Multimode Wind Tunnel Flow Field System Monitoring Based on KPLS

In a wind tunnel process, Mach number is the most important parameter. However, it difficult to measure directly, especially in multimode operation leading difficulty process monitoring. Thus, necessary indirectly by utilizing data-driven methods, and based on which, monitor status of process. this paper, therefore, flow field system monitoring strategy proposed. Since strongly nonlinear system, kernel partial least squares method, which can efficiently handle regression problem, utilized. Firstly, predicted method. Secondly, statistics, i.e., Hotelling T2 statistic square prediction error, SPE statistic, their control limits, are proposed be applied basis number. Finally, real where mode analysis division necessary. After division, single-mode processes modeled predicted, respectively, both monitored online. Satisfactory results were achieved compared with those