Coal bed gas low-yield well yield increase treatment process based on fuzzy evaluation and grey correlation analysis

Abstract

The invention relates to the technical field of coal bed gas development, in particular to a coal bed gas low-yield well yield increasing treatment process based on fuzzy evaluation and grey correlation analysis. Comprising the steps of data acquisition and preprocessing, evaluation index system construction, grey correlation analysis, fuzzy comprehensive evaluation, potential grading, process package intelligent matching, multi-well priority ranking, dynamic updating after treatment and the like. A potential score is constructed based on the correlation coefficient and the membership function, and quantitative evaluation of the yield increasing potential of the low-yield well is achieved; jointly driving process package matching according to the correlation degree and the score, and forming a coupling chain of factor identification and process decision; a construction priority index is constructed through the correlation degree of the potential score and the key factor, and optimal configuration of multi-well resources is achieved; after treatment is implemented, evaluation weight and process selection are automatically adjusted according to dynamic response, and a closed loop of diagnosis, decision making, implementation and feedback is formed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of coalbed methane development, in particular to a coalbed methane low-yield well stimulation treatment process based on fuzzy evaluation and grey correlation analysis. BACKGROUND

[0002] As an important unconventional natural gas resource, coalbed methane plays a key role in the current energy structure transformation and clean energy development in China. With the continuous expansion of coalbed methane development, the reservoir conditions of different regions and different coal seams differ significantly, resulting in large fluctuations in single well productivity. In actual production, a considerable number of coalbed methane wells exhibit low yield or long-term low yield, which seriously affects the overall development degree of the block and the realization of the yield target. The causes of low-yield wells are multi-source and coupled, including insufficient coalbed gas content, low permeability, poor development of fracture system, unfavorable geostress conditions, and mismatch between fracturing engineering parameters and drainage system configuration, among other factors. Therefore, systematic and scientific stimulation treatment research for low-yield wells is an important technical requirement for improving the efficiency of coalbed methane development.

[0003] Existing coalbed methane low-yield well treatment methods mainly rely on single process measures or operational experience, such as refracturing, water hydrolysis control and plugging, well pattern interference treatment, and drainage system optimization. However, these methods generally have technical limitations: on the one hand, grey correlation analysis is commonly used to quantitatively identify main control factors, but the results are difficult to reflect the coupling relationship between multiple indicators; on the other hand, fuzzy comprehensive evaluation can make multi-factor comprehensive judgments on reservoir and engineering parameters, but existing applications are mostly limited to single well evaluation and have not formed a systematic technical chain with actual stimulation process selection. In addition, low-yield well treatment schemes are mostly based on experience and lack quantitative decision-making models supported by an index system; multi-well treatment lacks objective priority ranking, which may lead to unreasonable allocation of construction resources; at the same time, existing technologies lack a feedback update mechanism based on real-time production dynamics after treatment, making it difficult to achieve closed-loop optimization of evaluation, decision-making, implementation, and re-evaluation:

[0004] In summary, the existing technology at least has the following problems:

[0005] (1) The coupling relationship of coalbed methane wells is complex, and existing methods are difficult to accurately identify the main control factors of productivity and quantify the stimulation potential of single wells;

[0006] (2) Stimulation process selection relies too much on human experience and lacks a model-based decision-making method linked to the evaluation index system, resulting in possible deviation of process package matching from actual needs;

[0007] (3) Multi-well treatment lacks quantitative ranking basis, which is not conducive to reasonable arrangement of construction sequence under resource constraints;

[0008] (4) The existing technology lacks a dynamic closed-loop mechanism, and cannot update the index weight, membership function and evaluation result in real time based on the production changes after the implementation of the management, resulting in difficulty in continuously improving the management effect;

[0009] (5) Currently, there is no integrated method for identifying main control factors, evaluating production increase potential, matching process packages and optimizing construction priority of coalbed methane low-yield wells by coupling gray correlation analysis and fuzzy comprehensive evaluation. SUMMARY

[0010] The present application provides a coalbed methane low-yield well production increase management process based on fuzzy evaluation and gray correlation analysis. By constructing a multi-dimensional evaluation index system, performing weight fusion fuzzy comprehensive evaluation, classifying low-yield well potential, intelligently matching production increase measures, optimizing multi-well construction sequence, and automatically iterating and updating dynamic data after management, the present application realizes intelligent closed-loop control of the whole process of low-yield well production increase management.

[0011] To achieve the above-mentioned purpose, the present application provides the following technical scheme: a coalbed methane low-yield well production increase management process based on fuzzy evaluation and gray correlation analysis, which constructs a multi-dimensional evaluation index system covering resources, physical properties, engineering and dynamic response; uses gray correlation analysis to quantitatively identify key main control factors affecting productivity and their importance ranking; combines fuzzy comprehensive evaluation, integrates gray correlation weight and objective entropy weight, and calculates the quantitative score of single-well production increase potential; classifies the potential according to the score and matches targeted production increase process package; further optimizes multi-well construction sequence by comprehensively considering potential score and key factor correlation degree; and finally updates model parameters according to dynamic production data feedback after management, forming a closed loop of "evaluation-decision-implementation-feedback-reoptimization".

[0012] The present scheme realizes intelligent decision-making of the whole process from low-yield well cause diagnosis, potential quantification, measure optimization to construction optimization and effect iteration, and improves the pertinence of management measures, resource allocation efficiency and sustainability of production increase effect.

[0013] Further, by combining the use of Z-score standardization and Min-max normalization and automatically selecting according to the characteristics of the index, the comparability and analysis reliability of multi-source heterogeneous data are ensured.

[0014] Further, by using gray correlation analysis to identify main control factors based on high-yield wells, accurate data support is provided for subsequent weight allocation and process targeting.

[0015] Further, triangular or trapezoidal membership functions are used and their boundaries are dynamically adjusted according to block data, so that the fuzzy evaluation can adapt to the characteristics of different geological blocks, improving the objectivity and accuracy of the evaluation.

[0016] Furthermore, by using a linear fusion of grey relational weights and entropy weights to determine the comprehensive weights, the importance of factors to production capacity is reflected, while also taking into account the information content of the indicator data itself, making the weight allocation more scientific and reasonable.

[0017] Furthermore, based on a clear threshold for production increase potential, the system is categorized and process package selection rules are established, enabling model-based and rule-based decision-making from potential evaluation to matching governance measures, thus overcoming the limitations of human experience.

[0018] Furthermore, by constructing a priority index of comprehensive production potential and key control factors to rank multiple wells, a quantitative basis for construction arrangements under limited resources was provided, thus optimizing the overall governance efficiency.

[0019] Furthermore, based on the dynamic data after governance, the correlation degree was recalculated and the model parameters were adjusted, and an adaptive update closed-loop optimization mechanism was established, which enabled the governance strategy to continuously evolve with production dynamics and ensure long-term effectiveness.

[0020] Compared with the prior art, the beneficial effects of this application are:

[0021] 1. It has achieved accurate identification of the main factors controlling production capacity and quantitative evaluation of production potential. It objectively quantifies the impact of each indicator on production capacity through grey relational analysis, and uses fuzzy comprehensive evaluation to integrate multiple factors to obtain a quantitative score for a single well. This solves the problems of traditional methods in handling the coupling relationship of multiple factors and subjective evaluation, and provides a reliable basis for scientific decision-making.

[0022] 2. A model-driven intelligent matching method for production enhancement processes was established. Potential is classified based on quantitative scoring, and process packages such as refracking, drainage optimization, and unblocking and water control are automatically matched according to the correlation rules of the main control factors. This transforms process selection from relying on experience to data-based model decision-making, improving the pertinence and effectiveness of measures.

[0023] 3. It provides a quantitative optimization basis for the treatment order of multiple wells. By constructing a priority index that integrates the production potential score and the correlation of key control factors, the treatment order of multiple low-yield wells is objectively sorted, achieving the optimal allocation under the condition of limited operating resources, and improving the overall treatment efficiency and economic benefits of the block.

[0024] 4. A dynamic closed-loop optimization mechanism was constructed to improve the sustainability of governance effectiveness. By collecting dynamic production data after governance, recalculating the grey relational degree, and updating the membership function and weights, the evaluation model can adapt to changes in production status, realize iterative optimization of governance strategies, and form a self-improving closed loop.

[0025] 5. An integrated technology chain for increasing production in low-yield wells has been formed. For the first time, grey relational analysis, fuzzy comprehensive evaluation, process matching logic, construction sequencing algorithm and dynamic feedback mechanism have been organically coupled to form a complete and systematic process method from diagnosis, evaluation, decision-making, implementation to re-optimization. Attached Figure Description

[0026] Figure 1 This is a diagram showing the overall process steps of this application;

[0027] Figure 2 This is a schematic diagram illustrating the coupling of grey relational analysis and fuzzy comprehensive evaluation in this application;

[0028] Figure 3 This is a pie chart showing the distribution of potential wells in this application.

[0029] Figure 4 This is a bar chart comparing the treatment of high-potential wells before and after the application. Detailed Implementation

[0030] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0031] Please see Figures 1-4 This application provides a further detailed description of the technical solution in conjunction with the accompanying drawings:

[0032] Example 1: Overall Process Flow

[0033] This application provides a coalbed methane production enhancement technology based on fuzzy evaluation and grey relational analysis. By constructing a unified evaluation index system and sequentially performing steps such as data preprocessing, grey relational analysis, fuzzy comprehensive evaluation, production potential classification, process package matching, construction sequence optimization and dynamic updating, the systematization, quantification and closed-loop optimization of the production enhancement scheme for low-yield wells are achieved.

[0034] The process for increasing coalbed methane production in low-yield wells based on fuzzy evaluation and grey relational analysis includes the following steps:

[0035] S1) Data acquisition and preprocessing; S2) Construction of evaluation index system; S3) Grey relational analysis; S4) Fuzzy comprehensive evaluation; S5) Grading of production potential; S6) Matching of process packages; S7) Optimization of construction sequence; S8) Dynamic update.

[0036] Each step forms a closed-loop connection of data flow and control flow, realizing a closed-loop process for the entire governance solution from evaluation and selection to optimization.

[0037] Step S1: Data Acquisition and Preprocessing

[0038] Data sources include geological parameters, reservoir properties, engineering construction parameters, and production dynamics data from the acquisition well, mainly including: gas content (m³ / t); coal thickness (m); permeability (mD); porosity (%); burial depth (m); number of fracturing stages, sand volume, liquid volume, and construction discharge rate; daily gas volume, daily water volume, bottom hole flowing pressure, and casing pressure; and gas production curve morphology data (early growth rate, plateau period, and late decline coefficient).

[0039] All data is stored in a structured manner in a data matrix. .

[0040] Data quality processing:

[0041] 1) Missing value handling: If a certain indicator is missing in a well, it will be filled in according to the following rules: Random forest imputation is used for non-linear indicators; linear interpolation is used for missing time series; permeability and porosity are imputed using KNN (k=5).

[0042] The specific formula is as follows: For missing values ​​in time series:

[0043]

[0044] For multidimensional indicators:

[0045]

[0046] in , The distance is Euclidean.

[0047] 2) Outlier identification: using Law:

[0048]

[0049] If the value is out of the question, it will be considered an outlier and either removed or replaced with the median.

[0050] 3) Normalization:

[0051] Different normalization methods are used to process different types of indicators, so that all indicators are mapped to comparable dimensions. The normalization method is automatically selected according to the distribution characteristics of the indicators: Z-score is used for indicators that conform to normal distribution, and Min-max normalization is used for indicators with clear upper and lower limits or extreme value sensitivity.

[0052] The normalization methods used include:

[0053] Maximum index:

[0054]

[0055] Minimum Indicator:

[0056]

[0057] Volatility indicator: Z-score

[0058]

[0059] For time series features, trend features, maximum mutation index, and decline coefficient are extracted as inputs for dynamic response indicators. After preprocessing, a data matrix with a uniform scale is formed. .

[0060] Step S2: Component Evaluation Index System

[0061] Based on the main controlling factors of coalbed methane well productivity, the indicators are divided into four categories:

[0062] 1) Resource indicators R include: gas content, coal thickness, and burial depth.

[0063] Define a vector of resource indicators: .

[0064] 2) Physical and stress properties P include: permeability, porosity, in-situ stress gradient, and coal rank.

[0065] Define the vector of physical properties and stress indices: .

[0066] 3) Engineering modification index E includes: number of fracturing stages, amount of sand added, amount of liquid, construction discharge, and initial fracturing pressure.

[0067] Define the engineering modification index vector: .

[0068] 4) Dynamic response index D includes: daily gas volume, daily water volume, bottom hole pressure, pressure recovery parameters, and gas production curve characteristic parameters.

[0069] Define a dynamic response index vector: .

[0070] Each category of indicators consists of multiple sub-indicators, forming a structured indicator set: This indicator system is used to support subsequent grey relational analysis and fuzzy comprehensive evaluation.

[0071] Step S3: Grey Relational Analysis

[0072] The average index of high-yield wells within the block is used as a reference sequence. Construct a comparison sequence: ,

[0073] 1) Difference sequence calculation: Calculate the difference between each index of each well and the reference sequence. ;

[0074] 2) Correlation coefficient The calculation is performed using the standard form employed by grey system theory: Among them, the resolution coefficient Used to adjust for the significance of differences in correlation coefficients, the value is usually between 0 and 1. In order to achieve a balance between correlation resolution and stability.

[0075] 3) Grey relational degree Calculate the average correlation coefficient for each well: , Used to characterize the importance of key factors controlling production capacity. This characterizes the degree of consistency between the characteristics of this well and those of high-yield wells, and is used for subsequent weight calculations.

[0076] Step S4: Fuzzy Comprehensive Evaluation

[0077] 1) Membership Function Construction: For each indicator, a triangular or trapezoidal membership function is used based on its data distribution to map the numerical indicator to multiple levels (such as low, medium, and high) of fuzzy membership. For example, the form of a trapezoidal membership function is:

[0078]

[0079] The triangular membership function is a special case of the trapezoidal function, meaning that when b=c, the function always takes the value 1 on the interval [b, c]. The boundary parameters (a, b, c, d) of the membership function are automatically determined based on the quantiles of historical samples: when the sample size N≥30, the 10%, 30%, 70%, and 90% quantiles of historical samples are used as the points a, b, c, and d for the trapezoidal function, respectively, to achieve dynamic adaptation to the data distribution; when the sample size N<30, a sliding window sample is constructed using data from adjacent wells within the past 3 years; if it is still less than 30, expert experience parameters are used as the initial boundary, and automatically corrected in subsequent dynamic updates.

[0080] 2) Membership matrix of indicators: The membership degrees of all indicators form a matrix R:

[0081]

[0082] 3) Weight vector

[0083] According to the fusion rules, it is obtained from the grey relational degree. Obtained by the entropy weight method Then, they are fused together: .

[0084] Among them, the fusion coefficient Used to adjust the relative importance of the two weights, after multiple experiments, when At the same time, it can better balance the identification of the main controlling factors (grey relational analysis) and the amount of information in the indicators (entropy weight).

[0085] 4) Evaluation results (production potential score) were obtained using a weighted fuzzy comprehensive method: This yields a production potential score for each well.

[0086] Step S5: Grading of Production Potential

[0087] Based on ratings Threshold execution well classification: High-potential well; Medium potential well; Low-potential wells. This classification result is used for subsequent process screening.

[0088] Step S6: Process Package Matching

[0089] The system has multiple built-in process packages, including "refracking process package, drainage system optimization package, unblocking or water control package, and well network interference control package".

[0090] Based on the characteristics of grey relational With rating Values ​​and indicators are used as logical judgment conditions to select appropriate process packages:

[0091] Refracking process package: Decision logic: Measures: Add sand, fill holes, and use slug fracturing.

[0092] Optimization package for sampling and extraction system: When the dynamic response correlation degree is ≥0.7: adopt: step-by-step differential pressure control; stable sampling and extraction; reverse sampling and extraction.

[0093] Unblocking or water control treatment package: When the correlation of abnormal water production is ≥0.75: use: acidification; cleaning fluid replacement; nano-interface regulator.

[0094] Well network interference control package: When the correlation of inter-well interference index is ≥0.7: adopt: neighboring well collaborative pressure stabilization; pressure control production; competitive drainage regulation.

[0095] The main measures of the refracking process package include: proppant strength: 80–150 t / m; number of fracturing stages: 5–10 stages; flow rate: 2.0–5.0 m³ / min; fluid volume: 800–1600 m³; temporary plugging and diverting agent: 40–80 mesh fiber or 20–40 mesh proppant can be selected; post-pressure stabilization and drainage strategy: initial bottomhole flowing pressure is controlled at 60–75% of the original pressure. All the above parameters are within feasible ranges, and engineers can adjust them within these ranges according to the specific block.

[0096] The main measures of the drainage system optimization package include: step-by-step differential pressure control: differential pressure is gradually increased, with each level increasing by 0.3–0.7 MPa; stable drainage: stable casing pressure range: ±0.15 MPa; reverse drainage: reverse drainage differential pressure: 0.2–0.5 MPa; duration 24–72 h.

[0097] The main measures of the unblocking or water control treatment package include: cleaning fluid replacement volume: 20–60 m³; acidification fluid type: weak acid (3–8 wt% acetic acid or citric acid), slow-release acid, or cleaning acid; nano-interface regulator concentration: 0.05–0.3 wt%, used to improve water and liquid lock conditions; precipitation control measures: implemented when daily water production is >60 m³ and the water-to-air ratio is >0.015.

[0098] The main measures of the well network interference control package include: pressure difference control between adjacent wells: the maximum difference shall not exceed 0.8 MPa; coordinated pressure stabilization mode: the target difference of casing pressure between adjacent wells shall not exceed 0.2 MPa; competitive drainage control: when the influence coefficient is >0.6, the drainage intensity of high-yield wells shall be reduced by 10-20%.

[0099] Step S7: Optimize the construction sequence

[0100] Based on production potential scoring and key control factors Calculate priority index :

[0101]

[0102] in: The correlation degree of key control factors; its value is the average correlation degree of the top 3 indicators in the grey correlation degree ranking (i.e., the three factors that have the greatest impact on production capacity); As a weighting factor, Take 0.4–0.6.

[0103] according to The values ​​are sorted from high to low, and the sorted values ​​form a treatment and construction sequence.

[0104] Step S8: Dynamic Update

[0105] After the treatment is implemented, the system continuously monitors production dynamic data for 7 to 90 days. An update is triggered when one of the following conditions is met: the production growth rate is less than 10% on the 30th day of monitoring after the modification; the daily water volume exceeds the historical maximum by 20%; or the bottom well pressure deviates from the predicted curve by more than 15%.

[0106] The system automatically re-executes steps S1 to S7 until the change in the production potential score between two consecutive evaluations is less than 5%.

[0107] Update mechanism: Re-collection Recalculate and ; Readjust membership function parameters a, b, c, d; Update Re-output and .

[0108] Through the cyclical execution of steps S1 to S8 above, when any treatment well meets the dynamic update trigger condition after the implementation of the measures, the system automatically collects the latest production dynamic data (including daily gas production, daily water production, bottom hole flowing pressure, casing pressure, and pressure recovery curves for 7 to 90 days after treatment), re-executes data preprocessing, grey relational analysis, and fuzzy comprehensive evaluation, and automatically updates the following core parameters:

[0109] The correlation sequence between the grey relational coefficient and each indicator;

[0110] Membership function boundary parameters (a, b, c, d);

[0111] Indicator weight vector (The grey relational weights and entropy weights are re-integrated).

[0112] Single-well production potential rating and priority index ;

[0113] The updated model is used to guide the re-evaluation of the potential of low-yield wells, the re-matching of process packages, and the optimization of the construction sequence in the next round, thereby realizing a complete closed-loop optimization mechanism of "evaluation-decision-implementation-feedback-re-evaluation".

[0114] This mechanism ensures that the governance plan adapts and adjusts itself to the dynamic evolution of production in the block, thereby continuously improving the targeting, timeliness, and overall economic benefits of production enhancement measures, and ultimately forming a closed-loop governance mechanism for intelligent production enhancement of low-yield coalbed methane wells that can be dynamically iterated and self-improving.

[0115] Example 2: Application of Vertical Well Groups in Baiyanghe Mining Area, Fukang, Xinjiang

[0116] This embodiment focuses on 198 vertical coalbed methane wells put into production in the Baiyanghe mining area of ​​Fukang, Xinjiang in recent years. Among them, 36 are high-yield wells (stable daily gas production ≥ 2500 m³ / d), 89 are low-yield wells (daily gas production < 700 m³ / d and stable production period < 150 days), and the rest are medium-yield wells. The coal seams are mainly the Jurassic Badaowan Formation, with a burial depth of 520–950 m and a coal thickness of 3.8–7.2 m.

[0117] Step S1: Data Acquisition and Preprocessing

[0118] Five wells were randomly selected, and 21 indicators were collected from each well. These five wells covered three types: high-yield, medium-yield, and low-yield. Some of the raw data are shown in Table 1, illustrating the differences between the indicators in the actual data and providing a data foundation for subsequent preprocessing and analysis.

[0119] Table 1. Examples of partial raw data (5 wells randomly selected)

[0120] Hash Gas content (m³ / t) Coal thickness (m) Burial depth (m) Permeability (mD) Sand addition (t) Fracturing segment number Daily gas production (m³ / d) Daily water production (m³ / d) BK-07 (high yield) 19.6 6.8 612 0.76 712 9 4280 8.2 BK-89 (low yield) 11.2 4.3 891 0.08 385 6 280 46.5 BK-112 (medium yield) 15.8 5.5 735 0.31 520 7 1120 21.3 BK-156 (low yield) 9.8 4.1 928 0.05 410 5 190 58.7 BK-203 (high yield) 18.3 6.4 658 0.62 680 8 3150 12.6

[0121] As shown in Table 1, high-yield wells (BK-07, BK-203) outperform low-yield wells (BK-89, BK-156) in key indicators such as gas content, coal thickness, permeability, and sand addition, which directly reflects the comprehensive impact of geological and engineering factors on production capacity.

[0122] Missing values ​​were handled as follows: 7.3% of the permeability was missing, and KNN (k=5) was used for imputation; linear interpolation was used for missing daily water production data.

[0123] Outlier handling: using Two outliers with sand additions exceeding 880 t were removed and replaced with the 95th percentile of 850 t for that block.

[0124] Normalization: Min-Max normalization is used for indicators such as gas content, coal thickness, permeability, and sand addition, which are considered "better the larger the better"; Minimization is used for indicators such as burial depth, in-situ stress gradient, and daily water production, which are considered "better the smaller the better"; and Z-score normalization is used for the decline rate of the gas production curve.

[0125] After processing, a standardized evaluation index matrix of 198×21 is formed. .

[0126] Step S2: Construct an evaluation index system for low-yield coalbed methane wells

[0127] Based on the geological and engineering characteristics of the Baiyanghe mining area, an evaluation index system comprising 21 indicators across 4 categories was ultimately established, as detailed in Table 2. This system comprehensively covers the resource, physical, engineering, and dynamic factors affecting coalbed methane well productivity, providing structured input for subsequent grey relational analysis and fuzzy evaluation.

[0128] Table 2 Evaluation Index System for Baiyanghe Mining Area

[0129] Primary indicators Secondary indicators (21 in total) Indicator attributes Resource-based indicators (R) Gas content, coal thickness, burial depth Positive / negative indicators Physical and stress-based indicators (P) Permeability, porosity, in-situ stress gradient, coal rank Positive / negative indicators Engineering modification indicators (E) Fracturing segment number, sand addition, sand addition intensity (t / m), fluid volume, construction discharge, initial fracture pressure Positive indicators Dynamic response indicators (D) Daily gas production, daily water production, bottom-hole flowing pressure, casing pressure, bottom-hole flowing pressure decline rate, peak time of gas production curve, gas production curve decline rate, pressure recovery rate Positive / negative indicators

[0130] This indicator system includes both static geological parameters (such as gas content and coal rank) and dynamic production response (such as the decline rate of the gas production curve), realizing a multi-dimensional comprehensive evaluation from geology to engineering to production.

[0131] All indicators form a standardized evaluation indicator matrix of 198×21. This provides a unified data foundation for subsequent grey relational analysis and fuzzy evaluation.

[0132] Step S3: Grey Relational Analysis

[0133] The daily average gas production sequence of 36 high-yield wells was used as a reference sequence. resolution coefficient Grey relational analysis was performed on various indicators of 89 low-yield wells. The indicators were sorted in descending order of relational degree to obtain the ranking of the main factors controlling production capacity. The top 8 indicators are shown in Table 3.

[0134] Table 3 Ranking of the main factors controlling the production capacity of low-yield wells in Baiyanghe Mining Area (Top 8)

[0135] Ranking Indicator Grey correlation degree Belonging category 1 Sand addition intensity (t / m) 0.927 Engineering modification 2 Permeability (mD) 0.889 Physical properties 3 Daily water production (m³ / d) 0.862 Dynamic response 4 Gas content (m³ / t) 0.847 Resource-based 5 Fracturing segment number 0.831 Engineering modification 6 Construction discharge (m³ / min) 0.809 Engineering modification 7 Burial depth (m) 0.791 Resource-based 8 Bottom-hole flowing pressure decline rate 0.773 Dynamic response

[0136] Table 3 shows that engineering modification indicators (sand addition intensity, number of fracturing stages, and construction discharge rate) and reservoir physical property indicators (permeability) play a dominant role in the production capacity control of low-yield wells in this mining area. This result verifies that grey relational analysis can effectively identify the factors with the most significant impact on production capacity, providing a key decision-making basis for subsequent process package matching.

[0137] Step S4: Fuzzy Comprehensive Evaluation

[0138] The weights are obtained by linearly fusing grey relational weights and entropy weights, resulting in a final weight vector. (First 8 items): =[0.148, 0.132, 0.119, 0.106, 0.097, 0.089, 0.081, 0.074, …]

[0139] The membership functions employ a combination of trapezoidal and triangular bounding boxes, with their boundary parameters automatically determined based on the distribution characteristics of historical data from recent years for this block. Some indicator membership function boundaries are shown in Table 4.

[0140] Table 4. Membership function boundaries for some indicators (units and values)

[0141] Indicator Low Medium High Gas content [0, 12] [10, 18] [16, 25] Permeability [0, 0.1] [0.08, 0.4] [0.35, 1.2] Sand addition intensity [0, 60] [55, 100] [95, 150] Daily water production [80, +∞) [30, 90] [0, 40]

[0142] As shown in Table 4, the membership function boundaries are dynamically set based on the actual data distribution. For example, the boundary for the "high" gas content level is [16, 25] m³ / t, which matches the gas content characteristics of high-yield wells in this mining area. This data-driven boundary determination method ensures that the fuzzy evaluation accurately reflects the actual situation of this block, avoiding the subjectivity of manually setting boundaries.

[0143] The production potential scores of 89 low-yield wells were calculated. Range 0.18–0.91, average 0.61.

[0144] Steps S5-S6: Potential Classification and Process Package Matching

[0145] Based on the production potential score calculated in step S4, and according to the production potential grading standard ( High-potential well; Medium potential well; 89 low-yield wells were classified into categories (low-potential wells). Simultaneously, based on the grey relational analysis of each well obtained in step S3 and the above classification results, and according to the process package selection rules, a corresponding production enhancement process package was matched for each well. The overall statistical results of classification and process package matching are shown in Table 5.

[0146] Table 5. Statistics on the Matching of Classification Results and Measures

[0147] Potential level Scoring interval Well number (mouth) Proportion Main matching technology package High-potential well S≥0.75 18 20.2% Re-fracturing (100%) + production optimization (72%) Medium-potential well 0.5≤S<0.75 48 53.9% Production optimization (88%) + plugging removal or water control (65%) Low-potential well S<0.5 23 25.9% Postpone treatment or convert to monitoring well

[0148] As shown in Table 5, among the 89 low-yield wells, 18 were identified as high-potential wells (20.2%), and their main remediation direction is re-fracture combined with drainage optimization; 48 wells (53.9%) have medium potential, and their main remediation direction is drainage optimization and unblocking or water control measures; 23 wells (25.9%) have low potential, and it is recommended to postpone remediation or convert them into monitoring wells. Through quantitative scoring and rule matching, the potential of low-yield wells can be effectively differentiated, and clear remediation decisions can be provided for each type of well, avoiding the blindness of manual experience.

[0149] Step S7: Optimize the construction sequence

[0150] Priority index The calculation formula is: , the first 10 wells Values ​​range from 0.81 to 0.93, and treatment is carried out sequentially.

[0151] Step S8: Post-treatment effects and dynamic model updates

[0152] After implementing "temporary plugging, diversion, re-fracture, and proppant addition with a strength of 120 t / m and stepped pressure control for drainage" on 18 high-potential wells:

[0153] The average daily gas production per well increased by 1980 m³, representing a growth rate of 386%.

[0154] The highest daily gas production from a single well reached 8100 m³;

[0155] Even after 90 days of treatment, the average daily gas production remained above 3400 m³.

[0156] Data was recollected 30 days after treatment, and the model was automatically updated: the correlation between sand addition intensity and treatment volume increased to 0.941, the correlation between construction discharge volume and daily water production decreased to 0.804. After the update, the model achieved a prediction accuracy of 94.7% for the next round of 12 wells, forming a closed-loop optimization.

[0157] The grey relational analysis in this application is not only used to identify key factors affecting production capacity, but also forms a coupling relationship with the process package matching module. Through the logical mapping between the correlation coefficient and process requirements, a closed-loop chain from "factor identification to process matching" is realized, which is different from the traditional practice in the prior art where grey relational analysis is only used for single-factor identification.

[0158] The dynamic membership function boundary combined with the fusion method of grey relational weight and entropy weight in this application enables the evaluation model to automatically adjust the membership function and weight vector according to the geological attributes and dynamic response of different blocks, avoiding the bias caused by manual experience setting, which is an improvement at the parameter structure level.

[0159] The construction priority index proposed in this application Through potential scoring Correlation with key factors The combination of these factors forms a quantifiable multi-well construction sequencing logic. Compared with the existing technology that relies on human experience to determine the construction sequence, this formula has a clear calculation structure and reproducibility, and represents a structural optimization of the construction decision-making process.

[0160] Therefore, the three combinations proposed in this application constitute a linkage mechanism between grey relational analysis, fuzzy evaluation, process matching, construction sequencing and dynamic feedback, thereby producing a synergistic effect superior to the prior art.

[0161] Although embodiments of this application have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of this application, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A coalbed methane production enhancement technology based on fuzzy evaluation and grey relational analysis, characterized in that, Includes the following steps: S1: Data Acquisition and Preprocessing. This involves acquiring geological parameters, reservoir properties, fracturing engineering parameters, drainage and production process parameters, and production dynamics data for the target well. Missing values ​​are filled in, outliers are removed, and the data is normalized to form an evaluation index matrix. : S2: Construct an evaluation index system for low-yield coalbed methane wells. The index system includes four categories: resource indicators, reservoir physical properties and stress indicators, engineering transformation indicators, and production dynamic response indicators. S3: Perform grey relational analysis, using the high-yield well production capacity sequence as a reference sequence, calculate the correlation coefficient and correlation degree of each indicator relative to the reference sequence, and obtain the correlation degree ranking of the main control factors of production capacity. S4: Perform fuzzy comprehensive evaluation, determine the index weights based on the grey relational results, and calculate the fuzzy membership matrix of each index based on the membership function to obtain the production potential score of a single well. S5: Classify the production potential of low-yield wells and classify the target wells into high-potential wells, medium-potential wells or low-potential wells according to the production potential score. S6: Match production enhancement treatment process packages. Based on the potential level and its main controlling factors, select at least one of the following: refracking process package, drainage system optimization package, unblocking or water control treatment package, well network interference treatment package, or water-bearing treatment package. S7: Develop a construction implementation sequence, calculate a priority index based on the production potential score and the correlation between key factors, and optimize the treatment order of multiple wells; S8: Dynamic updates after implementation; After the governance is implemented, the grey relational degree and fuzzy evaluation weights are adjusted according to the changes in well gas production, water production and pressure, so as to update and optimize the governance model.

2. The coalbed methane low-yield well production enhancement treatment technology based on fuzzy evaluation and grey relational analysis according to claim 1, characterized in that: The data normalization process in step S1 includes a combination of Z-score normalization and Min-max normalization, and the normalization method is automatically selected according to the indicator type.

3. The coalbed methane low-yield well production enhancement technology based on fuzzy evaluation and grey relational analysis according to claim 1, characterized in that: The evaluation index system in step S2 includes the following sub-indicators: Resource indicators: gas content, coal thickness, burial depth; Physical and stress parameters: permeability, porosity, in-situ stress gradient, coal rank; Engineering modification indicators: number of fracturing stages, amount of sand added, amount of liquid, construction discharge rate, and initial fracturing pressure; Dynamic response indicators: daily gas volume, daily water volume, bottom hole pressure, gas production curve morphology parameters, and pressure recovery parameters.

4. The coalbed methane low-yield well production enhancement treatment technology based on fuzzy evaluation and grey relational analysis according to claim 1, characterized in that: The grey relational degree calculation in step S3 uses the relational coefficient formula:

5. Among them: The resolution coefficient is set to 0.

5. To compare the absolute difference between the sequence and the reference sequence, and These are the minimum and maximum values ​​among all differences; And calculate the grey relational degree based on this. The sequence of controlling factors is obtained, and its calculation formula is as follows:

6. Among them: The number of evaluation indicators, Used to characterize the overall similarity between the i-th well and the reference sequence and the importance of its main production control factors.

7. The coalbed methane low-yield well production enhancement treatment technology based on fuzzy evaluation and grey relational analysis according to claim 1, characterized in that: The fuzzy membership function in step S4 is a triangular or trapezoidal function, and the upper and lower limits of the membership degree are automatically adjusted according to the block data.

8. The coalbed methane low-yield well production enhancement treatment technology based on fuzzy evaluation and grey relational analysis according to claim 1, characterized in that: The index weight vector in step S4 Grey relational weights Weights of the entropy weight method Linear fusion yields: ， in: .

9. The coalbed methane low-yield well production enhancement treatment technology based on fuzzy evaluation and grey relational analysis according to claim 1, characterized in that: The production potential grading criteria in step S5 are as follows: High-potential well; Medium potential well; Low-potential well.

10. The coalbed methane low-yield well production enhancement treatment technology based on fuzzy evaluation and grey relational analysis according to claim 1, characterized in that: The process package selection rules in step S6 include: When the grey relational degree of fracturing-related indicators is ≥ 0.8 and the production potential score is ≥ 0.75, a re-fracturing process package is matched. When the correlation of dynamic response indicators is ≥ 0.7, the matching and sampling system optimization package is applied. When the correlation coefficient of abnormal water production indicators is ≥ 0.75, a deblocking or water control treatment package is matched. When the correlation of inter-well interference indicators is ≥ 0.7, a well network interference mitigation package is matched.

11. The coalbed methane low-yield well production enhancement treatment technology based on fuzzy evaluation and grey relational analysis according to claim 1, characterized in that: The well sequence priority index in step S7 Calculate using the following formula:

12. Among them: Score the potential for increased production; Correlation degree of key controlling factors; The value ranges from 0.4 to 0.

6.

13. The coalbed methane low-yield well production enhancement treatment technology based on fuzzy evaluation and grey relational analysis according to claim 1, characterized in that: The dynamic update in step S8 includes: Based on the changes in gas production, water production, and bottom hole pressure 7–90 days after treatment, the grey relational degree is recalculated. The membership function boundary is automatically adjusted based on the updated indicator data. Update weight vector This forms a new production potential score; This serves as a guide for the implementation of the next round of governance measures, forming a closed-loop optimization mechanism.

Images