Method and system for predicting well testing productivity of fractured-vuggy oil and gas reservoir

Abstract

The invention discloses a method for predicting well testing productivity of a fracture-vug type oil and gas reservoir, which comprises the following steps that according to well testing interpretation data of a fracture-vug type oil and gas reservoir and geological data of an area where the reservoir is located, a plurality of characteristic parameters related to the productivity are preferably selected and a corresponding influence degree coefficient is calculated; wherein the characteristic parameters comprise original pressure, shaft skin, reservoir rock permeability, sweep efficiency, damping coefficient, karst cave volume, crack volume and total volume; historical sample data of all kinds of characteristic parameters are preprocessed, on the basis, a prediction model used for predicting the fracture-cavity type oil and gas reservoir well testing productivity is trained in combination with the influence degree coefficient, and the prediction model is a one-dimensional convolutional neural network model; and on the basis of the real-time sample data set of the to-be-predicted well with respect to the plurality of characteristic parameters, well test interpretation productivity prediction is carried out by using the trained prediction model. The method is high in prediction result precision, low in dependence degree on geological information and more flexible in operation compared with a traditional method.

Description

technical field [0001] The invention relates to the technical field of well test interpretation, in particular to a method and system for predicting well test productivity of fracture-cavity oil and gas reservoirs. Background technique [0002] Most of the production capacity prediction of oil and gas fields is to carry out regression analysis and fitting on the existing production data first, and then use the equation obtained by the regression method to predict the production capacity. The traditional analysis process for production data is usually based on material balance conditions and three ideal assumptions, that is, a vertical well, formation boundary closure, and flow reaching the boundary flow stage. In the field of production data analysis technology, the decline curve analysis method was first proposed by Arps. This method uses a simple form of mathematical function to fit the production change curve. The curve is mainly in three forms: exponential, hyperbolic an...

AI Technical Summary

Problems solved by technology

In the field of production data analysis technology, the decline curve analysis method was first proposed by Arps. This method uses a simple form of mathematical function to fit the production change curve. The curve is mainly in three forms: exponential, hyperbolic and harmonic. The Arps method is suitable for steady-state bottomhole flow and can explain the monotonic variation of data, but it is not suitable for complex bottomhole flow, nor can it explain the non-monotonic variation of data

In addition, Fetkovich combined the Arps method with the unsteady flow formula in well test analysis, and extended the Arps method to the unsteady flow stage, but there are errors in the productivity prediction results

[0003] Since the productivity evaluation of oil and gas wells using productivity well test interpretation data is a commonly used method at home and abroad, but because the productivity test requires that the pressure at each flow rate must be stable, in the actual productivity test, especially for carbonate rocks In fracture-cavity reservoirs, the bottom hole pressure is often difficult to stabilize, which will lead to errors in productivity prediction results

Images