基于BP神经网络的飞机燃油量测量改进算法

摘要

The calculation method of fuel quantity measurement is one key technology for the design of aircraft fuel system.Inter⁃polation method using dummy sensor has the shortage of requiring huge database for mass data of airborne volume characteristic.Based on the BP neural network,an improved calculation method for aircraft fuel quantity gauge is applied in the paper,focusing on the method of the determination of adaptive boundary value and volume data compression,and making a simulation on the fuel tank of an aircraft. The simulation results show that it reduces the amount of data greatly in the airborne computer,providing a guaranteed high precision of the calculation.%油量测量计算方法是飞机燃油系统设计的关键技术。针对基于虚拟传感器插值方法存在机载体积特性数据库容量庞大的不足，提出基于BP神经网络的改进计算方法。重点讨论了自适应边界值确定与体积特性数据压缩的方法，并以某型飞机油箱为例进行仿真。仿真结果表明，在保证计算精度的前提下，可极大地减少机载计算机的数据量。