科研之友
微信
新浪微博
Facebook
分享链接
摘要
In view of the problem that the portable X-ray fluorescence (pxrf) method is not accurate enough for the determination of heavy metals in soil, a general modeling method and process based on the relationship between the measured value of Pxrf and the characteristics of the site parameters are proposed, which is used to build a Pxrf correction model suitable for the characteristics of each site, so as to improve the measurement accuracy of Pxrf. Taking a typical mining area in Beijing as an example, By analyzing the correlation of potentially related factors (i. e., soil organic matter content, water content, and actual pollutant content) and different mathematical transformations of these factors (i. e., index, logarithm, and power function) with the measured values of Pxrf, we identified the influencing variables, influence degrees, and influence forms that significantly affect the measured values of Pxrf. Based on the identification of influencing factors, we established the mathematical model of Pxrf to achieve the purpose of correcting the measured value of pxrf and improving the accurate value. The results showed that the measured values of Pb, Cu and Zn were significantly correlated with the actual concentrations, and the correlation coefficients were 0.75~0.82; the correlation coefficients of Sr, Fe and Mn were 0.25~0.52, with slightly poor correlation. The influence of water content on the detection of Pxrf was in order of Sr, Mn, Cu, Pb, Zn, the correlation coefficient range was 0.1~0.7. The Pxrf of all heavy metals is weakly affected by the content of soil organic matter, and the correlation coefficient was lower than 0.5. After the model transformation, the correlation between the Pxrf of heavy metals Zn, Pb, Cu and the actual concentration increases (increased by 4%~13%); other heavy metals had no significant change. The model verification results showed that the correlation between the Pxrf of four heavy metals Pb, Zn, Cu, Mn Compared with the original Pxrf measurement, the relative error of the corrected Pxrf value in the model could be reduced from 32.4%~56.6% to 4%~29.9%.
- 单位
