Single well foam flooding average production scoring, prediction, evaluation method and model establishment method

Abstract

This invention provides a method for scoring, predicting, and evaluating the average production of foam drainage in a single well, as well as a method for establishing a model. The method for establishing the prediction model includes: acquiring raw sample data from multiple single wells; normalizing the raw sample data to obtain intermediate sample data; calculating the intermediate sample data using partial least squares to obtain quantitative parameters of the correlation between each influencing factor and the average production of foam drainage; obtaining a score for the comprehensive average production of foam drainage in each single well based on the intermediate sample data and the quantitative parameters of correlation; and determining the prediction model based on the comprehensive average production of foam drainage in each single well and its score. This prediction model can predict the average production of foam drainage in a single well quickly, is highly practical, and has high accuracy. Furthermore, based on this prediction model, it is possible to further evaluate whether a single well has the potential for increased gas production through foam drainage and whether it can be selected as a candidate well, which helps improve the accuracy and success rate of process implementation.

Description

Technical Field

[0001] This invention relates to the field of drainage gas production in gas field development, specifically to a method for scoring, predicting, evaluating, and establishing a model for single-well drainage average production. Background Technology

[0002] Foam drainage gas production technology is an important auxiliary production technology developed for water-producing gas fields. In the later stages of gas field development, many gas wells accumulate liquid because the produced water is not completely and promptly removed. During gas well production, this accumulated liquid significantly impacts the production and flow period of low-pressure, low-yield gas wells, easily leading to reduced production and, in severe cases, shutdown. When using foam drainage gas production technology, the liquid is distributed within a foam film, providing a larger surface area. This reduces gas slippage and forms a low-density gas-liquid mixture. In low-yield gas wells, the foam effectively lifts the liquid to the surface, thus ensuring normal gas well production.

[0003] To determine whether a single well is suitable for foam drainage gas production technology, an effect prediction and evaluation must be conducted before implementing the measures. However, existing assessments and predictions are mostly based on human assessments and predictions, which are prone to errors. The accuracy and success rate of the technology implementation are low, which is not conducive to significantly improving production efficiency.

[0004] Chinese patent document CN111914209A discloses a fuzzy comprehensive evaluation method for the effect of foam drainage gas production based on the entropy method. Targeting the characteristics of foam drainage gas production wells, it selects indicators to evaluate the effect of foam drainage gas production, constructs an evaluation index system for the effect of foam drainage gas production following the principle of combining qualitative and quantitative methods, considers the interrelationships between the evaluation indicators, establishes a membership matrix using linear analysis, determines the index weights using the entropy method, and calculates the comprehensive evaluation index of the effect of foam drainage gas production in gas wells using the fuzzy comprehensive evaluation method. This invention is mainly based on fuzzy mathematics and focuses on the effect evaluation after foam drainage gas production. However, it does not perform effect prediction and evaluation before the implementation of measures in individual wells, and therefore cannot predict the effect after the measures or provide data support for well selection.

[0005] Chinese patent document CN105160071B discloses a method for identifying well conditions suitable for horizontal wells with simultaneous gas and liquid production. Using a multivariate regression statistical method, it utilizes production data including well inclination angle, gas-liquid ratio, wellhead casing pressure difference, and the difference between critical fluid-carrying flow rate and gas production to predict the maximum and average pressure drop gradients in the wellbore. This accurately guides the implementation of drainage gas production processes. The process is simple, data is obtained quickly and timely, and it meets production needs. Furthermore, the testing depth is not affected by the well inclination angle. However, this invention is based on multivariate regression statistical methods, which result in non-unique calculation formulas with multiple solutions. Moreover, this invention primarily provides evaluation indicators to guide the types of drainage gas production processes implemented, but does not predict or evaluate the effectiveness of these processes. Therefore, current predictions of single-well gas production after foam drainage gas production are mostly based on manual comprehensive judgment, which is inconvenient to calculate, lacks practicality, and requires further improvement in assessment and prediction accuracy. Summary of the Invention

[0006] The purpose of this invention is to address at least one of the aforementioned shortcomings of the existing technology. For example, one objective of this invention is to provide a method for establishing a single-well foam drainage average production prediction model, and a single-well foam drainage average production prediction method. Based on data statistics, this method uses a partial least squares algorithm to predict the gas volume of a single well in foam drainage gas production. The calculation is fast, highly practical, and avoids the uncertainty and erroneous predictions caused by purely manual evaluation. Furthermore, this invention provides a single-well foam drainage average production evaluation method. Based on the predicted gas volume of a single well in foam drainage gas production, this method evaluates whether a single well has the potential for increased production through foam drainage gas production, thereby screening candidate wells. This improves the accuracy and success rate of process implementation, saves reagent costs, and significantly increases production benefits.

[0007] To achieve the above objectives, this invention proposes a scoring method for the average production of a single well's comprehensive foam drainage, the method comprising:

[0008] Obtain raw sample data from multiple single wells. The raw sample data includes the average production of the comprehensive drainage of a single well (average production during the measures) and multiple influencing factors, including the production before the measures, the casing pressure before the measures, and several factors affecting the drainage and gas production effect.

[0009] The original sample data is normalized to obtain intermediate sample data;

[0010] The partial least squares method was used to calculate the correlation quantification parameters between each influencing factor and the average yield of bubble discharge. The correlation quantification parameters include weights and weight coefficients.

[0011] Based on intermediate sample data and correlation quantification parameters, the score of the average production of each single well was obtained.

[0012] A second aspect of this invention also proposes a method for establishing a prediction model for the average production rate of a single well's bubble discharge, the method comprising:

[0013] Obtain raw sample data from multiple single wells. The raw sample data includes the average production of single well comprehensive drainage and multiple influencing factors, including production before measures, casing pressure before measures, and several factors affecting drainage and gas production effects.

[0014] The original sample data is normalized to obtain intermediate sample data;

[0015] The partial least squares method was used to calculate the correlation quantification parameters between each influencing factor and the average yield of bubble discharge. The correlation quantification parameters include weights and weight coefficients.

[0016] Based on intermediate sample data and correlation quantification parameters, the score of the average comprehensive bubble drainage production of each single well is obtained;

[0017] The prediction model is determined based on the average production and score of each individual well's comprehensive drainage.

[0018] Normalization is a common technique in data preprocessing, designed to scale data proportionally so that it falls within a small, specific range, typically [0, 1] or [-1, 1]. When dealing with features of different dimensions or magnitudes, normalization can help improve the convergence speed and performance of algorithms.

[0019] In a preferred embodiment of this scheme, the partial least squares algorithm model is written using Matlab software.

[0020] In a preferred embodiment of this scheme, the step of obtaining the score of the average production of each single well based on intermediate sample data and correlation quantification parameters includes:

[0021] Based on intermediate sample data and correlation quantification parameters, scores were obtained for each influencing factor affecting the average production of single-well bubble drainage.

[0022] Based on the scores of each influencing factor for each individual well, the score of the average comprehensive bubble discharge production of each individual well is obtained.

[0023] In a preferred embodiment of this scheme, obtaining the scores of each influencing factor affecting the average production of single-well bubble drainage based on intermediate sample data and correlation quantification parameters includes:

[0024]

[0025] Where i is the i-th influencing factor, Score well_i This represents the score of the i-th influencing factor, dimensionless. For the i-th influencing factor, the normalized result of its original sample data is pi, ω i Let i = 1, 2, ..., n, and n be the weight and coefficient of the i-th influencing factor influencing the average daily output of foam drainage during the injection cycle.

[0026] In this plan, the injection cycle refers to the period during which the foaming and drainage process is carried out.

[0027] In a preferred embodiment of this scheme, the step of obtaining the average comprehensive foaming rate of each single well based on the scores of each influencing factor of each single well includes:

[0028] For each individual well

[0029] Where i is the i-th influencing factor, Score well This represents a dimensionless fraction representing the average production of a single well's combined drainage capacity. The normalized result of the i-th influencing factor is dimensionless, pi, ω i Let i and n be the weights and coefficients of the i-th influencing factor influencing the average yield of foam excretion during the injection cycle, respectively. They are dimensionless, i = 1, 2, ..., n.

[0030] In a preferred embodiment of this solution, n ≥ 5;

[0031] Furthermore, multiple influencing factors include pre-measure output, pre-measure pressure, production method, unobstructed flow rate, and throttling method.

[0032] In a preferred embodiment of this solution, the factors affecting the drainage gas extraction effect include, but are not limited to, production method, annular liquid level before the measure, porosity, permeability, gas saturation, unobstructed flow rate, and throttling method.

[0033] In this scheme, during the normalization process, the original units of some influencing factors need to be converted to dimensionless units to eliminate the influence of units on data analysis. Among them, the original unit of the annular liquid level before the conversion is m; porosity, permeability, and gas saturation are %, without units; the original unit of the unobstructed flow rate before the conversion is 10,000 cubic meters / day; the throttling method has no unit, including throttling devices and unobstructed flow; the production method has no unit, including continuous production, intermittent production, intermittent open production, and long-term closed state, etc.

[0034] In a preferred embodiment of this scheme, when the production method is used as an influencing factor, continuous production is assigned a value of 0, intermittent production is assigned a value of 0.5, and plunger production is assigned a value of 1, in order to determine the original sample data of the production method, and / or

[0035] When the throttling method is taken as an influencing factor, throttling production is assigned a value of 1, and unrestricted production is assigned a value of 0, in order to determine the original sample data of the throttling method.

[0036] In a preferred embodiment of this solution, the normalization preprocessing for each factor includes:

[0037]

[0038] in, This represents the normalized result of the original sample data of a certain influencing factor of the j-th well, or the normalized result of the original sample data of the average production of the comprehensive bubble discharge of the single well, x. max The maximum value of its original sample data, x min Let j be the minimum value of its original sample data, j = 1, 2, ..., m.

[0039] In a preferred embodiment of this scheme, determining the prediction model based on the average production rate and score of each individual well's comprehensive bubble drainage includes:

[0040] Based on the average production and score of each individual well's comprehensive drainage, a linear regression model is constructed to determine the prediction model.

[0041] In a preferred embodiment of this solution, the prediction model includes:

[0042] q g_calc =aScore well +b;

[0043] Where qg_calc is the predicted average production of a single well, in 10,000 cubic meters per day; Scorewell is the comprehensive average production score of a single well, dimensionless; a and b represent the empirical coefficients of the fit, in 10,000 cubic meters per day.

[0044] In another aspect, the present invention also proposes a method for predicting the average production of bubble discharge from a single well, the method comprising: using the model established by the above method to predict the average production of bubble discharge from a target single well.

[0045] In a preferred embodiment of this solution, the prediction method includes the following steps:

[0046] The predicted average production of foam discharge per well is calculated based on the prediction model.

[0047] Based on the predicted average production of foam discharge from a single well, the relative error ε of the well is calculated. When ε ≤ 30%, the prediction is successful.

[0048] The formula for calculating the relative error ε is:

[0049] Where, q g qg_cale represents the average production rate of a single well, and qg_cale is the predicted average production rate of a single well.

[0050] Furthermore, this invention also proposes a method for evaluating the average production of single-well bubble drainage, comprising the following steps:

[0051] Based on the above prediction method, the predicted value of the average production of bubble discharge in a single well is obtained;

[0052] Calculate Δq 差值 When Δq 差值 If the output is greater than or equal to 0.1 million cubic meters per day, the well is considered to have the potential for increased gas production through foam drainage.

[0053] Δq 差值 The calculation formula is: Δq difference = qg_calc - q before the measure;

[0054] Where, q 措施前 qg_cale represents the pre-treatment production rate in 10,000 cubic meters per day, and qg_cale represents the predicted average production rate of a single well in 10,000 cubic meters per day.

[0055] Compared with the prior art, the beneficial effects of the present invention include at least one of the following:

[0056] (1) This invention proposes a method for predicting the average production of single-well foam drainage. Based on the analysis and statistics of previous single-well geological parameters and current production data, the production before the measures and the casing pressure before the measures are determined. Then, factors such as production mode, annular liquid level before the measures, porosity, permeability, gas saturation, unobstructed flow rate, and throttling method are selected to construct multiple original sample data affecting the average production of foam drainage. The original sample data are normalized and preprocessed to obtain intermediate sample data. Then, based on the partial least squares algorithm model, the score of the comprehensive average production of foam drainage of each single well is obtained. Finally, a prediction model of the gas volume of single wells for foam drainage gas production is established through linear regression.

[0057] This model can predict the production increase effect of single-well foam drainage gas production measures, guide whether a single well is suitable for this process, and avoid the uncertainty and erroneous prediction caused by simple human evaluation and prediction. Moreover, the dynamic and static parameters required to build this model are all conventional parameters, which are easy to obtain and applicable to actual field conditions.

[0058] (2) Key factors affecting well production: This invention also proposes a method for evaluating the average production of a single well using foam drainage, which obtains the predicted gas volume of a single well using a prediction model, qg_cale; and calculates the predicted value q. g_c al c Compared with the output q before the measures 措施前 The difference is used to evaluate whether the well has the potential to increase gas production through foam drainage, so as to select backup wells, thereby improving the accuracy and success rate of process implementation, saving reagent costs, and significantly improving production benefits.

[0059] (3) This scheme uses the PLS partial least squares method to obtain the weights and coefficients of different influencing factors, providing a quantitative basis for well selection for drainage and gas production. The calculation is fast and the practicality is strong. Attached Figure Description

[0060] The above and other objects and / or features of the present invention will become clearer from the following description taken in conjunction with the accompanying drawings, in which:

[0061] Figure 1 The diagram shows a regression curve of the average production of each well and its fraction, representing an exemplary embodiment of the single-well bubble discharge average production prediction method of the present invention.

[0062] Figure 2 The diagram illustrates the process of establishing a single-well average production prediction model for an exemplary embodiment of the single-well bubble discharge average production prediction method of the present invention. Detailed Implementation

[0063] In the following sections, the single-well average production scoring, prediction, evaluation method, and model building method of the present invention will be described in detail with reference to exemplary embodiments.

[0064] The overall concept of this invention is to first obtain multiple influencing factors affecting the average production of foam drainage gas production, then perform normalization preprocessing on the original sample data of each influencing factor, and obtain the weight and coefficient of each influencing factor influencing the daily average production (average production of foam drainage gas production) during the injection cycle based on the partial least squares algorithm model. Then, the influencing factors are scored based on the normalization preprocessing results, as well as the weight and coefficient of the influencing factors, to obtain the comprehensive average production score of foam drainage gas production for a single well. Finally, based on the comprehensive average production score of foam drainage gas production for a single well and the actual gas volume of multiple single wells in foam drainage gas production, a prediction model for the gas volume of single wells in foam drainage gas production is constructed.

[0065] Exemplary Example 1

[0066] To achieve the above objectives, this invention proposes a scoring method for the average production of a single well's comprehensive foam drainage, the method comprising:

[0067] Obtain raw sample data from multiple single wells. The raw sample data includes the average production of single well comprehensive drainage and multiple influencing factors, including production before measures, casing pressure before measures, and several factors affecting drainage and gas production effects.

[0068] The original sample data is normalized to obtain intermediate sample data;

[0069] The partial least squares method was used to calculate the correlation quantification parameters between each influencing factor and the average yield of bubble discharge. The correlation quantification parameters include weights and weight coefficients.

[0070] Based on intermediate sample data and correlation quantification parameters, the score of the average production of each single well was obtained.

[0071] In this exemplary embodiment, the partial least squares algorithm model is written using Matlab software.

[0072] In this exemplary embodiment, the step of obtaining the score of the average production of each single well based on intermediate sample data and correlation quantification parameters includes:

[0073] Based on intermediate sample data and correlation quantification parameters, scores were obtained for each influencing factor affecting the average production of single-well bubble drainage.

[0074] Based on the scores of each influencing factor for each individual well, the score of the average comprehensive bubble discharge production of each individual well is obtained.

[0075] In this exemplary embodiment, obtaining the scores of each influencing factor affecting the average production of single-well bubble drainage based on intermediate sample data and correlation quantification parameters includes:

[0076]

[0077] Where i is the i-th influencing factor, and Scorewell_i represents the score of the i-th influencing factor, which is dimensionless. For the i-th influencing factor, the normalized result of its original sample data is dimensionless, pi, ω i Let i and n be the weights and coefficients of the i-th influencing factor influencing the average yield of foam excretion during the injection cycle, respectively. They are dimensionless, i = 1, 2, ..., n.

[0078] In this exemplary embodiment, the step of obtaining the average comprehensive foam discharge production score of each individual well based on the scores of each influencing factor of each individual well includes:

[0079] For each individual well

[0080] Where i is the i-th influencing factor, and Scorewell represents the fraction of the average production of the comprehensive bubble drainage of a single well, which is dimensionless. The normalized result of the i-th influencing factor is dimensionless, pi, ω i Let i and n be the weights and coefficients of the i-th influencing factor influencing the average yield of foam excretion during the injection cycle, respectively. They are dimensionless, i = 1, 2, ..., n.

[0081] In this exemplary embodiment, the factors affecting the drainage gas extraction effect include, but are not limited to, production method, annular liquid level before the measure, porosity, permeability, gas saturation, unobstructed flow rate, and throttling method.

[0082] In this exemplary embodiment, when production mode is considered as an influencing factor, continuous production is assigned a value of 0, intermittent production is assigned a value of 0.5, and plunger production is assigned a value of 1, in order to determine the original sample data for production mode, and / or

[0083] When the throttling method is taken as an influencing factor, throttling production is assigned a value of 1, and unrestricted production is assigned a value of 0, in order to determine the original sample data of the throttling method.

[0084] In this exemplary embodiment, the normalization preprocessing includes, for each factor:

[0085]

[0086] in, x represents the normalized result of the original sample data of a certain influencing factor of the j-th single well. max The maximum value of its original sample data, x min Let j be the minimum value of its original sample data, j = 1, 2, ..., m.

[0087] Exemplary Example 2

[0088] A method for establishing a single-well average production prediction model, the method comprising:

[0089] Obtain raw sample data from multiple single wells. The raw sample data includes the average production of single well comprehensive drainage and multiple influencing factors, including production before measures, casing pressure before measures, and several factors affecting drainage and gas production effects.

[0090] The original sample data is normalized to obtain intermediate sample data;

[0091] The partial least squares method was used to calculate the correlation quantification parameters between each influencing factor and the average yield of bubble discharge. The correlation quantification parameters include weights and weight coefficients.

[0092] Based on intermediate sample data and correlation quantification parameters, the score of the average comprehensive bubble drainage production of each single well is obtained;

[0093] The prediction model is determined based on the average production and score of each individual well's comprehensive drainage.

[0094] Figure 2 A schematic diagram of the process for establishing a prediction model for the average production of single-well bubble drainage is shown.

[0095] Normalization is a common technique in data preprocessing, designed to scale data proportionally so that it falls within a small, specific range, typically [0,1] or [-1,1]. When dealing with features of different dimensions or magnitudes, normalization can help improve the convergence speed and performance of algorithms.

[0096] In this exemplary embodiment, the partial least squares algorithm model is written using Matlab software.

[0097] In this exemplary embodiment, the step of obtaining the score of the average production of each single well based on intermediate sample data and correlation quantification parameters includes:

[0098] Based on intermediate sample data and correlation quantification parameters, scores were obtained for each influencing factor affecting the average production of single-well bubble drainage.

[0099] Based on the scores of each influencing factor for each individual well, the score of the average comprehensive bubble discharge production of each individual well is obtained.

[0100] In this exemplary embodiment, obtaining the scores of each influencing factor affecting the average production of single-well bubble drainage based on intermediate sample data and correlation quantification parameters includes:

[0101]

[0102] Where i is the i-th influencing factor, and Scorewell_i represents the score of the i-th influencing factor, which is dimensionless. For the i-th influencing factor, the normalized result of its original sample data is dimensionless, pi, ω i Let i and n be the weights and coefficients of the i-th influencing factor influencing the average yield of foam excretion during the injection cycle, respectively. They are dimensionless, i = 1, 2, ..., n.

[0103] In this exemplary embodiment, the step of obtaining the average comprehensive foam discharge production score of each individual well based on the scores of each influencing factor of each individual well includes:

[0104] For each individual well

[0105] Where i is the i-th influencing factor, and Scorewell represents the fraction of the average production of the comprehensive bubble drainage of a single well, which is dimensionless. The normalized result of the i-th influencing factor is dimensionless, pi, ω i Let i and n be the weights and coefficients of the i-th influencing factor influencing the average yield of foam excretion during the injection cycle, respectively. They are dimensionless, i = 1, 2, ..., n.

[0106] In this exemplary embodiment, the factors affecting the drainage gas extraction effect include, but are not limited to, production method, annular liquid level before the measure, porosity, permeability, gas saturation, unobstructed flow rate, and throttling method.

[0107] In this exemplary embodiment, when the production method is taken as an influencing factor, continuous production is assigned a value of 0, intermittent production is assigned a value of 0.5, and plunger production is assigned a value of 1, and / or when the throttling method is taken as an influencing factor, throttling production is assigned a value of 1, and unimpeded production is assigned a value of 0.

[0108] In this exemplary embodiment, the normalization preprocessing includes, for each factor:

[0109]

[0110] in, This represents the normalized result of the original sample data of a certain influencing factor of the j-th well, or the normalized result of the original sample data of the average production of the comprehensive bubble discharge of the single well, x. max The maximum value of its original sample data, x min Let j be the minimum value of its original sample data, j = 1, 2, ..., m.

[0111] In this exemplary embodiment, determining the prediction model based on the average production rate of each individual well and its score includes:

[0112] Based on the average production and score of each individual well's comprehensive drainage, a linear regression model is constructed to determine the prediction model.

[0113] In this exemplary embodiment, the prediction model includes:

[0114] q g_calc =aScore well +b;

[0115] Where qg_cakc is the predicted average production of bubble discharge from a single well; Scorewell is the comprehensive average production score of bubble discharge from a single well, which is dimensionless; a and b represent the empirical coefficients of the fit, which are dimensionless.

[0116] Exemplary Example 3

[0117] A method for predicting the average production of a single well's bubble discharge, the method comprising: using a model established by the above method to predict the average production of a target single well's bubble discharge.

[0118] In this exemplary embodiment, the prediction method includes the following steps:

[0119] The predicted average production of foam discharge per well is calculated based on the prediction model.

[0120] Based on the predicted average production of foam discharge from a single well, the relative error ε of the well is calculated. When ε ≤ 30%, the prediction is successful.

[0121] The formula for calculating the relative error ε is:

[0122] Where, q g qg_calc represents the average production rate of a single well, and qg_calc represents the predicted average production rate of a single well.

[0123] Exemplary Example 4

[0124] A method for evaluating the average production of single-well bubble drainage includes the following steps:

[0125] Based on the above prediction method, the predicted value of the average production of bubble discharge in a single well is obtained;

[0126] Calculate Δq 差值 When Δq 差值 If the output is greater than or equal to 0.1 million cubic meters per day, the well is considered to have the potential for increased gas production through foam drainage.

[0127] Δq 差值 The calculation formula is: Δq difference = qg_calc - q before the measure;

[0128] Where, q 措施前 The output before the measures are implemented is 10,000 cubic meters per day, and qg_cakc is the predicted average output of a single well, 10,000 cubic meters per day.

[0129] Exemplary Example 5

[0130] A method for predicting the average production of a single well's bubble discharge includes the following steps:

[0131] S1. Obtain raw sample data from multiple single wells. The raw sample data includes the average production of single well comprehensive drainage and n influencing factors. The n influencing factors include the production before the measures, the casing pressure before the measures, and several factors affecting the drainage and gas production effect.

[0132] S2. Obtain the raw sample data of m single wells, and perform normalization preprocessing on the raw sample data to obtain intermediate sample data.

[0133] Normalization preprocessing includes:

[0134] in, This represents the normalized result of the original sample data for influencing factor j, or the normalized result of the original sample data for the average production of single-well comprehensive foam drainage. max x is the maximum value of the sample data. minLet j be the minimum value of the sample data, where j = 1, 2, ..., m.

[0135] S3. Use Matlab software to write a PLS partial least squares algorithm model, input intermediate sample data into the partial least squares algorithm model, and obtain the correlation quantification parameters of each influencing factor with the average yield of foaming and excretion. The correlation quantification parameters include the weight and weight coefficient of the daily average yield during the injection cycle.

[0136] S4. Based on intermediate sample data and correlation quantification parameters, the scores of each influencing factor affecting the average production of single-well bubble drainage are obtained, including:

[0137]

[0138] Where i is the i-th influencing factor, and Scorewell_i represents the score of the i-th influencing factor, which is dimensionless. For the i-th influencing factor, the normalized result of its original sample data is dimensionless, pi, ω i Let i and n be the weights and coefficients of the i-th influencing factor influencing the average yield of foam excretion during the injection cycle, respectively. They are dimensionless, i = 1, 2, ..., n.

[0139] S5. Based on the scores of each influencing factor for each individual well, the score for the average comprehensive foam discharge production of each individual well is obtained, including:

[0140] For each individual well

[0141] Where i is the i-th influencing factor, and Scorewell represents the fraction of the average production of the comprehensive bubble drainage of a single well, which is dimensionless. The normalized result of the i-th influencing factor is dimensionless, pi, ω i Let i and n be the weights and coefficients of the i-th influencing factor influencing the average yield of foam excretion during the injection cycle, respectively. They are dimensionless, i = 1, 2, ..., n.

[0142] S6. Obtain and construct a linear regression model based on the comprehensive average foam drainage production score of m single wells in the implementation of the main control factor measures, and the comprehensive average foam drainage production of each single well, to obtain the prediction model of gas volume of single wells for foam drainage gas production.

[0143] The prediction model includes: q g_calc =aScore well +b, where qg_calc is the predicted gas volume of a single well in foam drainage gas production, dimensionless; Scorewell is the comprehensive average production score of a single well in foam drainage, dimensionless; a and b represent the empirical coefficients of the fit, dimensionless.

[0144] S7. Based on the predicted gas volume of a single well using foam drainage gas production, calculate the relative error ε of the well. When ε ≤ 30%, the prediction is successful.

[0145] The formula for calculating the relative error ε is:

[0146] Where, q g qg_calc represents the actual gas volume of a single well in foam drainage gas production, while qg_calc represents the predicted gas volume of a single well in foam drainage gas production.

[0147] In this exemplary embodiment, n = 7, and the seven factors are production method, pre-measure annular liquid level, porosity, permeability, gas saturation, unobstructed flow rate, and throttling method. Specifically, when production method is used as the evaluation index, continuous production is assigned a value of 0, intermittent production is assigned a value of 0.5, and plunger production is assigned a value of 1; when throttling method is used as the evaluation index, throttling production is assigned a value of 1, and unobstructed production is assigned a value of 0.

[0148] Exemplary Example 6

[0149] This exemplary embodiment provides a method for predicting the average production of a single well's bubble discharge.

[0150] The method includes the following steps:

[0151] Step 1: Determine the main controlling factors, namely, the gas production rate and casing pressure before the measures are implemented. Then, select the remaining n evaluation indicators A that affect the drainage gas production effect. j , j = 1, 2, ..., n.

[0152] Step 2: Due to the difference in dimensions of the parameters, for the production method, this parameter is assumed to be 0 for continuous production, 0.5 for intermittent production, and 1 for plunger production. For the throttling method, this parameter is assumed to be 1 for throttling production and 0 for unobstructed production. The remaining parameters are processed according to their own values, and the maximum and minimum values ​​are the range of values ​​of each parameter in the selected sample.

[0153] Different parameters are preprocessed by normalization: Where, x max x is the maximum value of the sample data. min This represents the minimum value of the sample data.

[0154] Step 3: Use Matlab software to write a PLS partial least squares algorithm model to calculate the weight and coefficient of the average daily output within different injection cycles.

[0155] Step 4: Based on the normalization results of different parameters and the calculation results of weights and coefficients, define the influencing factors of the average daily production of a single well's bubble discharge and calculate the score. Define the score of the influencing factor of a certain bubble discharge average production of a single well as:

[0156] Where, p i To calculate the weights of the average daily output within different injection cycles for the partial least squares algorithm model, ω i The partial least squares algorithm model is used to calculate the coefficient of average daily production under different parameters within the injection cycle. The comprehensive average production score of a single well is defined as follows:

[0157]

[0158] Step 5: Based on the regression curve of gas well score and average daily production from foam drainage, which shows a clear linear relationship, a single-well gas production prediction model for foam drainage gas production is established as follows: q g_calc =aScore well +b.

[0159] Step 6: Verify the calculated gas volume q of a single well in foam drainage gas production. g_calc Compared with the actual value q g A prediction is considered successful if the relative error is less than or equal to 30%; otherwise, the prediction fails. The relative error ε is:

[0160] Step 7: Calculation of gas volume q for single well in foam drainage gas production g_calc Compared with the output q before the measures 措施前 The difference Δq 差值 Greater than or equal to 0.1 million cubic meters / day, Δq 差值 =q g_calc -q 措施前 A value of ≥0.1 indicates the well has the potential to increase gas production through foam drainage and can be included in the candidate well list.

[0161] In the above technical solution, in step 1, when n=7, the evaluation parameters are production method, annular liquid level before the measure, porosity, permeability, gas saturation, unobstructed flow rate, and throttling method.

[0162] To better understand the exemplary embodiments described above, further explanation is provided below with specific examples.

[0163] Example 1

[0164] Step 1: Analyze and statistically analyze the geological parameters of the single wells in the early stage and the current production data to obtain the original sample data of multiple single wells. The original sample data includes the average production of the single well comprehensive drainage (average production during the measures) and multiple influencing factors, including the production before the measures, the casing pressure before the measures, and several factors that affect the drainage and gas production effect.

[0165] In this example embodiment, the factors affecting the gas extraction effect of drainage include the production method, the annular liquid level before the measure, porosity, permeability, gas saturation, unobstructed flow rate, and throttling method.

[0166] Step 2: Since the parameters of each influencing factor have different dimensions, for the influencing factor of production method, the original sample data assumes that continuous production is 0, intermittent production is 0.5, and plunger production is 1. For the influencing factor of throttling method, the original sample data assumes that throttling production is 1 and unobstructed production is 0. The sample parameters of the other influencing factors are processed according to their own values, and the maximum and minimum values ​​are selected as the range of values ​​of each parameter in the selected sample, as shown in Table 2.

[0167] The original sample data is shown in Table 1. The original sample data of the 10 categories in Table 1 are normalized respectively to obtain the intermediate sample data as shown in Table 3.

[0168] Normalization preprocessing for raw sample data of a specific category includes:

[0169] Where j is the j-th single well; This represents the normalized result of the original sample data of a certain influencing factor of the j-th well, or the normalized result of the original sample data of the comprehensive average production of the single well (average production in the measures); x max It is the maximum value of its original sample data; x min Let j be the minimum value of its original sample data; j = 1, 2, ..., m.

[0170] Table 1 Original Sample Data

[0171]

[0172]

[0173] Table 2. Extreme values ​​of different parameters in the original sample data.

[0174]

[0175] Table 3 Normalized Data Table

[0176]

[0177]

[0178] Step 3: Use Matlab2017b software to write the PLS partial least squares algorithm model and input the normalized sample data results into the program for calculation.

[0179] The multi-factor analysis results of the average daily output during the injection cycle were obtained through calculation, that is, the correlation quantitative parameters between each influencing factor and the average output of foam were obtained. The correlation quantitative parameters include weights and weight coefficients. The results are shown in Table 4.

[0180] Table 4. Summary of the correlation between various influencing factors and the average yield of foamed eggplant.

[0181]

[0182] Step 4: Based on the intermediate sample data and correlation quantification parameters, obtain the scores of each influencing factor affecting the average production of single well foam drainage; based on the scores of each influencing factor for each single well, obtain the score of the comprehensive average production of single well foam drainage.

[0183] The scores of each factor affecting the average production of single-well bubble drainage are as follows:

[0184] Where i is the i-th influencing factor, and Scorewell_i represents the score of the i-th influencing factor, which is dimensionless. For the i-th influencing factor, the normalized result of its original sample data is dimensionless, P. i ω i Let i and n be the weights and coefficients of the i-th influencing factor influencing the average yield of foam excretion during the injection cycle, respectively. They are dimensionless, i = 1, 2, ..., n.

[0185] The average production score of a single well is:

[0186] Where i is the i-th influencing factor, and Scorewell represents the fraction of the average production of the comprehensive bubble drainage of a single well, which is dimensionless. The normalized result for the i-th influencing factor is dimensionless, P i ω i Let i and n be the weights and coefficients of the i-th influencing factor influencing the average yield of foam excretion during the injection cycle, respectively. They are dimensionless, i = 1, 2, ..., n.

[0187] The average production score of each individual well is shown in Table 5.

[0188] Table 5. Summary of Average Production Scores for 25 Single Wells

[0189] hashtag Average production score Average yield (comprehensive average yield of foaming and draining) in the measures: qg (10,000 cubic meters / day) Well 1 0.02 0.17 Well 2 0.06 0.61 Well 3 0.05 0.39 Well 4 0.02 0.26 Well 5 0.01 0.21 Well 6 0.004 0.17 Well 7 0.002 0.19 Well 8 0.05 0.33 Well 9 0.01 0.03 Well 10 0.02 0.30 Well 11 0.25 1.77 Well 12 0.02 0.26 Well 13 0.18 1.00 Well 14 0.05 0.19 Well 15 0.01 0.17 Well 16 0.06 0.34 Well 17 0.06 0.58 Well 18 0.13 0.68 Well 19 0.10 0.61 Well 20 0.12 0.99 Well 21 0.20 1.07 Well 22 0.02 0.30 Well 23 0.02 0.35 Well 24 0.05 0.53 Well 25 -0.01 0.03

[0190] Step 5: Based on the clear linear relationship between the average production of each single well and its fractional regression curve, a prediction model for the gas volume of a single well in foam drainage gas production is established as follows: q g_calc =aScore well +b;

[0191] From Table 5 and Figure 1 It can be seen that the calculated score and the actual average yield of the comprehensive foaming process have a clear linear relationship, with a correlation coefficient R². 2 The value is 0.907, and the linear relationship expression is:

[0192] q g_calc =5.5997 Score well +0.1276.

[0193] Step 6: Verify the predicted value q of the average production rate of a single well. g_ca l c Average production of single-well integrated foam drainage q g A prediction is considered successful if the relative error is less than or equal to 30%; otherwise, the prediction fails. The relative error ε is:

[0194]

[0195] Step 7: Select 5 more wells that are different from those in Table 4 above to conduct model verification. The data of the individual wells are shown in Table 6.

[0196] Table 6 Single Well Data Table

[0197]

[0198] Note: In the production mode column, 0 represents continuous, 0.5 represents intermittent, and 1 represents plunger; in the throttle column, 1 represents throttle and 0 represents unobstructed.

[0199] Step 8: Perform normalization preprocessing as described in Step 2 of Example 1. Normalization preprocessing includes: Where j is the j-th single well; This represents the normalized result of the original sample data of a certain influencing factor of the j-th well, or the normalized result of the original sample data of the comprehensive average production of the single well (average production in the measures); x max It is the maximum value of its original sample data; x min The minimum value of its original sample data; j = 1, 2, ..., m, the maximum and minimum values ​​of single well data are shown in Table 7, and the formula is used. The normalized data table is shown in Table 8.

[0200] Table 7 Maximum and Minimum Values ​​of Single Well Data

[0201]

[0202]

[0203] Table 8. Normalized Data for Single Wells

[0204]

[0205] Step 9: Based on Step 3 of Example 1, the weights and coefficients of each influencing factor on the average production of bubble discharge are shown in Table 9; the comprehensive average production score of a single well is obtained using the formula given in Step 4 of Example 1. Combining Tables 8 and 9, the average comprehensive bubble discharge production score of these 5 wells can be calculated, as shown in Table 10.

[0206] Table 9 Summary of the correlation weights and coefficients of each influencing factor on the average yield of foamed eggplant.

[0207]

[0208] Table 10 Summary of Average Production Scores for Integrated Bubble Discharge from 5 Single Wells

[0209]

[0210] Step 10, based on the prediction model q for single-well gas volume of foam drainage gas production obtained in Step 5 of Example 1 above. g_calc =5.5997 Score well +0.1276, the prediction results are shown in Table 11. Five out of five single wells were predicted successfully, with a success rate of 100%.

[0211] Table 11. Predicted Production and Results for 5 Individual Wells

[0212]

[0213]

[0214] Example 2

[0215] Based on the gas production prediction of a single well for foam drainage gas production shown in Example 1, this example evaluates whether a single well has the potential to increase gas production through foam drainage gas production in order to screen candidate wells.

[0216] The specific steps are as follows: Based on the predicted average production value qg_calc of single well bubble discharge in Table 5, calculate Δq. 差值 :

[0217] Δq difference = qg_calc - q before the measure, where q 措施前 The gas production per well before the measures were implemented is 10,000 cubic meters per day.

[0218] When Δq 差值 Greater than or equal to 0.1 million cubic meters / day, Δq 差值 =q g_calc -q 措施前 A value of ≥0.1 indicates the well has the potential to increase gas production through foam drainage and can be included in the candidate well list.

[0219] According to Δq 差值Based on the calculation results, the alternative recommendations for 5 single wells are shown in Table 12.

[0220] Table 12 Recommended Wells

[0221]

[0222] Although the present invention has been described above in conjunction with exemplary embodiments and accompanying drawings, those skilled in the art should understand that various modifications can be made to the above embodiments without departing from the spirit and scope of the claims.

Claims

1. A scoring method for the average production of a single well's comprehensive foam drainage, characterized in that, The method includes: Obtain raw sample data from multiple single wells. The raw sample data includes the average production of single well comprehensive drainage and multiple influencing factors, including production before measures, casing pressure before measures, and several factors affecting drainage and gas production effects. The original sample data is normalized to obtain intermediate sample data; The partial least squares method was used to calculate the correlation quantification parameters between each influencing factor and the average yield of bubble discharge. The correlation quantification parameters include weights and weight coefficients. Based on intermediate sample data and correlation quantification parameters, the score of the average production of each single well was obtained.

2. A method for establishing a single-well average production prediction model, characterized in that, The method includes: Obtain raw sample data from multiple single wells. The raw sample data includes the average production of single well comprehensive drainage and multiple influencing factors, including production before measures, casing pressure before measures, and several factors affecting drainage and gas production effects. The original sample data is normalized to obtain intermediate sample data; The partial least squares method was used to calculate the correlation quantification parameters between each influencing factor and the average yield of bubble discharge. The correlation quantification parameters include weights and weight coefficients. Based on intermediate sample data and correlation quantification parameters, the score of the average comprehensive bubble drainage production of each single well is obtained; The prediction model is determined based on the average production and score of each individual well's comprehensive drainage.

3. The method according to claim 1 or 2, characterized in that: The step of obtaining the score of the average comprehensive foaming production of each single well based on intermediate sample data and correlation quantification parameters includes: Based on intermediate sample data and correlation quantification parameters, scores were obtained for each influencing factor affecting the average production of single-well bubble drainage. Based on the scores of each influencing factor for each individual well, the score of the average comprehensive bubble discharge production of each individual well is obtained.

4. The method according to claim 3, characterized in that: The scores for each influencing factor affecting the average production of single-well bubble drainage, obtained based on intermediate sample data and correlation quantification parameters, include: Where i is the i-th influencing factor, Score well_i This represents the score of the i-th influencing factor, dimensionless. For the i-th influencing factor, the normalized result of its original sample data is dimensionless, pi, ω i Let i and n be the weights and coefficients of the i-th influencing factor influencing the average yield of foam excretion during the injection cycle, respectively. They are dimensionless, i = 1, 2, ..., n.

5. The method according to claim 4, characterized in that: The score for obtaining the average comprehensive foam discharge production of each single well based on the scores of each influencing factor includes: For each individual well Where i is the i-th influencing factor, Score well This represents a dimensionless fraction representing the average production of a single well's combined drainage capacity. The normalized result of the i-th influencing factor is dimensionless, pi, ω i Let i and n be the weights and coefficients of the i-th influencing factor influencing the average yield of foam excretion during the injection cycle, respectively. They are dimensionless, i = 1, 2, ..., n.

6. The method according to claim 5, characterized in that: The n≥5; Furthermore, multiple influencing factors include pre-measure output, pre-measure pressure, production method, unobstructed flow rate, and throttling method.

7. The method according to claim 2, characterized in that: The factors that affect the effectiveness of drainage gas extraction include, but are not limited to, production method, annular liquid level before the measures, porosity, permeability, gas saturation, unobstructed flow rate, and throttling method.

8. The method according to claim 7, characterized in that: When production mode is considered as an influencing factor, continuous production is assigned a value of 0, intermittent production is assigned a value of 0.5, and plunger production is assigned a value of 1, to determine the original sample data for production mode, and / or When the throttling method is taken as an influencing factor, throttling production is assigned a value of 1, and unrestricted production is assigned a value of 0, in order to determine the original sample data of the throttling method.

9. The method according to claim 2, characterized in that: For each factor, the normalization preprocessing includes: in, This represents the normalized result of the original sample data of a certain influencing factor of the j-th well, or the normalized result of the original sample data of the average production of the comprehensive bubble discharge of the single well, x. max The maximum value of its original sample data, x min Let j be the minimum value of its original sample data, j = 1, 2, ..., m.

10. The method according to claim 2, characterized in that: The determination of the prediction model based on the average production and score of each individual well's comprehensive foam drainage includes: Based on the average production and score of each individual well's comprehensive drainage, a linear regression model is constructed to determine the prediction model.

11. A method for predicting the average production of a single well's bubble discharge, characterized in that, The method includes: using a model established by the method of any one of claims 2 to 10 to predict the average production of a target single well.

12. The method for predicting the average production of a single well's bubble discharge according to claim 11, characterized in that, The prediction method includes the following steps: The predicted average production of foam discharge per well is calculated based on the prediction model. Based on the predicted average production of foam discharge from a single well, the relative error ε of the well is calculated. When ε ≤ 30%, the prediction is successful. The formula for calculating the relative error ε is: Where, q g The average daily production of a single well is 10,000 cubic meters. g_calc This is the predicted average daily production of a single well.

13. A method for evaluating the average production of a single well's bubble discharge, characterized in that, Includes the following steps: According to the prediction method of claim 11 or claim 12, a predicted value of the average production of bubble discharge in a single well is obtained; Calculate Δq 差值 When Δq 差值 If the output is greater than or equal to 0.1 million cubic meters per day, the well is considered to have the potential for increased gas production through foam drainage. Δq 差值 The calculation formula is: Δq 差值 =q g_calc -q 措施前 ; in, q 措施前 The output before the measures, 10,000 cubic meters / day, q g_calc This is the predicted average daily production of a single well.

Images