传感技术学报2025，Vol.38Issue(4)：579-587,9.DOI:10.3969/j.issn.1004-1699.2025.04.002
基于LSTM的固冲发动机双冗余压强传感器故障诊断研究
Research on Fault Diagnosis of Dual-Redundant Pressure Sensors of Solid Fuel Ramjet Based on LSTM
摘要
Abstract
The pressure in the gas generator is an important feedback variable for the thrust control loop of solid fuel ramjet,and its pre-cise measurement is crucial for ensuring the safety of the engine.In response to the limited reference data and difficulty in fault localiza-tion of dual-redundant pressure sensors,a fault diagnosis method based on long short-term memory(LSTM)and state observer is pro-posed.Firstly,the working principle of the gas generator is introduced and a fault model for the pressure sensor is established.Then,to increase redundancy,the state observer is used to obtain the pressure estimate and construct the system's nominal pressure reference sig-nal.Subsequently,fault labels are set,and corresponding faults are injected into the two sensor models to collect time-series data for each label.Detection and classification of sensor faults are achieved via LSTM.Simulation results indicate that the proposed method can effectively identify the faults of both single-sensor and dual-sensors,with an average accuracy of mathematical simulation over 95％and an average accuracy of hardware in-loop over 90％simulation.It also has good generalization ability.关键词
压强传感器/故障诊断/长短时记忆/双冗余/固体冲压发动机:状态观测器Key words
pressure sensor/fault diagnosis/long short-term memory/dual redundancy/solid fuel ramjet/state observer分类
信息技术与安全科学引用本文复制引用
邹易君,曾庆华,肖云雷,张宗宇,王宏福,叶宵宇,王欢..基于LSTM的固冲发动机双冗余压强传感器故障诊断研究[J].传感技术学报,2025,38(4):579-587,9.基金项目
国家自然科学基金项目(61174120) （61174120）
