摘要
针对影响因素众多、各因素之间耦合关系复杂,相对油耗呈显著非线性问题,提出了一种基于BP网络的神经元组合线性方法;该方法实现了复杂非线性关系的逼近,并利用MIV算法进行影响因素结构分析,以及各因素微小变化对巡航段油耗的贡献;实验结果对比表明:该方法建立的模型预测精度较高,泛化能力较强,对实际飞行中航线飞行油耗估计以及影响因素评估具有参考价值。
Due to the complex relations among the various factors, the aircraft fuel consumption is significantly nonlinear. This paper propose a method of neurons＇ linear combination based on BP neural network. The method achieved the approaching of complicated nonlinear relationship, and MIV algorithm is introduced to analyse the structure of factors, further, the experiment studied the contribution caused by small changes of each factor on cruise segment fuel consumption. The simulation results suggests that the model built up based on this method performing with high precision and better generalization abilities, and it has reference value in actual aircraft fuel consumption and structure analysis of the factors.
出处
《计算机测量与控制》
2015年第6期2135-2138,共4页
Computer Measurement &Control
基金
国家科技支撑计划(2012BAC20B03)
民航局科技基金项(MHRD201121)
中央高校基本科研业务费(ZXH2012D015
ZXH2012G004
3122013J004)
