针对具有多维状态变量、多种工作模式和故障模式的复杂工程系统,提出一种基于综合健康指数(synthesized health index,SHI)与相关向量机(relevance vector machine,RVM)的系统级失效预测方法。在离线训练阶段,先根据有限失效历史数据建立各工作模式下的健康评估模型,并据此获得各历史退化轨迹的SHI序列;然后再使用RVM对这些序列进行回归处理,进而辨识出与回归曲线最为匹配的函数模型。在线预测阶段,先运用健康评估模型计算当前设备的SHI序列并进行RVM回归,再拟合出离线阶段确定的函数模型并添加时变噪声;最后,外推预测出系统剩余使用寿命的概率密度分布。该方法成功应用到涡轮发动机的失效预测案例。
The online diagnosis for aircraft system has always been a difficult problem. This is due to time evolution of system change, uncertainty of sensor measurements, and real-time requirement of diagnostic inference. To address this problem, two dynamic Bayesian network(DBN) approaches are proposed. One approach prunes the DBN of system, and then uses particle filter(PF) for this pruned DBN(PDBN) to perform online diagnosis. The problem is that estimates from a PF tend to have high variance for small sample sets. Using large sample sets is computationally expensive. The other approach compiles the PDBN into a dynamic arithmetic circuit(DAC) using an offline procedure that is applied only once, and then uses this circuit to provide online diagnosis recursively. This approach leads to the most computational consumption in the offline procedure. The experimental results show that the DAC, compared with the PF for PDBN, not only provides more reliable online diagnosis, but also offers much faster inference.
Particle filtering (PF) is being applied successfully in nonlinear and/or non-Gaussian system failure prognosis. However, for failure prediction of many complex systems whose dynamic state evolution models involve time-varying parameters, the traditional PF-based prognosis framework will probably generate serious deviations in results since it implements prediction through iterative calculation using the state models. To address the problem, this paper develops a novel integrated PF-LSSVR framework based on PF and least squares support vector regression (LSSVR) for nonlinear system failure prognosis. This approach employs LSSVR for long-term observation series prediction and applies PF-based dual estimation to collaboratively estimate the values of system states and parameters of the corresponding future time instances. Meantime, the propagation of prediction uncertainty is emphatically taken into account. Therefore, PF-LSSVR avoids over-dependency on system state models in prediction phase. With a two-sided failure definition, the probability distribution of system remaining useful life (RUL) is accessed and the corresponding methods of calculating performance evaluation metrics are put forward. The PF-LSSVR framework is applied to a three-vessel water tank system failure prognosis and it has much higher prediction accuracy and confidence level than traditional PF-based framework.
