Basic Research On The Application Of Rock Forming Element Logging In Resource Exploration

The detector size of neutron gamma instrument for rock forming element is small,and the measurement accuracy of gamma spectrum is low.Based on the conventional spectral analysis method,the calculation error of element dry weight is large.In this paper,the optimized spectrally weighted least squares method is proposed.This method combines the full-spectrum weighted least squares method and the spectrally weighted least squares method.The full-spectrum weighted least-squares method is used to screen the formation elements.According to the combination of formation elements,the spectrally weighted least square method of specific energy channel is used to resolve the spectrum,and the spectral analysis effect is obviously better than the conventional spectral analysis method.In addition,this paper studies the data processing method of element logging based on deep learning.Based on the relationship between the element mixed spectrum,the element dry weight and the formation minerals,A large number of samples are generated which take the mixed spectrum as input data and the element dry weight as output data,as well as the element dry weight as input data and mineral content as output data.The learning model is established through a deep neural network,and the element dry weight and mineral content are obtained based on the learning model.The research results show that the element dry weight calculated by the optimized spectral weighted least squares method is most accurate than that calculated by the full spectrum weighted least squares and deep learning method.Both matrix transformation method and deep learning method can realize the transformation from elements to minerals accurately.The element measurement accuracy of the rock forming element logging meets the requirements of the mineral industry and has an indicative role in the mineral exploitation.