COIL operation process monitoring by kernel principal component analysis

A laser output process monitoring method for the chemical oxygen iodine laser (COIL) is proposed based on the kernel 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, the T2 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 simulations and experimental tests.