Research On Key Technologies Of EDXRF Spectrometer And Spectral Analysis Software Design

With the increase of the export of domestic electrical and electronic products,the detection of hazardous substances in such products has become prominent.And the demand of testing substances in daily chemical industry,environment protection,food safety and other industries is also increasing.Energy Dispersive X-ray Fluorescence spectrometer is an ideal instrument to detect the hazardous substances in the listed fields,which has been favored by all walks of life.With XRF spectral analysis technology becoming relatively mature today,the inspection ability of the spectrometer mainly relies on the improvement and innovation of the spectral analysis algorithmThis paper firstly studies the spectral preprocessing algorithms,including filtering and background subtraction.For spectral filtering,based on the wavelet multi-resolution analysis and combined with the characteristics of soft and hard thresholds,a weighting factor is designed and the optimal wavelet tree filtering method is devised.To solve the problem of time-consuming of traditional background subtraction algorithms with wavelet analysis,a double-threshold rule is designed to less the iterative times.Besides,quarter coefficient method is also proposed to determine the threshold.In qualitative analysis,the threshold parameter peak search algorithm and the overlapped peak calibration algorithm are separately studied for the isolated and overlapped spectral segments.The former is the combination of derivative method and Gaussian fitting peak search method,based on the principle of adjusting the threshold parameters for rough peak finding.For the latter,based on the characteristics of the Gaussian function,the fourth-order derivative-based overlapped peak correction algorithm is studied,and the calculation method of the ratio of the peak-to-peak widths is given in the thesis.The extraction of fluorescence intensity is the most important step in quantification analysis.Gaussian peak fitting method and trapezoidal area method are discussed in this paper.The latter is a reasonable simplification on the basis of the former,and single peak fluorescence intensity calculation method in overlapped peak is introduced as well.A hybrid programming method of C # and MATLAB combined with multi-threading technology is adopted on the.NET Framework platform for PC analysis software development.Through the joint debugging of hardware and software,this paper takes the determination of cadmium in rice flour as an example to verify and evaluate the algorithms.The results show that the relative error of qualitative is less than 1% and the relative error of quantitative determination of cadmium is less than 10%.In addition,the performance indicators such as the detection limit,the limit of quantification and the repeatability of the cadmium element are also quantified,which shows that the whole spectrometer system can accurately measure the cadmium element and has the advantages of easy operation,high accuracy and high real-time response.