摘要
针对飞机在飞行时油箱因受震动引起油面起伏不平,导致原有静止状态时的计算模型产生较大测量误差,提出采用BP神经网络的预测飞机剩余油量;但由于BP神经网络存在学习效率低、收敛速度慢和易陷入局部极小等局限,采用改进粒子群算法优化BP神经网络的训练;将改进PSO-BP算法用于飞机剩余油量的测量,实验结果表明,与传统BP学习算法比较,改进PSO-BP算法具有训练时间短,相对误差小,控制精度高等优点,有效地提高了油量测量的精度。
When the aircraft in the flight, fuel level is rise and fall because of tanks＇ vibration, which result in that calculate model of static condition produces bigger measurement error. BP neural network algorithm is put forward to calculate the remaining fuel of the airplane. However, because BP neural network has the limitations, which are lower learning efficiency, slow convergence and the local extreme values, a kind of improved PSO algorithm is adopted to optimize the training of the BP neural network. Finally we apply the PSO-- BP algo- rithm to measure the aircraft remaining fuel volume, the results of experiments indicate that compared to the traditional BP algorithm, the PSO--BP algorithm has advantages of lower training time, lower relative error and higher control accuracy, and it also has enhanced the measurement accuracy of the fuel volume.
出处
《计算机测量与控制》
CSCD
北大核心
2012年第6期1452-1454,1457,共4页
Computer Measurement &Control
