Multi-element wave peak-based laser-induced breakdown spectroscopy rock fragment type identification method

Abstract

The invention provides a rock fragment type identification method. The method comprises the following steps: 1, measuring the spectra of N standard rock fragments in a range of 200-950nm by using a laser-induced breakdown spectrograph; 2, carrying out spectral line identification on the spectra of the N standard rock fragments, and finding out the wave peak spectral lines of a preset number of elements with the top contents; 3, combining the wave peak spectral lines of the selected elements in the standard rock fragments to form a characteristic spectrum as the input value of a BP nerve network, and training the BP nerve network to obtain a BP nerve network structure; 4, measuring the spectrum of a rock fragment to be identified in the range of 200-950nm by using the laser-induced breakdown spectrograph; 5, carrying out spectral line identification on the spectrum of the rock fragment to be identified to obtain spectral lines of elements same to the elements in step 2; and 6, combining the spectral line wave peaks of the selected elements in the rock fragment to be identified as the input value of the BP nerve network, and identifying the rock fragment to be identified by using the obtained BP nerve network to obtain an identification result.

Description

technical field [0001] The invention relates to the technical field of cuttings type identification, in particular to a method for identifying cuttings types based on multi-element peaks in laser-induced breakdown spectroscopy. Background technique [0002] In recent years, with the rapid development of new oil drilling technology, the cuttings returned from the bottom of the well have been very finely divided, even powdered, which makes the traditional description of cuttings very difficult. [0003] At present, the methods of laser-induced breakdown spectroscopy applied to rock type identification mainly include: The first method is to use laser-induced breakdown spectroscopy combined with partial least squares discriminant analysis to realize automatic identification of rock samples. The experimental model adopted is mainly There are two types: one is the full-spectrum model, which has relatively high recognition results, but requires more data to be processed, and has a ...

AI Technical Summary

Problems solved by technology

[0002] In recent years, with the rapid development of new oil drilling techniques, the cuttings returned from the bottom of the well have been very finely divided, even in powder form, making the traditional description of cuttings very difficult

[0003] At present, the methods of laser-induced breakdown spectroscopy applied to rock type identification mainly include: The first method is to use laser-induced breakdown spectroscopy combined with partial least squares discriminant analysis to realize automatic identification of rock samples. The experimental model adopted is mainly There are two types: one is the full-spectrum model, the recognition result of this method is relatively high, but more data needs to be processed, and the background noise has a greater impact

The second method is the peak intensity and ratio model (feature model). This method is relatively simple in data processing, but the recognition accuracy rate has declined.

Among these methods, although some recognition effects are better, data processing is time-consuming, while other methods have poor recognition effects.

Images