混沌蝙蝠算法UAV网络控制系统故障诊断策略

Fault Diagnosis Strategy of UAV Network Control System Based on Chaotic Bat Algorithms

针对总线条件下具有双通道多包传输的短时变时延无人机(Unmanned Aerial Vehicle,UAV)发动机控制系统,提出一种基于改进蝙蝠优化算法(Bat Algorithm,BA)的发动机故障检测算法,快速变化的延迟被认为是系统不确定性的一种。残差信号的传递函数与噪声和故障信号的传递函数之比用作目标函数。通过改进的Bat算法对其进行优化,以获得最佳的观察增益矩阵,使得该故障观测器系统不仅能够抑制噪声,而且对提升对故障敏感程度。最后将该方法在网络控制系统半物理平台上进行仿真,仿真结果说明所提出的方法能够同时抑制噪声信号和放大故障信号,从而提高了故障诊断的正确率,降低了虚警率。

Aiming at the engine control system of the short and varied time delay unmanned aerial vehicle(UAV)with two channels and multiple packets transmission under the bus condition,an engine fault detection algorithm based on the improved bat algorithm(BA)is proposed.The fast changing delay was considered as one of the system uncertainties.The ratio of transfer function of residual signal to transfer function of noise and fault signal is used as objective function.The improved bat algorithm is used to optimize the system to obtain the best observation gain matrix,which makes the system not only suppress noise,but also improves the sensitivity to fault.Finally,the method is simulated on the semi physical platform of the network control system.The simulation results showed that the proposed method can suppress noise signal and amplify fault signal at the same time,thus improving the accuracy of fault diagnosis and reducing the false alarm rate.