General pressure measurement fluidity interpretation method

Abstract

The invention belongs to the field of electric digital data processing, and particularly relates to a general pressure measurement fluidity interpretation method. According to the method, multiple advanced algorithms and technologies such as numerical simulation, orthogonal experimental design, neural network algorithm and iterative computation are comprehensively applied. The algorithms and technologies are mutually matched and complemented to form a complete pressure measurement fluidity interpretation system. The numerical simulation provides fine geological and fluid models, the orthogonal experimental design defines main control factors, the neural network algorithm realizes intelligent prediction, and the iterative calculation further optimizes the fluidity interpretation result. The mode of comprehensively applying multiple algorithms not only improves the accuracy and efficiency of pressure measurement fluidity explanation, but also enhances the reliability and stability of the explanation result, can better meet the accurate explanation requirements of pressure measurement fluidity under different probe types, reservoir characteristics and pumping parameters, and has good application prospects. And the method can better adapt to the requirements of fluidity interpretation under complex geological conditions.

Description

Technical Field

[0001] The present invention belongs to the field of electrical digital data processing, and particularly relates to a general pressure-measuring mobility interpretation method. Background Art

[0002] At present, for the interpretation of pressure-measuring mobility, the main interpretation models are mainly based on the seepage of single-phase fluid, that is, the fluid in the formation pores is assumed to be single-phase fluid, and considering different influencing factors of the testing process, the relationship between formation and instrument parameters and the testing pressure response is obtained by using analytical or numerical methods. The commonly used pressure-measuring mobility interpretation methods now include the quasi-steady-state pressure drop method, the spherical flow pressure build-up method, the radial flow pressure build-up method, the formation flow rate analysis method, the automatic fitting analysis method, the neural network method, the least square method, etc. Among them, the quasi-steady-state pressure drop method, the spherical flow pressure build-up method, and the radial flow pressure build-up method are the most commonly used methods to determine the formation mobility. The quasi-steady-state pressure drop method uses the quasi-steady-state solution of the spherical flow equation to obtain the permeability value, and the permeability is the parameter to be measured, and repeated measurements are required to determine the critical permeability value. Therefore, this method is generally only used for qualitative interpretation; both the spherical flow and the radial flow pressure build-up methods use the straight line segment that appears on the double logarithmic curve of pressure and time during the pressure build-up in the later stage of the test to obtain the formation permeability, which requires an extended test time, increasing the risk of the instrument getting stuck, and ignoring the dynamic pressure changes in the initial and middle stages of the test. When the straight line segment on the double logarithmic curve is difficult to appear, the data cannot be interpreted, resulting in a low accuracy of pressure-measuring mobility interpretation.

[0003] The analysis results of pressure-measuring mobility are affected not only by the interpretation model used, but also by the probe type, the time and speed of pressure measurement extraction, the formation permeability size, etc. The pressure-measuring tools actually used in the field mainly include MDT of Schlumberger, RCI of Baker Hughes, etc. The structural differences of different instruments are relatively large, and they have different probe area sizes, resulting in a large difference in the contact form with the formation; in order to improve the working efficiency, the pressure measurement process is often short, about dozens of seconds, and the fluid extraction speed is large, resulting in large pressure fluctuations, which have a greater impact on the pressure measurement interpretation results; the low-permeability formation reservoir is dense, the fluid seepage ability is poor, the pressure propagation range is limited, and it is difficult to reach the steady / quasi-steady state flow, and the flow form is mainly non-steady state, resulting in poor reliability of the pressure-measuring mobility interpretation results for dense reservoirs.

[0004] Chinese Patent Document CN112147051A, a method for standardizing pressure-measured mobility based on the permeability distribution pattern, includes: 1) Selecting the test data of a certain formation testing instrument in a certain area as the standard, and establishing the mobility and its corresponding permeability data volume for different wells and different depths; 2) Establishing the corresponding relationship data between the test mobility of the well to be standardized and the reservoir permeability, and making the frequency distribution map of the permeability data; 3) According to the same permeability interval as the well to be standardized, and making the frequency distribution map of the permeability data of each well, respectively calculating the similarity coefficient between the permeability distribution pattern of the well to be standardized and the permeability distribution pattern of each standard well; 4) Finding the well or sub-data volume with a permeability distribution pattern similar to that of the well to be standardized data and the standard test results, and performing standardization processing on the well to be standardized data. However, this method is only applicable to standardizing the mobility data measured by different instruments when the permeability distribution patterns are similar; in the case of large differences in permeability distribution patterns, it is impossible to accurately standardize the mobility, especially in complex geological conditions such as low-permeability and tight formations, its applicability is limited; in addition, this method mainly relies on the similarity of the permeability distribution pattern to standardize the mobility, this similarity may be somewhat subjective, and it does not fully consider the main influencing factors of pressure-measured mobility, lacking in-depth excavation of dynamic data, resulting in limited interpretation accuracy; it also has defects such as strong data dependence, single data processing method, and insufficient treatment of differences between different instruments.

[0005] The Chinese literature "Using the MDT Pressure Drop Mobility Trend Translation Method to Assist Well Logging Interpretation Calibration" found through the analysis of the MDT pressure drop mobility of 30 cored wells of different types in the Pearl River Mouth Basin that there is a very good consistency between the MDT pressure drop mobility trend and the permeability trend. Therefore, without considering the dimension, the MDT pressure drop mobility trend can be translated left and right to reach the calibrated permeability. After analyzing the pressure drop mobility scale range for the key wells of each oilfield, this scale range is applied to other wells in this oilfield and oilfields of the same type. Using the MDT pressure drop mobility trend translation method to calibrate the well logging interpretation permeability of the section without core or with unqualified core analysis can further improve the well logging interpretation accuracy and reserve evaluation accuracy, make up for the lack of core data, and minimize cost and increase efficiency to the greatest extent. This method mainly relies on the consistency between the MDT pressure drop mobility trend and the permeability trend for calibration. Although this method of trend translation can improve the well logging interpretation accuracy to a certain extent, it does not fully consider the main influencing factors of pressure-measured mobility, such as probe area, pressure measurement time, permeability, pressure measurement speed, and fluid viscosity, which may lead to certain deviations in the calibration results; in addition, this method mainly focuses on the trend translation of data, and does not conduct in-depth research and solution on the complexity of the seepage law of low-permeability reservoirs and the reliability of pressure-measured mobility interpretation results, and the ability to interpret pressure-measured mobility in special geological conditions such as low-permeability reservoirs is limited.

[0006] Therefore, the currently published related research mainly focuses on the calibration of existing pressure-measured mobility interpretation models, failing to fully consider the influences of probe types, pumping parameters, reservoir seepage capacity, etc., and no general precise mobility interpretation method has been formed. Moreover, currently, the mobility interpretation methods of most wireline formation testing instruments on the market are mainly based on simplified theoretical models, usually without considering the influences of complex factors such as pumping time and speed. When dealing with low-permeability oil reservoirs or contaminated reservoirs, these traditional methods often lead to a significant decrease in interpretation accuracy due to the neglect of these key factors. For example, the neglect of the influences of pressure-measuring time and speed may lead to misjudgments of reservoir dynamic responses. Therefore, this simplified method is difficult to meet the requirements of high-precision reservoir evaluation under complex geological conditions. The present invention intends to clarify the main controlling factors of pressure-measured mobility interpretation results through orthogonal experimental design and propose a general precise pressure-measured mobility interpretation method. Summary of the Invention

[0007] In view of the above problems, the present invention provides a general pressure-measured mobility interpretation method. First, the present invention uses numerical simulation and orthogonal experimental design to clarify the main controlling factors of pressure-measured mobility; secondly, according to the differences in reservoir characteristics and tool types, a series of parameters are set to establish a sample library of G correction coefficients for pressure-measured mobility interpretation; finally, the neural network algorithm is used, through a large number of trainings and tests, to construct a relationship model between G correction coefficients and pumping and reservoir parameters to achieve precise interpretation of pressure-measured mobility.

[0008] To achieve the above object, the present invention adopts the following technical solutions: A general pressure-measured mobility interpretation method, comprising the following steps: S1. Establish a fine reservoir numerical simulation model for the pumping pressure measurement process; S2. Based on the fine reservoir numerical simulation model established in step S1, determine the main controlling factors of pumping pressure-measured mobility through orthogonal experimental design; S3. Based on the main controlling factors of pumping pressure-measured mobility determined in step S2, use the fine reservoir numerical simulation model established in step S1 to establish a database model of G correction coefficients for pumping pressure-measured mobility; S4. Based on the database model of G correction coefficients for pumping pressure-measured mobility established in step S3, use the neural network algorithm, through training and verification, to construct a prediction model of G correction coefficients based on the neural network algorithm; S5. Based on the G correction coefficient prediction model established in step S4, establish a general precise mobility interpretation method for reservoirs based on the iterative method.

[0009] Preferably, step S1 includes establishing a multi-layer radial grid geological model and a multi-phase multi-component seepage model module, a model operation module, and a model result processing module.

[0010] Further preferably, the module for establishing a multi-layer radial grid geological model and a multi-phase and multi-component seepage model is based on the technological parameters of the oil and gas reservoir, fluid properties, probe size, and pumping dynamics, and uses grid encryption, fluid zoning, and multi-phase fluid state equations to establish a refined multi-layer radial grid geological model and a multi-phase and multi-component seepage model to simulate the fluid flow during the pumping pressure measurement process; The model operation module is to establish a probe model, and in combination with the above-established multi-layer radial grid geological model and multi-phase and multi-component seepage model, quickly generate a refined reservoir numerical simulation model that truly reflects the pumping pressure measurement process; The model result processing module is to run the refined reservoir numerical simulation model of the pumping pressure measurement process generated above and directly output relevant dynamic curves.

[0011] Preferably, the orthogonal experimental design in step S2 refers to considering the actual situation on site, selecting possible influencing factors of the pressure measurement mobility, setting the variation range of the influencing factors, and using the orthogonal experimental design method to determine the experimental scheme.

[0012] Further preferably, the possible influencing factors of the pressure measurement mobility include one or more combinations of reservoir permeability, fluid viscosity, pressure measurement extraction time and speed, and probe area.

[0013] More preferably, the possible influencing factors of the pressure measurement mobility include reservoir permeability, fluid viscosity, pressure measurement extraction time and speed, and probe area.

[0014] Further preferably, in step S2, based on the refined reservoir numerical simulation model of the pumping pressure measurement process established in step S1, the experimental scheme is simulated and calculated: the interpreted mobility is calculated by applying the area integration method, and based on the actual fluid mobility, the mobility calculation deviation and range of each scheme are obtained.

[0015] Still preferably, in step S2, according to the calculated range size, the main controlling influencing factors of the pressure measurement mobility interpretation result are determined: the ranges are sorted from largest to smallest, the influencing factor with the smallest range is excluded, and the remaining influencing factors are used as the main controlling influencing factors of the pressure measurement mobility interpretation result.

[0016] Preferably, step S3 is specifically: according to the actual situation on site, a series of parameters are set for the main controlling influencing factors of the pressure measurement mobility determined in step S2, the Latin hypercube sampling method is introduced to determine the simulation scheme, and the refined reservoir numerical simulation model of the pumping pressure measurement process established in step S1 is used for simulation. The interpreted mobility is calculated by applying the area integration method, and based on the actual fluid mobility, the G correction coefficient of each scheme is calculated, and a database model between the main controlling influencing factors of the pressure measurement mobility and the G correction coefficient is constructed, where G = actual fluid mobility / interpreted mobility.

[0017] Further preferably, the neural network algorithm described in step S4 is ANN.

[0018] Preferably, step S5 includes: 1) Determine the parameter values of the G correction coefficient prediction model established in step S4 and input the determined parameter values into the G correction coefficient prediction model established in step S4 to obtain the G correction coefficient. Use the area integration method to calculate the fluid mobility, and this fluid mobility is the interpreted mobility ( δ, v, t, k ). The parameters include the probe area, the pressure measurement extraction speed, the pressure measurement extraction time, and the permeability. The interpreted mobility ( δ, v, t, k ) δ is the probe area, v is the pressure measurement extraction speed, t is the pressure measurement extraction time, k is the permeability; 2) Calculate the fluid mobility according to the permeability described in 1), and this fluid mobility is the true reservoir mobility; 3) Compare the interpreted mobility ( δ, v, t, k ) with the true reservoir mobility: When the set termination condition is not met, perform iterative calculations until the set termination condition is reached; 4) Export the final fluid mobility value to achieve the interpretation of the general fluid mobility.

[0019] Further preferably, the parameter values of the G correction coefficient prediction model established in step S4 are determined according to the following method: For the actual pump pressure measurement curve, determine its probe area according to the used wireline formation testing tool and construction technology; Analyze the speed and time data extracted during the actual pressure measurement process to determine the pressure measurement extraction speed and pressure measurement extraction time input into the model; According to the data in the pressure build-up section, perform data point encryption and extended regression processing on it, and comprehensively consider the core-taking, logging, and well logging data to give the initial permeability of the formation in the test section.

[0020] Further preferably, the termination condition is that the interpretation error is less than 3%, and the interpretation error is the error between the interpreted mobility ( δ, v, t, k ) and the true reservoir mobility.

[0021] The present invention also provides the application of the general pressure measurement mobility interpretation method for the development of a general pressure measurement mobility interpretation module, the interpretation of effective point mobility in the oilfield, or the interpretation of tight point mobility in the oilfield.

[0022] Compared with the prior art, the present invention has the following beneficial effects: (1) Through the orthogonal experimental design in step S2, the main factors affecting the pressure measurement mobility interpretation results during the pump pressure measurement process can be systematically analyzed and determined. This method avoids the previous empirical or one-sided selection of influencing factors, thus providing a more scientific and accurate basis for subsequent model establishment and mobility interpretation. Combining with the fine reservoir numerical simulation model established in step S1, the orthogonal experimental design can be carried out in a highly realistic simulation environment, making the determined main control factors more in line with the actual pressure measurement situation, thereby significantly improving the accuracy of pressure measurement mobility interpretation.

[0023] (2) Based on the determined main control factors in step S3, the Latin hypercube sampling method is used to construct a sample library of the G correction coefficient for pressure measurement mobility. This process not only considers the comprehensiveness of the main control factors but also ensures the diversity and representativeness of the samples through advanced sampling techniques. The establishment of the sample library provides a rich and accurate data basis for the subsequent prediction model, enabling the model to better learn and capture the complex relationships between pressure measurement mobility and various influencing factors. Combined with steps S1 and S2, the construction process of the sample library makes full use of the fineness of the numerical simulation model and the scientific nature of the orthogonal experimental design to ensure the quality and reliability of the data.

[0024] (3) In step S4, a neural network algorithm is used to establish a prediction model for the G correction coefficient. The neural network algorithm has strong nonlinear fitting ability and generalization ability, and can automatically learn and extract features from a large amount of sample data, so as to achieve accurate prediction of the G correction coefficient. Combining steps S1 - S3, the establishment of the prediction model is carried out on the basis of fine numerical simulation, scientific experimental design and comprehensive sample library, making the model have higher accuracy and reliability. At the same time, the high computing ability of the neural network algorithm also greatly improves the computing efficiency of pressure measurement mobility interpretation and can quickly process and interpret actual pressure measurement data.

[0025] (4) The present invention comprehensively utilizes a variety of advanced algorithms and technologies such as numerical simulation, orthogonal experimental design, neural network algorithm and iterative calculation. These algorithms and technologies cooperate with each other and complement each other to form a complete pressure measurement fluidity interpretation system. Numerical simulation provides a fine geological and fluid model, orthogonal experimental design clarifies the main control factors, neural network algorithm realizes intelligent prediction, and iterative calculation further optimizes the fluidity interpretation results. This comprehensive use of multiple algorithms not only improves the accuracy and efficiency of pressure measurement fluidity interpretation, but also enhances the reliability and stability of the interpretation results, and can better meet the needs of accurate interpretation of pressure measurement fluidity under different probe types, reservoir characteristics and pumping parameters. At the same time, the method proposed in the present invention is particularly suitable for pressure measurement fluidity interpretation under complex geological conditions such as low permeability and denseness. Under these geological conditions, traditional pressure measurement fluidity interpretation methods are often difficult to obtain accurate results. The present invention, by comprehensively considering multiple influencing factors, establishes a fine numerical simulation model and an intelligent prediction model, which can better meet the needs of fluidity interpretation under complex geological conditions.

[0026] (6) In view of the current complex seepage laws of low-permeability reservoirs, the large differences in flow interpretation results of different instrument methods for the same type of reservoirs, and the insufficient interpretation of dynamic data mining such as pumping pressure drop and pressure recovery, the present invention innovatively proposes a universal pressure measurement flow interpretation method based on a neural network algorithm. The method is easy to understand and simple and quick to apply. It can quickly obtain accurate pressure measurement flow according to the actual pumping pressure measurement dynamics of the target block, and can be used to subsequently calculate the production capacity of the layer section. It has a strong guiding significance for the flow interpretation and production capacity evaluation of tight low-permeability reservoirs.

[0027] (7) Based on the orthogonal experimental design to clarify the main controlling factors of the pressure measurement flow rate interpretation results, the present invention integrates multiple algorithms such as artificial neural networks and iterative calculations to propose a universal pressure measurement flow rate interpretation method that is suitable for different probe types, reservoir characteristics and pumping parameters. This method solves the problem that the pressure measurement flow rate interpretation results of different instruments are very different and dense points cannot be explained. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 This is a schematic diagram of a numerical simulation model for pumping and pressure measurement in a certain oil reservoir in Example 1 of the present invention; Figure 2 This is a schematic diagram of the pumping process of the cable formation test in Example 1 of the present invention; Figure 3 This is a schematic diagram of the near-wellbore mud contamination zone in Example 1 of the present invention; Figure 4 This is a schematic diagram of oil-water relative permeability in Example 1 of the present invention; Figure 5 This is a schematic diagram of a multi-layer radial grid geological model with probes according to Example 1 of the present invention; Figure 6 It is the pressure drop - recovery graph in Embodiment 1 of the present invention; Figure 7 It is the graph for ranking the influence degrees of various factors on the interpretation result of pressure - testing mobility in Embodiment 1 of the present invention; Figures 8(a) and 8(b) are respectively the schematic diagrams of the training and testing results of the G - coefficient correction prediction model based on neural network in Embodiment 1 of the present invention; Figure 9 It is the technical route flowchart of the general pressure - testing mobility interpretation method in Embodiment 1 of the present invention. Specific Embodiments

[0029] The present invention will be further described in detail below in combination with specific embodiments and the accompanying drawings of the specification, but the embodiments of the present invention are not limited thereto.

[0030] Embodiment 1 S1. Establish a fine reservoir numerical simulation model for the pump - pumping pressure - testing process; Considering the current on - site pump - pumping pressure - testing process, an oil reservoir numerical simulation method is adopted to establish a fine reservoir numerical simulation model for simulating the pump - pumping pressure - testing process. The model establishment can be mainly divided into three modules: establishing a multi - layer radial grid geological model and a multi - phase multi - component seepage model, model operation, and model result processing. Establishing a multi - layer radial grid geological model and a multi - phase multi - component seepage model is mainly for establishing geological models and fluid property parameters, that is, based on the oil and gas reservoir, fluid properties, probe size, and pump - pumping dynamic process parameters, using grid encryption, fluid zoning, and multi - phase fluid state equations to establish a fine multi - layer radial grid geological model and a multi - phase multi - component seepage model to simulate the fluid flow during the pump - pumping pressure - testing process; the model operation module is to establish a probe model, and combined with the established multi - layer radial grid geological model and multi - phase multi - component seepage model, quickly generate a fine reservoir numerical simulation model that truly reflects the pump - pumping pressure - testing process; the model result processing module can directly output the pump - pumping fluid volume and the bottom - hole flowing pressure dynamic curve, as Figure 1 , and a schematic diagram of the pump - pumping pressure - testing numerical simulation model of a certain oil reservoir is given.

[0031] Specifically, using Eclipse numerical simulation software, by constructing a multi - layer radial grid geological model and based on the multi - phase multi - component seepage model, the dynamic laws during the pressure - testing pump - pumping process are simulated.

[0032] The parameters of the multi - layer radial grid geological model are as follows: (1) Model shape and size: The model adopts a cylindrical shape with a radius of 1000 cm and a thickness of 50 cm. This size design can fully consider the fluid flow characteristics in the near - wellbore area and the heterogeneity of the formation, while taking into account the calculation efficiency and accuracy, as Figure 2 shown, and a schematic diagram of the pump - pumping process of the wireline formation tester is given.

[0033] (2) Mesh generation: Centered on the wellbore, multi-layer radial meshes are used for generation, with the number of meshes being 10,000. In the near-wellbore area, the drilling fluid filtrate invades, with a radius less than 50 cm, and the solid-phase contamination area has a radius of 1 cm, which is described by local mesh refinement, with the mesh size being 0.1 cm. Local mesh refinement can more accurately capture the fluid flow and pressure changes in the near-wellbore area and improve the simulation accuracy.

[0034] (3) Permeability and fluid distribution settings In the virgin formation, the permeability is set to 0.01 - 10.0 mD according to the actual reservoir data, and the porosity is 0.104. A solid-phase contamination zone is set within a 1 cm length range around the wellbore. The permeability within this zone is significantly reduced, and the reduction degree is controlled by the solid-phase contamination degree, which is 10% - 90% of the permeability of the virgin formation; in the near-wellbore area, within a radius of 50 cm, a filtrate invasion zone is set, with the permeability being 80% of the virgin formation, as Figure 3 shown, a schematic diagram of the near-wellbore mud contamination zone is given.

[0035] The model contains two fluids: crude oil and drilling fluid filtrate. The composition of the crude oil is C1 - C6 + , N 2 , CO 2 , and the viscosity range is set to 0.205 - 7.5 cP. The drilling fluid filtrate mainly consists of formation water, with a viscosity of 0.204 cP. In the near-wellbore area, the fluid in the filtrate invasion zone is mainly the drilling fluid filtrate, while in the area far from the wellbore, the fluid is mainly crude oil. The model considers the influence of water saturation on fluid flow, and sets the initial water saturation of the virgin formation to 0.4, while due to the invasion of the drilling fluid in the contaminated area, the water saturation increases to 0.8. The difference in water saturation also affects the relative permeability and flow resistance changes of the reservoir fluids, making it closer to the actual reservoir situation.

[0036] The establishment of the multi-phase multi-component seepage model is as follows: The original reservoir pressure of the model is 38.5 MPa, and the original reservoir temperature is 120 °C. Considering the complex phase change characteristics during the reservoir fluid pressure measurement process, a multi-phase multi-component seepage simulation model is adopted, comprehensively considering factors such as fluid viscosity, density, relative permeability, and reservoir heterogeneity. The establishment of the model is based on Darcy's seepage law and combines multi-phase flow theory to accurately simulate the fluid flow behavior in the actual reservoir. The viscosity of the oil-phase fluid is relatively high, about 0.205 - 7.5 cP, and the relative permeability of the oil phase decreases with the increase of water saturation. The viscosity of the water-phase fluid is relatively low, about 0.2 cP, and the relative permeability 𝑘𝑟𝑤 of the water phase increases with the increase of water saturation, as Figure 4 shown.

[0037] The establishment of the probe model is as follows: (1) Probe size: The probe size is set according to the actual reservoir wireline formation testing conditions, and the size range is 0.04 - 5.97 in 2 .

[0038] (2) Probe position: The well is located at the center of the model, and the probe depth is in the middle of the model. Take the pressure at the contact point between the probe and the formation as the wireline formation testing pressure. By changing the pressure measurement extraction time and pressure measurement extraction speed parameters in the model, the simulated pressure change curves under different conditions can be obtained. As Figure 5 shown, a multi-layer radial grid geological model with a probe is given.

[0039] S2. Based on the fine reservoir numerical simulation model established in step S1, determine the main influencing factors of the pump pressure measurement mobility through orthogonal experimental design; Considering the actual situation on site, select the reservoir permeability, fluid viscosity, pressure measurement extraction time and speed, and probe area as the influencing factors of the pressure measurement mobility. Set the variation range of the parameters of each factor, use the orthogonal experimental design method, select the orthogonal design table L16(45), determine the experimental scheme, a total of 16 groups of experiments, each factor has 4 levels. Table 1 is the statistical table of the orthogonal experimental design scheme for the main influencing factors of the pressure measurement mobility.

[0040] Table 1 Orthogonal experimental design scheme for the main influencing factors of the pressure measurement mobility

[0041] Use the fine reservoir numerical simulation model of the pump pressure measurement process established in step S1 to simulate and calculate the above-designed scheme. Apply the area integration method to calculate the interpreted mobility, and take the actual fluid mobility as the benchmark to obtain the mobility calculation deviation and range of each scheme.

[0042] The specific simulation calculation is as follows: Input the reservoir permeability, fluid viscosity, pressure measurement extraction time and speed, and probe area. Use the fine reservoir numerical simulation model of the pump process established in step S1, and calculate the interpreted mobility by the area integration method.

[0043] The area integration method is a method for calculating the fluid mobility by integrating the area of the pressure measurement curve. The area integration method integrates the pressure over time, and this method is a further improvement based on the standard mobility calculation formula. Assume that the experimental core is an infinite formation and is completely saturated with liquid, and has a constant porosity, permeability, and storage coefficient ΦCt. There is a circular wellbore passing through the formation in the vertical direction. When the wellhead pressure > formation pressure, no filtrate invades the formation; when the wellhead pressure < formation pressure, the filtrate flows into the intermediate container through the probe. According to the area integration method calculation formula, only the pressure signal during the period from the start of pressure drop during pump extraction to the pressure recovery and balance after pump shutdown needs to be integrated over time and the total volume of the pumped fluid needs to be calculated. AsFigure 6 As shown, a pressure drop - recovery graph is given. This method can be applied to various situations in mobility testing: non - constant pumping speed, mud - wellbore adhesion, and anisotropic formations. The main calculation formulas are Formulas (1) - (3).

[0044] (1) (2) From Formulas 1 - 2, it can be deduced that: (3) Where, k is the permeability, mD ; μ is the fluid viscosity, cP ; P fl is the flowing pressure, MPa ; P i is the initial formation pressure, MPa ; t 2 is the start time of pumping, s ; t max is the time when the pressure recovers and stabilizes after stopping the pump, s ; V T is the total pumped fluid volume at time T , cm 3 ; V t is the total pumped fluid volume at time t , cm 3 ; V is the total pumped fluid volume during the experimental test cm 3 ; G Rf is the experimental correction coefficient value; Δp is the pressure drop during the pressure measurement and extraction process, MPa ; Δt is the time during the pressure measurement and extraction process, s ; is the flow efficiency function; is the perforation hole radius, cm ; is the wellbore radius, cm ; is the horizontal permeability of the formation, mD ; is the vertical permeability of the formation, mD .

[0045] In the actual cable formation test process, the correction coefficient value G Rf Only equipment factors and some formation factors such as permeability are considered, and other formation factors such as fluid viscosity and process factors are not considered, resulting in a large error between the interpreted fluidity and the actual fluidity. Therefore, the present invention will add other formation factors and process factors to the following experiments. G Rf The influence of the value can form a more accurate pressure measurement and flow interpretation method.

[0046] The actual fluid flow rate is calculated by formula (4): (4) in, is the fluidity, mD / cP ; k is the permeability, mD ; μ is the fluid viscosity, cP ; The permeability and fluid viscosity are artificially set in the fine reservoir numerical simulation model based on the target area data, specifically the permeability and fluid viscosity input into the model for each experimental scheme.

[0047] like Figure 7 , the influence of each factor on the fluidity interpretation results is ranked: Through the differential analysis of different experimental schemes, the influence of various factors on the interpretation accuracy of pressure flow is clarified, and the specific order is: probe area > pressure measurement extraction time > reservoir permeability > pressure measurement extraction speed > fluid viscosity. The influencing factors with the smallest differential are eliminated, and the remaining influencing factors are used as the main controlling factors of the interpretation results of pressure measurement flow. A larger probe area can reduce fluid flow resistance and improve test efficiency, while a smaller probe area may lead to larger flow interpretation errors, especially in low permeability reservoirs, where vacuum phenomenon is prone to occur; pressure measurement time and pressure measurement speed are also important influencing factors, which together determine the total amount of fluid extraction and flow state, and thus affect the shape of the pressure measurement curve and the flow interpretation results; although reservoir permeability and fluid viscosity have a certain influence on flow interpretation, they are relatively small among the above factors. S3, based on the main influencing factors of the pumping pressure measurement fluidity determined in step S2, using the fine reservoir numerical simulation model established in step S1, establish a pumping pressure measurement fluidity G correction coefficient database model; For the main influencing factors of the pressure flow rate determined in step S2, namely the probe area, pressure extraction speed and time, and reservoir permeability, the range of their variation is set to 0.1~10in according to the actual situation on site. 2, 0.01~10 cc / s, 5~50 s, and 0.001~10 mD. The Latin hypercube sampling method is introduced to determine the simulation scenarios, with a total of 1000 groups. Using the numerical simulation model building method established in step S1, simulations are carried out separately. The interpreted mobility is calculated using the area integration method. Based on the actual fluid mobility, the G correction coefficient of each scenario, i.e., G( δ, v, t, k ), is calculated, and then a database model between the main influencing factors of the pressure-measured mobility and the G correction coefficient is constructed.

[0048] Specifically, the database model consists of independent variables and dependent variables. The independent variables are the pressure-measured extraction rate, pressure-measured extraction time, probe size, and reservoir permeability, and the dependent variable is the G correction coefficient. Among them, G = actual fluid mobility / interpreted mobility.

[0049] The interpreted mobility is the same as formula (3).

[0050] The actual fluid mobility is the same as formula (4). The fluid viscosity can be obtained by running the fine reservoir numerical simulation model according to the proportion of each component set in the fine reservoir numerical simulation model and the corresponding reservoir temperature and pressure conditions.

[0051] S4. Based on the database model of the G correction coefficient of the pump pressure-measured mobility established in step S3, using the neural network algorithm, through training and verification, a G correction coefficient prediction model based on the neural network algorithm is constructed; The core formula of the G correction coefficient prediction model is shown in (5). A key coefficient G( δ, v, t, k ) is introduced. The calculation formula of the interpreted mobility ([[]] , v, t, ) is as follows: (5) Among them, k is the permeability, mD ; μ is the fluid viscosity, cP ; P f （t） is t the flowing pressure at time MPa ; P i is the initial formation pressure, MPa ; t 0 is the start time of pump extraction, s ; V f （T） is T the total volume of pumped fluid at time cm 3 ;V f （t 0 ） is t 0 the total fluid volume pumped by the master pump at a moment, cm 3 ; G ( δ, v, t, k ) is a dimensionless correction coefficient, which is used to adjust and correct the interpretation results of pressure-measuring mobility to improve the accuracy of interpretation.

[0052] The calculation method of the actual fluid mobility is the same as formula (4).

[0053] Based on the database model of the G correction coefficient for pump-pressure-measuring mobility established in S3, using the artificial neural network algorithm ANN, taking 70% of the data in the database as the training set and the remaining 30% as the validation set, setting reasonable neural network input parameters for training to ensure that the training accuracy reaches more than 98%. The training results are shown in Figure 8(a), and the training accuracy can reach 99.7%. Based on this, according to the trained model, using the validation set to detect the accuracy of the training model to ensure that the prediction accuracy reaches more than 95%. As shown in Figure 8(b), the accuracy reaches 98%, making the model have a high credibility. By comparing with the experimental results, the results show that the mobility interpretation error of the established model is 1.2% - 4.6%, which verifies the accuracy and reliability of the established G correction coefficient prediction model. Among them, the interpretation error is the error between the interpreted mobility ( δ, v, t, k ) and the actual fluid mobility.

[0054] S5. Based on the G correction coefficient prediction model established in step S4, establish a precise interpretation method for the general mobility of the reservoir based on the iterative method; For the actual pump-pressure-measuring curve, determine the probe area size according to the used wireline formation testing tool and construction technology; analyze the extraction speed and time data extracted during the actual pressure measurement process to determine the pressure measurement extraction speed and pressure measurement extraction time input to the model; according to the data in the pressure build-up section, perform data point encryption and extended regression processing, comprehensively integrate core, logging, and well logging data, give the initial permeability of the formation in the test section, and calculate the fluid mobility based on this permeability. This fluid mobility is the true mobility of the reservoir; input the above parameters, namely the probe area, pressure measurement extraction speed, pressure measurement extraction time, and permeability, into the G correction coefficient prediction model established in S4 to obtain the G correction coefficient, and use the area integral method to calculate the fluid mobility. This fluid mobility is the interpreted mobility ( δ, v, t, k ), and compare the interpreted mobility ( δ, v, t, k ) with the true mobility of the reservoir: If the set termination condition is not met, then perform iterative calculations until the set termination condition is reached, and export the final fluid mobility size to realize the interpretation of the general fluid mobility.

[0055] The termination condition is: the interpretation error is less than 3%.

[0056] The interpretation error is the error between the interpreted mobility ( δ, v, t, k ) and the true reservoir mobility.

[0057] The true reservoir mobility is calculated according to formula (4), where the actual fluid viscosity is obtained from the data of on-site pump sampling and fluid samples, and the physical properties such as the viscosity of the fluid are obtained through PVT analysis in the ground laboratory.

[0058] The formula for the interpreted mobility ( δ, v, t, k ) is the same as (5).

[0059] Based on the area integral method and seepage theory, the present invention establishes a brand-new general pressure-measuring mobility accurate interpretation model, aiming to improve the accuracy and reliability of mobility interpretation. The core formula of the new model is shown in (5), where a key coefficient G ( δ, v, t, k ) is introduced. This coefficient comprehensively considers the reservoir characteristic permeability k and the pressure-measuring process parameters in the well condition, including the probe area δ, the pressure-measuring extraction speed v and the pressure-measuring time t of the four main control factors. By incorporating these factors into the model, the new model can more comprehensively reflect the flow characteristics of reservoir fluids during the actual pressure measurement process, thus providing a more accurate theoretical basis for mobility interpretation.

[0060] Based on this new model, this study establishes a complete set of general pressure-measuring mobility interpretation methods. The technical route of this method is as Figure 9 shown, covering the whole process from data acquisition, model parameter determination to mobility interpretation. It mainly includes the following steps: (1) Data acquisition: Data acquisition is the basis of the entire interpretation process, and its quality directly affects the accuracy of the final interpretation result. The original data is obtained through on-site pressure measurement tests, including key parameters such as the pressure change curve, pressure-measuring time, and pressure-measuring speed. At the same time, combined with laboratory analysis, the characteristic parameters of the reservoir are obtained, such as permeability, porosity, and fluid viscosity.

[0061] (2) Model parameter determination: Determine the key parameters in the model, including the reservoir characteristic parameter permeability and the pressure-measuring process parameters probe size, extraction speed, and extraction time. Through these parameters, the coefficient G in the model is calculated, which comprehensively reflects the influence of reservoir characteristics, pressure-measuring technology, and well conditions on mobility.

[0062] (3) Mobility interpretation: After determining the model parameters, use the area integral method to process the pressure measurement curve, and combine the seepage theory to perform mobility interpretation on the processed pressure measurement curve to calculate the true mobility of the reservoir.

[0063] (4)Result verification: Evaluate the accuracy of the interpretation results by comparing and verifying with the actual mobility data. If there is a large deviation between the interpretation results and the actual data, adjust the model parameters and perform the interpretation again until the interpretation results are consistent with the actual data.

Claims

1. A general pressure measurement flow interpretation method, characterized in that: The following steps are involved: S1. Establish a fine reservoir numerical simulation model for the pumping and pressure measurement process; S2, based on the fine reservoir numerical simulation model established in step S1, determine the main influencing factors of pumping pressure measurement fluidity through orthogonal experimental design; S3, based on the main influencing factors of the pumping pressure measurement fluidity determined in step S2, using the fine reservoir numerical simulation model established in step S1, establish a pumping pressure measurement fluidity G correction coefficient database model; S4, based on the pump pressure measurement flow G correction coefficient database model established in step S3, using the neural network algorithm, through training and verification, to construct a G correction coefficient prediction model based on the neural network algorithm; S5. Based on the G correction coefficient prediction model established in step S4, an accurate interpretation method for universal reservoir fluidity based on an iteration method is established.

2. The universal pressure measurement flow interpretation method according to claim 1, characterized in that: The orthogonal experimental design described in step S2 refers to selecting possible influencing factors of pressure measurement flowability based on consideration of the actual situation on site, setting the range of variation of the influencing factors, and determining the experimental scheme using the orthogonal experimental design method; step S2 is based on the fine reservoir numerical simulation model of the pumping pressure measurement process established in step S1, and simulates and calculates the experimental scheme: the area integration method is used to calculate the interpreted flowability, and the flow calculation deviation and differential of each scheme are obtained based on the actual fluid flowability.

3. The universal pressure measurement flow interpretation method according to claim 2, characterized in that: The possible influencing factors of the pressure measurement fluidity include one or more combinations of reservoir permeability, fluid viscosity, pressure measurement extraction time and speed, and probe area.

4. The universal pressure measurement flow interpretation method according to claim 2, characterized in that: In step S2, the main controlling factors of the pressure flow interpretation result are determined according to the calculated difference size: the differences are sorted from large to small, the factor with the smallest difference is eliminated, and the remaining factors are used as the main controlling factors of the pressure flow interpretation result.

5. The universal pressure measurement flow interpretation method according to claim 1, characterized in that: Step S3 is specifically as follows: according to the actual situation on site, a series of parameters are set for the main influencing factors of the pressure measurement fluidity determined in step S2, and the Latin hypercube sampling method is introduced to determine the simulation scheme, and the fine reservoir numerical simulation model of the pumping pressure measurement process established in step S1 is used for simulation, and the interpreted fluidity is calculated by the area integration method. Based on the actual fluid fluidity, the G correction coefficient of each scheme is calculated, and a database model between the main influencing factors of the pressure measurement fluidity and the G correction coefficient is constructed, wherein G=actual fluid fluid fluidity / interpreted fluidity.

6. The universal pressure measurement flow interpretation method according to claim 1, characterized in that: The neural network algorithm in step S4 is ANN.

7. The universal pressure measurement flow interpretation method according to claim 1, characterized in that: Step S5 includes: 1) Determine the parameter value of the G correction coefficient prediction model established in step S4 and input the determined parameter value into the G correction coefficient prediction model established in step S4 to obtain the G correction coefficient. Use the area integration method to calculate the fluid mobility. This fluid mobility is the interpreted mobility ( δ,v,t,k ), the parameters include probe area, pressure extraction speed, pressure extraction time, permeability, the interpreted mobility ( δ,v,t,k ), δ is the probe area, v To measure the pressure extraction speed, t is the pressure extraction time, k is the permeability; 2) Calculate the fluid mobility according to the permeability described in 1), and this fluid mobility is the real fluidity of the reservoir; 3) Explain the fluidity ( δ,v,t,k ) is compared with the actual fluidity of the reservoir: if the set termination condition is not met, it is iterated until the set termination condition is met; 4) Derive the final fluid flow size and realize the interpretation of universal fluid flow.

8. The universal pressure measurement flow interpretation method according to claim 7, characterized in that: The parameter values ​​of the G correction coefficient prediction model established in the input S4 are determined according to the following method: For the actual pumping pressure test curve, the probe area is determined according to the cable formation test tool and construction technology used; Analyze the speed and time data extracted during the actual pressure measurement process to determine the pressure measurement extraction speed and pressure measurement extraction time input into the model; According to the pressure recovery section data, data point encryption and extended regression processing are performed on it, and the initial permeability of the test section formation is given by integrating the coring, logging and well logging data.

9. The universal pressure measurement flow interpretation method according to claim 7, characterized in that: The termination condition is that the interpretation error is less than 3%, and the interpretation error is the interpretation flow ( δ,v,t,k ) and the true reservoir fluidity.

10. The universal pressure measurement flow interpretation method according to any one of claims 1 to 9, applied in the development of a universal pressure measurement flow interpretation module, flow interpretation of effective points in a mine or flow interpretation of dense points in a mine.

Images