论文部分内容阅读
微粒碰撞噪声检测（PIND：ParticleImpactNoiceDetection）是一种被用来识别密封元器件中可移动多余物（微粒）的无损检测手段。这些多余物会引起元器件短路或随机短路失效。由于PIND测试局限于人工判别，所以其结论具有主观性，判别的精度亦低于50％。本文研究了航天继电器中多余物的自动检测系统，提出自动检测系统的构成框图，并就框图中各个组成部分的实现进行了探讨，着重研究了继电器信号的特征提取问题。
Particle Impact Noise Detection (PIND: ParticleImpactNoiceDetection) is a non-destructive testing method used to identify movable parts (particles) in sealed components. These extraneous materials can cause short circuit or random short circuit failure. Because the PIND test is limited to manual discrimination, its conclusion is subjective and its accuracy is less than 50%. In this paper, we study the automatic detection system of redundant objects in aerospace relay, propose the block diagram of automatic detection system, discuss the realization of each component in the block diagram, and focus on the feature extraction of relay signals.