Logging reservoir identification and prediction method based on machine learning

A technology of reservoir identification and machine learning, applied in the field of machine learning, can solve problems such as poor applicability of logging interpretation models, complex pore structure, and strong reservoir heterogeneity
CN113642772APending Publication Date: 2021-11-12CHONGQING UNIVERSITY OF SCIENCE AND TECHNOLOGY
7 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Current Assignee / Owner
CHONGQING UNIVERSITY OF SCIENCE AND TECHNOLOGY
Publication Date
2021-11-12

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The invention provides a logging reservoir identification and prediction method based on machine learning. The invention aims at solving the problems that when a conventional well logging interpretation method is used for tight gas low-porosity and low-permeability reservoir fluid recognition, due to the fact that the reservoir is high in heterogeneity and the pore structure is complex, a well logging interpretation model is poor in applicability and low in accuracy, and a traditional single machine learning algorithm is prone to falling into the problems of overfitting, local optimum and the like. According to the invention, the classic ensemble learning algorithm xgboost is adopted, the gas testing logging result and the conventional logging data are combined, the tight sandstone reservoir is recognized, the prediction precision can be effectively improved, the time for testing personnel to judge the reservoir is shortened, and therefore a large amount of cost is saved.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The invention belongs to the field of machine learning and relates to a method for identifying and predicting well logging reservoirs based on machine learning. Background technique

[0002] Integrated learning is a branch direction of machine learning. It mainly forms a strong learner by combining multiple weak learners, and adopts an appropriate combination strategy to synthesize the prediction results of each learner to obtain a more comprehensive and reliable judgment, thereby improving The accuracy of the model's predictions. Compared with a single machine learning algorithm, it is not easy to fall into problems such as overfitting and local optimum. Among them, xgboost is a more classic algorithm for integrated learning. xgboost is based on the idea of ​​GBDT, using first-order and second-order derivative information. It can also be said that it is a kind of GBDT because it is also based on Gradient and Boosting ideas, but The difference from t...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More