Multi-dimensional intelligent crude oil property rapid identification method and device
By using a multi-dimensional intelligent identification method and a deep learning model to automatically identify crude oil properties, the problem of high time consumption and cost of traditional methods is solved, and rapid and accurate crude oil property identification is achieved, improving the interpretation accuracy and work efficiency.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CNPC BOHAI DRILLING ENG
- Filing Date
- 2024-12-03
- Publication Date
- 2026-06-05
AI Technical Summary
Traditional methods for identifying crude oil properties are time-consuming and costly, rely on the experience of geologists, and cannot quickly and accurately identify crude oil properties.
A multi-dimensional intelligent identification method is adopted, which combines gas measurement parameters and crude oil density through a deep learning model. The deep learning algorithm is used to automatically identify the properties of crude oil, including data collection, preprocessing, feature extraction and model training, and output the identification results.
Under different crude oil physical properties, the on-site interpretation compliance rate exceeded 60%, the comprehensive interpretation and evaluation compliance rate exceeded 85%, and the work efficiency was improved by 30%.
Smart Images
Figure CN122157845A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of petroleum exploration technology, specifically a multi-dimensional intelligent method and device for rapid identification of crude oil properties. Background Technology
[0002] Currently, traditional methods for identifying crude oil properties rely on the experience of geologists and complex laboratory tests, which are not only time-consuming but also costly. With the development of deep learning technology, it has become possible to automatically identify crude oil properties using machine learning models. Summary of the Invention
[0003] To address the shortcomings of existing technologies, this invention aims to provide a multi-dimensional intelligent crude oil property rapid identification method and device, so as to achieve the purpose of rapidly identifying crude oil properties under different crude oil property conditions.
[0004] To achieve the above objectives, the technical solution adopted by this invention is as follows: A multi-dimensional intelligent crude oil property rapid identification method, comprising the following steps:
[0005] S1. Well data analysis to obtain crude oil density;
[0006] S2. Determine the relationship between gas measurement parameters and crude oil density;
[0007] S3. Filter gas measurement parameters;
[0008] S4. Use Equation 1 to perform calculations to obtain the properties of crude oil;
[0009]
[0010] In the formula, i is the parameter used, j and K are the number of operations, and H is the exponent.
[0011] As a limitation of the present invention, S1. Well data analysis to obtain crude oil density through years of oil testing verification;
[0012] S2. Statistically analyze the gas logging parameters and crude oil density of the oil layer in the test wells over the years, and plot data points using crude oil density and different gas logging parameters to determine the relationship between gas logging parameters and crude oil density.
[0013] As a further limitation of the present invention, in step S2, the relationship between the gas measurement parameters and the crude oil density is as follows:
[0014] Crude oil density is less than 0.77 g / cm³ 3 At that time, the humidity ratio of the gas measured was less than 10, the equilibrium ratio was greater than 30, the hydrocarbon slope was greater than 2, and the characteristic ratio was greater than 1.0.
[0015] Crude oil density greater than 0.89 g / cm³ 3At that time, the humidity ratio of the gas measured was less than 10, the equilibrium ratio was greater than 10, the hydrocarbon slope was greater than 2, and the characteristic ratio was less than 0.6.
[0016] Crude oil density is between 0.8 and 0.89 g / cm³. 3 At that time, the humidity ratio of gas measurements was concentrated in the range of 10-30, the balance ratio in the range of 5-20, the hydrocarbon slope in the range of 1-2, and the characteristic ratio in the range of 0.4-0.8.
[0017] This invention also provides a multi-dimensional intelligent crude oil property rapid identification device, the technical solution of which is as follows:
[0018] This method is used to implement a multi-dimensional intelligent rapid identification method for crude oil properties, and it includes the following modules:
[0019] The data collection module collects data generated during the logging process, including crude oil density, peak value, and non-hydrocarbon gas content.
[0020] The data preprocessing module standardizes the collected data to eliminate differences between different wells;
[0021] The feature extraction module uses deep learning algorithms to extract features from the preprocessed data;
[0022] The model training module uses the extracted features to train the deep learning model;
[0023] The crude oil property identification module applies the trained model to new logging data to automatically identify and classify crude oil properties.
[0024] The results output module outputs the recognition results in the form of graphics or text.
[0025] By adopting the above-described technical solution, the beneficial effects achieved by this invention compared to the prior art are as follows:
[0026] This invention rapidly identifies crude oil properties using parameters such as density, peak value, and non-hydrocarbon gas content in well logging technology under different crude oil physical property conditions. The on-site interpretation accuracy rate exceeds 60%, the comprehensive interpretation and evaluation accuracy rate exceeds 85%, and the efficiency of comprehensive interpretation and evaluation work is improved by 30%. Attached Figure Description
[0027] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.
[0028] Figure 1 This is a well logging composite diagram of well 1 in Embodiment 1 of the present invention;
[0029] Figure 2 This is a combined logging diagram of two wells from Embodiment 1 of the present invention;
[0030] Figure 3This is a comprehensive logging diagram of three wells from Embodiment 1 of the present invention;
[0031] Figure 4 This is a graph showing the relationship between crude oil density and gas humidity values in Example 1 of the present invention.
[0032] Figure 5 This is a graph showing the relationship between crude oil density and gas balance ratio in Example 1 of the present invention;
[0033] Figure 6 This is a graph showing the relationship between crude oil density and hydrocarbon slope in Example 1 of the present invention.
[0034] Figure 7 This is a graph showing the relationship between crude oil density and gas measurement characteristic ratio in Example 1 of the present invention;
[0035] Figure 8 This is an overall flowchart of Embodiment 2 of the present invention. Detailed Implementation
[0036] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustrative and understanding purposes only and are not intended to limit the scope of the invention.
[0037] Example 1: A Multidimensional Intelligent Crude Oil Property Rapid Identification Method
[0038] S1. Well data analysis to obtain crude oil density
[0039] This involves analyzing well data on crude oil density obtained through years of oil testing. The analysis reveals significant differences in gas logging parameters for different crude oil densities. (Reference) Figure 1-3 As shown, the density of the condensate oil crude oil is 0.67 g / cm³. 3 Light crude oil has a density of 0.83 g / cm³. 3 The density of heavy crude oil is 0.93 g / cm³. 3 The density of medium-quality crude oil is 0.83-0.93 g / cm³. 3 .
[0040] S2. Determine the relationship between gas measurement parameters and crude oil density.
[0041] Historical data were collected on gas logging parameters and crude oil density of oil layers in oil testing wells. Data points were plotted using crude oil density and different gas logging parameters, and referenced. Figure 4-7 The relationship between gas measurement parameters and crude oil density was determined as follows:
[0042] Crude oil density is less than 0.77 g / cm³ 3 At that time, the humidity ratio of the gas measured was less than 10, the equilibrium ratio was greater than 30, the hydrocarbon slope was greater than 2, and the characteristic ratio was greater than 1.0.
[0043] Crude oil density greater than 0.89 g / cm³ 3 At that time, the humidity ratio of the gas measured was less than 10, the equilibrium ratio was greater than 10, the hydrocarbon slope was greater than 2, and the characteristic ratio was less than 0.6.
[0044] Crude oil density is between 0.8 and 0.89 g / cm³. 3 At that time, the humidity ratio of gas measurements was concentrated in the range of 10-30, the balance ratio in the range of 5-20, the hydrocarbon slope in the range of 1-2, and the characteristic ratio in the range of 0.4-0.8.
[0045] S3. Filtering gas measurement parameters
[0046] By analyzing the relationship between the gas measurement parameters and crude oil density in step S2, gas measurement parameters with strong data patterns are selected. As crude oil density increases, the gas humidity parameter gradually increases; when the crude oil density exceeds 0.89 g / cm³... 3 The humidity parameter gradually decreases. The characteristic ratio parameter gradually decreases as the crude oil density gradually increases. The equilibrium ratio parameter gradually decreases as the crude oil density gradually increases, except when the crude oil density is greater than 0.89 g / cm³. 3 As time progresses, this parameter gradually increases. The density parameters with strong correlations are peak value, humidity ratio, hydrocarbon slope, characteristic ratio, and equilibrium ratio.
[0047] S4. Using the deep learning algorithm in Equation 1, perform sensitive parameter calculations to obtain the properties of crude oil.
[0048]
[0049] In the formula, i is the parameter used, j and K are the number of operations, and H is the exponent.
[0050] This embodiment can accurately characterize logging technical parameters under different crude oil physical property conditions based on the data range of the P[i] parameter.
[0051] Crude oil properties in conclusion P[i] Explanation accuracy rate Condensate oil oil layer ≥0.9557 98.387％ Light oil oil layer 0.4957-0.9999 82.485％ medium oil oil layer 0.5092-0.9903 92.523％ Heavy oil oil layer ≥0.7785 99.000％
[0052] Example 2: A Multidimensional Intelligent Crude Oil Property Rapid Identification Device
[0053] This embodiment is used to implement a multi-dimensional intelligent crude oil property rapid identification method, which includes the following modules:
[0054] The data collection module collects data generated during the logging process, including parameters such as crude oil density, peak value, and non-hydrocarbon gas content.
[0055] The data preprocessing module standardizes the collected data using the Z-score standardization method to ensure that the data are on the same scale, thereby eliminating differences between different wells and improving the model's generalization ability.
[0056] The feature extraction module uses deep learning algorithms, such as convolutional neural networks (CNN), to automatically extract features from the preprocessed data, that is, to extract the gas measurement parameters selected in step S3. Feature extraction uses an autoencoder structure to learn the inherent representation of the data in an unsupervised manner.
[0057] The model training module uses the extracted features to train a deep learning model, performing calculations as per step S4. The learning model can be a multilayer perceptron (MLP), a recurrent neural network (RNN), or other network structures suitable for classification tasks. The training strategy employs cross-validation and early stopping techniques to prevent overfitting. The model training process is as follows: First, the model is trained using historical logging data; then, the model parameters are iteratively optimized until the data training accuracy reaches 100%.
[0058] The crude oil property identification module applies a trained model to new logging data to automatically identify and classify crude oil properties. The identification process involves inputting new logging data into the model, which then calculates and outputs the crude oil properties through forward propagation.
[0059] The results output module outputs the recognition results in the form of graphics or text for practical applications and further analysis.
[0060] Application data is real-time data that has been preprocessed in the data preprocessing module and then processed according to the classification criteria to obtain the results, which can be directly applied in practice.
[0061] It should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art can still modify the technical solutions described in the above embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A multi-dimensional intelligent method for rapid identification of crude oil properties, characterized in that... Includes the following steps: S1. Well data analysis to obtain crude oil density; S2. Determine the relationship between gas measurement parameters and crude oil density; S3. Filter gas measurement parameters; S4. Use Equation 1 to perform calculations to obtain the properties of crude oil; In the formula, i is the parameter used, j and K are the number of operations, and H is the exponent.
2. The multidimensional intelligent crude oil property rapid identification method according to claim 1, characterized in that: S1. Analysis of well data on crude oil density obtained through years of oil testing verification; S2. Statistically analyze the gas logging parameters and crude oil density of the oil layer in the test wells over the years, and plot data points using crude oil density and different gas logging parameters to determine the relationship between gas logging parameters and crude oil density.
3. The multidimensional intelligent crude oil property rapid identification method according to claim 2, characterized in that: In step S2, the relationship between gas measurement parameters and crude oil density is as follows: Crude oil density is less than 0.77 g / cm³ 3 At that time, the humidity ratio of the gas measured was less than 10, the equilibrium ratio was greater than 30, the hydrocarbon slope was greater than 2, and the characteristic ratio was greater than 1.
0. Crude oil density greater than 0.89 g / cm³ 3 At that time, the humidity ratio of the gas measured was less than 10, the equilibrium ratio was greater than 10, the hydrocarbon slope was greater than 2, and the characteristic ratio was less than 0.
6. Crude oil density is between 0.8 and 0.89 g / cm³. 3 At that time, the humidity ratio of gas measurements was concentrated in the range of 10-30, the balance ratio in the range of 5-20, the hydrocarbon slope in the range of 1-2, and the characteristic ratio in the range of 0.4-0.
8.
4. A multi-dimensional intelligent crude oil property rapid identification device, characterized in that: The method for rapid identification of multidimensional intelligent crude oil properties according to any one of claims 1-3 includes the following modules: The data collection module collects data generated during the logging process, including crude oil density, peak value, and non-hydrocarbon gas content. The data preprocessing module standardizes the collected data to eliminate differences between different wells; The feature extraction module uses deep learning algorithms to extract features from the preprocessed data; The model training module uses the extracted features to train the deep learning model. The crude oil property identification module applies the trained model to new logging data to automatically identify and classify crude oil properties. The results output module outputs the recognition results in the form of graphics or text.