摘要
建立了基于粒子群优化的轴流压气机机匣压力支持向量机预测模型。利用支持向量机的强大非线性映射能力,实现了对某型轴流压气机机匣压力时间序列的非线性预测,并运用粒子群优化算法对支持向量机的重要参数进行了优化,增强了预测模型对混沌动力学的联想和泛化推理能力,提高了预测的精度和稳定性。而针对发动机台架试验数据的预测结果证明了方法的有效性,这一结果对于轴流压气机内部流动特性研究及稳定性监控具有重要意义。
Based on particle swarm optimization and support vector machines,a forecasting model for compressor casing wall pressure is presented.The strong nonlinear mapping capability of the support vector machines is used to implement nonlinear forecasting of the measured time series of compressor casing wall pressure.Particle swarm optimization is employed to optimize important parameters of support vector machines.The association and generalization capabilities of forecasting model on chaos dynamics are increased,thus the forecasting precision and stability are improved.Forecasting results based on the experimental data has proved the effectiveness of the forecasting model,which may serve as a promising stability monitoring method for axial flow compressors.
出处
《燃气涡轮试验与研究》
2011年第2期20-22,48,共4页
Gas Turbine Experiment and Research
关键词
压气机
粒子群优化
支持向量机
时间序列预测
compressor
particle swarm optimization
support vector machines
time series forecasting
