CN117145450A - Physical-data collaborative driving unconventional gas well yield prediction method and system - Google Patents

Physical-data collaborative driving unconventional gas well yield prediction method and system Download PDF

Info

Publication number
CN117145450A
CN117145450A CN202311114106.XA CN202311114106A CN117145450A CN 117145450 A CN117145450 A CN 117145450A CN 202311114106 A CN202311114106 A CN 202311114106A CN 117145450 A CN117145450 A CN 117145450A
Authority
CN
China
Prior art keywords
production
data
pressure
time
gas well
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202311114106.XA
Other languages
Chinese (zh)
Inventor
任文希
段又菁
郭建春
曾小军
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Southwest Petroleum University
Original Assignee
Southwest Petroleum University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Southwest Petroleum University filed Critical Southwest Petroleum University
Priority to CN202311114106.XA priority Critical patent/CN117145450A/en
Publication of CN117145450A publication Critical patent/CN117145450A/en
Pending legal-status Critical Current

Links

Classifications

    • E—FIXED CONSTRUCTIONS
    • E21—EARTH DRILLING; MINING
    • E21B—EARTH DRILLING, e.g. DEEP DRILLING; OBTAINING OIL, GAS, WATER, SOLUBLE OR MELTABLE MATERIALS OR A SLURRY OF MINERALS FROM WELLS
    • E21B47/00—Survey of boreholes or wells
    • E—FIXED CONSTRUCTIONS
    • E21—EARTH DRILLING; MINING
    • E21B—EARTH DRILLING, e.g. DEEP DRILLING; OBTAINING OIL, GAS, WATER, SOLUBLE OR MELTABLE MATERIALS OR A SLURRY OF MINERALS FROM WELLS
    • E21B47/00—Survey of boreholes or wells
    • E21B47/06—Measuring temperature or pressure
    • E—FIXED CONSTRUCTIONS
    • E21—EARTH DRILLING; MINING
    • E21B—EARTH DRILLING, e.g. DEEP DRILLING; OBTAINING OIL, GAS, WATER, SOLUBLE OR MELTABLE MATERIALS OR A SLURRY OF MINERALS FROM WELLS
    • E21B47/00—Survey of boreholes or wells
    • E21B47/06—Measuring temperature or pressure
    • E21B47/07—Temperature
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/02—Agriculture; Fishing; Mining
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Abstract

The application discloses a method and a system for predicting the yield of an unconventional gas well driven by physical-data cooperation, wherein the method comprises the following steps: s1: acquiring production data of a target unconventional gas well, and preprocessing the production data; s2: based on a dynamic drainage area calculation model and a material balance equation coupled with the dynamic drainage area, establishing an objective function with the minimum accumulated yield error as a target; s3: performing history fitting to obtain model parameters; s4: calculating the average pressure of the dynamic drainage areas at different moments; s5: calculating gas production indexes corresponding to gas wells at different moments; s6: drawing a gas production index-material balance pseudo-time curve, and obtaining a characteristic straight line; s7: obtaining fitting parameters according to the slope and the inclined distance of the characteristic straight line; s8: and calculating a gas production index corresponding to any time tj, and carrying out yield prediction based on a recurrence principle. The application realizes the prediction of the yield of the unconventional gas well driven by physical-data cooperation and can provide technical support for the development of the unconventional gas well.

Description

Physical-data collaborative driving unconventional gas well yield prediction method and system
Technical Field
The application relates to the technical field of oil and gas reservoir exploitation, in particular to a physical-data collaborative driving unconventional gas well yield prediction method and system.
Background
Accurate prediction of production of unconventional gas wells is the basis for developing and adjusting completion strategies and development schemes. The unconventional gas reservoir generally develops micro-nano pores, has the characteristic of ultralow permeability, and can realize scale benefit development only by adopting a multi-stage fracturing and horizontal well technology. In addition to free gas, adsorbed gas may also be present in the unconventional gas reservoir nanopores. In addition, the gas seepage mechanism in the micro-nano pore is complex, besides the conventional viscous flow, the gas seepage mechanism also comprises multiple flow modes such as Knudsen diffusion and surface diffusion, and the like, so that a great challenge is provided for the conventional seepage theory, and the difficulty in accurately predicting the yield of an unconventional gas well is great. In addition, multi-section fracturing not only can generate artificial cracks, but also can activate natural cracks, and two types of cracks can interweave into a complex crack network, so that the difficulty in accurately predicting the yield of an unconventional gas well is further increased. Currently, unconventional gas well production prediction methods can be broadly divided into four categories: experience decremental model, transient flow analysis, numerical simulation and artificial intelligence model, and the advantages and disadvantages of various methods are as follows:
(1) Experience decremental model: the empirical reduction model is a model based on empirical/semi-empirical equations, and has no explicit physical meaning per se. It generally determines model parameters by fitting historical yield data and predicts future yields by extrapolation. The experience decremental model only needs historical output data, does not need additional data, has low cost and convenient calculation, and can realize quick prediction of future output. However, the empirical depletion model relies solely on production data for analysis and generally assumes that the well production regime is unchanged. The production gas well is limited or frequently shut in due to the influence of factors such as adjacent well fracturing, ground pipeline construction and the like, and the production system is complex. Thus, the empirical reduction model is difficult to adapt to unconventional gas wells with complex production regimes.
(2) Transient flow analysis: the analytical model and the semi-analytical model form the basis of transient flow analysis (Rate-Transient Analysis). Transient flow analysis is a practical reservoir parameter inversion and production prediction method, similar to well test analysis, both of which are established based on seepage theory. However, well testing analysis investigated the response of gas well pressure, while transient flow analysis investigated the change in gas well production. The greatest characteristic of transient flow analysis is high calculation efficiency, but the transient flow analysis has some defects: 1) The partial model does not consider non-reformed regions between adjacent hydraulic fractures; 2) The model has strong multi-solution property; 3) Cracks are too simplified, such as flat cracks; 4) It is necessary to artificially assume the shape of the drainage area.
(3) Numerical simulation: numerical simulation is the only method for quantitatively describing natural gas flow rules in heterogeneous reservoirs. The numerical simulation method can comprehensively utilize data information in various fields such as geology, oil and gas reservoirs, indoor experiments, well logging and the like. The method is characterized in that the method is carried out on grids obtained through geological modeling, a seepage equation and a material balance equation are further solved, and finally, the seepage rule of underground natural gas and the yield change of a well are obtained. The factors considered by the numerical simulation are comprehensive, but the modeling and calculation are time-consuming. In addition, the numerical simulation requires more input data, the data quality requirement is high, and errors of the input data can cause serious errors. For unconventional gas well numerical modeling, to obtain accurate production prediction results, it is also necessary to: 1) Accurately representing the position and distribution of an underground fracture network; 2) Accurately simulate the seepage process of gas in the matrix, the crack and between the matrix and the crack. However, there is currently no established method for accurately identifying and characterizing a fracture network in the subsurface, nor is there a recognized matrix, matrix-fracture coupled percolation model.
(4) Artificial intelligence model: some scholars try to establish an unconventional gas well yield prediction model by adopting artificial intelligence, but the model is mostly a model driven by pure data, and the quality of a training sample is directly related to the precision of the yield prediction model and the mobility of a prediction result. In addition, pure data driven models have poor interpretability, and stability and generalization capability are to be improved.
Disclosure of Invention
Aiming at the problems, the application aims to provide a physical-data collaborative driven unconventional gas well yield prediction method and a system, which are based on continuous quasi-steady-state flow assumption and combine dynamic leakage flow volume calculation, material balance theory, decreasing rule analysis and recursion principle to realize the unconventional gas well yield prediction driven by physical-data collaborative.
The technical scheme of the application is as follows:
in one aspect, a method for predicting production of an unconventional gas well driven by physical-data cooperation is provided, comprising the steps of:
s1: acquiring production data of a target unconventional gas well, and preprocessing the production data to obtain preprocessed production data; the production data includes daily yield q g Cumulative yield G p Bottom hole flow pressure p wf ；
S2: according to the preprocessed production data, establishing an objective function with the minimum accumulated yield error as a target based on a dynamic drainage area calculation model and a material balance equation of a coupled dynamic drainage area;
s3: according to the objective function, performing history fitting to obtain model parameters; the model parameters include reservoir pressure p prior to production of the gas well i Formation compressibility c f Fitting parameter a 1 ；
S4: based on the model parameters, the average pressure of the dynamic drainage areaCalculating the average pressure of the dynamic drainage area at different moments according to the calculation formula of (1);
s5: combining the bottom hole flow pressure p according to the average pressure of the dynamic drainage areas at different moments wf Calculating gas production indexes J corresponding to gas wells at different moments;
s6: drawing a gas production index-material balance pseudo-time curve in a double-logarithmic coordinate system according to gas production indexes J corresponding to gas wells at different moments, and obtaining a characteristic straight line of the gas production index-material balance pseudo-time curve;
s7: obtaining fitting parameter a according to the slope and the slope distance of the characteristic straight line 2 And fitting parameter b 2 ；
S8: according to the fitting parameter a 2 And the fitting parameter b 2 Calculating any time t j Corresponding gas production index J (t) j ) And yield prediction is performed based on the recurrence principle.
Preferably, in step S2, the dynamic drain area calculation model is:
wherein: v p For the pore volume of the dynamic drainage zone, m 3 /m 3 ；q g For daily output, 10 4 m 3 /d；φ i Reservoir pressure p before production for gas well i The corresponding pseudo pressure, MPa; phi (phi) p For the average pressure of the dynamic drainage areaThe corresponding pseudo pressure, MPa; c (C) ti Is the comprehensive compression coefficient under the initial condition, 1/MPa; t is t a Time-fitting for material balance, d;
wherein the pseudo pressure phi is calculated by the following formula:
wherein: phi is the pseudo pressure and MPa; mu (mu) i The viscosity of methane under the initial condition is Pa.s; z is Z i Is the methane compression factor under the initial condition, and is dimensionless; p is p i Reservoir pressure, MPa, prior to production for a gas well; p is reservoir pressure, MPa; p' is the pressure integral variable, MPa; mu is the viscosity of methane and Pa.s; z is a methane compression factor, dimensionless;
the material equilibrium pseudo-time t a The calculation is performed by the following formula:
wherein: t is time, d; c t To synthesize the compression coefficient, MPa -1 The method comprises the steps of carrying out a first treatment on the surface of the t' is a time integral variable, d;
the comprehensive compression coefficient c t The calculation is performed by the following formula:
V L ＝19.825×TOC 0.4958 (6)
p L ＝1.7192×TOC -0.462 (7)
wherein: c g The compression coefficient of methane is 1/MPa; c f Is the stratum compression coefficient, 1/MPa;is the porosity of the reservoir, dimensionless; b (B) gi Is the gas volume coefficient under the initial condition, and is dimensionless; ρ r Is rock density, t/m 3 ；V L Is Langmuir volume, m 3 /t；p L Is Langmuir pressure, MPa; TOC is the organic carbon content, dimensionless.
Preferably, the methane compression factor Z is calculated by the formula:
wherein: t (T) r For the visual comparison of temperature, dimensionless; p is p r The contrast pressure is seen, and the dimensionless is achieved; p is p c Is the critical pressure of methane, MPa; t is the temperature, K; t (T) c Is the critical temperature of methane, K.
Preferably, the methane viscosity μ is calculated by the formula:
c′＝2.447-0.2224b′ (14)
wherein: a ', b ', c ' are intermediate parameters; ρ is methane density, kg/m 3 The method comprises the steps of carrying out a first treatment on the surface of the M is the molar mass of methane and mol/kg; t is the temperature, K; r is an ideal gas constant, J/mol/K.
Preferably, in step S2, the material balance equation of the coupling dynamic drainage area is:
wherein: g p For cumulative yield, 10 4 m 3 ；B g For the average pressure of the dynamic drainage areaCorresponding gas volume coefficient, dimensionless; />The average pressure of the dynamic drainage area is MPa; t is the temperature, K; p is p sc The pressure is MPa corresponding to the ground standard state; t (T) sc The temperature K is the temperature corresponding to the ground standard state;
the average pressure of the dynamic drainage areaThe calculation is performed by the following formula:
wherein: a, a 1 Is a fitting parameter.
Preferably, in step S2, the objective function is:
wherein: obj is the objective function; n is the sample size of the cumulative yield data; g p (t j ) At t j Calculated cumulative yield at time, 10 4 m 3 ；G pr (t j ) At t j Actual cumulative output at time, 10 4 m 3 。
Preferably, in step S5, the gas production index J is calculated by the following formula:
preferably, in step S8, at any time t j Corresponding gas production index J (t) j ) The calculation is performed by the following formula:
wherein: t is t a (t j ) For any time t j The corresponding material balance is simulated for time d.
Preferably, the step S8 specifically includes the following substeps:
s81: according to the fitting parameter a 2 And the fitting parameter b 2 Calculating a gas production index corresponding to the current moment;
s82: adopting explicit time dispersion, and predicting the next time t by using the gas production index corresponding to the current time j+1 Is produced by the method:
s83: according to the next time t j+1 Calculation of the gas production of (2) the next time t j+1 Corresponding cumulative yield:
G p (t j+1 )＝G p (t j )+q g (t j+1 )×Δt (24)
s84: according to the next time t j+1 Calculating the next time t corresponding to the accumulated output j+1 Corresponding average pressure in dynamic drain region:
s85: according to the next time t j+1 Calculating the next time t according to the average pressure of the corresponding dynamic drainage area j+1 Corresponding substance balance simulationTime:
s86: according to the next time t j+1 Calculating the next time t according to the corresponding material balance quasi-time j+1 The corresponding gas production index:
s87: and repeating the steps S81-S86 to obtain the yield prediction result of the target unconventional gas well.
In another aspect, there is provided a physical-data co-driven unconventional gas well production prediction system comprising:
the acquisition module is used for acquiring production data of the target unconventional gas well and preprocessing the production data;
the fitting parameter acquisition module is used for processing the preprocessed production data to obtain fitting parameters a 2 And fitting parameter b 2 ；
A yield prediction module for predicting the yield of the plant according to the fitting parameter a 2 And the fitting parameter b 2 And carrying out yield prediction by combining a recurrence principle.
The beneficial effects of the application are as follows:
the method simplifies the description of the heterogeneity of the reservoir, avoids the difficult problems of accurate identification and characterization of the slotted network, breaks through the defect that the analysis solution/semi-analysis solution productivity model needs to presuppose the shapes of the cracks and the drainage areas, overcomes the defect of large numerical simulation calculation amount, and solves the problems of undefined physical meaning and easy accumulation of prediction errors of the pure data driving model.
Drawings
In order to more clearly illustrate the embodiments of the application or the technical solutions of the prior art, the drawings which are used in the description of the embodiments or the prior art will be briefly described, it being obvious that the drawings in the description below are only some embodiments of the application, and that other drawings can be obtained according to these drawings without inventive faculty for a person skilled in the art.
FIG. 1 is a schematic flow diagram of a method for predicting production of a physical-data co-driven unconventional gas well according to the present application;
FIG. 2 is a schematic illustration of the void volume of the corresponding drainage area at different times;
FIG. 3 is a schematic diagram of a non-conventional gas well production prediction flow based on the recurrence principle of the present application;
FIG. 4 is a graph showing the results of a gas production index-material balance pseudo-time curve according to one embodiment;
FIG. 5 is a graph showing the cumulative yield prediction for the first 860 days of an embodiment;
FIG. 6 is a graph showing the cumulative yield prediction at 369 days after one embodiment;
FIG. 7 is a graph showing the cumulative yield prediction for 20 years for one embodiment.
Detailed Description
The application will be further described with reference to the drawings and examples. It should be noted that, without conflict, the embodiments of the present application and the technical features of the embodiments may be combined with each other. It is noted that all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs unless otherwise indicated. The use of the terms "comprising" or "includes" and the like in this disclosure is intended to cover a member or article listed after that term and equivalents thereof without precluding other members or articles.
In one aspect, as shown in FIG. 1, the application provides a physical-data co-driven unconventional gas well production prediction method comprising the steps of:
s1: acquiring production data of a target unconventional gas well, and preprocessing the production data to obtain preprocessed production data; the production data includes daily yield q g Cumulative yield G p Bottom hole flow pressure p wf 。
In a specific embodiment, if a pressure gauge is installed downhole, the bottom hole pressure p may be obtained directly wf The method comprises the steps of carrying out a first treatment on the surface of the If not provided with pressureThe downhole flow pressure p is calculated by using wellhead pressure and production data through a Beggs-Brill method by a force gauge wf This is the prior art, and the specific method is not described here again. And preprocessing the production data, namely deleting the production data of the well closing stage.
S2: and establishing an objective function with the minimum accumulated yield error as a target based on a dynamic drainage area calculation model and a material balance equation coupled with the dynamic drainage area according to the preprocessed production data.
In a specific embodiment, the dynamic drain region calculation model is:
wherein: v p For the pore volume of the dynamic drainage zone, m 3 /m 3 ；q g For daily output, 10 4 m 3 /d；φ i Reservoir pressure p before production for gas well i The corresponding pseudo pressure, MPa;mean pressure for dynamic drainage area->The corresponding pseudo pressure, MPa; c (C) ti Is the comprehensive compression coefficient under the initial condition, 1/MPa; t is t a Time-fitting for material balance, d;
wherein the pseudo pressure phi is calculated by the following formula:
wherein: phi is the pseudo pressure and MPa; mu (mu) i The viscosity of methane under the initial condition is Pa.s; z is Z i Is the methane compression factor under the initial condition, and is dimensionless; p is p i Reservoir pressure, MPa, prior to production for a gas well; p is reservoir pressure, MPa; p' is the pressure integral variable, MPa; mu is the first partAlkane viscosity, pa.s; z is a methane compression factor, dimensionless;
the material equilibrium pseudo-time t a The calculation is performed by the following formula:
wherein: t is time, d; c t To synthesize the compression coefficient, MPa -1 The method comprises the steps of carrying out a first treatment on the surface of the t' is a time integral variable, d;
the comprehensive compression coefficient c t The calculation is performed by the following formula:
V L ＝19.825×TOC 0.4958 (6)
p L ＝1.7192×TOC -0.462 (7)
wherein: c g The compression coefficient of methane is 1/MPa; c f Is the stratum compression coefficient, 1/MPa;is the porosity of the reservoir, dimensionless; b (B) gi Is the gas volume coefficient under the initial condition, and is dimensionless; ρ r Is rock density, t/m 3 ；V L Is Langmuir volume, m 3 /t；p L Is Langmuir pressure, MPa; TOC is the organic carbon content, dimensionless.
When no adsorbed gas was present, the Langmuir volume was set to 0.
The material balance equation of the coupling dynamic drainage area is as follows:
wherein: g p For cumulative yield, 10 4 m 3 ；B g For the average pressure of the dynamic drainage areaCorresponding gas volume coefficient, dimensionless; />The average pressure of the dynamic drainage area is MPa; t is the temperature, K; p is p sc The pressure is MPa corresponding to the ground standard state; t (T) sc The temperature K is the temperature corresponding to the ground standard state;
the average pressure of the dynamic drainage areaThe calculation is performed by the following formula:
wherein: a, a 1 Is a fitting parameter.
The objective function is:
wherein: obj is the objective function; n is the sample size of the cumulative yield data; g p (t j ) At t j Calculated cumulative yield at time, 10 4 m 3 ；G pr (t j ) At t j Actual cumulative output at time, 10 4 m 3 。
In the above embodiment, the derivation process of the dynamic drain region calculation model is as follows:
for unconventional gas reservoirs, the pressure wave propagation velocity is slow due to its low permeability, so unconventional gas wells are in an unstable flow state for long periods of time. The application approximates an unstable flow to a series of quasi-steady flows using a continuous quasi-steady method. As production proceeds, the pressure drop spread is continuously expanded, so that the range of the drainage area is also continuously expanded, and the pore volume of the drainage area is used for representing the range of the drainage area/pressure drop spread, as shown in fig. 3, so that the artificial assumption of the geometrical shape of the drainage area/crack by the traditional method is broken through.
For any moment, under quasi-steady state flow state, the pressure at each point in the drainage area is reduced at the same speed, and the corresponding yield q g The method comprises the following steps:
substituting it into the material balance pseudo-time t a The technical formula of (2) can be obtained:
further, the time integral variable t 'therein is converted into the pressure integral variable p', and thus, it is possible to obtain:
in connection with the pseudo pressure calculation method shown in the formula (2), the equation can be expressed as:
thus, the pore volume v of the dynamic drain region p Can be expressed as a dynamic drain region calculation model shown in the formula (1).
In a specific embodiment, the methane compression factor Z is calculated by:
wherein: t (T) r For the visual comparison of temperature, dimensionless; p is p r The contrast pressure is seen, and the dimensionless is achieved; p is p c Is the critical pressure of methane, MPa; t is the temperature, K; t (T) c Is the critical temperature of methane, K.
In a specific embodiment, the methane viscosity μ is calculated by:
c′＝2.447-0.2224b′ (14)
wherein: a ', b ', c ' are intermediate parameters; ρ is methane density, kg/m 3 The method comprises the steps of carrying out a first treatment on the surface of the M is the molar mass of methane and mol/kg; t is the temperature, K; r is an ideal gas constant, J/mol/K.
In the above embodiment, the calculation method of the methane compression factor Z and the methane viscosity μ is only a preferred calculation method of the present application, and other methods in the prior art may be adopted to obtain the above two parameters when the present application is used.
S3: according to the objective function, performing history fitting to obtain model parameters; the model parameters include reservoir pressure p prior to production of the gas well i Formation compressibility c f Fitting parameter a 1 。
S4: based on the model parameters, the average pressure of the dynamic drainage areaThe calculation formula of (2) calculates the average pressure of the dynamic drainage area at different moments.
S5: combining the bottom hole flow pressure p according to the average pressure of the dynamic drainage areas at different moments wf And calculating gas production indexes J corresponding to the gas wells at different moments.
In a specific embodiment, the gas production index J is calculated by:
s6: and drawing a gas production index-material balance pseudo-time curve in a double-logarithmic coordinate system according to the gas production index J corresponding to the gas well at different moments, and obtaining a characteristic straight line of the gas production index-material balance pseudo-time curve.
In the application, the data points of the gas production index-material balance pseudo-time curve can be converged into a straight line with a negative slope, namely the characteristic straight line, and the relationship between the gas production index and the material balance pseudo-time is represented by adopting a linear equation by combining the ideas of data driving and decreasing analysis:
log 10 (J)＝-b 2 ×log 10 (t a )+log 10 (a 2 ) (31)
equation (31) is essentially a power function and can therefore be transformed into:
in this way, step S7 can obtain the fitting parameter a according to the slope and the slope distance of the characteristic line 2 And fitting parameter b 2 。
S7: obtaining fitting parameter a according to the slope and the slope distance of the characteristic straight line 2 And fitting parameter b 2 。
S8: according to the fitting parameter a 2 And the fitting parameter b 2 Calculating any time t j Corresponding gas production index J (t) j ) And yield prediction is performed based on the recurrence principle.
In a specific embodiment, as shown in fig. 3, step S8 specifically includes the following sub-steps:
s81: according to the fitting parameter a 2 And the fitting parameter b 2 Calculating the current time t j Corresponding gas production index J (t) j )：
Wherein: t is t a (t j ) For the current time t j The corresponding material balance is simulated for time d.
S82: adopting explicit time dispersion, and predicting the next time t by using the gas production index corresponding to the current time j+1 Is produced by the method:
s83: according to the next time t j+1 Calculation of the gas production of (2) the next time t j+1 Corresponding cumulative yield:
G p (t j+1 )＝G p (t j )+q g (t j+1 )×Δt (24)
s84: according to the next time t j+1 Calculating the next time t corresponding to the accumulated output j+1 Corresponding average pressure in dynamic drain region:
s85: according to the next time t j+1 Calculating the next time t according to the average pressure of the corresponding dynamic drainage area j+1 The corresponding material balance simulation time:
s86: according to the next time t j+1 Calculating the next time t according to the corresponding material balance quasi-time j+1 The corresponding gas production index:
s87: and repeating the steps S81-S86 to obtain the yield prediction result of the target unconventional gas well.
In the embodiment, the physical-data collaborative driving unconventional gas well yield prediction method integrates the advantages of physical driving and data driving, on one hand, modeling difficulty can be reduced through data driving, modeling process is simplified, and on the other hand, generalization capability of a model can be improved through physical driving, and interpretability of the model is enhanced. The application adopts an embedding mechanism based on input data to realize physical-data collaborative driving, namely, the physical model is utilized to generate additional characteristics, and the additional characteristics and the actual data are used as the input of a data driving model, and in the embodiment, the modeling thought of the application is as follows:
(1) By combining continuous quasi-steady state assumption, dynamic drainage area calculation and a material balance equation, a capacity model based on physical driving is established, fitting parameters in the capacity model are determined through history fitting, and an empirical relation between the average pressure and the accumulated yield of the dynamic drainage area shown in a formula (19) is obtained.
(2) And combining model parameters obtained by fitting, calculating average pressure and material balance fitting time of the dynamic drainage areas corresponding to different moments by using an energy production model, and taking the average pressure and the material balance fitting time as two additional characteristics and taking historical production data as input of a data driving model.
(3) And constructing a data driving model based on gas production index calculation, descending rule analysis and characteristic straight line analysis, and obtaining the relationship between the gas production index and the material balance simulation time shown in the formula (22) through model training.
(4) And (3) carrying out yield prediction by utilizing a recurrence principle based on the empirical relation between the average pressure and the accumulated yield of the dynamic drainage area obtained in the step (1) and the relation between the gas production index and the material balance simulation time obtained in the step (3).
Therefore, the application breaks through the defect that the analysis solution/semi-analysis solution productivity model needs to presuppose the shapes of cracks and drainage areas, overcomes the defect of large numerical simulation calculation amount, solves the problems of undefined physical meaning and easy accumulation of prediction errors of the pure data driving model, and ensures that the yield result obtained by the prediction of the application is more in accordance with the actual working condition.
In another aspect, the present application also provides a physical-data co-driven unconventional gas well production prediction system comprising:
the acquisition module is used for acquiring production data of the target unconventional gas well and preprocessing the production data;
the fitting parameter acquisition module is used for processing the preprocessed production data to obtain fitting parameters a 2 And fitting parameter b 2 The method comprises the steps of carrying out a first treatment on the surface of the The fitting parameter a is obtained by adopting the steps S2-S7 in the physical-data collaborative driving unconventional gas well yield prediction method 2 And the fitting parameter b 2 。
A yield prediction module for predicting the yield of the plant according to the fitting parameter a 2 And the fitting parameter b 2 Carrying out yield prediction by combining a recurrence principle; i.e. using the physical-data according to the applicationStep S8 in the synergistically driven unconventional gas well production prediction method performs production prediction.
In a specific embodiment, taking a certain unconventional gas well A as an example, based on production data of the unconventional gas well A, the unconventional gas well yield prediction method driven by physical-data cooperation is adopted to predict the yield, and the unconventional gas well yield is compared with an empirical yield decreasing model and a time sequence neural network yield prediction model. In this example, a Logical Growth Model (LGM), an elongation index decreasing model (SEPD), and a Duong model were selected as representative of the empirical yield decreasing model; nonlinear autoregressive neural network models (NARs) and long short memory neural network models (LSTMs) are selected as representative of the time series neural network yield prediction models.
The historical production data of the A well is divided into a training set and a testing set according to the proportion of 7:3, namely, the historical fitting is carried out by using 70% of the production data, and then the model test is carried out by using the rest 30% of the data. In this embodiment, the gas production index-material balance pseudo-time curve of the present application is shown in fig. 4, and it can be seen from fig. 4 that the later data points can be converged into a straight line with a negative slope, i.e. a characteristic straight line. Based on the slope and intercept of the characteristic straight line, the relation between the gas production index and the material balance simulation time can be obtained, and fitting parameters of the application are shown in table 1:
table 1 fitting parameters and values of the application
p i (MPa) a 1 (m -3 ) c f (MPa -1 )
55.30 -6.06×10 -9 55.99×10 -4
Historical production data for well a was taken for 1229 days, and history fits were performed using the production data from the previous 860 days, and the fitting results for the present application, LGM model, SEPD model, and dunng model are shown in fig. 5. As can be seen from FIG. 5, the fitting effect of the present application is best, and the corresponding Root Mean Square Error (RMSE) is 15.25X10 4 m 3 The method comprises the steps of carrying out a first treatment on the surface of the The fitting effect of the LGM model and the SEPD model was inferior, and the corresponding RMSE was 71.99X10, respectively 4 m 3 And 81.04 ×10 4 m 3 The method comprises the steps of carrying out a first treatment on the surface of the The Duong model has the worst fitting effect, and the corresponding RMSE reaches 202.99 ×10 4 m 3 。
Further, the cumulative yield at 369 days after training was predicted using the trained model and compared with the actual cumulative yield data, the results are shown in FIG. 6. As can be seen from FIG. 6, the prediction effect of the present application is best, the prediction result is relatively close to the actual value, and the corresponding RMSE is 16.73X10 4 m 3 The method comprises the steps of carrying out a first treatment on the surface of the The predictive results of the LGM and Duong models were high, corresponding to RMSE of 151.3X10, respectively 4 m 3 And 677.31 ×10 4 m 3 The method comprises the steps of carrying out a first treatment on the surface of the The SEPD model has the worst prediction effect and is far smaller than the actual value, and the corresponding RMSE is as high as 2.39X10 7 m 3 。
Further, long-term integrated yield was predicted using the trained model for 20 years, 7300 days, and the results are shown in FIG. 7. As can be seen from fig. 7, the Duong model predicts an increase in cumulative yield with time and the trend of increasing gradually with time increases, which is not in line with the actual situation, and the Duong model gives a cumulative yield prediction result of 4.32 hundred million square after 20 years, which is not in line with the conventional knowledge. This is because the Duong model assumes that the gas well is in a linear flow regime for long periods of time, and if the later gas well flow regime changes, and the previous production data is used to predict production, a much greater result than the actual value will be obtained, i.e., an overly optimistically estimated production. The prediction results of the application are between the prediction results of the LGM model and the prediction results of the SEPD model, the cumulative yield prediction results given by the model of the application after 20 years are 0.3 hundred million square, and the prediction results given by the LGM model and the SEPD model are 0.56 hundred million square and 0.24 hundred million square respectively. This also indirectly verifies the correctness of the model of the application, since the predicted results of the SEPD model are generally more conservative.
In summary, the present application is able to more accurately predict unconventional gas well production. Compared with the prior art, the application has obvious progress.
The present application is not limited to the above-mentioned embodiments, but is intended to be limited to the following embodiments, and any modifications, equivalents and modifications can be made to the above-mentioned embodiments without departing from the scope of the application.

Claims (10)

1. A method for predicting production of an unconventional gas well driven by physical-data cooperation, comprising the steps of:
s1: acquiring production data of a target unconventional gas well, and preprocessing the production data to obtain preprocessed production data; the production data includes daily yield q g Cumulative yield G p Bottom hole flow pressure p wf ；
S2: according to the preprocessed production data, establishing an objective function with the minimum accumulated yield error as a target based on a dynamic drainage area calculation model and a material balance equation of a coupled dynamic drainage area;
s3: according to the objective function, performing history fitting to obtain model parameters; the model parameters include reservoir pressure p prior to production of the gas well i Formation compressibility c f Fitting parameter a 1 ；
S4: based on the model parameters, on the average of the dynamic drainage areaPressure ofCalculating the average pressure of the dynamic drainage area at different moments according to the calculation formula of (1);
s5: combining the bottom hole flow pressure p according to the average pressure of the dynamic drainage areas at different moments wf Calculating gas production indexes J corresponding to gas wells at different moments;
s6: drawing a gas production index-material balance pseudo-time curve in a double-logarithmic coordinate system according to gas production indexes J corresponding to gas wells at different moments, and obtaining a characteristic straight line of the gas production index-material balance pseudo-time curve;
s7: obtaining fitting parameter a according to the slope and the slope distance of the characteristic straight line 2 And fitting parameter b 2 ；
S8: according to the fitting parameter a 2 And the fitting parameter b 2 Calculating any time t j Corresponding gas production index J (t) j ) And yield prediction is performed based on the recurrence principle.
2. The method for predicting production of a physical-data co-driven unconventional gas well according to claim 1, wherein in step S2, the dynamic drainage area calculation model is:
wherein: v p For the pore volume of the dynamic drainage zone, m 3 /m 3 ；q g For daily output, 10 4 m 3 /d；φ i Reservoir pressure p before production for gas well i The corresponding pseudo pressure, MPa;mean pressure for dynamic drainage area->The corresponding pseudo pressure, MPa; c (C) ti Is the comprehensive compression coefficient under the initial condition, 1/MPa; t is t a Time-fitting for material balance, d;
wherein the pseudo pressure phi is calculated by the following formula:
wherein: phi is the pseudo pressure and MPa; mu (mu) i The viscosity of methane under the initial condition is Pa.s; z is Z i Is the methane compression factor under the initial condition, and is dimensionless; p is p i Reservoir pressure, MPa, prior to production for a gas well; p is reservoir pressure, MPa; p' is the pressure integral variable, MPa; mu is the viscosity of methane and Pa.s; z is a methane compression factor, dimensionless;
the material equilibrium pseudo-time t a The calculation is performed by the following formula:
wherein: t is time, d; c t To synthesize the compression coefficient, MPa -1 The method comprises the steps of carrying out a first treatment on the surface of the t' is a time integral variable, d;
the comprehensive compression coefficient c t The calculation is performed by the following formula:
V L ＝19.825×TOC 0.4958 (6)
p L ＝1.7192×TOC -0.462 (7)
wherein: c g The compression coefficient of methane is 1/MPa; c f Is of formation pressureShrink coefficient, 1/MPa;is the porosity of the reservoir, dimensionless; b (B) gi Is the gas volume coefficient under the initial condition, and is dimensionless; ρ r Is rock density, t/m 3 ；V L Is Langmuir volume, m 3 /t；p L Is Langmuir pressure, MPa; TOC is the organic carbon content, dimensionless.
3. The method for predicting production of a physical-data co-driven unconventional gas well of claim 2 wherein the methane compression factor Z is calculated by:
wherein: t (T) r For the visual comparison of temperature, dimensionless; p is p r The contrast pressure is seen, and the dimensionless is achieved; p is p c Is the critical pressure of methane, MPa; t is the temperature, K; t (T) c Is the critical temperature of methane, K.
4. The method for predicting production of a physical-data co-driven unconventional gas well of claim 2, wherein the methane viscosity μ is calculated by:
c′＝2.447-0.2224b′ (14)
wherein: a ', b ', c ' are intermediate parameters; ρ is methane density, kg/m 3 The method comprises the steps of carrying out a first treatment on the surface of the M is the molar mass of methane and mol/kg; t is the temperature, K; r is an ideal gas constant, J/mol/K.
5. The method for predicting production of a physical-data co-driven unconventional gas well of claim 2, wherein in step S2, the mass balance equation of the coupled dynamic drainage zone is:
wherein: g p For cumulative yield, 10 4 m 3 ；B g The gas volume coefficient corresponding to the average pressure p of the dynamic drainage area is dimensionless;the average pressure of the dynamic drainage area is MPa; t is the temperature, K; p is p sc The pressure is MPa corresponding to the ground standard state; t (T) sc The temperature K is the temperature corresponding to the ground standard state;
the average pressure of the dynamic drainage areaThe calculation is performed by the following formula:
wherein: a, a 1 Is a fitting parameter.
6. The method for predicting production of a physical-data co-driven unconventional gas well of claim 5, wherein in step S2, the objective function is:
wherein: obj is the objective function; n is the sample size of the cumulative yield data; g p (t j ) At t j Calculated cumulative yield at time, 10 4 m 3 ；G pr (t j ) At t j Actual cumulative output at time, 10 4 m 3 。
7. The method for predicting production of a physical-data co-driven unconventional gas well of claim 1, wherein in step S5, the gas production index J is calculated by:
8. according to claim 1The physical-data collaborative driving unconventional gas well yield prediction method is characterized in that in step S8, at any time t j Corresponding gas production index J (t) j ) The calculation is performed by the following formula:
wherein: t is t a (t j ) For any time t j The corresponding material balance is simulated for time d.
9. The method for predicting production of a physical-data co-driven unconventional gas well of claim 8, wherein step S8 specifically comprises the sub-steps of:
s81: according to the fitting parameter a 2 And the fitting parameter b 2 Calculating a gas production index corresponding to the current moment;
s82: adopting explicit time dispersion, and predicting the next time t by using the gas production index corresponding to the current time j+1 Is produced by the method:
s83: according to the next time t j+1 Calculation of the gas production of (2) the next time t j+1 Corresponding cumulative yield:
G p (t j+1 )＝G p (t j )+q g (t j+1 )×Δt (24)
s84: according to the next time t j+1 Calculating the next time t corresponding to the accumulated output j+1 Corresponding average pressure in dynamic drain region:
s85: according to the next time t j+1 Corresponding dynamic drainage area average pressure gaugeCalculate the next time t j+1 The corresponding material balance simulation time:
s86: according to the next time t j+1 Calculating the next time t according to the corresponding material balance quasi-time j+1 The corresponding gas production index:
s87: and repeating the steps S81-S86 to obtain the yield prediction result of the target unconventional gas well.
10. A physical-data co-driven unconventional gas well production prediction system, comprising:
the acquisition module is used for acquiring production data of the target unconventional gas well and preprocessing the production data;
the fitting parameter acquisition module is used for processing the preprocessed production data to obtain fitting parameters a 2 And fitting parameter b 2 ；
A yield prediction module for predicting the yield of the plant according to the fitting parameter a 2 And the fitting parameter b 2 And carrying out yield prediction by combining a recurrence principle.
CN202311114106.XA 2023-08-31 2023-08-31 Physical-data collaborative driving unconventional gas well yield prediction method and system Pending CN117145450A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311114106.XA CN117145450A (en) 2023-08-31 2023-08-31 Physical-data collaborative driving unconventional gas well yield prediction method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311114106.XA CN117145450A (en) 2023-08-31 2023-08-31 Physical-data collaborative driving unconventional gas well yield prediction method and system

Publications (1)

Publication Number Publication Date
CN117145450A true CN117145450A (en) 2023-12-01

Family

ID=88907465

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311114106.XA Pending CN117145450A (en) 2023-08-31 2023-08-31 Physical-data collaborative driving unconventional gas well yield prediction method and system

Country Status (1)

Country Link
CN (1) CN117145450A (en)

Similar Documents

Publication Publication Date Title
CN107506948B (en) Shale oil gas comprehensive yield analysis method based on dynamic drainage volume
CN102041995B (en) System for monitoring complicated oil deposit flooding conditions
CN111441767B (en) Oil reservoir production dynamic prediction method and device
Gherabati et al. The impact of pressure and fluid property variation on well performance of liquid-rich Eagle Ford shale
CN104899411B (en) A kind of reservoir productivity prediction model method for building up and system
CN110952976B (en) Single-well exploitation stable yield potential evaluation method in gas reservoir development mode
CN109543206A (en) The economic calorific requirement optimization method of heavy crude heat extraction horizontal well reservoir
CN115526108B (en) Landslide stability intelligent dynamic prediction method based on multisource monitoring data
CN111927417A (en) Shale gas staged fracturing horizontal well group reserve utilization condition evaluation method
CN116127675A (en) Prediction method for maximum recoverable reserve of shale oil horizontal well volume fracturing
CN116658155A (en) Shale gas well yield prediction method
CN106501146B (en) Method and device for determining physical upper limit of tight oil reservoir
CN107130952A (en) A kind of three-level composite petroleum method for increasing
Yan et al. A prediction model for pressure propagation and production boundary during coalbed methane development
CN116122801A (en) Shale oil horizontal well volume fracturing compressibility comprehensive evaluation method
CN114580261A (en) Recoverable storage prediction model establishing method, recoverable storage prediction method and recoverable storage prediction device
CN117145450A (en) Physical-data collaborative driving unconventional gas well yield prediction method and system
CN107704646B (en) Modeling method after compact reservoir volume modification
RU2657917C1 (en) Method of update of the geological-gas flue model of the gas deposit according to the data of operation
Rafiei Improved oil production and waterflood performance by water allocation management
CN111101930B (en) Single-well exploitation production-increasing potential evaluation method in gas reservoir development mode
CN115705452A (en) Novel recovery ratio prediction method for middle and later stages of integrated sandstone reservoir development
Hu et al. Quick assessment to ascertain technical rational well spacing density in artificial water flooding oilfield
Chen et al. GINI coefficient: An effective way to evaluate inflow profile equilibrium of horizontal wells in Shengli Oil Field
CN112558179A (en) Intelligent inversion method of far field stress state

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination