Abstract:

Accurately evaluating missile products with the characteristics of "long-term storage and one-time use" is a prerequisite and key to achieving safe storage， reliable use， and reasonable maintenance， which has significant military and economic benefits. However， existing storage reliability evaluation methods do not consider the phenomenon of "upside down" and incomplete data of missile products， resulting in low evaluation accuracy. According to the three characteristics of missile product life data， a minimum Chi-square estimation method based on order preserving regression is proposed in this paper. First， isotonic regression method is used to solve the problem of "upside down" in stored information. Then， the storage reliability is evaluated by the minimal Chi-square estimation method， and the goodness of fit is determined based on chi-square statistics. Finally， using the missile to store information naturally， the traditional method is compared with the method established. The results show that the establishment method significantly improves the rationality and accuracy of missile storage reliability evaluation and lays a foundation for scientific life determination and maintenance of equipment products.

Key words: assessment of storage reliability, regular inspection, binary data, incomplete data, isotonic regression, minimum Chi-square estimation

摘要：

准确评估具有“长期贮存、一次使用”特点的导弹产品是实现安全贮存、可靠使用以及合理维护的前提与关键，具有重大的军事经济效益。然而现有贮存可靠度评估方法未考虑导弹产品数据的“倒挂”、不完全等现象，存在评估精度较低的问题。针对导弹产品寿命“倒挂”、成败型和不完全3个特点，提出一种基于保序回归的极小卡方估计方法。利用保序回归方法解决贮存信息中存在的“倒挂”问题，通过极小卡方估计方法评估贮存可靠度，并基于卡方统计量判断拟合优度。利用导弹自然贮存信息，对传统方法和新建立方法进行对比分析。结果表明提出的方法显著提高了导弹贮存可靠度评估的合理性与准确性，为导弹产品的科学定寿和维修延寿奠定了基础。

关键词: 贮存可靠度评估, 定期检测, 成败型数据, 不完全数据, 保序回归, 极小卡方估计