Leakage speed probability distribution prediction method and system based on machine learning

A probability distribution and machine learning technology, applied in the field of oil and gas well engineering, can solve problems such as single leak rate value, not very accurate

Pending Publication Date: 2021-07-06
CHINA UNIV OF PETROLEUM (BEIJING)
View PDF5 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

There are also leak rate predictions, but most of the predictions are a single leak rate v

Method used

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
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Leakage speed probability distribution prediction method and system based on machine learning
  • Leakage speed probability distribution prediction method and system based on machine learning
  • Leakage speed probability distribution prediction method and system based on machine learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0033] This embodiment discloses a method for predicting leakage rate probability distribution based on machine learning, such as figure 1 shown, including the following steps:

[0034] S1 performs dimensionality reduction processing on mud logging data.

[0035] The process of dimension reduction processing is as follows: according to the comprehensive logging data and leakage records of multiple wells drilled, the leakage rate is used as a label to analyze the correlation between the comprehensive logging parameters and the leakage rate in the leakage records, and the Pearson correlation analysis algorithm, The random forest algorithm and the recursive elimination feature algorithm screen the comprehensive mud logging parameters, and select the characteristic parameter with the highest correlation with leakage rate as the characteristic parameter of leakage rate prediction.

[0036] S2 normalizes the mud logging data after dimension reduction processing, and uses the normal...

Embodiment 2

[0053] Based on the same inventive concept, this embodiment takes the Mishrif reservoir in the H oilfield in the Middle East as an example to further illustrate the scheme in Embodiment 1:

[0054] H Oilfield is a gentle syncline structure with a dip angle of less than 5°. There are 9 sets of reservoirs developed from top to bottom, mainly sandstone reservoirs and limestone reservoirs.

[0055] The data of 45 wells with known leakage were used to train the mixed density neural network, and then a new well HF-P1 in the wing of H Oilfield was used as the test well, and the trained mixed density neural network model was used to diagnose leakage and predict the probability distribution of leak rate.

[0056] S1 performs dimensionality reduction processing on mud logging data.

[0057] The training data is the mud logging data of 45 wells in the Mishrif reservoir of the H Oilfield in the Middle East. A total of 22 sets of characteristic parameters related to leakage and 1 set of le...

Embodiment 3

[0065] Based on the same inventive concept, this embodiment discloses a leak rate probability distribution prediction system based on machine learning, including:

[0066] Dimensionality reduction module, used for dimensionality reduction processing of mud logging data;

[0067] The model training module is used to normalize the mud logging data processed by dimension reduction, and use the normalized data to train the mixed density neural network model;

[0068] The uncertainty calculation module is used to calculate the uncertainty of the mixed density neural network model, and obtain the variance of the mixed density neural network model;

[0069] The prediction module is used to adjust the leak rate range according to the variance to obtain the probability distribution of the leak rate.

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

PUM

No PUM Login to view more

Abstract

The invention belongs to the field of oil and gas well engineering, and relates to a leakage speed probability distribution prediction method and system based on machine learning, and the method comprises the following steps: S1, carrying out the dimension reduction processing of logging data; s2, normalizing the logging data subjected to dimension reduction processing, and training a mixed density neural network model by adopting the normalized data; s3, calculating the uncertainty of the mixed density neural network model, and obtaining the variance of the mixed density neural network model; and S4, adjusting a leakage speed range according to the variance, and obtaining a leakage speed probability distribution condition. According to the method, the leakage speed range can be accurately predicted, reference is provided for well drilling parameter optimization in combination with the uncertainty of the prediction result, and leakage control is achieved.

Description

technical field [0001] The invention relates to a leakage rate probability distribution prediction method and system based on machine learning, belonging to the field of oil and gas well engineering. Background technique [0002] Lost circulation is a very common accident in drilling engineering. It will not only cause the loss of drilling fluid, but also may cause other downhole complex accidents such as pipe sticking, thereby increasing the drilling cost and drilling cycle. Therefore, the leakage problem is a major problem in drilling engineering. [0003] Previously, many experts and scholars have proposed models and methods for the prediction, diagnosis and control of the leakage in drilling. These models can be mainly divided into two categories, namely empirical models and mathematical models. The establishment method of the empirical model is simple, and it is highly adaptable to specific problems. Through correlation analysis, the binary regression method is used 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

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06Q10/04G06F17/18G06N3/04G06N3/08G06Q50/02
CPCG06Q10/04G06Q50/02G06F17/18G06N3/04G06N3/08
Inventor 庞惠文樊永东金衍王汉青
Owner CHINA UNIV OF PETROLEUM (BEIJING)
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap