Failure and reliability prediction of engine systems using iterated nonlinear filters based state-space least square support vector machine method

Failure and reliability prediction in engine systems have attracted much attention over the past decades. However, this task remains challenging due to the stochastic nature and dynamic uncertainty of failure and reliability time series data. Two novel approaches for reliability prediction are developed in this study by integrating least square support vector machine (LSSVM) and the iterated nonlinear filters for updating the reliability data accurately. In the presented methods, a nonlinear state-space model is first formed based on the LSSVM and then the iterated nonlinear filters are employed to perform dynamic state estimation iteratively on reliability data with stochastic uncertainty. The suggested approaches are demonstrated with two illustrative examples from the previous literature and compared with the existing neural networks (NNs) and SVMs models. The experimental results reveal that the proposed models can result in much better reliability prediction performance than other technologies.