Research On Optimization And Monitoring Of Parameters For New Atomic Fluorescence Spectrometer

Atomic fluorescence spectrometer is an instrument that obtains the elemental content by measuring the fluorescence intensity of the measured element.It is mainly used for trace and ultra-trace detection.In order to solve the problem of scattering interference and spectral interference in traditional non-dispersive atomic fluorescence spectrometer,our team developed a new atomic fluorescence spectrometer for dispersion-non-dispersion,and hope to reduce the spectral interference and scattering interference of atomic fluorescence spectrometer by dispersion detection technology.In this paper,the parameters optimization of the new atomic fluorescence spectrometer is carried out.Taking the standard solution of Bi element as an example,the instrument operating parameters when using the dispersion detection system for sample determination are optimized.These instrument parameters include carrier gas flow rate,shielding gas flow rate,atomizer height,acid concentration,reducing agent concentration,and so on.In this paper,a flame detection system is built to detect the flame in the atomizer.A flame detection software was developed using Qt,OpenCV and C++.This software could control the camera to capture the flame image,then display the high-definition image of the flame in real time on the computer,also,it could store the images and videos of the flame.In this paper,the monitoring of the working parameters of the new atomic fluorescence spectrometer is carried out.It was found in experiments that some operating parameters of the instrument,such as carrier gas flow rate,shielding gas flow rate,atomizer height,acid concentration,and reducing agent concentration,are not only affecting the fluorescence intensity,but also affect the state of the flame in the atomizer.The flame images can be used to monitor the abnormality of some working parameters of the instrument.Using the flame images collected by the flame detection system,the convolutional neural network algorithm in deep learning is used to build the model of the relationship between the flame images and the corresponding instrument parameters.The model built can determine whether the parameters are abnormal through the flame images.