摘要
针对航空活塞发动机排气门卡阻故障,经过对故障机理的分析,提出了一种利用神经网络对排气门导套与气门杆的配合间障进行预测,以间接预测排气门卡阻故障的方法。将影响排气门积垢速率的因素设定合理的特征值,以这些特征值和发动工作时间作为输入向量,配合间隙作为输出向量,分别建立了GRNN神经网络和BP神经网络预测模型。预测实例表明,GRNN神经网络预测模型具有较高的预测精度、稳定的网络以及较快的收敛速度,预测性能优于BP神经网络模型,预测结果可作为评估排气门卡阻故障发生概率的有效依据。
For aviation piston engine exhaust valve jamming failure,after analyzing failure mechanism, this paper puts forward a method to predict the assembled clearance of exhaust valve guide and valve stem base on neural network, the method can indirect predict the exhaust valve jamming failure. The reasonable characteristic values are set for the factors that will affect exhaust valve fouling velocity, these characteristic values and the working time of engine are served as input vector, the assembled clearance is served as output vector, using the above-mentioned vectors a GRNN neural network and a BP neural network prediction model are established. The prediction example shows, the GRNN neural network prediction model has high prediction accuracy, network stability and fast convergence, it＇s prediction performance is superior to the BP neural network, it＇s prediction results can be used to effectively assess the probability of exhaust valve jamming failure.
出处
《机械》
2014年第6期12-16,共5页
Machinery
基金
中国民航飞行学院重点科学基金资助项目:维修信息驱动的通用飞机故障预测与健康管理(PHM)技术的研究(ZJ2012-04)
