【摘要】 针对液体火箭发动机试验台健康性能退化的问题,提出一种模糊综合评判与自适应相关向量机(Adaptive Relevance Vector Machine,ARVM)相结合的试验台健康评估与故障预测方法。利用模糊综合评判策略对多传感器在单时序点上的数据进行融合,通过去模糊化方法得到健康度随时间变化的曲线;借助相空间重构方法构造三维输入输出模型,获取不同时序点健康度之间的相关关系,并以此为输入输出,建立试验台健康度预测模型,该模型以健康度的发展来表征试验台健康性能变化;将该方法应用于某型发动机增压系统健康评估中,验证了其有效性。更多还原

【Abstract】 Aiming at the degradation of health performance of LRE test-stand, a method of health assessment and fault prediction, based on fuzzy evaluation and Adaptive Relevance Vector Machine(ARVM), is proposed. Firstly, the fuzzy-comprehensive evaluation strategy is used to fuse the multi-sensor data at single timing point, to obtained the changing curve of the health degree over time. Secondly, the 3 D input and output model is constructed by means of phase space reconstruction to acquire the correlation between the health of different timing points, then, the establishment of test bed health prediction model, the model is used to characterized the health performance of the test bed. Finally, the method is applied to the health prediction of the fuel booster system of a certain engine test bed to verify its effectiveness.更多还原