论文部分内容阅读
飞机舵面出现损伤时,为了更准确的预测状态参量变化情况,提出了一种改进的序贯最小优化支持向量回归(SMO-SVR,Sequential Minimal Optimization Support VectorRegression)预测方法.采用改进C-C平均方法对多元时间序列进行相空间重构,以确定最优嵌入维数m和延迟时间τd.根据所求m和τd建立加权SVR预测模型,并调整了SMO算法的停机准则.利用区间自适应粒子群算法(IAPSO,Interval Adaptive Particle Swarm Optimization)优化SVR参数,以提高参数优化速度.为了验证改进算法的有效性,针对飞机方向舵损伤故障趋势进行了预测和分析,并与径向基函数神经网络(RBFNN,Radial Basis Function Neural Net-work)方法进行了对比,仿真结果表明SMO-SVR预测模型具有很好的预测能力.
In order to predict the change of state parameters more accurately, an improved prediction method of Sequential Minimal Optimization Support Vector Regression (SMO-SVR) is proposed in this paper.A modified CC average method Multivariate time series to reconstruct the phase space to determine the optimal embedding dimension m and delay time τd.Weighted SVR prediction model is established according to m and τd and the downtime rules of SMO algorithm are adjusted.Using interval adaptive particle swarm optimization (IAPSO, Interval Adaptive Particle Swarm Optimization) to improve the parameter optimization speed.In order to verify the effectiveness of the improved algorithm, the prediction and analysis of the aircraft rudder damage fault trend is carried out and compared with radial basis function neural network (RBFNN, Radial Basis Function Neural Net-work method. The simulation results show that the SMO-SVR prediction model has good predictive ability.