Permeability calculation method and device based on rock physics constraint neural network

Abstract

The embodiment of the invention discloses a permeability calculation method and device based on a rock physics constraint neural network, and the method comprises the steps: extracting corresponding depth point logging parameters based on a rock core sample, and carrying out the preprocessing of the logging parameters and corresponding rock core data, so as to construct a first sample set and a second sample set; constructing and training according to the first sample set to obtain a first neural network; constructing and training according to the second sample set to obtain a second constraint neural network, wherein the second constraint neural network is composed of an input layer, a constraint layer, a hidden layer and an output layer, and the constraint layer is a network layer designated to add constraint elements, the constraint element is obtained through calculation according to an output result of the first neural network, is input to a constraint layer of the second constraint neural network and restrains calculation of the second constraint neural network; acquiring logging parameters of a to-be-predicted well, and respectively inputting the logging parameters into the first neural network and the second constraint neural network to obtain the permeability of the to-be-predicted well. The method improves the prediction accuracy of the permeability of the whole well section.

Description

technical field [0001] Embodiments of the present invention relate to the technical field of data processing in petroleum exploration, and in particular to a method and device for calculating permeability based on a petrophysically constrained neural network. Background technique [0002] Permeability is a measure of the ability of rock to allow fluid to pass through under differential pressure. It is an extremely important parameter in the exploration and development of oil and gas fields. By establishing a mathematical model of permeability and other petrophysical parameters and using the mathematical model to calculate permeability, it is currently the only available The whole well section permeability evaluation method. The permeability mathematical model can be roughly divided into two types. One is the classic porosity-permeability formula, such as the Timur formula, which has low calculation accuracy due to regional applicability and other reasons; the other is the pe...

AI Technical Summary

Problems solved by technology

Due to the high cost of formation test method and core test measurement method and the limited depth range of measurement, this method is limited to the calibration of well logging permeability calculation

At present, the most commonly used method is still to use conventional logging data to establish a formula or model, and apply the method of experimental analysis and calibration to calculate the permeability. Due to the empirical nature of the formula or model, the influence of human factors, regional differences, and complicated and cumbersome model establishment steps, it is often difficult to As a result, the final calculation model cannot be widely used or the calculation accuracy is insufficient, and some methods of using machine learning algorithms to predict reservoir permeability require more samples of core permeability, and because it is only a simple and blunt application of machine learning algorithms for model training Prediction does not take into account the influence of core distribution, so this method is not effective in permeability calculation, and the calculation accuracy is generally low

Images