Complex system health condition estimation using tree-structured simple recurrent unit networks

Abstract Modern production has stricter requirements for the reliability of complex systems; thus, it is meaningful to estimate health systems. A system diverse observation features and internal structures, which have been difficult study with regard condition estimation. To describe continuous gradually changing time-based characteristics a system’s condition, this develops feature selection model based on information amount stability. Then, tree analysis designed according selected relevant features, developed using expert knowledge, node weight calculated by correlation coefficient generated during process. Using simple recurrent unit (SRU), time series machine learning algorithm that achieves high operating efficiency, results are combined establish tree-structure SRU (T-SRU) Finally, NASA turbofan engine data used verification. Results show proposed T-SRU can more accurately improve execution efficiency networks approximately 46%.