Deep coal reservoir top and bottom plate fracturing packer well logging evaluation method

Abstract

The invention provides a deep coal reservoir top and bottom plate fracturing packer well logging evaluation method. Comprising the following steps: step 1, determining a lithology-thickness index; 2, determining a fracture development index; 3, determining the stress difference between the roof and floor and the coal seam; 4, determining a sound wave impedance difference coefficient; 5, determining the fracture extension pressure gradient; 6, building a top and bottom plate fracturing packer evaluation black box model based on the random forest; 7, determining a deep coal reservoir top and bottom plate fracturing packer risk grade standard; 8, deep coal reservoir fracturing construction is guided, and yield increase and efficiency improvement are achieved. The method provided by the invention provides a solid technical support for evaluating the packer property of the top and bottom plates of the deep coal reservoir by combining geophysical logging information with a machine learning algorithm, so that the overall development effect of the deep coal bed gas reservoir is promoted, and the method has good popularization and application values.

Description

Technical Field

[0001] The present invention relates to the technical field of reservoir evaluation in the development of unconventional coalbed methane reservoirs; in particular, it relates to a logging evaluation method for the fracturing sealing performance of the roof and floor of deep coal reservoirs. Background Art

[0002] Deep coalbed methane resources have gradually become the focus of current exploration and development due to successive high-yields. The evaluation of the effectiveness and sealing performance of deep coal seams and their roof and floor is a key link among them. The free gas content in deep coal reservoirs is relatively high, but the sealing performance of the roof and floor directly affects the fracturing effect and the preservation and migration of free gas. A poor-sealing roof may lead to the flow-through of fracturing fluid or gas escape, which not only reduces the recovery rate but may also cause environmental problems. Through logging evaluation, the quality of the sealing performance can be accurately identified, guiding the optimization of the fracturing process and improving the gas production per well.

[0003] The existing evaluations of the sealing performance of unconventional gas reservoirs mainly target the middle and shallow coal seams in the conventional oil and gas fields. First, the traditional logging methods have limited ability to identify complex lithologies and fractures, failing to achieve the fine quantitative dynamic characterization of the lithologies of the roof and floor, being difficult to distinguish lithologies with similar physical properties, and prone to misjudgment of the sealing performance. Second, the existing methods have insufficient quantitative analysis of the mechanical properties of the roof and floor, affecting the accuracy of the dynamic prediction of the sealing performance. Third, the multi-index fusion modeling based on lithology, fracture parameters, mechanical properties, etc. is not comprehensively considered, resulting in a large error in the risk level evaluation of the sealing performance of the roof and floor and lacking a general quantitative index. Therefore, it is urgent to clarify the main controlling factors of the sealing performance of the roof and floor and establish a logging evaluation method for the fracturing sealing performance of the roof and floor of deep coal reservoirs based on machine learning algorithms to provide technical support for perforation section selection and fracturing construction.

[0004] Existing conventional methods cannot meet the requirements of logging evaluation for the fracturing sealing performance of the roof and floor of deep coal reservoirs. In the conventional oil and gas field, a patent with the patent number CN 202210673198.4 and the title "Method, Device and System for Establishing and Evaluating a Sealing Performance Evaluation Model for Carbonate Rock Interlayers" discloses a method for evaluating the sealing performance of carbonate rock interlayers through numerical simulation means. It is a method that, based on the barrier time of carbonate rock interlayers, screens multiple sensitive parameters from the possible influencing parameters of interlayer sealing performance, determines multiple groups of calculation examples to form a calculation example library, and each group of calculation examples includes the values of each sensitive parameter and the corresponding interlayer barrier time; uses the calculation example library to train a selected machine learning model to obtain an evaluation model for evaluating the sealing performance of carbonate rock interlayers; and uses the sealing time as the characterization parameter of interlayer sealing performance to reasonably and quantitatively evaluate the sealing performance of carbonate rock interlayers. The model uses permeability as the main discrimination parameter, but the pore-throat structure of carbonate rock interlayers is complex (throat radius < 0.5 μm), and static permeability testing is difficult to accurately characterize the dynamic sealing ability, which may underestimate the actual risk of sealing failure. A patent with the patent number CN201910533536.2 and the title "A Method for Segmenting and Fracturing Horizontal Wells Drilled Along the Roof and Floor of Coal Seams" discloses a method of directional perforation and segmented fracturing. It is a method that uses the roof of the structural coal seam as the target layer for gas drainage, drills a horizontal wellbore at a specific position therein, perforates the roof to communicate the horizontal wellbore with the coal seam; the purpose of strengthening and exploitation is to implement strengthening transformation on the roof and coal seam, so as to extract and collect coalbed methane in the structural coal seam. Although this method can maintain the original fractures in the coal seam and facilitate subsequent coal seam mining, the ground stress difference between the roof and floor is not fully considered. If the relationship between the vertical stress and the horizontal principal stress does not meet a specific threshold, the fracture propagation across layers is limited, and the cross-interface sealing performance is difficult to guarantee.

[0005] Practice has found that geophysical logging information can effectively reflect the reservoir quality and fracturability of deep coal seams and their roof and floor, and has low cost and good effect. The evaluation of the fracturing sealing performance of the roof and floor needs to comprehensively consider geological engineering indicators, including lithology-thickness characteristics, fracture development degree, stress difference level, etc. The purpose is not only to seal the escape of coalbed methane, but also to control the leakage of construction pressure during the fracturing operation and improve the fracturing efficiency.

[0006] Accordingly, it is necessary to construct an index reflecting lithology-thickness characteristics for the roof and floor of deep coal reservoirs, establish a fracture development index method, construct a stress difference model between the roof and floor and the coal seam, construct an acoustic impedance difference coefficient model, and establish a fracture extension pressure gradient method. Based on the above five basic indicators, a machine learning algorithm is introduced to construct a logging evaluation model for the fracturing sealing performance of the roof and floor of deep coal reservoirs, providing a quantitative basis for predicting the dynamic risk level of the sealing performance. Finally, a risk classification and evaluation standard is established based on the production capacity, fracturing construction parameters, and fracturing monitoring data corresponding to single wells in the block. From the existing methods, there is still no method that uses geophysical logging data and combines machine learning methods to evaluate the fracturing sealing performance of the roof and floor of deep coal reservoirs, which brings many inconveniences to the evaluation of the fracturing sealing performance of the roof and floor of deep coal reservoirs. Summary of the Invention

[0007] The object of the present invention is to provide a logging evaluation method for the fracturing sealing performance of the roof and floor of deep coal reservoirs.

[0008] The present invention is realized through the following technical solutions:

[0009] The present invention relates to a logging evaluation method for the fracturing sealing performance of the roof and floor of deep coal reservoirs, including the following steps: Step 1, determining the lithology-thickness index: The lithology-thickness index is determined by using formulas (1), (2), and (3):

[0010] Where the roof and floor are limestone:

[0011] (1)

[0012] Where the roof and floor are mudstone:

[0013] (2)

[0014] Where the roof and floor are sandstone:

[0015] (3)

[0016] In the above formulas (1), (2), and (3):

[0017] I L is the lithology-thickness index of limestone, dimensionless; I S is the lithology-thickness index of mudstone, dimensionless; I SS is the lithology-thickness index of sandstone, dimensionless; ω ρb is the weight of compensated density, dimensionless; ω GR is the weight of natural gamma, dimensionless; ω Δtc is the weight of acoustic travel time, dimensionless; ρ bnorm is the standardized compensated density, dimensionless; GR normis the standardized natural gamma, dimensionless; Δt cnorm is the standardized acoustic travel time difference, dimensionless; H L is the limestone thickness, m; H S is the mudstone thickness, m; H SS is the sandstone thickness, m;

[0018] Step 2, determine the fracture development index: Use Equation (7) to determine the fracture development index:

[0019] (7)

[0020] In Equation (7): I f is the fracture development index, dimensionless; R lld is the deep lateral resistivity, Ω·m; R lls is the shallow lateral resistivity, Ω·m;

[0021] Step 3, determine the stress difference between the roof and floor and the coal seam: Use Equation (8) to determine the stress difference between the roof and floor and the coal seam:

[0022] (8)

[0023] In the formula: Δσ is the stress difference between the roof and floor and the coal seam, MPa; σ Htb is the maximum horizontal principal stress of the roof and floor, MPa; σ Hc is the maximum horizontal principal stress of the coal seam, MPa;

[0024] Step 4, determine the acoustic impedance difference coefficient: Use Equation (14) to determine the acoustic impedance of the roof and floor and the coal seam:

[0025] (14)

[0026] In Equation (14): D p is the acoustic impedance difference coefficient, %; Z tb is the acoustic impedance of the roof and floor, g / cm 3 ·m / s; Z c is the acoustic impedance of the coal seam, g / cm 3 ·m / s;

[0027] Step 5, determine the fracture propagation pressure gradient: Use Equation (16) to determine the fracture propagation pressure gradient:

[0028] (16)

[0029] In Equation (16): G p is the fracture propagation pressure gradient, MPa / 100m; S t is the tensile strength, MPa; D ep is the formation depth, m;

[0030] Step 6, establish a black-box model for evaluating the fracturing sealing performance of the roof and floor based on random forest:

[0031] Take five indicators, namely the lithology-thickness index, fracture development index, stress difference between the roof and floor and the coal seam, acoustic impedance difference coefficient, and fracture extension pressure gradient, as independent variables Xi; determine the risk level types (Type I, Type II, Type III) of the fracturing sealing performance of the roof and floor based on the actual production capacity, geological characteristics, fracturing construction parameters, and fracturing monitoring data of a large number of wells in the study area, and quantify it as the dependent variable Y by comprehensively considering the fracturing fracture scale and post-fracturing production capacity;

[0032] The classification standard for the risk types of the fracturing sealing performance of the roof and floor corresponding to the value of the dependent variable Y: when Y > 16.8, the sealing level is Class I safety; when 14.2 < Y < 16.8, the sealing level is Class II risk; when Y < 14.2, calibrate the prediction results according to this standard;

[0033] Based on the random forest algorithm, train a large number of data sets to form a black-box model for evaluating the fracturing sealing performance of the roof and floor, and continuously optimize the model by adjusting the optimal combination of parameters; finally, apply the optimized model to predict new data to comprehensively test the reliability of the model.

[0034] Step 7, determine the risk level standard for the fracturing sealing performance of the roof and floor of deep coal reservoirs: Based on the determination process of the dependent variable Y in Step 6, establish a quantified risk level type table for the fracturing sealing performance of the roof and floor, so as to clarify the risk levels corresponding to the five basic indicators in different distribution intervals, and provide comprehensive technical support for fracturing operations.

[0035] Step 8, guide the fracturing construction and production increase and efficiency improvement of deep coal reservoirs.

[0036] Preferably, in Step 1, the determination method of the GR norm、 Δt cnorm、 ρ bnorm is as follows:

[0037] (4)

[0038] (5)

[0039] (6)

[0040] In the formula: ρ b is the compensated density of the roof and floor, g / cm 3 ; GR is the natural gamma of the roof and floor, API; Δt c is the acoustic travel time of the roof and floor, μs / ft.

[0041] Preferably, in step 3, the formula (8) also involves other parameters, and their determination methods are as follows:

[0042] (9)

[0043] In formula (9): The maximum horizontal principal stress, MPa;

[0044] α is the effective stress coefficient, dimensionless; σ v is the vertical stress, MPa; P p is the coal seam pore pressure, MPa; ε H is the strain in the direction of the maximum horizontal principal stress, mm; ε h is the strain in the direction of the minimum horizontal principal stress, mm; E tb is the Young's modulus of the roof and floor, MPa; μ is the Poisson's ratio of the roof and floor, dimensionless;

[0045] Among them:

[0046] (10)

[0047] In formula (10): represents the overlying strata pressure, MPa;

[0048] TVD is the formation vertical depth of the unmeasured density log, m; is the average density of the formation in the unlogged section (usually taken as 2.31); i is the density log value, g / cm 3 ; is the vertical depth sampling interval, m;

[0049] (11)

[0050] In formula (11): represents the formation pore pressure, MPa;

[0051] P w is the hydrostatic injection pressure, MPa; Δt n is the acoustic travel time under normal compaction trend, μs / ft; nformation is the compaction exponent, generally taken as 1.2 - 3.0;

[0052] (12)

[0053] In formula (12): E tb is the Young's modulus of the roof and floor, MPa;

[0054] β is the multiplication factor, generally 9.290304×10 7 , if Δt cIn units of μs / m, it is 30.472197×10 7 , dimensionless;

[0055] (13)

[0056] In Equation (13): represents the Poisson's ratio, dimensionless;

[0057] Δt s is the shear wave slowness, μs / ft.

[0058] Preferably, in step 4, the parameter determination method in Equation (14) is as follows:

[0059] (15)

[0060] In Equation (15): ρ btb is the density log value of the roof and floor, g / cm 3 ; V ptb is the acoustic velocity of the roof and floor, m / s; ρ bc is the density log value of the coal seam, g / cm 3 ; V pc is the acoustic velocity of the coal seam, m / s.

[0061] Preferably, in step 5, the parameter determination method in Equation (16) is as follows:

[0062] (17)

[0063] In Equation (17): V sh is the shale content of the roof and floor, %.

[0064] The method involved in the present invention uses five evaluation indexes, namely the lithology-thickness index, the fracture development index, the stress difference between the roof and floor and the coal seam, the acoustic impedance difference coefficient, and the fracture extension pressure gradient, to construct an evaluation model for the fracturing sealing performance of the roof and floor based on the random forest algorithm. Based on the actual production capacity, geological characteristics, fracturing construction parameters, fracturing monitoring data, etc., a risk grading standard is established, so as to effectively predict the fracturing sealing performance of the roof and floor of deep coal reservoirs, in order to provide a solid technical support for clarifying the problem of the sealing performance of the roof and floor of deep coal reservoirs, and lay a theoretical foundation for improving the overall development effect of deep coalbed methane reservoirs by combining geophysical logging data with machine learning algorithm evaluation means. It has the characteristics of simple and practical methods.

[0065] The present invention has the following advantages:

[0066] (1)The present invention uses geophysical logging data combined with machine learning algorithms to evaluate the fracturing sealing performance of the roof and floor of deep coal reservoirs, fully exploiting the rock mechanical properties and fracture development characteristics contained in the geophysical logging data. For the roof and floor of deep coal reservoirs, five basic indicators are constructed, including the lithology-thickness index, fracture development index, stress difference between the roof and floor and the coal seam, acoustic impedance difference coefficient, and fracture extension pressure gradient. Then, based on the random forest machine learning algorithm, with the five basic indicators as independent variables X i , and based on the actual production capacity, geological characteristics, fracturing construction parameters, fracturing monitoring data, etc. of numerous single wells in the block, the risk level types (Type I, Type II, Type III) of the fracturing sealing performance of the roof and floor are determined. The resulting dataset is used as the dependent variable Y. The black-box model established by this method is trained based on big data and has obvious professionalism and accuracy.

[0067] (2)The method involved in the present invention provides a solid technical support for clarifying the problem of the sealing performance of the roof and floor of deep coal reservoirs, and lays a theoretical foundation for the evaluation method of combining geophysical logging data with machine learning algorithms, thereby improving the overall development effect of deep coalbed methane reservoirs, and has good popularization and application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 is a flow chart of the logging evaluation method for the fracturing sealing performance of the roof and floor of deep coal reservoirs involved in the present invention;

[0069] Figure 2 is a black-box model diagram of the logging evaluation method for the fracturing sealing performance of the roof and floor of deep coal reservoirs involved in the present invention;

[0070] Figure 3 is an evaluation result diagram of the risk level types of the fracturing sealing performance of the roof and floor of deep coal reservoirs of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0071] The present invention will be described in detail below with reference to specific embodiments. It should be noted that the following embodiments are only further descriptions of the present invention, but the protection scope of the present invention is not limited to the following embodiments.

[0072] Embodiment

[0073] This embodiment relates to a logging evaluation method for the fracturing sealing performance of the roof and floor of deep coal reservoirs, as shown in Figure 1 and includes the following steps:

[0074] Step 1: Determination of the lithology-thickness index:

[0075] The lithology-thickness index comprehensively reflects the inhibitory ability of the lithology characteristics and thickness of the roof and floor rocks on the vertical extension of fractures. Thick roof and floor rocks increase the fracture propagation resistance through the "stress barrier effect" and reduce the risk of communicating with aquifers or adjacent coal seams. In the fracturing of deep coal reservoirs, a higher lithology-thickness index (i.e., high strength and large thickness of the roof and floor rocks) is more conducive to improving the sealing performance. Limestone has high strength and relatively large brittleness, and fracture propagation is prone to extend within the limestone layer. However, if the thickness is sufficient, a high-stress barrier can be formed to inhibit fracture cross-layer. In well logging, the response characteristics of high density, low gamma, and low acoustic time difference should be prominent; mudstone has strong plasticity, and the energy at the crack tip is easily absorbed. However, thin mudstone is easily hydraulically fractured, while thick mudstone seals fractures through plastic deformation. In well logging, the response characteristics of high gamma, high density, and low acoustic time difference should be prominent; sandstone has strong heterogeneity, and fractures are prone to turn along the sand-mud interface. It is necessary to accumulate a sufficient thickness (such as the total thickness of multiple interbeds > 15m) to form an effective seal. In well logging, parameters such as gamma, density, and acoustic time difference should be in the intermediate values between limestone and mudstone.

[0076] Therefore, in the fracturing design, it is necessary to preferentially select areas with a larger lithology-thickness index of the roof and floor to ensure the vertical sealing of fractures. Based on this, this study intends to propose the following well logging calculation model for the lithology-thickness index, and use equations (1), (2), and (3) to determine the lithology-thickness index:

[0077] The roof and floor are limestone:

[0078] (1)

[0079] The roof and floor are mudstone:

[0080] (2)

[0081] The roof and floor are sandstone:

[0082] (3)

[0083] In the formula: I L is the lithology-thickness index of limestone, dimensionless; I S is the lithology-thickness index of mudstone, dimensionless; I SS is the lithology-thickness index of sandstone, dimensionless; ω DEN is the weight of compensated density, dimensionless; ω GR is the weight of natural gamma, dimensionless; ω AC is the weight of acoustic time difference, dimensionless; is the standardized compensated density, dimensionless; is the standardized natural gamma, dimensionless; is the standardized acoustic time difference, dimensionless; is the thickness of limestone, m;H S is the thickness of mudstone, m; H SS is the thickness of sandstone, m.

[0084] The method for determining the parameters in the above formulas (1), (2) and (3) is as follows:

[0085] (4)

[0086] (5)

[0087] (6)

[0088] In formula (4), (5), (6): ρ b is the compensation density of the top and bottom plates, g / cm 3 ; GR is the natural gamma of the top and bottom plates, API; Δt c is the acoustic time difference between the top and bottom plates, μs / ft.

[0089] Step 2: Crack Development Index:

[0090] The fracture development index is a comprehensive parameter for evaluating the degree of development of natural fractures or microcracks in the roof and floor of coal reservoirs. By integrating logging data such as the difference in deep and shallow lateral resistivity, acoustic wave time difference and density, the density, connectivity and extension scale of the fractures in the roof and floor can be quantified. When fractures develop in deep coal reservoirs, the acoustic wave time difference increases to a certain extent, and the compensation density decreases. When mud invades natural fractures in the near-wellbore area, the shallow lateral resistivity decreases significantly, but the change in deep lateral resistivity is not obvious, so the difference in deep and shallow lateral resistivity increases, showing an obvious double-track phenomenon.

[0091] Therefore, when fracturing deep coal reservoirs, priority should be given to areas with smaller fracture development indexes in the roof and floor plates, so as to utilize the natural barrier effect of dense rock mass to limit the vertical extension of fractures and ensure the effectiveness and safety of fracturing. Based on this, this study proposes the following fracture development index logging calculation model, and uses formula (7) to determine the fracture development index:

[0092] (7)

[0093] In formula (7): I f is the crack development index, dimensionless; Δt c is the acoustic time difference, μs / ft; R lld is the deep lateral resistivity, Ω·m; R lls is the shallow lateral resistivity, Ω·m; the other parameters have the same meanings as above.

[0094] Step 3: Determine the stress difference between the roof and floor and the coal seam:

[0095] The stress difference between the roof and floor and the coal seam is defined as the difference between the maximum horizontal principal stress of the roof and floor and the maximum horizontal principal stress of the coal seam. This parameter is the key barrier controlling the vertical extension of hydraulic fractures in deep coal reservoirs. When the stress difference is large, the high stress difference between the roof and floor and the coal seam will form a natural "stress shielding" effect, forcing the fractures to mainly extend horizontally, thus inhibiting the fractures from penetrating the roof and floor and enhancing the sealing property. If the stress of the roof and floor is significantly higher than that of the coal seam (>8 MPa), the energy required for the vertical extension of the fractures increases significantly, and the fracture height is often limited within the coal seam, reducing the risk of communicating with the aquifer or adjacent non-target layers. On the contrary, if the stress difference is small (<5 MPa), the stress barrier effect between the coal seam and the roof and floor weakens, and the energy of the fracturing fluid is easily driven to break through the lithological interface, resulting in uncontrolled vertical extension and reducing the efficiency of reservoir stimulation.

[0096] Therefore, during the fracturing operation, this stress difference should be as large as possible to strengthen the sealing property. Accordingly, this study intends to propose the following logging calculation model for the stress difference between the roof and floor and the coal seam, and use Equation (8) to determine the stress difference between the roof and floor and the coal seam:

[0097] (8)

[0098] In Equation (8): Δσ is the stress difference between the roof and floor and the coal seam, MPa; σ Htb is the maximum horizontal principal stress of the roof and floor, MPa; σ Hc is the maximum horizontal principal stress of the coal seam, MPa.

[0099] The determination methods of other parameters in Equation (8) are as follows:

[0100] (9)

[0101] In Equation (9): α is the effective stress coefficient, dimensionless; σ v is the vertical stress, MPa; P p is the pore pressure of the coal seam, MPa; ε H is the strain in the direction of the maximum horizontal principal stress, mm; ε h is the strain in the direction of the minimum horizontal principal stress, mm; E tb is the Young's modulus of the roof and floor, MPa; μ is the Poisson's ratio of the roof and floor, dimensionless;

[0102] Among them:

[0103] (10)

[0104] In Equation (10): TVD is the formation vertical depth of the unmeasured density logging, m; is the average density of the formation in the unlogged section (usually taken as 2.31); i is the density logging value, g / cm 3 ; is the vertical depth sampling interval, m.

[0105] (11)

[0106] In Equation (11): P w is the hydrostatic injection pressure, MPa; Δt is the measured acoustic travel time difference between the roof and floor, μs / ft;

[0107] Δt n is the acoustic travel time difference under normal compaction trend, μs / ft; n formation is the compaction index, generally taken as 1.2 - 3.0; the meanings of other parameters are the same as above;

[0108] (12)

[0109] In Equation (12): β is the conversion factor, generally 9.290304×10 7 (if Δt c is in the unit of μs / m, then it is 30.472197×10 7 ), dimensionless; the meanings of other parameters are the same as above;

[0110] (13)

[0111] In Equation (13): Δt s is the shear wave travel time difference, μs / ft; the meanings of other parameters are the same as above.

[0112] Step Four: Acoustic impedance difference coefficient:

[0113] The acoustic impedance difference coefficient quantifies the acoustic impedance difference between the roof and floor and the coal seam (i.e., the ratio of the difference between the acoustic impedance of the roof and floor and the acoustic impedance of the coal seam to the acoustic impedance of the roof and floor), reflecting the interfacial barrier effect in rock mechanical properties between the two. The larger this coefficient, the more significant the acoustic impedance difference between the roof and floor and the coal seam, usually corresponding to denser and more rigid roof and floor rocks (such as high - density limestone, dense sandstone), while the acoustic impedance of the coal seam is relatively low. In the fracturing of deep coal reservoirs, a higher acoustic impedance difference coefficient means that an obvious wave impedance interface is formed between the roof and floor and the coal seam. This interface can inhibit the vertical propagation of fracturing cracks through the stress difference and reduce the risk of cracks breaking through the roof and floor. If this coefficient is low, the impedance between the roof and floor and the coal seam is close, the interfacial barrier effect is weakened, and the fracturing cracks are likely to invade the roof and floor along the mechanically weak surface, resulting in seal failure or communication with the aquifer. Therefore, in the fracturing operation, it is necessary to preferentially select the roof and floor areas with a larger acoustic impedance difference coefficient and utilize its strong interfacial barrier characteristics to achieve vertical plugging of cracks. This study intends to propose the following logging calculation model for the acoustic impedance difference coefficient and determine the acoustic impedance difference coefficient using Equation (14):

[0114] (14)

[0115] In formula (14): D p is the acoustic impedance difference coefficient, %; Z tb is the acoustic impedance of the roof and floor, g / cm 3 ·m / s; Z c is the acoustic impedance of the coal seam, g / cm 3 ·m / s.

[0116] The determination method of the parameters in formula (14) is as follows:

[0117] (15)

[0118] In the formula: ρ btb is the logging value of the density of the roof and floor, g / cm 3 ; V ptb is the acoustic velocity of the roof and floor, m / s; ρ bc is the logging value of the density of the coal seam, g / cm 3 ; V pc is the acoustic velocity of the coal seam, m / s; the meanings of other parameters are the same as above.

[0119] Step Five: Fracture extension pressure gradient:

[0120] The fracture extension pressure gradient is characterized by the ratio of the sum of the minimum horizontal principal stress and the tensile strength to the depth, which represents the comprehensive resistance of the stress required for the vertical extension of the fracture in the roof and floor and the rock mass strength. The larger this parameter is, the stronger the ability of the roof and floor rock strata to resist fracture propagation and the better the sealing property. In the fracturing of deep coal reservoirs, a higher fracture extension pressure gradient means that the roof and floor have a high-stress background or strong tensile strength (such as dense limestone, thick mudstone), and a higher net pressure is required for the fracturing fracture to break through the interface, thus inhibiting the vertical penetration of the fracture and reducing the risk of communicating with the aquifer or adjacent reservoirs. If this gradient is low, the roof and floor lithology is weak (such as fractured sandstone, thin shale) or the effect of in-situ stress shielding is insufficient, and the fracturing fluid is likely to drive the fracture to break through the roof and floor, resulting in ineffective settlement of the proppant or induced interlayer crossflow.

[0121] Therefore, when designing fracturing, it is necessary to select the roof and floor areas with a larger fracture extension pressure gradient and utilize their natural high-resistance characteristics to achieve vertical sealing of fractures. This study intends to propose the following logging calculation model for the fracture extension pressure gradient and use formula (16) to determine the fracture extension pressure gradient:

[0122] (16)

[0123] In formula (16): G pis the fracture extension pressure gradient, MPa / 100m; S t is the tensile strength, MPa; D ep is the formation depth, m; the meanings of other parameters are the same as above.

[0124] The determination method of the parameters in Equation (16) is as follows:

[0125] (17)

[0126] In the formula: V sh is the shale content of the roof and floor, %, and the meanings of other parameters are the same as above.

[0127] Step 6: Determination of the fracturing sealing evaluation model for the roof and floor of deep coal reservoirs:

[0128] Taking five indicators including the lithology-thickness index, fracture development index, stress difference between the roof and floor and the coal seam, acoustic impedance difference coefficient, and fracture extension pressure gradient as the independent variable X i ; Based on the actual production capacity, geological characteristics, fracturing construction parameters, fracturing monitoring data, etc. of a large number of wells in the study area, determine the risk level types (Type I, Type II, Type III) of the fracturing sealing of the roof and floor. Quantify it as the dependent variable Y on the basis of comprehensively considering the fracturing fracture scale and post-fracture production capacity. The classification standard of the risk type of the fracturing sealing of the roof and floor corresponding to the value of the dependent variable Y: When Y > 16.8, the sealing level is Class I safety; when 14.2 < Y < 16.8, the sealing level is Class II risk; when Y < 14.2, the sealing level is Class III failure, and calibrate the prediction results according to the standard. Based on the random forest algorithm, train a large number of data sets to form a black-box model for evaluating the fracturing sealing of the roof and floor; see Figure 2 shown. Continuously optimize the model by adjusting the optimal combination of parameters, and finally use the optimized model to predict new data to comprehensively test the reliability of the model.

[0129] Step 7: Establish a risk level type table for the fracturing sealing of the roof and floor of deep coal reservoirs:

[0130] Based on the determination process of the dependent variable Y in Step 6, establish a quantified risk level type table for the fracturing sealing of the roof and floor, as shown in Table 1, so as to clarify the risk levels corresponding to the five basic indicators in different distribution intervals and provide comprehensive technical support for fracturing operations.

[0131] Table 1

[0132]

[0133] As can be seen from the data in Table 1 above, the risk level types of the fracturing and sealing properties of the roof and floor of deep coal reservoirs are divided into three categories, including Category I (safe), Category II (risk), and Category III (failure). Among them, Category I indicates that the roof and floor have strong integrity, mainly continuous mudstone / dense limestone, with basically no developed fractures, large rock stiffness, difficult elastic deformation, poor compressibility, limited longitudinal fracture propagation, and being effectively restricted within the coal seam, which is a typical safe sealing layer; Category II indicates that the roof and floor have general integrity, mainly sand-mud interbeds, general fracture development, medium rock stiffness, general anti-deformation ability, medium compressibility, increased randomness in fracture direction, but when the vertical stress advantage is not obvious, fractures may penetrate the roof and floor, which is a typical risk sealing layer; Category III indicates that the roof and floor have poor integrity, mainly sandstone / fractured limestone, higher fracture development, poor rock stiffness, poor anti-deformation ability, good compressibility, and fractures extending along the natural fracture network. Even when the vertical stress advantage is obvious, fractures can basically penetrate the roof and floor, which is a typical failure sealing layer.

[0134] Based on the established evaluation model and risk level standard for the fracturing and sealing properties of the roof and floor of deep coal reservoirs based on random forest, on the basis of compiling a processing and interpretation program, the fracturing and sealing properties of the roof and floor of the main deep coal reservoirs of each well in the study area are predicted and evaluated.

[0135] See Figure 3 shown in the evaluation result diagram of the risk level type of the fracturing and sealing properties of the roof and floor of the deep coal reservoir of Well X-2A. The depth of the deep coal seam and its roof and floor of this well is 1951 - 1987 m, with a total thickness of 36 m. Among them, the coal seam section is 1964.25 - 1970 m, with a thickness of 5.75 m. The lithologic profile interpreted by well logging shows that the fixed carbon content in the coal seam section exceeds 85%, the coal body structure is mainly primary structural coal, and there is no obvious hole enlargement phenomenon; the depth of the coal seam roof is 1951 - 1964.5 m, mainly limestone, with no obvious sand-mud interbed phenomenon, the thickness of continuous mudstone / dense limestone is 13.5 m, the natural gamma curve value is between 32 and 124.6 API, the density curve value is between 2.37 and 2.75 g / cm 3 between, the hole diameter shows no obvious hole enlargement, the spontaneous potential value is between 40.9 and 63.9 mV, the acoustic time difference value is between 51 and 101 μs / ft, the porosity is distributed between 1.2 and 3.4%, and the deep and shallow dual lateral resistivity has a slightly amplitude double-track phenomenon, distributed between 15.9 and 1434 Ω·m. It can be seen that the roof limestone layer still has heterogeneity and is relatively dense as a whole; the depth of the coal seam floor is 1970 - 1987 m, mainly pure mudstone and sand-mudstone, and the interbed phenomenon is relatively serious. The thickness of the sand-mud interbed and intercalated limestone is 17 m, the natural gamma curve value is between 40 and 281 API, the density curve value is between 2.14 and 2.81 g / cm 3Between them, the wellbore diameter shows obvious hole enlargement in some pure mudstone layers, and the hole enlargement rate exceeds 40%. The spontaneous potential value is between 67.6 and 77.4 mV, the acoustic travel time value is between 48.6 and 95.2 μs / ft, the porosity is distributed between 2.58 and 5.06%, and the deep and shallow dual laterolog resistivity also has a slightly amplitude double-track phenomenon, distributed between 24 and 282 Ω·m. Due to the large lithology change of the floor, the continuity is not as strong as that of the roof, and the heterogeneity is relatively strong.

[0136] After being predicted by the random forest algorithm and combined with the risk level type standard shown in Table 1, two risk types can be evaluated for the fracturing sealing of the top and bottom plates of this section, including 5 sections of Class I safety, 1951 - 1954.5 m, 1956.125 - 1958.38 m, 1959.375 - 1964.25 m, 1970 - 1975 m, 1982.5 - 1986 m, mainly composed of continuous limestone and part of marl. The lithology-thickness index is between 0.69 and 0.92, the fracture development index is between 3.24 and 6.08, the stress difference between the top and bottom plates and the coal seam exceeds 10 MPa, the acoustic impedance difference coefficient is greater than 69.57%, and the fracture extension pressure gradient is significantly greater than 2.05 MPa / 100 m. Therefore, it is comprehensively judged to have typical Class I safety sealing; 4 sections of Class II risk, 1954.5 - 1956.125 m, 1958.38 - 1959.375 m, 1975 - 1982.5 m, 1986 - 1987 m, mainly composed of continuous mudstone and part of sandstone-mudstone interbeds. The lithology-thickness index is between 0.38 and 0.55, the fracture development index is between 7.99 and 10.27, the stress difference between the top and bottom plates and the coal seam is between 5.9 and 7.4 MPa, the average acoustic impedance difference coefficient is 56.82%, and the fracture extension pressure gradient is significantly between 1.48 and 1.89 MPa / 100 m. Therefore, it has typical Class II risk sealing. In September 2023, fracturing was carried out on the deep coal seam of Well X-2A. The on-site construction design plan judged that the capping conditions of the top and bottom plates of the coal reservoir were good, and the fracture extension height of the coal seam was significantly controlled. In order to maximize the fracture length and increase the sand addition amount, the smooth casing variable-viscosity and slippery water fracturing was adopted, with large displacement, large liquid volume, and large sand volume construction. The fracture pressure was 45.8 MPa, the working pressure was 45.3 - 58.7 MPa, the working displacement was 11.0 m 3 / min, and the net liquid volume entering the ground was 1200.4 m 3 , adding 43.4 m of 100-mesh quartz sand 3, the designed sand addition amount was completed at 16.79%, and the average sand ratio was 5.53%. The fracturing monitoring results showed that the vertical extension of the post-fracture cracks in the coal seam roof was controlled within 8 m and did not penetrate through to the aquifer to cause engineering risks. The vertical extension of the post-fracture cracks in the coal seam floor was only 12 m. Although it caused some loss of proppants, it did not communicate with the aquifer. Generally speaking, the sealing performance of the roof and floor was good. After the fracturing construction, drainage and pressure reduction were carried out for desorption, showing that the daily gas production exceeded 8000 cubic meters.

[0137] This fully demonstrates that the risk level of the fracturing sealing performance of the deep coal reservoir roof and floor predicted by this study based on the random forest algorithm is in good agreement with the actual fracturing conditions and the post-fracture monitoring results. At the same time, it also further shows that the close combination of machine learning algorithms and geophysical logging data can play a key and core role in the evaluation of the fracturing sealing performance of the roof and floor.

[0138] Those skilled in the art should understand that due to the inevitable influence of the borehole diameter expansion of the deep coal reservoir and its roof and floor, both the acoustic travel time and compensated density logging will be affected by environmental factors such as borehole diameter expansion. The accuracy of parameters such as the lithology-thickness index, fracture development index, stress difference between the roof and floor and the coal seam, acoustic impedance difference coefficient, and fracture extension pressure gradient directly calculated using logging data is difficult to guarantee. To ensure the accuracy of the evaluation of the fracturing sealing performance of the deep coal reservoir roof and floor and the feasibility of the method, it is very necessary to correct the environmental influence such as borehole diameter expansion for its logging data.

[0139] The specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art can make various deformations or modifications within the scope of the claims, which does not affect the essence of the present invention.

Claims

1. A logging method for evaluating the fracturing isolation of the top and bottom plates of deep coal reservoirs, characterized in that: The following steps are involved: Step 1: Determine the lithology-thickness index: Use formulas (1), (2), and (3) to determine the lithology-thickness index: Among them, the top and bottom plates are limestone: （1） The top and bottom plates are mudstone: （2） The top and bottom plates are sandstone: （3） In the above formulas (1), (2) and (3): I L is the lithology-thickness index of limestone, dimensionless; I S is the lithology-thickness index of mudstone, dimensionless; I SS is the lithology-thickness index of sandstone, dimensionless; ω DEN is the weight for compensating density, dimensionless; ω GR is the weight of natural gamma, dimensionless; ω AC is the weight of the acoustic time difference, dimensionless; ρ bnorm is the normalized compensation density, dimensionless; GR norm is the normalized natural gamma, dimensionless; Δt cnorm is the standardized acoustic time difference, dimensionless; H L is the thickness of limestone, m; H S is the thickness of mudstone, m; H SS is the thickness of sandstone, m; Step 2: Determine the crack development index: Use formula (7) to determine the crack development index: （7） In formula (7): I f is the crack development index, dimensionless; R lld is the deep lateral resistivity, Ω·m; R lls is the shallow lateral resistivity, Ω·m; Step 3, determine the stress difference between the roof and the bottom plate and the coal seam: Use formula (8) to determine the stress difference between the roof and the bottom plate and the coal seam: （8） Where: Δσ is the stress difference between the roof and the bottom plate and the coal seam, MPa; σ Htb is the maximum horizontal principal stress of the top and bottom plates, MPa; σ Hc is the maximum horizontal principal stress of the coal seam, MPa; Step 4, determine the acoustic impedance difference coefficient: Use formula (14) to determine the acoustic impedance of the roof and floor plates and the coal seam: （14） In formula (14): D p is the acoustic impedance difference coefficient, %; Z tb is the acoustic impedance of the top and bottom plates, g / cm 3 ⋅m / s; Z c is the acoustic impedance of the coal seam, g / cm 3 ⋅m / s; Step 5, determine the fracture extension pressure gradient: Use formula (16) to determine the fracture extension pressure gradient: （16） In formula (16): G p is the fracture extension pressure gradient, MPa / 100m; S t is the tensile strength, MPa; D ep is the formation depth, m; Step 6: Establish a black box model for evaluating the sealing performance of roof and floor fracturing based on random forest: The five indicators of lithology-thickness index, fracture development index, stress difference between roof and floor plates and coal seams, acoustic impedance difference coefficient, and fracture extension pressure gradient are used as independent variables Xi; the risk level type of roof and floor plate fracturing isolation is determined based on the actual production capacity, geological characteristics, fracturing construction parameters, and fracturing monitoring data of a large number of well data in the study area, and it is quantified as the dependent variable Y on the basis of comprehensive consideration of the fracturing scale and post-fracturing production capacity; Based on the random forest algorithm, a large number of data sets are trained to form a black box model for evaluating the fracturing and sealing properties of the roof and floor plates, and the model is continuously optimized by adjusting the optimal combination of parameters; Step 7, determine the risk level standard of the roof and floor fracturing isolation of deep coal reservoirs: based on the determination process of the dependent variable Y in step 6, establish a quantitative roof and floor fracturing isolation risk level type table, so as to clarify the corresponding risk levels of the five basic indicators in different distribution intervals; Step 8: Guide the fracturing construction of deep coal reservoirs and increase production and improve efficiency.

2. The logging evaluation method for the fracturing isolation of the top and bottom plates of deep coal reservoirs according to claim 1, characterized in that: In step 1, the GR norm、 Δt cnorm、 ρ bnorm The method of determining is as follows: （4） （5） （6） Where: b is the compensation density of the top and bottom plates, g / cm 3 ; GR is the natural gamma of the top and bottom plates, API; Δt c is the acoustic time difference between the top and bottom plates, μs / ft.

3. The logging evaluation method for the fracturing isolation of the top and bottom plates of deep coal reservoirs according to claim 1, characterized in that: In step 3, the formula (8) also involves other parameters, which are determined as follows: （9） In formula (9): Maximum horizontal principal stress, MPa; α is the effective stress coefficient, dimensionless; σ v is the vertical stress, MPa; P p is the pore pressure of the coal seam, MPa; ε H is the strain in the direction of maximum horizontal principal stress, mm; ε h is the strain in the direction of the minimum horizontal principal stress, mm; E tb is the Young's modulus of the top and bottom plates, MPa; μ is the Poisson's ratio of the top and bottom plates, dimensionless; in: （10） In formula (10): Indicates the overburden pressure, MPa; TVD is the vertical depth of the formation without density logging, m; is the average formation density of the unmeasured section (usually taken as 2.31); i is the density logging value, g / cm 3 ; is the vertical depth sampling interval, m; （11） In formula (11): Indicates formation pore pressure, MPa; P w is the static liquid injection pressure, MPa; Δt n is the acoustic time difference under normal compaction trend, μs / ft; n formation is the compaction index, generally 1.2-3.0; （12） In formula (12): represents the Young’s modulus of the top and bottom plates, MPa; β is the multiplication factor, generally 9.290304×10 7 (If Δt c In μs / m, it is 30.472197×10 7 ), dimensionless; （13） In formula (13): represents Poisson's ratio, dimensionless; Δt s is the shear wave time difference, μs / ft.

4. The logging evaluation method for the fracturing isolation of the top and bottom plates of deep coal reservoirs according to claim 1, characterized in that: In step 4, the method for determining the parameters in equation (14) is as follows: （15） In formula (15): btb is the top and bottom plate density logging value, g / cm 3 ; V ptb is the top and bottom plate sound wave velocity, m / s; ρ bc is the coal seam density logging value, g / cm 3 ; V pc is the acoustic wave velocity in coal seam, m / s.

5. The logging evaluation method for the fracturing isolation of the roof and floor of a deep coal reservoir according to claim 1, characterized in that: In step 5, the method for determining the parameters in equation (16) is as follows: （17） In formula (17): V sh is the mud content of the top and bottom plates, %.

Images