关键词: XRD, 储层, 物性评价, 多层感知器, 神经网络

Abstract: Rapid and accurate evaluation of reservoir physical properties is the key to oil test and fracturing stage selection. In order to study the distribution interval of high-quality reservoir in Shangwuerhe Formation of Permian in Fukang Sag, Jundong, the application of multi-layer perceptron method in the evaluation of reservoir physical properties with XRD logging has been explored. By collecting a large amount of XRD logging data, five characteristic minerals are selected to construct a multi-layer perceptron neural network model. Through verification, it is found that the established model can better fit the relationship between reservoir physical property parameters and mineral content. The trained model achieves high prediction accuracy on the test data set, which provides a new way to evaluate reservoir physical properties quickly and accurately. The application of wells KT 5 and F 47 has achieved a good evaluation effect, which has a certain popularization and application value.

Key words: XRD, reservoir, physical property evaluation, multi-layer perceptron, neural network