Liu, Zihuai and Zhang, Linli and Li, Yonggang and Luo, Qiwu and Yang, Chunhua and Wang, Yibo and Gui, Weihua (2023) Multiple Harmonic Information Fusion for Residual Oxygen Concentration Analysis of Pharmaceutical Glass Vials. IEEE Transactions on Instrumentation and Measurement, 72: 7005811. pp. 1-11. ISSN 0018-9456
Full text not available from this repository.Abstract
Wavelength modulation-based tunable diode laser absorption spectroscopy (TDLAS/WMS) is widely used in various manufacturing processes. On the aseptic preparation filling production line, the seal quality of pharmaceutical glass vials should be strictly inspected in situ, where the second harmonic signal is usually selected for headspace residual oxygen measurement due to its prominent amplitude natures, while higher order harmonics still contain reliable concentration information but are often ignored. In this article, we first design a laser injection signal with dynamic modulation depth and then propose a new methodology of multiple harmonic information fusion (MHIF), which is simple but promotes the mining of more effective concentration information from the absorption spectrum. Based on the fusion analysis results, an MHIF-guidance gas concentration inversion model is designed. To verify the performance of the proposed inversion model, a TDLAS/WMS-based residual oxygen detection platform is built. In this system, we use the harmonic signals from the first to eighth cycles together to perform precise measurement of the oxygen concentration in pharmaceutical glass vials. The results show that the stability of the new MHIF-based feature can exceed that of the traditional second harmonic features by 24.7%. Moreover, the new inversion model can increase the detection accuracy by over 10% compared to the traditional inversion model.