A Bidirectional LSTM Prognostics Method Under Multiple Operational Conditions

TLDR

Modern engineered systems generally operate under complex conditions, yet most AI‑based prognostic methods lack models that effectively use operational condition data for remaining useful life prediction. This paper develops a novel prognostic method based on bidirectional long short‑term memory (BLSTM) networks. The BLSTM‑based architecture integrates multiple sensor signals and operational condition data through two BLSTM networks—one extracting raw sensor features, the other learning higher‑level features from operational signals and sensor‑derived features—followed by fully connected layers and a linear regression output for RUL prediction. The method is validated on aircraft turbofan engines and outperforms other state‑of‑the‑art prognostic approaches.

Abstract

Modern engineered systems generally work under complex operational conditions. However, most of the existing artificial intelligence (AI)-based prognostic methods still lack an effective model that can utilize operational conditions data for remaining useful life (RUL) prediction. This paper develops a novel prognostic method based on bidirectional long short-term memory (BLSTM) networks. The method can integrate multiple sensors data with operational conditions data for RUL prediction of engineered systems. The proposed architecture based on BLSTM networks includes three main parts: first, one BLSTM network is used to directly extract features hidden in the multiple raw sensors signals; second, another BLSTM network is employed to learn higher features from operational conditions signals and the learned features from the sensors signals; and, third, fully connected layers and a linear regression layer are stacked to generate the target output of the RUL prediction. Unlike other AI-based prognostic methods, the developed method can simultaneously model both sensors data and operational conditions data in a consolidated framework. The proposed approach is demonstrated through a case study on aircraft turbofan engines, and comparisons with other popular state-of-the-art methods are also presented.