Accurate and stable measurement of ash in coal by X-ray fluorescence spectrometry based on partial least squares

Abstract

Coal resources play a very important role in the energy structure, especially in the field of thermal power generation. However, in thermal power generation, coal resources can’t be used efficiently for various reasons, resulting in a waste of energy. Due to the advantages of many kinds of elements and rapid measurement, X-fluorescence spectroscopy (XRF) can rapid measurement a variety of metal elements in coal. Combined with partial least squares (PLS), the relationship between a variety of elements and ash in coal can be established, and the ash value in the measured coal can be quickly obtained, which can effectively guide the coal combustion process of power plant, so as to improve the utilization rate of coal. The experimental results show that under the PLS model, XRF technology has good measurement results for the ash content of 45 calibration set coal samples, in which the determination coefficient of the fitting curve (R2) reaches 0.946, and the root mean square error (RMSE) is only 1.177%. The accuracy of the model is further predicted under a cross validation. 15 validation set coal samples are brought into the model for prediction. According to the prediction results, the R2 reaches 0.982, and RMSE is only 0.726%. Finally, four of the 60 samples are randomly selected for stability test, and the relative standard deviation (RSD) is less than 1%. Therefore, the quantitative analysis method of XRF method based on PLS model has high accuracy and stability, which can provide guidance for rapid online measurement of coal ash and coal blending.