辨识技术在高空螺旋桨建模与验证中的应用  

Application of Identification Technology in High-altitude Propeller Modeling and Verification

作　　者：孙康文[1] 梁高铭[1] 祝明[1] 

机构地区：[1]北京航空航天大学航空科学与工程学院,北京100191

出　　处：《计算机仿真》2013年第12期53-56,共4页Computer Simulation

摘　　要：研究飞艇螺旋桨动力系统优化设计问题,平流层飞艇高空螺旋桨性能应有准确模型与实验数据。缺乏大量的实验数据支持,准确性难以保证。在对螺旋桨特性基本模型进行分析的基础上,将自适应神经模糊推理系统(ANFIS)进行合理改进,并引入到高空螺旋桨性能仿真建模中。在明确系统输入和输出的基础上,结合有限的风洞试验数据建立螺旋桨仿真模型进行验证。分析中为保证有限实验数据的合理利用,需首先进行有效试验数据的选取和输入空间的划分,并根据推理系统的结构分析输出参数的精度是否满足应用要求,再结合BP神经网络拟合训练方法进行训练以满足自适应神经模糊推理系统方法的应用。利用上述仿真模型进行了推力F、功率P和效率η的仿真数据与试验数据的对比分析,验证了模型在应用中的有效性和准确性,可为进一步的平流层飞艇能源动力系统优化设计提供参考。When the optimization design of airship propeller power system has been studied, the effective performance analysis of the stratospheric airship high altitude propeller should have an accurate model with experimental data. So, if we lack of a large number of experimental data, the accuracy is difficult to guarantee. An improved adaptive network -based fuzzy inference system （ANFIS） was introduced into the simulation model of high altitude propeller performance via the analysis of the basic working principle of the propeller. Based on the clear relationship between system＇s input and output, the construction and verification of the propeller simulation model were completed by using the limited wind tunnel test data. In order to ensure the rational use of limited experimental data, the paper presented the effective test data selection and the input space partition firstly. Then, whether the precision of the out- put parameter satisfies the application requirements was decided according to the analysis of inference system＇s structure. At last, the BP neural network fitting training method was employed for data training to meet ANFIS method. Finally, the simulation of the thrust F, the power P and the efficiency η was performed. Compared with the test data, the simulation results demonstrate that the method is valid and accurate. Thus, the method can provide some reference for further stratospheric airship energy power system detailed design.

关 键 词：平流层飞艇 螺旋桨 系统辨识 自适应神经模糊推理系统 验证 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]