Method and system for measuring composition and property of formation fluid
Provided are a method and system for measuring a composition and property of a formation fluid. The method includes: acquiring a measuring model used for measuring a composition and a property of a formation fluid; using a signal measured by a sensor on a downhole hydrocarbon formation tester in real-time as input data and inputting the input data into the measuring model; processing the input data by the measuring model; and directly outputting a processing result as data on the composition and the property of the real-time formation fluid, or parsing the data on the composition and the property of the real-time formation fluid according to the processing result.
1. A method for measuring composition and property of formation fluid, comprising:
acquiring a measuring model used for measuring a composition and a property of a formation fluid;
using a signal measured by a sensor on a downhole hydrocarbon formation tester in real-time as input data and inputting the input data into the measuring model;
processing the input data by the measuring model; and
directly outputting a processing result as data on the composition and the property of the real-time formation fluid, or parsing the data on the composition and the property of the real-time formation fluid according to the processing result;
wherein acquiring the measuring model used for measuring the composition and the property of the formation fluid comprises retrieving a pre-created and pre-trained measuring model, or creating and training the measuring model in real-time.
2. The method for measuring composition and property of formation fluid according to claim 1 , wherein before using the signal measured by the sensor on the downhole hydrocarbon formation tester in real-time as the input data of the measuring model and inputting the input data into the measuring model, the method further comprises:
loading the pre-trained measuring model into the downhole hydrocarbon formation tester to input the signal measured by the sensor on the downhole hydrocarbon formation tester in real-time into the measuring model during a real-time logging process of the downhole hydrocarbon formation tester.
3. The method for measuring composition and property of formation fluid according to claim 2 , wherein creating and training the measuring model comprises:
establishing a database about compositions and properties of various reservoir fluids and measurement signals of downhole sensors;
extracting a sample data set from the database, and training pre-created machine learning models by the sample data set; and
acquiring an optimal machine learning model from trained machine learning models as the measuring model.
4. The method for measuring composition and property of formation fluid according to claim 3 , wherein establishing the database about the compositions and the properties of the various reservoir fluids and the measurement signals of the downhole sensors comprises:
acquiring single-phase reservoir fluid samples meeting a preset requirement in one or more of following manners: surface sampling and downhole sampling; the single-phase reservoir fluid samples comprise oil phase reservoir fluid sample, gas phase reservoir fluid sample and water phase reservoir fluid sample;
performing single-stage flash on the single-phase reservoir fluid sample at standard atmospheric pressure and room temperature to separate equilibrium flashed gas and liquid;
performing gas chromatography analysis on the flashed gas and the flashed liquid respectively to obtain a composition and a first property of the single-phase reservoir fluid sample by use of mass balance calculations; performing a fluid pressure volume temperature (PVT) test on the single-phase reservoir fluid sample under a first preset pressure and a first preset temperature to obtain a second property of the single-phase reservoir fluid sample; wherein the first preset pressure is greater than the standard atmospheric pressure and the first preset temperature is greater than the room temperature; and
adding data related to the composition, the first property and the second property of the single-phase reservoir fluid sample as a part of big data of the formation fluid into the database.
5. The method for measuring composition and property of formation fluid according to claim 2 ,
wherein the first property comprises any one or more of the following: gas-oil ratio, American Petroleum Institute (API) gravity, molecular weight, sulfur content, carbon content, hydrogen content, Watson K value, SARA content and paraffin content; wherein SARA refers to saturated hydrocarbon, aromatic hydrocarbon, resin and asphaltene; and
the second property comprises any one or more of the following: bubble point, dew point, constant composition expansion (CCE) test characteristics, differential liberation (DL) test characteristics, constant volume depletion (CVD) test characteristics, separator test characteristics, density, viscosity, conductivity, compressibility coefficient, formation volume factor, paraffin wax formation condition and asphaltene onset condition.
6. The method for measuring composition and property of formation fluid according to claim 5 , wherein the measuring model is obtained by training pre-created machine learning models based on big data about compositions and properties of various reservoir fluids and measurement signals of downhole sensors as sample data sets.
7. The method for measuring composition and property of formation fluid according to claim 4 , further comprising:
performing of any one or more of the following:
measuring volume, temperature and pressure of the flashed gas when the gas chromatography analysis is performed on the flashed gas and adding the volume, the temperature and the pressure of the flashed gas into the database;
measuring volume, temperature, pressure, density and molecular weight of the flashed liquid when the gas chromatography analysis is performed on the flashed liquid and adding the volume, the temperature, the pressure, the density and the molecular weight of the flashed liquid into the database;
performing any one or more of the following measurements on the single-phase reservoir fluid sample under the first preset pressure and the first preset temperature: continuous near-infrared spectrum, nuclear magnetic resonance (NMR), acoustic wave, fluorescence and dielectric constant measurements, and results of the measurements are added into the database.
8. The method for measuring composition and property of formation fluid according to claim 7 , wherein the measuring model is obtained by training pre-created machine learning models based on big data about compositions and properties of various reservoir fluids and measurement signals of downhole sensors as sample data sets.
9. The method for measuring composition and property of formation fluid according to claim 4 , further comprising adding different drilling mud filtrates into different single-phase reservoir fluid samples, and performing the PVT test on the single-phase reservoir fluid samples in which corresponding drilling mud filtrates are added.
10. The method for measuring composition and property of formation fluid according to claim 4 wherein the measuring model is obtained by training pre-created machine learning models based on big data about compositions and properties of various reservoir fluids and measurement signals of downhole sensors as sample data sets.
11. The method for measuring composition and property of formation fluid according to claim 3 , wherein extracting the sample data set from the database, and training the pre-created machine learning models by the sample data set comprises:
preprocessing and standardizing the sample data set, wherein the preprocessing comprises any one or more of the following: denoising, outlier removal and smoothing;
dividing the preprocessed and standardized sample data set into a first data set and a second data set;
training each of the machine learning models using a preset machine learning method based on the first data set and the second data set, and obtaining an optimal trained machine learning model among all the trained models as the measuring model;
wherein the first data set serves as an input data vector/matrix of the machine learning model, and the second data set serves as a target data vector/matrix of an output data vector/matrix of the machine learning model.
12. The method for measuring composition and property of formation fluid according to claim 11 , wherein the preset machine learning method comprises a supervised machine learning method;
training each of the machine learning models using the preset machine learning method based on the first data set and the second data set, and obtaining the optimal trained machine learning model among all the trained models as the measuring model comprises:
performing following operations respectively on each function in a predefined function set: inputting a standardized first data set into each untrained machine learning model; calculating the output data vector/matrix according to the first data set and a function currently loaded in the machine learning model;
comparing a plurality of calculated output data vectors/matrices with the target data vector/matrix respectively, and determining a first function corresponding to the first output data vector/matrix with the smallest deviation with respect to the target data vector/matrix and a current coefficient of the first function from target data vectors/matrix in a plurality of comparison results; and
using a machine learning model loaded with the first function as an optimal trained machine learning model, wherein the loaded first function has the current coefficient.
13. The method for measuring composition and property of formation fluid according to claim 12 , further comprising supplementing an output data vector/matrix whose deviation with respect to the target data vector/matrix meets a preset deviation threshold to the database in one or more of following processes:
during a training process of the machine learning models and an implementation process of measuring the composition and the property of the formation fluid.
14. The method for measuring composition and property of formation fluid according to claim 3 , wherein the measuring model is obtained by training pre-created machine learning models based on big data about compositions and properties of various reservoir fluids and measurement signals of downhole sensors as sample data sets.
15. The method for measuring composition and property of formation fluid according to claim 2 , wherein the measuring model is obtained by training pre-created machine learning models based on big data about compositions and properties of various reservoir fluids and measurement signals of downhole sensors as sample data sets.
16. The method for measuring composition and property of formation fluid according to claim 1 , wherein the measuring model is obtained by training pre-created machine learning models based on big data about compositions and properties of various reservoir fluids and measurement signals of downhole sensors as sample data sets.
17. A system for measuring composition and property of formation fluid, comprising: a processor and a computer readable storage medium, wherein instructions are stored in the computer readable storage medium, and the processor is configured to execute the instructions to perform the method for measuring composition and property of formation fluid according to claim 1 .