A method for intelligently identifying multiphase flow characteristics of a condensate gas well

Abstract

The application relates to a method for intelligently identifying the multiphase flow characteristics of a condensate gas well, which comprises the following steps: acquiring the gas-liquid equilibrium constant and the liquid-solid equilibrium constant of a condensate gas system under a certain temperature and pressure condition of the condensate gas well; establishing a multiphase flow model flow equation of the condensate gas well; establishing a multiphase flow model energy conservation equation of the condensate gas well; solving the multiphase flow model of the condensate gas well considering the coupling process of the phase change behavior and the multiphase flow, wherein the gas-liquid equilibrium constant and the liquid-solid equilibrium constant and the multiphase flow model flow equation of the condensate gas well are used to participate in the solving of the multiphase flow model energy conservation equation of the condensate gas well, and the multiphase flow parameters in the wellbore flow process are obtained; the flow pattern identification diagram established according to the gas-liquid two-phase momentum conversion speed is identified and trained by using an LVQ neural network, and the identification network of the multiphase flow flow pattern characteristics is obtained, and the multiphase flow characteristics in the wellbore of the condensate gas well are intelligently identified. The application solves the coupling problem of the dynamic change of the multiphase components and the complex flow in the wellbore of the condensate gas well.

Description

TECHNICAL FIELD

[0001] The present application relates to a method for intelligently identifying the multiphase flow characteristics of a condensate gas well, solving the coupling problem of dynamic changes of gas-liquid-solid components and complex multiphase flow in the wellbore of a condensate gas well, especially considering the heat and mass transfer characteristics caused by the phase change of the condensate gas system, and the technical problem of intelligent identification of the multiphase flow characteristics of the condensate gas well, in particular, a method for intelligently identifying the multiphase flow characteristics of a condensate gas well based on the LVQ algorithm in machine learning. BACKGROUND

[0002] At present, in the global basin range, condensate gas reservoirs account for as high as 68% in the giant gas fields with geological reserves exceeding 1x1012m3. The production of condensate gas and light oil has rapidly increased with the innovation of condensate gas reservoir development methods, and the global production capacity has reached 12.03 million barrels per day. Condensate gas reservoirs are a special type of complex oil and gas reservoirs between oil reservoirs and pure gas reservoirs, with complex phase state and flow characteristics. Condensate gas reservoirs are an important oil and gas resource, with abundant reserves, and the output of natural gas and condensate oil has wide application in energy, chemical industry, transportation and other fields, with high economic benefits, so condensate gas reservoirs have high exploitation value. The advantages of condensate gas reservoirs have made them an important part of global oil and energy development.

[0003] However, during the production process in the wellbore of a condensate gas reservoir, the phenomenon of reverse condensation often occurs, especially when heavy components exist in the condensate gas well and wax crystal particles are precipitated, forming complex multiphase flow of gas, liquid and solid mixed flow, which leads to problems such as wellbore plugging, flow energy loss and difficulty in identifying flow characteristics. Therefore, in recent years, the coupling solution of component dynamic changes and complex multiphase flow in condensate gas wells, especially the intelligent identification of the multiphase flow characteristics of condensate gas wells, has become a problem to be solved in engineering. However, existing knowledge is limited to the establishment and modification of single phase equilibrium theory or gas-liquid two-phase flow model, and the coupling process of gas-liquid-solid dynamic equilibrium and multiphase flow model has not been considered, especially the heat and mass transfer characteristics caused by the phase change of the condensate gas system are ignored, which directly affects the accurate identification of the multiphase flow characteristics in the condensate gas well. Therefore, an intelligent method for identifying the multiphase flow characteristics of a condensate gas well based on the coupling of phase equilibrium and multiphase flow model is invented, which fully considers the precipitation of wax crystal particles and the dynamic change of gas-liquid ratio of condensate liquid system caused by the reverse condensation phenomenon, and solves the coupling of the multiphase flow characteristics of the condensate gas well and the phase change process of the gas-liquid-solid components, breaking through the limitations of traditional methods that ignore the heat and mass transfer problems caused by the phase change of the condensate gas system, especially considering the complexity of the multiphase dynamic flow of the condensate gas well and the slow identification of flow patterns relying on experience charts, and it is particularly necessary to design an intelligent method for identifying the multiphase flow characteristics of a condensate gas well. SUMMARY

[0004] The application aims to provide a method for intelligently identifying multiphase flow characteristics of a condensate gas well.

[0005] The technical solution adopted by the application to solve the technical problem is that the method for intelligently identifying multiphase flow characteristics of a condensate gas well comprises the following steps:

[0006] (I) acquiring gas-liquid equilibrium constants and liquid-solid equilibrium constants of a condensate gas system under a certain temperature and pressure condition of the condensate gas well

[0007] (II) establishing a multiphase flow model flow equation of the condensate gas well, which comprises a liquid-phase continuity equation of the condensate gas system, a gas-phase and solid-phase continuity equation of the condensate gas system, a liquid-phase conservation type momentum equation of the condensate gas system, and a gas-phase and solid-phase conservation type momentum equation of the condensate gas system

[0008] (III) establishing a multiphase flow model energy conservation equation of the condensate gas well

[0009]

[0010] In the formula, c p is the specific heat capacity of a micro-unit control body at constant pressure, J / (kg·K); α is the thermal diffusivity, m 2 / s; e is the specific internal energy of the micro-unit control body, J / kg;

[0011] (IV) solving the multiphase flow model of the condensate gas well considering the coupling process of phase change behavior and multiphase flow; the wellbore model of the condensate gas well is divided into a one-dimensional equidistant node grid, and the acquired gas-liquid equilibrium constants and liquid-solid equilibrium constants and the multiphase flow model flow equation are used to solve the multiphase flow model energy conservation equation of the condensate gas well, the multiphase flow parameters at the i+1th node at a certain height of the wellbore are solved first, then the multiphase flow parameters at each node are solved in turn at the same interval, the multiphase flow parameters at all nodes of the entire wellbore are solved, the multiphase flow parameters in the wellbore flow process are obtained, and the multiphase flow parameters include phase volume fraction, mass flow rate, flow velocity, temperature, pressure, and density

[0012] ​(V) intelligently identifying the multiphase flow characteristics in the condensate gas wellbore; the flow pattern recognition graph established according to the phase volume fraction, mass flow, flow velocity, temperature, pressure and density in the wellbore flow process obtained in step (IV) is identified and trained by using the LVQ neural network to obtain the identification network of the multiphase flow pattern characteristics, the flow pattern of any two-phase momentum conversion velocity is judged and a conclusion is quickly given, and the automatic intelligent identification of the component change, temperature and pressure distribution, wax precipitation characteristics and multiphase flow characteristics of the condensate gas well is realized.

[0013] The step one in the above scheme is specifically:

[0014] When the fugacity of each alkane component in the gas phase is equal to the fugacity of the liquid phase, it indicates that the gas-liquid two phases reach the equilibrium state, the PR state equation based on the SRK equation is used to calculate the fugacity of the gas phase and the fugacity of the liquid phase, and the gas-liquid equilibrium constant of the condensate gas system under a certain temperature and pressure condition is obtained according to the definition of the gas-liquid equilibrium constant.

[0015]

[0016] In the formula, is the mole fraction of the i-th component in the gas phase; is the mole fraction of the i-th component in the liquid phase; A V and A M are-0.623 and-0.53, respectively; λ is 0.36; is the activity coefficient of the i-th component in the liquid phase; R is the ideal gas constant, which is 8.3143 kJ / (kmol·K); T is the system temperature, K; V cal is the volume obtained by the PR state equation, m 3 ; the values of a i , b i are as follows:

[0017]

[0018] In the formula, P represents the system pressure, Pa; P c represents the critical pressure, Pa; T c represents the critical temperature, K; T r is the reduced temperature, that is, the ratio of the system temperature to the critical temperature; M is the molar mass of the pure substance, g / mol;

[0019] When the fugacity of each alkane component in the solid phase is equal to the fugacity of the liquid phase, it indicates that the liquid-solid two phases reach the equilibrium state, the activity coefficient method is used to calculate the fugacity of the solid phase and the liquid phase, and the liquid-solid equilibrium constant of the condensate gas system under a certain temperature and pressure condition is obtained according to the definition of the liquid-solid equilibrium constant.

[0020]

[0021] wherein, is the mole fraction of the i-th component in the solid phase; is the activity coefficient of the i-th component in the solid phase; ΔH f , ΔH tr is the enthalpy of fusion and the enthalpy of solid-solid transition, J / mol; T f , T tr is the melting temperature and the solid-solid transition temperature, K; ΔC p is the heat fusion difference between solid and liquid phases, cal / (g·K).

[0022] The step two in the above scheme is:

[0023] The mass increment in the micro-element control body in the wellbore flow process is equal to the mass change caused by the inflow and outflow of the condensate gas system in the control body, and the condensate gas system should also supplement the mass loss caused by the liquefaction of gas phase components and the mass increment caused by the dissolution of wax crystals, and the continuity equation of the condensate gas system liquid phase is constructed as follows:

[0024]

[0025] wherein, g, l, s respectively represent the gas phase, the liquid phase and the solid phase; ρ l is the density of the liquid phase, kg / m 3 ; α l is the phase volume fraction of the liquid phase, %; u l is the velocity of the liquid phase, m / s; z is the length of the micro-element control body, m; Δm li is the mass change rate from the liquid phase to the i-th phase, kg / (m 3 ·s), which is equal to A is the cross-sectional area of the condensate gas well, m 2 ; (G l )1 is the mass flow rate of the liquid phase at the inlet of the micro-element control body, kg / s; (G l )2 is the mass flow rate of the liquid phase at the outlet of the micro-element control body, kg / s;

[0026] In the gas phase and the solid phase of the condensate gas system, the mass increment caused by the change of multi-component alkanes is considered, and the continuity equations of the gas phase and the solid phase of the condensate gas system are constructed:

[0027]

[0028] The gas-liquid-solid flow in the micro-element control body in the body is analogous to the single-phase flow in the horizontal pipeline, that is, the viscous force and the interfacial drag force are combined into the tangential stress, which is equivalent to the friction stress causing the single-phase flow resistance in the horizontal pipeline, the momentum equation of the liquid phase conservation type of the condensate gas system is constructed by combining equation (2-2):

[0029]

[0030] The momentum equations of the gas phase and the solid phase of the condensate gas system are constructed again:

[0031]

[0032] In the formula, λ fl is the hydraulic friction coefficient of the liquid phase; D l is the hydraulic radius of the liquid phase, m.

[0033] The step two in the above scheme is specifically:

[0034] The mass increment of the micro-element control body in the wellbore multiphase flow process is equal to the mass change caused by the inflow and outflow of the control body, the condensate gas system should also supplement the mass loss caused by the liquefaction of the gas phase components and the mass increment introduced by the dissolution of the wax crystals, the time for the condensate gas system to reach the phase equilibrium is ignored, the equation of the liquid phase of the condensate gas system is constructed and simplified as follows:

[0035]

[0036] In the formula, g, l and s respectively represent the gas phase, the liquid phase and the solid phase; ρ l is the density of the liquid phase, kg / m 3 ; α l is the phase volume fraction of the liquid phase, %; u l is the velocity of the liquid phase, m / s; z is the length of the micro-element control body, m; Δm li is the mass change rate of the liquid phase into the i-th phase, kg / (m 3 ·s), which is A is the cross-sectional area of the condensate gas well, m 2 ; (G l )1 is the mass flow rate of the liquid phase at the inlet of the micro-element control body, kg / s; (G l )2 is the mass flow rate of the liquid phase at the outlet of the micro-element control body, kg / s;

[0037] The mass increment caused by the change of the multi-component alkane components in the gas phase and the solid phase of the condensate gas system is considered, and the continuity equations of the gas phase and the solid phase of the condensate gas system are constructed as follows:

[0038]

[0039] For the momentum conservation equation, the micro-element control body in the flow is analyzed, and the non-conservative correlation between the force of the fluid micro-cluster and its motion state is obtained for the liquid phase in the condensate gas system:

[0040]

[0041] In the formula, t is time, s; τ is viscous force, Pa; F D is interfacial drag, N / m 3 ; g is the acceleration of gravity, 9.8 m / s 2 ;

[0042] The gas-liquid-solid flow in the micro-element control body is analogous to the single-phase flow in the horizontal pipeline, that is, the viscous force and the interfacial drag are combined into the tangential stress, which is equivalent to the friction stress causing the friction of the single-phase flow in the horizontal pipeline, so that:

[0043]

[0044] In the formula, λ fl is the hydraulic friction coefficient of the liquid phase; D l is the hydraulic radius of the liquid phase, m.

[0045] At the same time, the material derivative of the liquid phase velocity is transformed, and there is:

[0046]

[0047] Combined with the continuity equation in formula (2-2), the conservative momentum equation of the liquid phase of the condensate gas system is constructed:

[0048]

[0049] Similarly, the conservative momentum equations of the gas phase and the solid phase of the condensate gas system are constructed:

[0050]

[0051] The third step in the above scheme is specifically:

[0052] When constructing the energy equation, the condensate gas system is regarded as a mixture. According to the principle of energy conservation, the micro-element control body of the condensate gas system will cause energy change due to the work done by gravity, pressure and viscous force, and there is a temperature difference between the condensate gas system and the formation to drive the formation of two kinds of heat loss including heat conduction, convective heat transfer, and also including the heat formed by the phase change of the system. Combined with the total energy change rate of the condensate gas system, the non-conservative energy equation is constructed:

[0053]

[0054] In the formula, E is total energy, J; q is the heat energy density of heat transfer in the condensate gas system, W / m3 ; φ is the heat exchange between the condensate gas system and the outside world, W; h ji is the enthalpy of phase transition from the jth phase to the ith phase, J / kg;

[0055] According to the Fourier heat conduction law, the total heat transfer between the condensate gas system and the formation outside the wellbore is represented as:

[0056] φ = -πD in K (T-T a -g T h) dz (3-2)

[0057] In the formula, D in is the inner diameter of the condensate gas well, m; K is the total heat transfer coefficient per unit area, W / (m 2 ·K); T a is the surface temperature at the wellhead of the condensate gas well, K; g T is the geothermal gradient, K / m; h is the current wellbore depth of the microelement control body, m;

[0058] Based on the series thermal resistance network model, the heat transfer process of the condensate gas system is regarded as a series of heat conduction thermal resistance and convective heat transfer thermal resistance. According to the multilayer cylindrical wall heat transfer model, the heat transfer area is combined with the total heat transfer coefficient, then the total heat transfer and the total heat transfer coefficient are:

[0059]

[0060] In the formula, K * is the total heat transfer coefficient of the microelement control body, W / (m 2 ·K); h f is the convective heat transfer coefficient, W / (m 2 ·K); r m is the radius of the mth layer of heat conduction medium in the heat transfer direction, m; λ m is the thermal conductivity of the mth layer of heat conduction medium, W / (m·K), and the formula for calculating the convective heat transfer coefficient is as follows:

[0061]

[0062] In the formula, λ is the thermal conductivity of the microelement control body, W / (m·K); Re is the Reynolds number of the microelement control body, W / (m·K); Pr is the Prandtl number of the microelement control body, W / (m·K);

[0063] Similarly, the material derivative of the total energy of the system is transformed, and the conservation type energy equation for the condensate gas system flowing in the wellbore is constructed as:

[0064]

[0065] In the formula, the relevant parameters of the infinitesimal control body are obtained by volume fraction weighted average, in addition, considering the correlation between the temperature and specific enthalpy of the condensate gas system, the diffusion term on the right side of formula (3-5) is converted into specific internal energy gradient, and the energy equation is simplified into specific enthalpy equation:

[0066]

[0067] The step four in the above scheme is specifically:

[0068] According to the finite difference method, the condensate gas wellbore model is divided into a one-dimensional equidistant node grid, the discrete methods of the convection term and the diffusion term of the unknown quantities such as flow velocity, density, and phase volume fraction are respectively a first-order upwind difference format and a central difference format, then the discrete forms of the continuity equation, the momentum equation, and the energy equation at the i+1th node are as follows:

[0069]

[0070] Considering that there is a nonlinear coupling of multiple unknown quantities in each equation of formula (4-1), based on the prediction correction strategy of the separation algorithm, it is assumed in advance that there is no phase transfer of mass and heat between each phase in the condensate gas system, that is, the density and phase volume fraction of each phase do not change, then the calculation results of temperature and pressure at the i+1th node can be obtained as follows:

[0071]

[0072] After decoupling the multiple variables in formula (4-2), the phase state transition in the condensate gas system is ignored, which does not conform to the actual process, so according to the temperature and pressure at the i+1th node of formula (4-2), the mole fraction of each alkane component in the gas-liquid-solid three phases at the i th node of the flow front is used as the initial value for iteration, and according to the gas-liquid equilibrium constant and the liquid-solid equilibrium constant obtained in step (1), a gas-liquid-solid flash model is constructed as follows:

[0073]

[0074] Assuming that the initial gas phase mole number is 1 and the liquid-solid two-phase mole numbers are both 0, the mole fraction of each alkane component in the gas-liquid-solid three phases is continuously iterated by formula (4-3) by using the dichotomy method until the material conservation principle is satisfied:

[0075]

[0076] In the formula, eps x is the calculation accuracy of the mole fraction of the multi-component alkane phase equilibrium process;

[0077] According to the mole fraction of each alkane component in the gas-liquid-solid three-phase, the corrected gas-liquid equilibrium constant and the liquid-solid equilibrium constant are obtained, when the accuracy requirement in formula (4-4) is met, the phase equilibrium parameters of the multi-component alkane at the i+1 node are obtained, otherwise, formula (4-2) and formula (4-3) need to be repeated until the accuracy requirement in formula (4-5) is met;

[0078]

[0079] In the formula, eps K is the equilibrium constant calculation accuracy of the multi-component alkane phase equilibrium process;

[0080] The volume fraction, density, mass flow rate and flow velocity of the gas, liquid and solid phases at the outlet node are calculated according to formula (4-6), and the correction is as follows:

[0081]

[0082] At the same time, since the flow process of the condensate gas system in the wellbore should satisfy the phase continuity equation in formula (4-1), the velocity and density of the gas, liquid and solid phases at the i+1 node are obtained in turn according to the difference form of the continuity equation and are twice corrected, and the corrected results are combined with the uncorrected wellbore temperature and pressure distribution of the condensate gas system in formula (4-2) to construct the relative error, and iteration is carried out until the corrected temperature and pressure can meet the accuracy requirement in formula (4-7):

[0083]

[0084] In the formula, eps P is the pressure calculation accuracy in the multiphase flow model, Pa; eps T is the temperature calculation accuracy in the multiphase flow model, K.

[0085] The fifth step in the above scheme is specifically:

[0086] According to the phase volume fraction, mass flow rate, flow velocity, temperature, pressure and density obtained by solving step (four), a data analysis program is used to realize automatic intelligent identification of the component change, temperature and pressure distribution, and wax precipitation characteristics of the multiphase flow of the condensate gas well; the flow pattern identification diagram established for the gas-liquid two-phase momentum conversion velocity is identified and trained by using the LVQ neural network, the LVQ neural network used is composed of three layers of neurons, i.e., the input layer, the competitive layer and the output layer, the flow pattern identification diagram of the wellbore depth range is obtained according to step (four) and the Hewitt and Roberts flow pattern diagram, 100 two-dimensional data points of the gas-liquid two-phase momentum conversion velocity at different depths are taken from the flow pattern identification diagram by using the point taking program, and the taken multiphase flow two-phase momentum conversion velocity is taken as the feature vector x=(x1,x2) T80 flow pattern recognition figures are randomly selected as training set data, and the rest 20 flow pattern recognition figures are selected as test set data;

[0087] Each input neuron of LVQ neural network corresponds to a characteristic value, and the extracted two-phase momentum reduced velocity characteristic index is used as the characteristic value for identifying flow pattern, i.e. 2 neurons of input layer, the number of linear output layer neurons is determined as 5, corresponding to five class labels of annular flow, slug flow, bubbly flow, ring-slug flow and ring-slug flow, and the number of competitive layer neurons is 5, wherein each competitive layer neuron is connected to only one linear output layer neuron;

[0088] The weight ω between the input layer and the competitive layer is initialized ij , and the learning rate η is set as 0.1, the weight represents the flow pattern division interval of the predicted two-phase momentum reduced velocity, and the feature vector is used as the input vector X=(x1, x2) T , which is sent to the input layer, and the weight between the input layer and the competitive layer forms a weight coefficient matrix W=(W1, W2, W3, W4, W5) T , and the distance between the competitive layer neuron and the input vector is calculated:

[0089]

[0090] In the formula, x j is the gas-liquid phase momentum reduced velocity input vector; w ij is the weight between the neuron j of the input layer and the neuron i of the competitive layer;

[0091] The distance d between the input vector and the competitive layer neuron is compared, and when d is the smallest, it indicates that the neuron of the competitive layer wins, and the winning competitive layer neuron outputs 1, and the unwinning output is 0, as shown in the following formula:

[0092]

[0093] The class label of the linear output layer neuron connected to the winning competitive layer neuron is recorded as C i , and the class label corresponding to the input two-phase momentum reduced velocity feature vector is recorded as C x , if C i =C x , the weight is adjusted by the following method:

[0094]

[0095] Otherwise, the weight is updated by the following method:

[0096]

[0097] The algorithm of the LVQ neural network is trained and optimized by mean square error, and a flow pattern feature recognition network of multiphase flow is obtained.

[0098] Beneficial effects:

[0099] (1) The dynamic phase equilibrium model of multi-component alkanes is constructed based on the thermodynamic equilibrium theory, and in the establishment process of the gas-liquid phase equilibrium model, the influence of the multi-component alkanes of the gas in the condensate gas wellbore on the selection of the gas state equation is considered, the PR state equation is selected, which is more suitable for the gas-liquid two-phase in the condensate gas well, and the accuracy of the calculation result of the equilibrium constant is greatly improved. Similarly, since the non-ideality of the solid phase wax crystal particles has a great effect on the calculation result of the liquid-solid phase equilibrium model, the regular solution model and the Wilson model are combined and unified, so that the calculation accuracy of the liquid-solid phase equilibrium model is improved. At the same time, the normal paraffin precipitation is considered, and the regular solution model is modified to obtain the solid phase excess enthalpy of the condensate gas system. The phase equilibrium model fully reflects the influence of the gas-liquid-solid component characteristics of the condensate gas wellbore on the calculation of the equilibrium constant, and improves the accuracy of the calculation of the equilibrium constant.

[0100] (2) In the construction of the gas-liquid continuity equation and the conservation type momentum equation, the mass loss caused by the liquefaction of the gas phase component in the gas-liquid and the mass increment caused by the dissolution of the wax crystal are considered, in the construction of the liquid-solid continuity equation, the mass loss is effectively introduced into the continuity equation in view of the mass increment caused by the change of the multi-component alkanes in the liquid-solid, which greatly increases the comprehensiveness and reliability of the continuity equation. In the establishment of the momentum conservation equation, since the velocity and the force in the one-dimensional steady flow only have one direction in the z-axis, the gas-liquid-solid flow in the microelement control body is analogous to the one-way flow in the horizontal pipeline, and the construction of the conservation type momentum equation is scientifically simplified.

[0101] (3) In the establishment of the conservation type energy equation of the condensate gas system flowing in the wellbore, based on the series thermal resistance network model, the heat transfer process of the condensate gas system is regarded as a series connection of multiple heat conduction thermal resistances and convective heat transfer thermal resistances, the heat transfer area in the multi-layer cylindrical wall is considered to be different from the structure radius of each layer of the wellbore, the heat transfer area and the total heat transfer coefficient are combined, the energy conservation equation is scientifically constructed, the problem of complex energy transfer in the condensate gas system wellbore is solved, and the establishment of the energy conservation equation is realized.

[0102] (Four) The application effectively constructs discrete forms of continuity equation, momentum equation and energy equation by using a difference format, after decoupling processing of multiple variables, the application couples the balance constant in step (1) to construct a gas-liquid-solid flash evaporation model, and continuously iterates the molar fraction of each alkane component in the gas-liquid-solid three-phase, realizes high-precision calculation of volume fraction, density, mass flow, flow rate and other parameters in the multiphase flow model, and more scientifically and intelligently represents the actual condensate gas wellbore temperature and pressure distribution.

[0103] (Five) The application realizes intelligent identification of the component change, wax precipitation characteristics and multiphase flow characteristics of the condensate gas well by applying data analysis program to the multiphase flow parameters calculated in step (Four). At the same time, the application uses the algorithm of LVQ neural network to intelligently identify the flow pattern characteristics in the multiphase flow process, which can automatically extract complex input characteristics, divide flow patterns and correct weights. The principle is clear and feasible, the method is scientific and reliable, and the application can break through the limitation of ignoring the heat and mass transfer problem caused by phase change of the condensate gas system in the traditional method, effectively provide an intelligent method for identifying the multiphase flow characteristics of the condensate gas well, and the scientificity, operability and practicality are strong. The application can provide a beneficial method and basis for deep revelation and effective identification of the phase change behavior of the multiphase flow of the oilfield development system, enrich and expand the coupling theory of multiphase flow and component dynamic change, and guide the development and design application of the condensate gas well of the oil and gas field.

[0104] (Six) The application fully considers the influence of gas-liquid-solid component change on the multiphase flow of the condensate gas well, is not limited to a single phase equilibrium theory or a gas-liquid two-phase flow model, and can identify the flow pattern of the condensate gas wellbore. The application uses convolutional neural network algorithm to intelligently predict the multiphase flow condition of the condensate gas well. The application considers the dynamic change of the gas-liquid ratio of the condensate liquid system caused by the precipitation of wax crystal particles and the anti-condensation phenomenon, and can more accurately identify the dynamic change of the complex flow in the condensate gas well under high gas-liquid ratio conditions. BRIEF DESCRIPTION OF DRAWINGS

[0105] Figure 1 is a schematic diagram of the phase change principle of the condensate gas well of the method.

[0106] Fig. 2 is a force analysis diagram of a unit microelement control body and fluid flow in a multiphase flow process, wherein Fig. 2(a) is a force analysis diagram of a microelement control body; and Fig. 2(b) is a force analysis diagram of unit fluid flow.

[0107] Figure 3 is a schematic diagram of the principle of the LVQ neural network.

[0108] Figure 4 shows the changes in mole fraction and volume fraction of the gas-liquid two-phase system and the density of the gas-liquid-solid two-phase system. Figure 4(a) shows the mole fraction and volume fraction of the gas-liquid two-phase system; Figure 4(b) shows the density change of the gas-liquid-solid phase. Figure 5 shows the changes in the momentum-converted velocity of the gas and liquid two-phase systems within a wellbore depth range of 1000–0 m. Figure 5(a) shows the changes in the momentum-converted velocity of the gas-liquid two-phase system; Figure 5(b) shows the multiphase flow pattern identification of the condensate gas well. In the figures: 1 Condensate gas system 2 Condensate gas wellbore 3 Gas phase 4 Liquid phase 5 Solid phase 6 Total heat transfer 7 Micro-element control volume 8 Gravity 9 Pressure 10 Viscous force 11 Input layer 12 Competition layer 13 Output layer Detailed implementation method:

[0109] The invention will be further described below with reference to the accompanying drawings:

[0110] like Figure 1 As shown, the condensate gas system 1 in the condensate gas wellbore 2 is selected for analysis. The tubing located in the wellbore is the main flow area of ​​the condensate gas system 1. The condensate gas extracted from the reservoir initially enters from the bottom of the tubing in a gaseous state. As the condensate gas system 1 flows, the temperature and pressure drop continuously occur, and the equilibrium state of its multi-alkane components continuously changes dynamically. The single gas phase 3 flow at the bottom of the wellbore gradually evolves into a complex three-phase flow of gas, liquid, and solid at the wellbore surface outlet.

[0111] As shown in Figure 2, a micro-element control volume 7 is selected for analysis. For the continuity equation, the mass increment within the micro-element control volume 7 should be constant equal to the mass change caused by the inflow and outflow of the condensate gas system 1. Considering the gas-liquid composition changes and wax crystal precipitation process of the condensate gas system 1 within the micro-element control volume 7, the mass loss caused by gas-liquid component liquefaction and the mass increment introduced by wax crystal dissolution should be added when constructing the continuity equation. During the flow of the condensate gas system 1 within the wellbore, there are gravity 8, pressure 9, viscous force 10, and interphase drag. Analyzing the control volume, the surface forces -p1A1n1 and -p2A2n2 acting on the control surface by the external fluid, and the pressure 9 (P) and viscous force 10 (T) on the side surface of the control surface, the multiphase flow of the condensate gas well is considered as a one-dimensional steady flow, i.e., velocity and force are considered in one direction along the Z-axis. The gas-liquid-solid flow within the micro-element control volume 7 is analogous to single-phase flow in a horizontal pipe. In the construction of the momentum equation, the viscous force 10 on the control surface and the interphase drag force between the fluid interfaces are combined into a tangential stress, which is equivalent to the frictional stress of single-phase flow.

[0112] like Figure 3As shown, the LVQ neural network is a forward input neural network, wherein when the input vector is sent to the network, the neuron of the competition layer 12 closest to the input mode is activated, the state of the neuron is "1", the state of other neurons of the competition layer 12 is "0", the state of the output layer 13 neuron connected with the activated neuron is also "1", the weight value is calculated and iteratively corrected, and the training of the LVQ neural network algorithm is completed.

[0113] The method for intelligently identifying the multiphase flow characteristics of the condensate gas well has the advantages that:

[0114] (1) Establishment of a gas-liquid-solid phase equilibrium model of the condensate gas well. When the fugacity of each alkane component in the gas phase 3 of the condensate gas system 1 is equal to the fugacity of the liquid phase 4, it indicates that the gas-liquid two-phase reaches an equilibrium state. The PR state equation based on the SRK equation is used to calculate the fugacity of the gas phase 3 and the fugacity of the liquid phase 4. According to the definition of the gas-liquid equilibrium constant, the gas-liquid equilibrium constant of the condensate gas system 1 under a certain temperature and pressure condition is obtained There are:

[0115]

[0116] In the formula, is the mole fraction of the i-th component in the gas phase 3; is the mole fraction of the i-th component in the liquid phase 4; A V and A M are-0.623 and-0.53, respectively; λ is 0.36; is the activity coefficient of the i-th component in the liquid phase 4; R is the ideal gas constant, which is 8.3143 kJ / (kmol·K); T is the system temperature, K; V cal is the volume obtained by the PR state equation, m 3 ; the values of a i , b i are as follows:

[0117]

[0118] In the formula, P represents the system pressure, Pa; P c represents the critical pressure, Pa; T c represents the critical temperature, K; T r is the reduced temperature, that is, the ratio of the system temperature to the critical temperature; M is the molar mass of the pure substance, g / mol.

[0119] Similarly, when the fugacity of each alkane component in the solid phase 5 is equal to the fugacity of the liquid phase 4, it indicates that the liquid-solid two phases reach the equilibrium state. The activity coefficient method is used to calculate the fugacity of the solid phase 5 and the liquid phase 4. According to the definition of the liquid-solid equilibrium constant, the liquid-solid equilibrium constant of the condensate gas system 1 under a certain temperature and pressure condition is obtained There are:

[0120]

[0121] In the formula, Xis the mole fraction of the ith component in the solid phase 5; is the activity coefficient of the component i in the solid phase 5; ΔH f , ΔH tr is the melting enthalpy and solid-solid transition enthalpy, J / mol; T f , T tr is the melting temperature and solid-solid transition temperature, K; ΔC p is the solid-liquid phase heat melting difference, cal / (g·K).

[0122] Thus, the establishment of the gas-liquid-solid phase equilibrium model of the condensate gas well is completed.

[0123] (II) Establishment of the flow equation of the multiphase flow model of the condensate gas well. For the continuity equation, the mass increment of the micro-element control body 7 in the wellbore flow process should be equal to the mass change caused by the inflow and outflow of the condensate gas system 1. The condensate gas system 1 should also supplement the mass loss caused by the liquefaction of the gas phase 3 components and the mass increment introduced by the dissolution of wax crystals. At the same time, it is considered that the time for the condensate gas system 1 to reach the phase equilibrium state is very short and can be ignored. The equation for the liquid phase 4 of the condensate gas system 1 is constructed and simplified as follows:

[0124]

[0125] In the formula, g, l, s respectively represent the gas phase 3, the liquid phase 4 and the solid phase 5; ρ l is the density of the liquid phase 4, kg / m 3 ; α l is the phase volume fraction of the liquid phase 4, %; u l is the velocity of the liquid phase 4, m / s; z is the length of the micro-element control body 7, m; Δm li is the mass change rate of the transformation from the liquid phase 4 to the ith phase, kg / (m 3 ·s), which is A is the cross-sectional area of the condensate gas well, m 2 ; (G l )1 is the mass flow rate of the liquid phase 4 at the inlet of the micro-element control body 7, kg / s; (G l )2 is the mass flow rate of the liquid phase 4 at the outlet of the micro-element control body 7, kg / s.

[0126] Considering the mass increment caused by changes in the polyalkane composition in gas phase 3 and solid phase 5 of condensate gas system 1, the continuity equations for gas phase 3 and solid phase 5 of condensate gas system 1 are as follows:

[0127]

[0128] Based on the momentum conservation equation, the flow of the infinitesimal control volume 7 is analyzed. Taking the liquid phase 4 in the condensate gas system 1 as an example, the non-conservative correlation between the forces acting on the fluid element and its motion state is as follows:

[0129]

[0130] In the formula, t is time (s); τ is viscous force (Pa); F D For interphase traction force, N / m 3 g is the acceleration due to gravity, taken as 9.8 m / s². 2 .

[0131] Analogizing the gas-liquid-solid flow within the micro-element control volume 7 to single-phase flow in a horizontal pipe, that is, combining the viscous force 10 and the interphase drag force into a tangential stress, which is equivalent to the frictional stress that causes frictional resistance in single-phase flow within a horizontal pipe, we have:

[0132]

[0133] In the formula, λ fl D is the hydraulic friction coefficient of liquid phase 4. l Let be the hydraulic radius of liquid phase 4, in meters.

[0134] Simultaneously, converting the mass derivative of the liquid phase 4 velocity, we have:

[0135]

[0136] Combining the continuity equation in equation (2-2), we construct a conserved momentum equation for the condensate gas system in phase 1 (liquid phase 4):

[0137]

[0138] Similarly, the conservation momentum equations for gas phase 3 and solid phase 5 of condensate gas system 1 are as follows:

[0139]

[0140] Thus, the flow equations for the multiphase flow model of condensate gas wells were established.

[0141] (Three) condensate gas well multiphase flow model energy conservation equation is established. In the construction of energy equation, the condensate gas system 1 is regarded as a mixture, according to the principle of energy conservation, the microelement control body 7 of condensate gas system 1 will cause energy change due to the work of gravity 8, pressure 9 and viscous force 10, and there is temperature difference between condensate gas system 1 and formation, which drives two kinds of heat loss including heat conduction, convective heat transfer, and also including the heat formed by phase change of the system, combined with the total energy change rate of condensate gas system 1, the non-conservative energy equation is constructed:

[0142]

[0143] In the formula, E is total energy, J; q is the heat energy density of heat transfer in condensate gas system 1, W / m 3 ; φ is the heat exchange between condensate gas system 1 and the outside world, W; h ji is the phase change enthalpy from the jth phase to the ith phase, J / kg.

[0144] According to Fourier heat conduction law, the total heat transfer 6 between condensate gas system 1 and the formation outside the wellbore can be expressed as:

[0145] φ = -πD in K (T-T a -g T h) dz (3-2)

[0146] In the formula, D in is the inner diameter of condensate gas well, m; K is the total heat transfer coefficient per unit area, W / (m 2 ·K); T a is the surface temperature of condensate gas well wellhead, K; g T is the geothermal gradient, K / m; h is the current wellbore depth of microelement control body 7, m.

[0147] Based on the series thermal resistance network model, the heat transfer process of condensate gas system 1 is regarded as a series of heat conduction thermal resistance and convective heat transfer thermal resistance, according to the multi-layer cylindrical wall heat transfer model, the heat transfer area and the total heat transfer coefficient are combined, then the total heat transfer 6 and the total heat transfer coefficient have:

[0148]

[0149] In the formula, K * is the total heat transfer coefficient of microelement control body 7, W / (m 2 ·K); h f is the convective heat transfer coefficient, W / (m 2 ·K); r m is the radius of the mth layer of heat conducting medium in the direction of heat transfer, m; λ mThe thermal conductivity coefficient of the mth layer of the thermal conductive medium is W / (m·K), and the calculation formula of the convective heat transfer coefficient is as follows:

[0150]

[0151] In the formula, λ is the thermal conductivity coefficient of the micro-element control body 7, W / (m·K); Re is the Reynolds number of the micro-element control body 7, W / (m·K); and Pr is the Prandtl number of the micro-element control body 7, W / (m·K).

[0152] Similarly, the material derivative of the total energy of the system is converted, so that the conservation type energy equation of the condensate gas system 1 flowing in the wellbore is constructed as:

[0153]

[0154] In the formula, the related parameters of the micro-element control body 7 are obtained by volume fraction weighted averaging. In addition, considering the correlation between the temperature and the specific enthalpy of the condensate gas system 1, the diffusion term on the right side of formula (3-5) is converted into the specific internal energy gradient, so that the energy equation is simplified into the specific enthalpy equation:

[0155]

[0156] In the formula, c p is the specific heat capacity at constant pressure of the micro-element control body 7, J / (kg·K); α is the thermal diffusivity, m 2 / s; and e is the specific internal energy of the micro-element control body 7, J / kg.

[0157] Thus, the establishment of the energy conservation equation of the multiphase flow model of the condensate gas well is completed.

[0158] (Four) Solution of the condensate gas well multiphase flow model considering the coupling process of phase change behavior and multiphase flow. According to the finite difference method, the model of the condensate gas wellbore 2 is divided into a one-dimensional equidistant node grid, and the discrete methods of the convection term and the diffusion term of the unknown quantities such as flow velocity, density, and phase volume fraction are respectively a first-order upwind difference format and a central difference format. Then, the discrete forms of the continuity equation, the momentum equation, and the energy equation at the i+1th node are as follows:

[0159]

[0160] Considering that there is a nonlinear coupling of multiple unknown quantities in each equation of formula (4-1), based on the prediction correction strategy of the separation algorithm, it is assumed that there is no mass and heat transfer between the phases in the condensate gas system 1, i.e., the density and the phase volume fraction of each phase do not change, so that the calculation results of the temperature and pressure at the i+1th node can be obtained as:

[0161]

[0162] After the decoupling of the multivariable, the phase transition in the condensate gas system 1 is ignored, which does not conform to the actual process. Therefore, according to the temperature and pressure at the i+1th node of formula (4-2), the molar fraction of each alkane component in the gas-liquid-solid three phases at the i th node of the flow front is used as the initial value of iteration, and the gas-liquid equilibrium constant and the liquid-solid equilibrium constant obtained in step (1) are used to construct the gas-liquid-solid flash model:

[0163]

[0164] Assuming that the initial gas phase 3 molar number is 1, the liquid-solid two-phase molar number is 0, and the bisection method is used to continuously iterate the molar fraction of each alkane component in the gas-liquid-solid three phases through formula (4-3) until the material conservation principle is satisfied.

[0165]

[0166] In the formula, eps x is the calculation accuracy of the molar fraction of the multicomponent alkane phase equilibrium process.

[0167] According to the molar fraction of each alkane component in the gas-liquid-solid three phases, the corrected gas-liquid equilibrium constant and liquid-solid equilibrium constant are obtained. When the accuracy requirement in formula (4-4) is met, the phase equilibrium parameters of the multicomponent alkane at the i+1th node can be obtained. Otherwise, formula (4-2) and formula (4-3) need to be repeated until the accuracy requirement in formula (4-5) is met.

[0168]

[0169] In the formula, eps K is the calculation accuracy of the equilibrium constant of the multicomponent alkane phase equilibrium process.

[0170] The volume fraction, density, mass flow, and flow velocity of the gas, liquid, and solid phases at the outlet node are calculated according to formula (4-6):

[0171]

[0172] At the same time, since the flow process of the condensate gas system 1 in the wellbore should satisfy the phase continuity equation in formula (4-1), the velocity and density of the gas, liquid, and solid phases at the i+1th node are obtained according to the difference form of the continuity equation and are twice corrected. The corrected results are combined with the uncorrected wellbore temperature and pressure distribution of the condensate gas system 1 in formula (4-2) to construct the relative error, and iteration is performed until the corrected temperature and pressure can meet the accuracy requirement of formula (4-7).

[0173]

[0174] In the formula, epsP Pa;eps is the pressure calculation accuracy in the multiphase flow model, Pa T Tcalc is the temperature calculation accuracy in the multiphase flow model, K.

[0175] Thus, the solution of the multiphase flow model of the condensate gas well considering the coupling process of the phase change behavior and the multiphase flow is completed.

[0176] (Five) Intelligent identification of the multiphase flow characteristics in the wellbore 2 of the condensate gas well. According to the parameters of the phase volume fraction, the mass flow, the flow velocity, the temperature, the pressure, the density and the like in the wellbore flow process obtained by the model solution of step (four), the data analysis program is applied to realize the automatic intelligent identification of the multiphase flow characteristics of the component change, the temperature and pressure distribution and the wax precipitation characteristics of the condensate gas well. The flow pattern identification diagram established for the gas-liquid two-phase momentum conversion velocity is identified and trained by using the LVQ neural network. The LVQ neural network used is composed of three layers of neurons, i.e., the input layer 11, the competitive layer 12 and the output layer 13. The flow pattern identification diagram of the wellbore partial depth range is obtained according to step (four) and the flow pattern diagram of Hewitt and Roberts. The two-dimensional data points of the gas-liquid two-phase momentum conversion velocity at 100 different depths are taken from the flow pattern identification diagram by using the point taking program. The extracted two-phase momentum conversion velocity is taken as the characteristic vector x = (x1, x2) T . The characteristic vectors extracted from 80 randomly selected flow pattern identification diagrams are taken as the training set data, and the characteristic vectors extracted from the remaining 20 flow pattern identification diagrams are taken as the test set data.

[0177] Each input neuron of the LVQ neural network corresponds to a characteristic value. The extracted two-phase momentum conversion velocity characteristic index is taken as the characteristic value for identifying the flow pattern, i.e., the two neurons of the input layer 11. The number of neurons of the linear output layer 13 is determined to be five, corresponding to the five class labels of the annular flow, the slug flow, the churn flow, the bubble flow and the ring-slug flow. The number of neurons of the competitive layer 12 is five, wherein each competitive layer 12 neuron is connected to only one linear output layer 13 neuron.

[0178] The weights ω ij and the learning rate η between the input layer 11 and the competitive layer 12 are initialized. The learning rate η is set to be 0.1. The weights represent the flow pattern division interval of the predicted two-phase momentum conversion velocity. The characteristic vector is taken as the input vector X = (x1, x2) T which is sent to the input layer 11. The weights between the input layer 11 and the competitive layer 12 can form a weight coefficient matrix W = (W1, W2, W3, W4, W5) T . The distance between the competitive layer 12 neuron and the input vector is calculated as follows:

[0179]

[0180] In the formula, x jis the input vector of two-phase momentum reduced velocity; ij is the weight between neuron j of input layer 11 and neuron i of competitive layer 12.

[0181] The size of the distance d between the input vector and the neuron of competitive layer 12 is compared, and when d is the smallest, it indicates that the neuron of competitive layer 12 wins, and the winning neuron of competitive layer 12 outputs 1, and the output of the non-winning neuron is 0, as shown in the following formula:

[0182]

[0183] The class label of the linear output layer 13 neuron connected with the winning neuron of competitive layer 12 is C i . The class label corresponding to the input two-phase momentum reduced velocity feature vector is C x , if C i =C x , the weight is adjusted as follows:

[0184]

[0185] Otherwise, the weight update is as follows:

[0186]

[0187] Considering that the algorithm of LVQ neural network also needs to be trained and optimized through a loss function, mean square error (MSE) is a commonly used loss function, which is the sum of squares of the difference between the predicted value and the target value, and its formula is as follows:

[0188]

[0189] Where N is the number of two-phase velocity samples at different depths in the training set; y i represents the actual value of the weight; represents the predicted value of the weight.

[0190] Set the accuracy requirement to 0.01, if the mean square error of the training set does not meet the accuracy requirement, then use the LVQ neural network algorithm to perform a new round of iteration correction on the weight, when the mean square error of the training set meets the accuracy requirement, the two-phase momentum reduced velocity feature indicators in the test set are used as the characteristic values for identifying the flow pattern, and the weights are obtained in sequence according to formulas (5-1) to (5-4), when the mean square error also meets the accuracy requirement, it indicates that the overfitting effect of the intelligent identification model of the condensate gas well multiphase flow pattern is weak, and the model training is completed, otherwise, the training set is re-inputted and the training is continued until the accuracy condition is met. After the flow pattern recognition of the test set is completed and the accuracy condition is met, the recognition network of the multiphase flow pattern characteristics can be determined, and the flow pattern of any two-phase momentum reduced velocity can be judged and a conclusion is quickly given.

[0191] Thus, the intelligent identification of the multiphase flow characteristics of the condensate gas well is completed.

[0192] The above steps are repeated, the gas-liquid ratio condition of the condensate gas well is changed, another kind of condensate gas well condition is determined, and the intelligent identification of the multiphase flow characteristics of different condensate gas wells is realized.

[0193] In the application, a i And b i Obtained by calculation of pure substance properties; the calculation volume V cal Obtained by PR state equation method; activity coefficient γ i Obtained by residual activity coefficient method and combined activity coefficient method; Van der Waals volume v wi Obtained by group volume increment method; the total heat transfer φ between the condensate gas system and the formation outside the wellbore is obtained by Fourier heat conduction law.

[0194] Embodiment

[0195] The method for intelligently identifying the multiphase flow characteristics of the condensate gas well is applied in a confidential manner, the application object is a condensate gas well with a bottom hole pressure of 120 MPa and a mass flow of 5 kg / s, the ground temperature is 285 K, the geothermal gradient is 3.29 K / 100 m, the oil jacket annulus heat transfer coefficient is 0.9 W / m*K, and the pipe material heat transfer coefficient is 45 W / m*K.

[0196] As shown in FIG. 4, in the condensate gas well wellbore, the change of the molar fraction of the gas phase and the liquid phase lags behind the phase volume fraction, with the continuous decrease of the gas-liquid ratio, the carbon number of the alkane component added to the liquid phase decreases, and at the same time, the liquid phase density shows a continuous decreasing trend.

[0197] In order to intelligently identify the flow pattern of the gas-liquid two-phase flow in the condensate gas well wellbore, the gas-liquid two-phase momentum conversion velocity change in the range of 1000-0 m of wellbore depth is obtained as shown in FIG. 5(a);

[0198] The results show that in the single gas phase flow stage, the gas phase momentum conversion velocity is always unchanged, but when the wellbore depth decreases to 846 m, the gas-liquid ratio decreases, the liquid phase is precipitated, with the gradual increase of the liquid phase molar fraction, the gas phase momentum conversion velocity rapidly decreases, and the liquid phase momentum conversion velocity rapidly increases, but since the content of the precipitated solid phase is less compared with the gas-liquid two-phase, the change range of the gas-liquid two-phase momentum conversion velocity can be regarded as the same. Based on this, in the production process of the wellbore with the same bottom hole pressure, bottom hole temperature and mass flow, the gas-liquid two-phase flow structure is identified by using the Hewitt and Roberts flow pattern diagram as shown in FIG. 5(b);

[0199] According to the flow pattern identification diagram, the two-phase momentum reduced velocity in the diagram is extracted as the input feature vector of the LVQ neural network by using a point extraction program, the distance between the competitive layer neuron and the input vector is calculated, and finally the weight iterative correction is carried out, so that the flow pattern of the two-phase momentum reduced velocity at a certain depth in the condensate gas well is quickly identified, and the intelligent identification of the multiphase flow pattern of the condensate gas well is realized.

[0200] The application well solves the problems of dynamic changes of gas-liquid-solid components in the multiphase flow process of the condensate gas well, especially solves the problems of mass imbalance calculation caused by solid phase flow and intelligent identification technology of multiphase flow characteristics. The gas-liquid-solid thermodynamic phase equilibrium model and the multiphase flow model are coupled to construct a gas-liquid-solid flashing model, and a method for obtaining the characteristic parameters of the multiphase flow of the condensate gas well is given. From the perspective of theoretical calculation, the intelligent identification of the multiphase flow pattern is carried out through the LVQ neural network learning, and the method for intelligently identifying the multiphase flow characteristics of the condensate gas well is effectively constructed. The method can realize the intelligent identification of the component density change, temperature and pressure drop, wax precipitation characteristics and flow pattern of the condensate gas well, and has clear principle, clear process, scientific method and strong operability and practicability. The method can provide a method and basis for the effective intelligent identification of the flow prediction and wellbore blockage in the production process of the condensate gas well in the oil and gas field, and guide and promote the development and design of the high-efficiency wax removal technology of the condensate gas well. Especially, the intelligent identification of the multiphase flow characteristics of the condensate gas well is realized through the neural network learning, which can provide an important basis for enriching and expanding the coupling theory of the multiphase flow and component dynamic change and the intelligent identification method of the multiphase flow characteristics.

[0201] The application mainly limits the research on the condensate gas well flow to the establishment and correction of single phase equilibrium theory or gas-liquid two-phase flow model, and does not consider the coupling process of gas-liquid-solid three-phase dynamic balance and multiphase flow model, especially ignores the mass imbalance in the liquid-solid phase balance calculation caused by the solid phase flow in the multiphase flow model, and does not pay too much attention to the influence of the gas-liquid ratio on the condensate gas wellbore flow and the phase equilibrium process. The application establishes an intelligent identification model of the flow characteristics of the condensate gas well under high gas-liquid ratio conditions by coupling the multiphase flow and the gas-liquid-solid component change, to provide a solution for describing the temperature and pressure distribution and flow pattern identification of the condensate gas well under high gas-liquid ratio conditions, and realize the accurate prediction and identification of the multiphase flow characteristics of the condensate gas well under high gas-liquid ratio conditions.

[0202] The first and second steps of the application are the construction of the gas-liquid, liquid-solid phase equilibrium model and the continuity equation, momentum conservation equation and energy conservation equation; the third step is to construct a gas-liquid-solid flashing model, and iteratively correct the related parameters until the accuracy requirement is met; the fourth step is intelligent deep learning, and the calculation results are characterized; and the fifth step is to realize the accurate identification method of the multiphase flow characteristics of the condensate gas well under high gas-liquid ratio.

Claims

1. A method for intelligently identifying multiphase flow characteristics of condensate gas wells, characterized in that... Includes the following steps: (i) Correlation to obtain the gas-liquid equilibrium constant of the condensate gas system under a certain temperature and pressure condition of the condensate gas well. and liquid-solid equilibrium constant (ii) Establish the flow equations for the multiphase flow model of condensate gas wells, including the liquid phase continuity equation of the condensate gas system, the gas phase and solid phase continuity equations of the condensate gas system, the liquid phase conservation momentum equation of the condensate gas system, and the gas phase and solid phase conservation momentum equations of the condensate gas system. (III) Establishing the energy conservation equation for a multiphase flow model of condensate gas wells: In the formula, c p α is the isobaric specific heat capacity of the infinitesimal control element, in J / (kg·K); α is the thermal diffusivity, in m³ / s. 2 / s; e is the specific internal energy of the infinitesimal control element, in J / kg; z is the length of the infinitesimal control element, in m. φ represents the heat exchange between the condensate gas system and the outside environment, expressed in W. T is the system temperature, in K. (iv) Solving the multiphase flow model of condensate gas well considering the coupling process of phase change behavior and multiphase flow; The condensate gas wellbore model is meshed using one-dimensional equally spaced nodes. The obtained gas-liquid equilibrium constant and liquid-solid equilibrium constant, as well as the flow equation of the multiphase flow model of condensate gas well, are used to solve the energy conservation equation of the multiphase flow model of condensate gas well. First, the multiphase flow parameters at the (i+1)th node at a certain height of the wellbore are calculated. Then, the multiphase flow parameters at each node are solved at equal intervals. The multiphase flow parameters at all nodes of the entire wellbore are solved to obtain the multiphase flow parameters in the flow process of the wellbore. The multiphase flow parameters include phase volume fraction, mass flow rate, flow velocity, temperature, pressure, and density. (V) Intelligent identification of multiphase flow characteristics in condensate gas wellbore; Based on the phase volume fraction, mass flow rate, flow velocity, temperature, pressure and density obtained in step (IV) during the flow process in the wellbore, the flow pattern identification diagram established by the momentum-conversion velocity of the gas and liquid two phases is trained by LVQ neural network to obtain the identification network of multiphase flow pattern characteristics. The flow pattern of any two phases is judged and a conclusion is given quickly, realizing the automatic intelligent identification of multiphase flow characteristics such as component changes, temperature and pressure distribution and wax precipitation characteristics in condensate gas wells.

2. The method for intelligently identifying multiphase flow characteristics of condensate gas wells according to claim 1, characterized in that: Step (I) specifically includes: Considering that the fugacity of the gas phase of each alkane component in the condensate gas system is equal to that of the liquid phase, indicating that the gas and liquid phases have reached equilibrium, the PR equation of state, which applies the concept of volume shift based on the SRK equation, is used to calculate the fugacity of the gas and liquid phases. According to the definition of the gas-liquid equilibrium constant, the gas-liquid equilibrium constant of the condensate gas system under a certain temperature and pressure conditions is obtained. In the formula, denoted as the mole fraction of the i-th component in the gas phase; A represents the mole fraction of the i-th component in the liquid phase. V and A M The values ​​of are -0.623 and -0.53 respectively; the value of λ is 0.36; Let be the activity coefficient of the i-th component in the liquid phase; R be the ideal gas constant, taken as 8.3143 kJ / (kmol·K); T be the system temperature in K; V be the system temperature in K. cal The volume obtained through the PR equation of state, in meters. 3 ;a i b i The method for determining the value is as follows: +35.3155(lnM) 2 In the formula, P represents the system pressure, with the unit being Pa; c T represents the critical pressure, with units of Pa; c T represents the critical temperature, with units of K; r The comparison temperature is the ratio of the system temperature to the critical temperature. M is the molar mass of the pure substance, expressed in g / mol; When the fugacity of the solid phase of each alkane component is equal to that of the liquid phase, it indicates that the liquid and solid phases have reached equilibrium. The activity coefficient method is used to calculate the fugacity of the solid and liquid phases. Based on the definition of the liquid-solid equilibrium constant, the liquid-solid equilibrium constant of the condensate gas system under a certain temperature and pressure conditions is obtained. In the formula, Let be the mole fraction of the i-th component in the solid phase; Let ΔH be the activity coefficient of component i in the solid phase; f ΔH tr T represents the enthalpy of fusion and the enthalpy of solid-solid transformation, expressed in J / mol. f T tr These are the melting temperature and solid-solid transition temperature, in Kelvin (K); ΔC p The solid-liquid phase thermal fusion difference is expressed in cal / (g·K).

3. The method for intelligently identifying multiphase flow characteristics of condensate gas wells according to claim 2, characterized in that: Step (II) is as follows: The mass increment within a small control element of length dz during wellbore flow is always equal to the mass change caused by the inflow and outflow of the condensate gas system. This condensate gas system should also be supplemented with the mass loss caused by liquefaction of gaseous components and the mass increment introduced by the dissolution of wax crystals. The liquid-phase continuity equation for the condensate gas system is constructed as follows: In the formula, g, l, and s represent the gas phase, liquid phase, and solid phase, respectively; ρ l This refers to the density of the liquid phase, expressed in kg / m³. 3 ;α l The volume fraction of the liquid phase, expressed in %; u l The velocity of the liquid phase is expressed in m / s. z is the length of the infinitesimal control volume, in meters (m). Δm li The mass change rate from liquid phase to the i-th phase is expressed in kg / (m³). 3 ·s), whose value is A represents the cross-sectional area of ​​the condensate gas wellbore, in meters (m²). 2 ;(G l )1 represents the mass flow rate of the liquid phase at the inlet of the micro-element control volume, in kg / s; (G l )2 represents the mass flow rate of the liquid phase at the outlet of the micro-element control volume, in kg / s; In the gas and solid phases of the condensate gas system, considering the mass increment caused by changes in the polyalkane composition, the gas and solid phase continuity equations of the condensate gas system are constructed as follows: By analogy between the gas-liquid-solid flow within the micro-element control body and the single-phase flow within a horizontal pipe, the viscous force and interphase drag force are combined into a tangential stress, which is equivalent to the frictional stress that causes frictional resistance in single-phase flow within a horizontal pipe. Combining this with equation (2-2), a liquid-phase conservation momentum equation for the condensate gas system is constructed: Reconstruct the gas- and solid-phase conservation momentum equations for the condensate gas system: In the formula, λ fl D is the coefficient of hydraulic friction of the liquid phase; l denoted as ρ, where ρ is the hydraulic radius of the liquid phase, in meters (m); g is the acceleration due to gravity, taken as 9.8 m / s². 2 .

Images