基于神经网络算法的液体火箭发动机实时故障检测方法研究

Real-time fault detection method based on neural network algorithm for LRE

以某液体火箭发动机为研究对象,将实时故障检测作为中心,分析了神经网络算法的特点及其实现步骤,利用Lab Windows/CVI与MATLAB混合编程的原理,实现和改进了基于神经网络的发动机实时故障检测方法,并用多次试车数据进行了检验。试车数据验证结果表明,该方法能及时、准确、有效地检测发动机稳态过程的故障。研究结果对发展未来液体火箭发动机的箭载故障检测系统具有重要的参考价值。

The real-time fault detection for a liquid-propellant rocket engine （LRE） is put on emphasis in this paper. The real-time fault detection based on the neural network algorithm is realized and ameliorated by means of analyzing the characteristic and realization steps of neural network （NN） algorithm, using the composite programming method with Lab Windows/CVI and MATLAB. The algorithm proposed in the paper has been tested and verified with a large amount of hot-fire test data of LRE. The results show that the method can detect the faults occurred in the steady-state process of LRE efficiently and accurately. The method presented can be valuable for the development of on-board fault detection system of LRE.