Aero-engine direct thrust control can not only improve the thrust control precision but also save the operating cost by reducing the reserved margin in design and making full use of aircraft engine potential performan...Aero-engine direct thrust control can not only improve the thrust control precision but also save the operating cost by reducing the reserved margin in design and making full use of aircraft engine potential performance.However,it is a big challenge to estimate engine thrust accurately.To tackle this problem,this paper proposes an ensemble of improved wavelet extreme learning machine(EW-ELM)for aircraft engine thrust estimation.Extreme learning machine(ELM)has been proved as an emerging learning technique with high efficiency.Since the combination of ELM and wavelet theory has the both excellent properties,wavelet activation functions are used in the hidden nodes to enhance non-linearity dealing ability.Besides,as original ELM may result in ill-condition and robustness problems due to the random determination of the parameters for hidden nodes,particle swarm optimization(PSO)algorithm is adopted to select the input weights and hidden biases.Furthermore,the ensemble of the improved wavelet ELM is utilized to construct the relationship between the sensor measurements and thrust.The simulation results verify the effectiveness and efficiency of the developed method and show that aero-engine thrust estimation using EW-ELM can satisfy the requirements of direct thrust control in terms of estimation accuracy and computation time.展开更多
基金supported by the National Natural Science Foundation of China (Nos.51176075,51576097)the Fouding of Jiangsu Innovation Program for Graduate Education(No.KYLX_0305)
文摘Aero-engine direct thrust control can not only improve the thrust control precision but also save the operating cost by reducing the reserved margin in design and making full use of aircraft engine potential performance.However,it is a big challenge to estimate engine thrust accurately.To tackle this problem,this paper proposes an ensemble of improved wavelet extreme learning machine(EW-ELM)for aircraft engine thrust estimation.Extreme learning machine(ELM)has been proved as an emerging learning technique with high efficiency.Since the combination of ELM and wavelet theory has the both excellent properties,wavelet activation functions are used in the hidden nodes to enhance non-linearity dealing ability.Besides,as original ELM may result in ill-condition and robustness problems due to the random determination of the parameters for hidden nodes,particle swarm optimization(PSO)algorithm is adopted to select the input weights and hidden biases.Furthermore,the ensemble of the improved wavelet ELM is utilized to construct the relationship between the sensor measurements and thrust.The simulation results verify the effectiveness and efficiency of the developed method and show that aero-engine thrust estimation using EW-ELM can satisfy the requirements of direct thrust control in terms of estimation accuracy and computation time.
文摘为精确预计锦界矿某工作面开采沉陷,首先结合该工作面的地质资料、采掘工作平面图及孔柱状图,采用FLAC3D软件建立了该工作面开采沉陷仿真模型,得到工作面推进100、300、500、700 m时的开采沉陷数据;其次基于该类数据对BP神经网络预计模型进行训练和验证,建立沉陷数据与工作面推进距离的非线性关联;然后用粒子群优化算法(Particle swarm optimization,PSO)对BP神经网络模型的结构参数和连接权值阈值进行优化,并引入遗传算法(Genetic algorithm,GA)中的自适应变异因子以一定概率初始化部分变量,以解决PSO算法易陷入局部最优解的问题,避免BP神经网络模型易陷入局部最小值、训练收敛速率低以及PSO算法易早熟收敛等问题。分别采用BP神经网络模型、PSO-BP神经网络模型以及所提模型进行试验对比,并引入偏差平方和(Sum of squares for total,SST)对各模型的预计精度进行评价,研究表明:在工作面分别推进100,300,500 m的情况下,BP神经网络模型的SST值分别为0.056,0.062,0.066,PSO-BP神经网络模型的SST值分别为0.049,0.054,0.048,所提模型的SST值分别为0.028,0.026,0.031,明显小于前两者,表明该模型有助于提高矿区开采沉陷预计精度,有一定的实用价值。
