基于KPCA的氧碘激光器出光过程监控方法

COIL operation process monitoring by kernel principal component analysis

根据氧碘化学激光器运行过程基本稳定的特点,利用核主元分析(KPCA)方法监控非线性过程的优势,提出了基于KPCA的激光器出光过程监控方法。该方法通过沿时间轴方向展开激光器正常出光的历史数据,建立了激光器出光过程中各时刻的KPCA模型,并在特征空间构建T2统计量和平方预测误差统计量对激光器出光过程进行监控。仿真计算和试验表明,该方法具有可靠的监控性能,能及时、准确地发现激光器出光过程中的异常状态。

A laser output process monitoring method for the chemical oxygen iodine laser （COIL） is proposed based on thekernel principal component analysis （KPCA）, which is well suit for the analysis of stable non-linear process. Based on the expansion of normal laser output data along the time axis, temporal KPCA models are established using this method. Furthermore, theT2 and SPE parameters are established in eigenspace to monitor laser output process. The feasibility of the monitoring method,which can detects various abnormal states of the laser output process promptly and accurately, is verified by theoretical simula-tions and experimental tests.