases and the consumer risk decreases, gradually approaching the nominal risk. Under Weibull distribution, the larger the sample size in the traditional scheme, the greater the probability that the electromechanical equipment will be accepted during the reliability demonstration test, and the larger the shape parameter of the Weibull distribution, the closer the real sampling characteristics to the ideal steep "step" shape. It can be used as a reference for conducting reliability demonstration tests of electromechanical equipment and reducing the risk of errors in engineering practice.

关 键 词：可靠性验证试验 统计方案 样本量 WEIBULL分布 机电设备 

分 类 号：TB114[理学—概率论与数理统计]