徐新华, 李伟光, 谢军龙. 舰船集成冷媒水系统损害检测[J]. 中国舰船研究, 2016, 11(1): 135-142. DOI: 10.3969/j.issn.1673-3185.2016.01.018
引用本文: 徐新华, 李伟光, 谢军龙. 舰船集成冷媒水系统损害检测[J]. 中国舰船研究, 2016, 11(1): 135-142. DOI: 10.3969/j.issn.1673-3185.2016.01.018
XU Xinhua, LI Weiguang, XIE Junlong. Damage detection of the warship integrated refrigerant system[J]. Chinese Journal of Ship Research, 2016, 11(1): 135-142. DOI: 10.3969/j.issn.1673-3185.2016.01.018
Citation: XU Xinhua, LI Weiguang, XIE Junlong. Damage detection of the warship integrated refrigerant system[J]. Chinese Journal of Ship Research, 2016, 11(1): 135-142. DOI: 10.3969/j.issn.1673-3185.2016.01.018

舰船集成冷媒水系统损害检测

Damage detection of the warship integrated refrigerant system

  • 摘要: 舰船集中冷媒水系统配置方式可以集中为舰船电子设备冷却提供保障,克服电子设备因自带冷却系统产生的寿命周期费用高、可靠性水平低等问题,能减少资源消耗,简化设备配置,优化系统设计,实现冷却资源的综合利用。集中冷媒水系统的损害检测及管理是舰船损害管制的一项重要内容。在分析舰船冷媒水系统集成及舰船损害管制的基础上,进一步探讨采用主元分析法(PCA)对冷媒水系统测量传感器的损害进行检测与诊断,并在水系统的模拟平台上对传感器的损害检测与诊断进行验证。根据水系统在正常运行条件下的系统压力传感器的测量数据建立的主元模型可解释系统方差95%以上,置信限为1.639 4。在某压力传感器损害(即故障)条件下的运行数据的Q-统计超出这一置信限,说明有压力传感器发生故障,且从Q-分布图可以看出,故障压力传感器的测量贡献率最高,占45%~75%,从而实现了故障传感器的诊断。提出一种平时与战时相结合的集中冷媒水管网的损害检测与诊断流程。

     

    Abstract: Warship integrated refrigerant systems may provide security for the cooling of electronic equip-ment, overcoming the problems of low reliability, short lifetime, and high costs etc. It may also reduce the resource consumption, optimize the system design, and realize the comprehensive utilization of cooling re-sources. Therefore, damage detection and management of this element is of vital importance to the warship damage control system. Based on the introduction and analysis of integration of warship refrigerant systems and warship damage control systems, this paper presents the damage detection and diagnosis of sensors of a refrigerant system by using Principal Component Analysis (PCA), which is further validated on the simula-tion platform of a water system. Particularly, the PCA model is established based on the normal operation data measured by pressure sensors of the water system. This model explains the systematic variance above 95%, and the confidence limit is 1.639 4. When one pressure sensor is damaged (i.e. fault), the Q-statistic of the operation data exceeds the confidence limit, indicating that some pressure sensor could be in fault. The Q-contribution shows that the measurement from the faulty sensor has the highest contribution from 45% to 75%. In addition, the faulty sensor can be isolated based on the Q-contribution plot. This paper fur-ther proposes the ways for damage detection and diagnosis of the pipe network of the warship integrated with refrigerant systems by considering the case under normal time and wartime together.

     

/

返回文章
返回