火箭发动机齿轮组故障检测数学模型

The Mathematical Model of Rocket Engine Gears Fault Detection

建立故障检测和特征频率提取数学模型,采用自适应BP神经网络算法对故障状况进行了仿真模拟运算。仿真结果表明新故障诊断算法结果优于传统BP神经网络算法,由原来的10000步降低至700步,有效提高了运算速度,同时运算精度也有所提高,检测准确置信度提高了10%,提高了故障检测的概率。研究成果为火箭发动机涡轮泵故障的早期发现与故障解决提供了算法理论的依据,有较好的工程推广运用性。

The fault detection and feature frequency extraction models were built mathematically. The simulation was builtbased on the real collected data. Simulation result shows that the new algorithm is better than the traditional algorithm withthe better detection performance and lower computing cost. The iteration steps were reduced from 10000 to 700, and theprecise detection convenience level is improved by 10%. Research result provides good theory base in the early fault diagnosis and discovery. It has good value in the application.