| Summary: | Qualitative and quantitative analysis on emission spectrographic detector
were important to identify impurities elements in nuclear fuel. In Indonesia, the
process is still using a densitometer based manual method which has limitations
due mechanical factors on instrument and expert chemists requirements. To
reduce measurement error on manual methods, we develops analytical methods by
processing scanned digital image of a spectrogram.
The aim of this research is to identify impurities on nuclear fuel elements
(boron and cadmium in uranium oxide). Image processing was done by cropping
digital image to separate each the spectra result of excitation processes and
identify impurities, which enlarged later to refine the distribution of signal
intensity of pixels. Signal processing methods for quantitative analysis conducted
with pulse peaks analysis and pulse wide analysis. Results obtained then
substituted into a standard curve obtained from regression analysis of sample
reference to calculate concentration of the element.
The results show that image processing can be used to identify and analyze
the elements in the detector, element concentrations value of image processing is
not different from the results of analysis of densitometer on statistically
significance level of 5%. The relative deviation value of boron analysis based on
extensive measurements of height and wide of the curve is 2,267% and 0,886%.
While the relative deviation value in the cadmium analysis is 1,718% and
17,119%.
|
|---|