论文部分内容阅读
为了实现航空发动机滑油压力、滑油温度、振动值在试飞中的趋势监控,采用神经网络方法对某型发动机大量试飞数据进行训练和验证,获得了这几个参数全过程较为准确的计算模型。计算模型应用于该型号另1台发动机参数趋势监控中,在应用前,利用有限架次试飞数据修正了这几个参数的计算模型,采用动态链接库形式实现计算模型与原有实时监控系统的协同工作,进行了模型计算结果和试飞结果趋势实时对比监控。结果表明:模型计算结果和试飞结果变化趋势吻合良好,说明了神经网络计算模型的准确性以及在关键参数趋势监控中的工程实用性。
In order to realize the trend monitoring of aeroengine oil pressure, oil temperature and vibration value during the flight test, neural network method was used to train and verify a large number of test data of a certain engine, and the more accurate calculation model of these parameters . The calculation model is applied to the trend monitoring of the other 1 engine parameter of this model. Before the application, the calculation models of these parameters are corrected by using the finite flight test data. The dynamic link library is used to realize the coordination between the calculation model and the original real-time monitoring system Work, carried out the model calculation results and test results real-time contrast monitoring. The results show that the model calculation results are in good agreement with the trend of flight test results, indicating the accuracy of the neural network model and engineering practicability in monitoring the trend of key parameters.